• Author / Uploaded
• Entratainment

Citation preview

title: author: publisher: isbn10 | asin: print isbn13: ebook isbn13: language: subject publication date: lcc: ddc: subject:

Gaskets and Gasketed Joints Bickford, John H. CRC Press 0824798775 9780824798772 9780585139357 English Gaskets, Joints (Engineering) 1997 TJ246.G38 1997eb 621.8/85 Gaskets, Joints (Engineering)

Page i

Gaskets and Gasketed Joints Edited by John H. Bickford Middletown, Connecticut

MARCEL DEKKER, INC.

NEW YORK BASEL HONG KONG

Page ii

Library of Congress Cataloging-in-Publication Data Gaskets and gasketed joints / edited by John H. Bickford. p. cm. Includes bibliographical references and index. ISBN 0-8247-9877-5 (acid-free paper) 1. Gaskets. 2. Joints (Engineering) I. Bickford, John H. TJ246.G38 1997 621.8'85--dc21 97-41590 CIP The publisher offers discounts on this book when ordered in bulk quantities. For more information, write to Special Sales/Professional Marketing at the address below. This book is printed on acid-free paper. Copyright ©1998 by MARCEL DEKKER, INC. All Rights Reserved. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission in writing from the publisher. MARCEL DEKKER, INC. 270 Madison Avenue, New York, New York 10016 http://www.dekker.com Current printing (last digit): 10 9 8 7 6 5 4 3 2 1 PRINTED IN THE UNITED STATES OF AMERICA

Page iii

PREFACE The bolted joint is a very complicated beast. So many variables affect its behavior that its performance in service is difficult to predict or explain. These problems arise for several reasons. First of all, bolts and nuts work properly only if frictional forces resist relative motion between themand friction is an unreliable ally. Literally hundreds of factors can affect the friction restraint between bolt and nut, so the resulting forces are economically impossible to control or predict. This makes the behavior of the fastener difficult to predict. Worse, it introduces significant uncertainty when we tighten the fasteners with an economically attractive tool (like a torque wrench). We can't predict exactly how much preload we'll create in the bolts for a given torque, yet the magnitude of that preload will probably dominate joint behavior. Second, joint behavior is strongly influenced by the elasticity of its partsthe nuts, bolts, and joint members. As a result, any attempt to explain this behavior must take elasticity into account. Gone are those happy, freshman days when we could treat everything as frictionless, rigid bodies. The difficulties increase significantly if the joint contains a gasket, because we now must also deal with the behavior within the joint of a body which is neither rigid nor elastic, but is elasto-plastic. It is partially elastic, but the degree of elasticity (its resilience or stiffness) is not a constant but is a function of the compressive stresses which have been placed on the gasket during assembly and after it was put in service. Furthermore, thanks to its partially

plastic nature, we must also take into account such things as permanent deformation and creep. Gaskets also escalate our concern for another reason. Although failure of many non-gasketed joints (airplane or bridge structures, for example) can be catastrophic, the failure of the majority of nongasketed joints is merely a nuisance. This is rarely true of the failurethrough leakageof gasketed joints. At best, this merely causes things like a large puddle on the floor or the temporary loss of horsepower in an engine; at worst, leakage can lead to pollution, fire,

Page iv

explosion, and/or the release of deadly chemicals. In short, the gasketed joint presents a real challenge to the engineering community. It is the intent of this book to help illuminate the darkness. Paradoxically, considering the fact that behavior of the gasketed joint is more complex than that of its non-gasketed cousin, we've probably learned more about gasketed joints in the last couple of decades than we have about the others, thanks to intensive research efforts in the United States and in Europe. Much of this work was sponsored by the Pressure Vessel Research Committee (PVRC) of the Welding Research Council (WRC), headquartered in New York. Most of the PVRC research was conducted at École Polytechnique in Montreal, but significant contributions have also been made by the Varysburg Gasket Laboratory in Albany, New York, by CETIM in France, and by the British Hydromechanics Research Agency (BHRA) in England. Further inputs have also been received from Germany and Japan. In any event, the main focus has been on large pressure vessel and piping joints of the sort found in refineries, petrochemical and power plants, and similar applications. Engineers who deal with smaller gasketed joints (in automobiles, for example) have told me that the basic behavior of those joints appears to be the same as that of the larger joints. Only the numbers change. Although many of the chapters in this book have been written by people involved with the PVRC and related efforts, I believe that most of the material will be of interest to those dealing with other types of gasketed joints as well. Eight of the fourteen chapters are devoted wholly or in part to other gasket types.

The book starts with a general discussion of the behavior of any type of gasketed joint by Dr. Jörg Latte of Istag, AG. This is followed by a full discussion of automotive gaskets by Daniel E. Czernik, of Fel-Pro in Skokie, Illinois, a leading manufacturer of such gaskets. Dan has been involved, from the beginning, in the PVRC research, so his comments are broad-based and authoritative. We then move on to a discussion of pressure vessel and other nonautomotive gaskets written by Dr. Latte and by Derek Coomber, President of Thermoseal Inc. Both men are associated with Klinger, GmbH, a leading European manufacturer of gaskets. The fourth chapter deals with chemical gaskets and is written by John Cocco of Loctite, in Connecticut (a leader in this field). Chapters 2, 3, and 4 together, therefore, define almost all of the commercially important types of gaskets in use today. With such a large number of types and configurations available, how can we identify the one best suited for a particular application? The next three chapters will help to answer that question. First, in Chapter 5, we have a long and detailed discussion of the new techniques developed at École Polytechnique in Montreal for evaluating and testing gaskets, thereby defining the characteristics that will play a large part in the selection process. This important chapter was written by Michel Derenne and Luc Marchand, and by consultant Jim Payne, in collaboration with André Bazergui. Each of these men played a major role in the development of these test procedures. The resulting

Page v

École tests are far more sophisticated than the simplistic tests which have been used in the past and some of them are currently being studied for possible use as economical and easy to use ASTM standards. The École tests, however, were originally designed to probe the mechanical and leakage behavior of gaskets, so this chapter also contains a lot of information about the behavior of the gasket. Furthermore, because a number of tests are required to fully reveal this behavior, École personnel and others involved with the PVRC research have also developed a number of rating factors which will someday improve the accuracy with which gasket manufacturers define gasket properties for gasket users. These rating factors are also defined and explained in Chapter 5. Knowing the mechanical and leakage behavior of many different types of gasket is useful when selecting a gasket for a particular application, but there are also many other factors you must consider when making a decision. J. Ronald Winter of Tennessee Eastman, who has been involved for many years in the PVRC activities, has developed a gasket selection procedure which incorporates the PVRC work but is primarily based on his many years of experience in a large petrochemical plant. His insightful chapter (Chapter 6) will help the reader apply the PVRC results to real-world pressure vessel and piping problems. As is evident from the length of his chapter, gasket selection involves a lot more than picking a part number from a manufacturer's catalog. It was inevitable that someone, in this age of the computer, would write a program to aid in gasket selection. Klinger has done so, as described in Chapter 7 by Reinhard Rödel. This program can be used to pick gaskets for any type of joint, including but not limited

to pressure vessel or piping joints. The selection procedure is based, in part, on the joint design procedures developed by the German engineering society Verein Deutscher Ingenieure (VDI). This interweaving of gasket selection and joint design is not just a convenience, it's a necessity. The gasket alone cannot prevent leakageit can do so only with the support and assistance of the joint members, i.e., of the joint itself. The VDI procedure gives us important insight into why this is so, and into how a good design can contribute. Because of its value as an educational tool, the VDI procedure lies at the core of the following chapter (Chapter 8) in which I summarize the procedure to show how gasket behavior affects the design of any type of gasketed joint. The VDI procedure is the most sophisticated bolted joint design procedure publicly available today. It takes all important aspects of joint design and the joint's often uncertain behavior into visible account. By contrast, the design procedures to be discussed next incorporate a variety of safety factors to account for behavioral uncertainties in an invisible manner. This VDI approach to the effect of behavior on design should help you understand and appreciate the simplified design rules presented in Chapters 9, 10, and 11. These chapters cover existing rules and new ones that have been

Page vi

proposed for and/or incorporated into the ASME Boiler and Pressure Vessel Codea legal and mandatory document for those involved in pressure vessel and piping design. These chapters were written by James R. Payne, long time chairman of many of the PVRC task groups and committees, and by authors Robert Schneider and A. E. Blach. In Chapter 9, Payne and Schneider describe the traditional Code rules for the design of circular, raisedface flanges, and then go on to describe the proposed new rules which take our new knowledge of gasket behavior into account. Payne's own contribution to this understandingand to the new rulesis especially significant. Chapters 10 and 11, by Blach, cover design procedures for noncircular joints and for flanges in full face contact. Blach has proposed similar rules for the Code itself. Although these three chapters deal exclusively with pressure vessel and piping joints, they could well serve as models for gasketed joint design rules in other industries since they are solidly based on a thorough understanding of gasket physicsan understanding gained by research efforts costing well over $750,000. It's worth noting that the traditional Boiler and Pressure Vessel Code was designed primarily to prevent catastrophic failure of pressurized equipmentsafety was the principal goal. More recent concerns about pollution and its effect on the environment, however, culminating in the laws dealing with fugitive emissions, have introduced leakage as an additional, important concern. The search for new rules, therefore, was in large part an attempt to define design procedures for leak-free joints. Unfortunately, the research has shown that there's no such thing as a truly leak-free joint (at least if the contained fluid is a light gas) forcing the joint designer to define an acceptable level of leakage when sizing the

joint members and bolts. Liquid joints can be made leak-free, but this will sometimes require uneconomically large and expensive joint members. In any event, the new design rules will deal with leakage for the first time. Selecting the best gasket and placing it in a properly designed joint are critically important steps, but we can't stop there. We must now assemble that joint (tighten those bolts) correctly because the PVRC tests have shown us that there will be a direct correlation between in-service leakage and the initial compressive stress with which we seat the gasket. This stress must be great enough to mate the gasket intimately with joint surfaces, filling all imperfections. Too little seating stress and fluid can leak along unplugged tool marks and the like. Too little stress can also lead to dramatic blowout of the gasket when the system is pressurized. Too much seating stress, however, can damage the gasket or can rotate the flange and partially unload it, in either case opening other leak paths. So getting the right seating stress is important and difficult. In fact, it's so difficult that the Code has traditionally chosen to remain almost mute on the subject of how to actually achieve a given seating stress. The proposed new rules do include a brief nod to the assembly uncertainties, as discussed in Chapter 9. Unlike the more complex VDI procedure, however, the Code still ignores most of the factors that affect seating stress. (These difficulties are covered indirectly,

Page vii

however, by a number of safety factors which have been incorporated in the rules.) This seating stress, of course, is created by tension (often called preload) in the tightened bolts and that tension is subject to many variables both during and after assembly. Even a properly tightened bolt can lose some or most of its tension when neighboring bolts are subsequently tightened, as explained in Chapter 12. Things like flange misalignment or gasket creep can also degrade seating stresses. Although it's often impossible to predict exact results, it helps to be aware of the potential problems, some of which, at least, can be avoided by careful attention to details. Chapter 12 also discusses a number of relatively uncommon preload control or assembly techniques that can help us deal with critical joints or chronic leakers. However, to paraphrase the poet Robert Burns, the best laid plans of mice and engineersor engineers and men, I can never remember whichaft gang a-gley. Whatever that means. I think it means trouble. It means that good gaskets, good joint designs, and good assembly practices do not always guarantee leak-free behavior. Leaks can and do open up in service. It's sometimes useful, therefore, to be able to carry out post-assembly inspections of the joint, to determine whether retained bolt loads meet your expectations and/or whether such things as thermal and pressure cycles are slowly unloading a gasket. Chapter 13 demonstrates a few ways to do this. Sometimes it's necessary to shut down an operating system to cure a problem, but, in petrochemical types of applications, at least, it's also common to use chemicals to plug in-service leaks. The last

chapter, Chapter 14, was written by Pat Kearns and describes the techniques and materials used for this purpose. In this book you can find authoritative reviews of gasket behavior, gasket materials and configurations, gasket tests and evaluation, gasket selection, the design and assembly of gasketed joints, and procedures for combating in-service leakage. You won't find explanations for all gasketed joint mysteries or answers to all gasketed joint problems, because neither exists. But you will find authoritative, state-of-the-art discussions of what we do know. It's a lot more than was known only a few years ago. I'm grateful, as must be obvious, to the many experts who contributed chapters to this work. I hope that my attempt to edit and knit together their manuscripts has not introduced errors or misunderstandings. I'm also grateful to and acknowledge the invisible help I've received over the years from so many others in the bolting world. Although they are too numerous to list here, I would not have been able to edit this book were it not for the education they provided me. Many are named in the references you'll find in this book and in the forthcoming The Handbook of Bolts and Bolted Joints (Marcel Dekker, Inc.). Finally, I'm grateful for the gentle support and assistance I've received from my production editors at Marcel Dekker, Inc., Jennifer Kelley and Rod Learmonth. They have kept me safely on course through a sometimes confusing voyage. JOHN H. BICKFORD

Page ix

CONTENTS Preface

iii

Contributors

xi

Part I: Gaskets 1. Gasket Behavior and Its Influence on the Safety of Flanged Joints Jörg Latte

1

2. Internal Combustion Engine Gaskets Daniel E. Czernik

35

3. Industrial Gaskets Jörg Latte and Derek Coomber

87

4. Chemical Gaskets John Cocco

123

Part II: Evaluating and Testing Gaskets 5. PVRC/MTI Technology for Characterizing Gaskets Used in Bolted Flanged Connections Michel Derenne, James R. Payne, Luc Marchand, and André Bazergui

137

Part III: Selecting a Gasket 6. Gasket SelectionA Flowchart Approach J. Ronald Winter

303

Page x

7. Reliable Gasket Calculations on a Personal Computer Reinhard Rödel

389

Part IV: The Design of Gasketed Joints 8. Introduction to the Design of Gasketed Joints John H. Bickford

403

9. ASME Flanged Joint Design RulesNew vs. Traditional 423 James R. Payne and Robert W. Schneider 10. Bolted Flanged Connections for Noncircular Pressure Vessels 487 A. E. Blach 11. Bolted Flanged Connections with Full-Face Gaskets A. E. Blach

507

Part V: The Gasketed Joint in Service 12. Assembling a Gasketed Joint John H. Bickford

523

13. In-Service Inspection of Gasketed Joints John H. Bickford

541

14. Stopping Leakage of In-Service Joints Pat Kearns

559

Index

589

Page xi

CONTRIBUTORS André Bazergui Director, École Polytechnique de Montréal, Montreal, Quebec, Canada John H. Bickford Consultant, Middletown, Connecticut A. E. Blach Consulting Engineer, Montreal, Quebec, Canada John Cocco Director, North American Engineering Center, Loctite Corporation, Rocky Hill, Connecticut Derek Coomber President, Thermoseal Inc., Sidney, Ohio Daniel E. Czernik Vice President, Advanced Technology Administration, Fel-Pro Inc., Skokie, Illinois Michel Derenne Professor, Department of Mechanical Engineering, École Polytechnique de Montréal, Montreal, Quebec, Canada Pat Kearns Manager of Manufacturing, Team, Inc., Alvin, Texas Jörg Latte Managing Director, Istag AG, Egliswil, Switzerland Luc Marchand Professor, Department of Mechanical Engineering, École Polytechnique de Montréal, Montreal, Quebec, Canada James R. Payne Principal Consulting Engineer, JPAC Inc., Long Valley, New Jersey Reinhard Rödel Managing Director, Klinger GmbH, Idstein, Germany Robert W. Schneider Consulting Engineer, R. W. Schneider Associates, Allentown, Pennsylvania

J. Ronald Winter Principal Engineering Mechanicist, Engineering and Construction Division, Eastman Chemical Company, Kingsport, Tennessee

PART I: GASKETS

Page 1

1 Gasket Behavior and Its Influence on the Safety of Flanged Joints JÖRG LATTE Istag AG, Egliswil, Switzerland 1. Introduction Flat gaskets for industry in the modern sense were invented by Austrian engineer Richard Klinger about 100 years ago, based on asbestos and rubber. Due to their success, these first products attracted numerous imitators, and soon many materials of various qualities were available. It quickly became necessary to be able to evaluate these materials. The chosen method at that time, because nothing better was available, was the determination of tensile strength, which today is of only limited importance (see later). The number of different sealing types was considerably increased with the development of asbestos-free gaskets. A summary tabulation of them is found in Chapter 3, Industrial Gaskets. Because of the confusion created by such a large number of offerings, industry wanted a universal gasket suitable for all applications. Such gaskets do not and cannot exist. This goal can only approximately be approached by increasing the application ranges of gasket materials.

Page 2

An industrial engineer and user of gaskets should always become critical when good or poor gaskets are discussed. The basic applicable approach is that there are no good or poor gasketsthere is only a correct or an incorrect selection of a gasket for a given application, since many other parameters play a part in this matter (see Chapters 3 and 6). Evaluation procedures in the form of test methods and standards have been developed over the last 40 years, accelerating in the last 15 years to assist in decision making with the newly developed asbestos-free gasket materials. The most important property criteria, suitable test procedures, and the consequences of an incorrect selection will be dealt with in the rest of the chapter. 2. Quality Evaluation Criteria for Sealing Materials The most important material properties for sealing materials are: 1. Loading capacity 2. Sealability 3. Elastic behavior of a sealing material 4. Capacity for chemical resistance against media The tensile strength of a material, expressed by the breaking or tensile strength, is not of particular importance for quality consideration, because no tensile forces are usually exerted on a gasket, only pressure forces. This is also expressed by the fact that high-grade conventional asbestos-containing gaskets (CAF materials or It materials) typically have tensile strengths according

to ASTM F 152 of at least 35 MPa (5.08 ksi) across the machine direction, whereas modern highgrade asbestos-free gaskets (CSF materials, compressed synthetic fibers) under the same conditions have a tensile strength of a maximum of 15 MPa (2.18 ksi), often even less than 10 MPa (1.45 ksi). In the case of unreinforced graphite gaskets, which are generally considered to be reliable, this value is even lower. While tensile strength was specified for asbestos-containing sealing materials in the old standards, this is generally not the case for current standards for asbestos-free materials. According to the current viewpoint, the determination of tensile strength is nevertheless important for production, since it is a simple and rapid method for demonstrating production stability and reproducibility. Other methods, apart from the determination of compressibility and recovery, are too time-consuming and expensive to be used to control production quality. As already mentioned, the importance of an evaluation criterion largely depends on the application. However, the load-bearing capacity of a sealing material is traditionally considered to be one of the most important evaluation

Page 3

criteria, if not indeed the most important one. An explanation of why this is still the case follows. However, we must also consider the sealability, or the characteristic leakage behavior, of a sealing material in the selection of the correct gasket. The two characteristics, loadbearing capacity and sealability, are frequently contrary to each other. In particular, high-duty fiber gaskets often present relatively high values during leakage tests at room temperature. In the evaluation of sealing quality, we must therefore evaluate how well a manufacturer has succeeded in balancing both properties. However, here, too, we must differentiate since the leakage rate for these sealing materials falls drastically on raising the temperature. Leakage investigations have systematically been made since the introduction of asbestos-free fiber gaskets. The leakage is normally measured as a function of the internal pressure at constant surface pressure. The minimum surface pressure necessary for a defined sealability in the operating state is derived from the results thus obtained. The first work in this direction was presented by Ernst Sauter [1]. Further investigations led to: 1. the PVRC-ROTT test procedure [2] 2. The German industrial standard DIN 28090 [3] These techniques are addressed elsewhere and therefore need not be discussed further here (see Chapter 5). Unfortunately, however, all the techniques mentioned are based only on measurements at room temperature, which in our opinion is insufficient, because the leakage may greatly decrease even at moderately elevated temperatures. Figure 1 shows this in the

example of a CSF-G fiber gasket based on glass fiber. This conformity is valid for all fiber gaskets and for some highly loadable PTFE gaskets. It can be demonstrated that, at constant surface pressure, temperature elevation acts on the sealability of a flange joint in a manner similar to elevating the surface pressure at room temperature, as shown in Fig. 2. The previously mentioned processes for determining design factors therefore require appropriate adaptation. Unfortunately, these considerations have not been widely accepted to date, perhaps partly because more intensive and more costly investigations must be conducted. The loading capacity and the characteristic leakage behavior as a function of the application temperature must therefore be considered as the first criterion for the selection of the correct gasket. In addition, consideration should also be given to chemical resistence, i.e., how a medium acts on a gasket and thereby influences the safety of the flange joint. An appropriate test procedure is given in German industrial standard DIN 28090, Part 3. In view of the large number of media and application parameters that come into consideration, however, it will take years before it is possible to make absolutely reliable statements about this.

Page 4

Figure 1 High-temperature sealability.

Page 5

Figure 2 Dependency of leak rates: (a) on temperature, with gasket pressure constant; (b) on gasket pressure, with temperature constant.

3. Using a Model to Understand Gasket Behavior A flange joint is shown in Chapter 3, Industrial Gaskets, and is repeated here as Fig. 3, which shows the forces acting on the gasket. We will use this representation as a starting point for further discussion. It is clear that we are schematizing this picture, since we consider only the pressure conditions. This can best be achieved in a function diagram by plotting the pressure conditions against the distance from the flange center. The installation surface pressure amounts to sv = 50 MPa (7.25 ksi) in our model. It is governed by the tightness class to be maintained in an operational condition. We thus obtain Fig. 4a. In the next step, the flange is placed under internal pressure Pi. This now causes hydrostatic end thrust HET, for which the following relation applies:

(1) where ai = internal area of the joint (mm2, in.2) ad = sealing area of the joint (mm2, in.2)

Page 6

Figure 3 Forces acting on a gasket.

We see from this the hydrostatic end thrust depends on the gasket and flange dimensions. The greater they are, the more rapidly we approach critical conditions by hydrostatic end thrust. This is particularly the case if the joint width and therefore sealing surface area ad is small in relation to the gasket diameter. For our model, we set as base HET relief = 10 MPa (1.45 ksi) (Fig. 4b); i.e., the surface relation is ai:ad = 2:1 at an internal pressure of 50 bar = 5 MPa (0.725 ksi). We see from our diagram that our model flange gaskets are still in a stable and safe condition. The operational surface pressure sVB at room temperature (RT) represented here is: sVB(RT) = sv - HET = 40 MPa (2)

Application conditions at low temperatures (T = RT) almost never lead to problems and need not be discussed further here. On the other hand, industrial processes at higher temperatures run differently. We will investigate how this works out for highcapacity and low-capacity gasket types at higher temperatures on our model, for example, at 300°C (572°F). Sealing materials exhibit a creep-relaxation behavior under pressure and temperature stress; i.e., they give way from the pressure exerted from outside

Page 7

Figure 4 Pressure model of a gasket joint: (a) during assembly; (b) after pressurizing.

(surface pressure) until a state of equilibrium is reached. This process is associated with a reduction in gasket thickness. A state of equilibrium is reached when the external surface pressure has decreased to the extent that the internal stress of the gasket is able to resist; i.e., the surface pressure and the stress as opposing force are equal, and the decrease in thickness is stopped. The properties indicated here are thickness-dependent, in that thin gaskets present a more favorable relaxation behavior, i.e., lesser

loss of thickness and lesser surface pressure loss. In this case, the comparison thickness is always the effective thickness, i.e., always after the installation of the gasket, and not the nominal thickness (thickness before installation).

Page 8

Figure 4 Continued.

The relaxation behavior is a characteristic property for all materials. For example: 1. Flexible graphite, after being compressed by the installation, shows only very little relaxation. 2. Fiber gaskets are found to be highly and weakly loadable (low and high relaxation), depending on the gasket type and the manufacturer. 3. PTFE gaskets usually exhibit a permanent creep; i.e., the decrease in thickness stops only when the outer forces (surface

pressure) have fallen almost to zero. (For exceptions, see the discussion of modified PTFE in Chapter 3.)

Page 9

For the model consideration started earlier, we first take up a highly loadable fiber gasket (Fig. 5a). It has a residual stress of more than 25 MPa (3.63 ksi) (see British Standard BS 7531, Definition of x-grade materials in 1.5-mm thickness) at 300°C (572°F) and 40 MPa (5.8 ksi) effective initial surface pressure; i.e., the previously mentioned equilibrium state is reached at 25 MPa (3.63 ksi). We observe that we are still on the safe side with our model flange in this case. At the thickness of 2.0 mm (0.787 in.) (usual in German industry), the remaining

Figure 5 Influence of creep relaxation on the pressure model: (a) for

high-grade gasket with high load-bearing characteristics; (b) for low-grade gasket with low load-bearing characteristics.

Page 10

Figure 5 Continued.

surface pressure would possibly still be only 20 MPa (2.9 ksi) and also safe in this case. Weakly loadable gaskets would have a residual surface pressure of perhaps 10 MPa (1.45 ksi) (Fig. 5b) under the same conditions: i.e., we approach relatively close to the critical area in which our system will probably leak. In this case, it would be possible to raise the safety margin by installing a thinner gasket. Since this gasket type usually presents high compressibility and therefore compensates better for out-of-parallel flanges and roughness or damage, a large gasket thickness usually is not necessary. If a

thicker gasket (e.g., 2.0 mm or 3.0 mm or 0.787 in. or 1.18 in.), is installed because of lack of knowledge, however, we are

Page 11

definitely in the danger area, since the remaining residual surface pressure is minimal or the gasket is ruined by overload (details follow). The impression might develop in our model consideration that weakly loadable gaskets are of low grade. This is by no means the case; rather it is due to the model parameters we selected. We presupposed a high internal pressure Pi, expressed by the relatively high value for the hydrostatic end thrust HET. If this is not the case, and the operational temperatures are much lower, this gasket type can be even better suited than a highly loadable fiber gasket. In our case, the weakly loadable gasket would be an incorrect choice for our current assumption (high internal pressure and high temperature). However, we would like to maintain this fiction in order to demonstrate certain effects. As already mentioned, PTFE gasket types normally exhibit a permanent creep. Such gaskets usually cannot be installed under the conditions portrayed here. However, the trend in the development of PTFE-based sealing materials has recently led to modifying this gasket type so that it shows significantly lower creep properties. A summary of the PTFE types currently available and their subdivision into various subgroups was presented at the 1996 FSA Conference [4]. According to this, at least one modified PTFE sealing material is currently available that would be usable under the model application parameters. Until now in the model considerations we always tacitly assumed ideal flanges, i.e., flanges with infinite stiffness. In the case of the torque applied to achieve the desired surface pressure, however, not only are the sealing materials deformed but also the flange itself,

identified as flange rotation. This is represented, somewhat exaggerated, in Fig. 6. It can be seen that the surface pressure on the internal radius of the gasket is considerably lower than that further outside. For example, we can demonstrate this with a fourbolt flange with the use of pressure-sensitive paper. At a theoretically applied surface pressure of 30 MPa (4.35 ksi) on average, a pressure of only 26 MPa (3.77 ksi) was measured on the internal diameter, but it was 37 MPa (5.37 ksi) on the outside diameter (Fig. 7). This has considerable consequences for our model consideration. The conditions in all cases can, of course, be represented only schematically. First a better sealing is to be expected at the outer edge of the gasket due to the elevated surface pressure in both cases for the highly loadable gasket and for the less highly loadable gasket. On the other hand, we recognize an altered relaxation behavior. The relaxation is less at the internal edge of the gasket than in the ideal case, but correspondingly higher at the outside edge. For the highly loadable gasket (Fig. 8), we recognize that this different relaxation behavior has practically no influence on the functionality of the gasket, since we are well within the safe range. However, this is (unfortunately) only an idealized representation. Depending on the flange type, especially in the case of large flanges, the effect presented here can lead to a situation in which the internal edge of the gasket

Page 12

Figure 6 Flange rotation.

Figure 7 Influence of flange rotation on the distribution of gasket pressure, with bolt torque even. Theoretical gasket pressure = 30 MPa (4.35 ksi). The gasket pressure is applied crosswise in four steps of 25, 30, 75, and 100%. The estimated effective gasket pressure = 26 MPa (3.77 ksi) at

inner diameter and 37 MPa (5.37 ksi) at outer diameter.

Page 13

Figure 8 Effect of flange rotation on the pressure model of a gasket joint for high-grade gaskets.

remains unloaded, and a gap between the gasket and the flanges may even form (Fig. 9). Since the effective internal diameter of the gasket and the effective gasket width are thereby altered, this affects the hydrostatic end thrust, which can increase by 50% or 100% or even more. With maintenance of the same installation pressure sv (which is unrealistic; see later), a definite additional hydrostatic end thrust DHET would result, which, because of the now-overall-lower surface pressure, also reduces the creep relaxation of the sealing material.

The maintenance of equal installation surface pressure sv preproposes, however, that the torque of the bolts would have to be reduced because of

Page 14

Figure 9 Flange rotation with change of effective gasket area.

the effectively reduced stressed area. This is unrealistic, however, because the maintenance person would not pay any attention to our model consideration but instead would tighten the bolts with all possible strength, as usual. Assuming that all other conditions remain the same as in our model, the effective installation surface pressure sv(eff) is thereby probably higher than it would have been calculated by the installation planner. It is absolutely realistic to assume that the pressed surface would become reduced by half in the case of large flanges because of flange rotation, and the average effective installation pressure could thereby become doubled. For the same reasons, the average effective installation pressure becomes extremely high at the outside edge of the gasket, but it would be zero at the effective internal edge. Many sealing materials

will not be able to withstand such high loads and will be overstressed. Weakly loadable gaskets are generally softer. In our model flange, they will therefore yield to bolting forces more strongly than will highly loadable gaskets. As a result, smaller flange rotations appear. The effective sealing flange width is not normally altered; i.e., the complicated considerations just discussed do not relate to them.

Page 15

It might now be possible to conclude that this gasket type is better suited for our model than are highly loadable gaskets. A glance at Fig. 10 reveals that with the inclusion of the relaxation at operational temperature we have to reckon with critical operational conditions. In our overall model consideration, we have advanced in steps from ideal conditions to more and more complicated conditions. We can now take one more

Figure 10 Effect of flange rotation on the pressure model of a gasket joint for low-grade gaskets.

Page 16

realistic step forward. But we wish only to outline the problem and not discuss it in detail, since this would lead too far and would be too complicated. The basic problems have already been discussed. The model cases discussed were still based on ideal conditions to a certain extent, in that the application of surface pressure took place uniformly by tightening the bolts. Unfortunately, practice has shown that the installation instructions of gasket manufacturers frequently are not observed. They mention that bolts should be tightened crosswise in at least four steps. Only by doing this can a reasonable distribution of surface pressure be achieved. Everyone knows that, for safety reasons, this rule is to be followed on any car when replacing wheels. Nevertheless, it is often found that, for the sake of convenience, installers attempt to completely tighten one bolt after another in series. Where this happens, instead of the uniform stress distribution shown in Fig. 7, at the same nominal surface pressure of 30 MPa (4.35 ksi) we find a significant unevenness in the stress distribution, as illustrated in Fig. 11. In the case of the first bolt tightened, a surface pressure of 60 MPa (8.7 ksi) is measured at the outside edge, and 40 MPa (5.8 ksi) is measured at the inside edge. Only surface pressures half as great can be applied in the case of the opposite bolt, 3, in spite of equal bolt torque.

Figure 11 Influence of rotation on the distribution of gasket pressure, with bolt torque even. Theoretical gasket pressure = 30 MPa (4.35 ksi). Estimated effective gasket pressure (MPa): At inner diameter At outer diameter

Bolt 1 40 60

Bolt 2 27 45

Bolt 3 25 30

Bolt 4 27 45

Page 17

If we transfer these conditions to our model diagrams, we recognize that insufficient surface pressure is quickly achieved by bolt 3, and leakage is to be expected. On the other hand, we must expect that, on installing a weakly loadable gasket, at bolt 1 the maximum loading capacity of the sealing material will be exceeded under the action of temperature, and the gasket will be crushed. A blowout may then develop. 4. Load-Bearing Characteristics of the Gasket In the discussion until now, we have learned that we must distinguish between highly and weakly loadable gasket types according to application conditions. The former are sought if demanding application conditions are involved, for example, high temperatures, high internal pressures, high flange pressures, and solid flange with high stiffness. On the other hand, weakly loadable gasket types are preferred at low internal pressures and temperatures, especially if in these cases the flange joints are weaker and therefore can be designed less expensively. Less loadable gasket types are usually softer and are better and sufficiently well adapted to the sealing surface of the flange, even in the case of nonuniform surface pressures. During our model considerations, we dodged the description of what highly and weakly loadable materials are and how this can be detected and defined by measuring techniques. But such knowledge is essential for the selection of a suitable gasket type. In the discussion of our model (Fig. 5), we already encountered British standard BS 7531 for obtaining an indication of relaxation at 300°C (572°F) under 40 MPa (5.8 ksi) effective initial surface

pressure. An identical procedure is described in German industrial standard DIN 52913, but where the initial surface pressure is set at 50 MPa (7.25 ksi). In addition, ASTM F 38 B should also be mentioned, although it was established mainly for automobile applications. In all three procedures, the surface pressure is applied with a central bolt, whose characteristic is also subject to temperature influences. It is also necessary to consider that the initial assembly surface pressure decreases because of the relaxation of the sealing material; thus it is not possible, on the basis of these procedures, to determine at which temperature what surface pressure was established. Moreover, the residual stress is measured after cooling at room temperature and not at the end or operating temperature. We thus see that we do not obtain any pure material characteristic with the three procedures mentioned and generally recognized techniques, but only get a comparison value that depends on various listed influences. However, the result is valuable, since it is makes a comparison among different materials possible; thus, materials with higher residual stress are always also more highly loadable. Tests according to ASTM F 38B usually do

Page 18

not give the residual stress but instead yield the percentage decrease in initial or installation surface pressure. In this case, the lower percentage value naturally indicates the more highly loadable material. If we make such a comparison, we may of course only compare values measured according to the same standard. A favorable result according to ASTM F 38B [measurement at 100°C, 23 MPa (212°F, 3.34 ksi)] does not automatically indicate that this material will also provide more favorable results at higher temperatures and initial surface pressures in tests according to BS 7531 and DIN 52913. Tests according to these three standards are suitable only for a limited (and only by experts), recognizable conclusion regarding applicability and application limits. However, we obtain pure material characteristics for loading capacity by a test procedure that was only defined in German industrial standard DIN 28090 in 1995 and based on the wellestablished Klinger hot-compression test. The same company also offers the only available test equipment for this procedure. In contrast to the earlier-mentioned techniques, this test measures the change in material under constant surface pressure, with the pressure maintained by a computer-controlled hydraulic press. The process variable is thus the change in thickness, occurring in all cases under pressure, or the effect of temperature. The test process consists of two partial steps (Fig. 12). In the first step, the designated surface pressure is applied at room temperature (left partial diagram of Fig. 12.) and the cold thickness decrease is determined (also better known as compressibility). This procedure corresponds to the installation of a gasket. In the second step (right

partial diagram of Fig. 12), the temperature is raised as the surface pressure is maintained, and the additional decrease in thickness as a function of the temperature is recorded. The reason for this hot compression is the creep, which is characteristic for every material and which the material reacts irreversibly to an outside force. This creep is normally expressed as a percentage based on the thickness after cold deformation. The following applies at: d0= initial thickness, nominal thickness dl

= thickness after application of the load at room temperature

dn= thickness after temperature action For the (cold) thickness decrease (compressibility): (3) and for the (hot) thickness decrease:

Page 19

Figure 12 Results of hot-compression test.

(4) We are now able to plot this hot compression at designated initial thickness for all operational states as a function of surface pressure s and temperature T: hot compression = f(s, T) (5) (For complete precision, we would also have to consider the heating rate and the flange surface roughness.)

Page 20

Figure 13 presents this functional dependence for a specified highly loadable fiber material, CSF-G in 2.0-mm (0.787-in.) initial thickness (other compressed fiber materials also based on glass fiber have other creep characteristics). The German industrial standard DIN 28090 or the related standard DIN 28091 for industrial terms of delivery permit a maximum hot compression of 15% for fiber gaskets, for which, it is known from experience, safe application conditions are given. We can now recognize all maximum permissible operational states for which this condition is applicable from Fig. 13. Since, on the other hand, hot compression permits a direct understanding of the loss in surface pressure to be expected, we are able to describe the conditions existing at the inner and outer edges of the gasket in Fig. 8, with consideration given to the flange and bolt characteristics. In these model considerations, we have also included more weakly loadable gasket types in addition to the highly loadable gasket (compare Fig. 13). Figure 14 presents the characteristic hot compression of the operational state for a medium-grade strongly loadable material. We see for our model in Fig. 5b that, at initial surface pressure sVB(RT) = 40 MPa (5.8 ksi), we must anticipate a hot compression of 35% at an operational temperature of 300°C (572°F). These are unsafe operational conditions according to the requirements of the German industrial standards cited. On the other hand, we can read the maximum permissi-

Figure 13 Hot compression as a function of gasket pressure and temperature for high-grade gaskets.

Page 21

Figure 14 Hot compression as a function of gasket pressure and temperature for medium-grade gaskets.

ble temperature Tmax from the same diagram as 180°C (356°F). If we realistically also include the model case of flange rotation, we must further reduce the maximum permissible temperature. We thus recognize that knowledge of the creep properties of a material provides us with an excellent selection criterion for exacting operational applications. As indicated in our model considerations, we must also deal with operational conditions that the sealing material cannot withstand and that therefore destroy its structure. In this case, a more or less continuous creeping of the sealing material does not take place; instead, the thickness changes abruptly. Such a case is presented in Fig. 15, for example, which corresponds to the test procedure of DIN 28090. In it, the surface pressure is increased in definite steps at a fixed test temperature [in the case of Fig. 15, 100°C

(212°F)dotted line]. This procedure has the disadvantage that the properties of materials based on rubber change due to the elevated temperature during the test duration of about 12 minutes. Actually, a stabilization of the loading capacity is produced, so a higher critical load limit is measured than is really applicable to the virgin material. Consequently, a better way is to elevate the temperature at constant surface pressure. The effect in this case is that we determine somewhat less favorable critical loading-capacity values, but we move further to the safe loading

Page 22

Figure 15 Crash point estimation.

side, in contrast to DIN 28090. This last-mentioned technique has already been applied by Ernst Sauter. Both Sauter and DIN 28090 now define the maximum permissible load (surface pressure) sBO of a sealing material as: 1. Critical surface pressure (load at crash point according to Fig. 15) with deduction of a safety margin of 20%: sBO = 0.8 × scritical

Page 23

2. Or, if it is not possible to determine any crash point, surface pressure at maximum 15% hot compression (see earlier) Sauter summarized the individual data thus obtained in a survey diagram (Fig. 16, lower partial diagram) as a function of the material thickness. The total diagram of Fig. 16 was published as the so-called Klinger graph in the technical brochure of this company, in which the upper partial diagram represents the minimum necessary surface pressure and the lower partial diagram the maximum permissible surface pressure in the operational state. The relations concerning the critical or maximum permissible loads presented here provide information on the correct selection of sealing materials, depending on the operational conditions in given cases. However, this applies only to the installation of gaskets free from errors, i.e., installation according to the installation guidelines of the manufacturer. If flange bolts are tightened on one side (see earlier), an overload could result on this side, with consequent destruction of the gasket and/or too low a flange load on the opposite side, which then may result in leakage. If an attempt is made to retorque the flange bolts after a very long operating time because of too low a surface pressure, fiber gaskets could crack because of the brittleness that has developed, causing a blowout. A retorque is therefore permissible only under controlled conditions in consultation with the manufacturer and in a time frame that depends on the operational temperature. The loading capacity of a sealing material can also be considerably reduced by the use of installation aids. Almost all manufacturers warn against fixing gaskets to the flange with the use of bolt lubricants or adhesives in order to facilitate installation. Figure 17

shows what can result from this. The crash points of various materials were tested once with and once without aids. A lubricant that every maintenance person has available for lubricating bolts (but unfortunately also misuses it) was selected as the aid. We observe that this leads to a reduction in load bearing of at least 2530% in all sealing materials. In the case of asbestos-free fiber gaskets (here as examples CSF-G and CSF-A), this can lead to a complete destabilization, because the critical load limit may fall by 7080%. 5. Long-Term Behavior, Creep Relaxation, and Thermal Effects Our determination and characterization of the load-bearing characteristics or creep relaxation was based on short time measurements of 1530 minutes. The knowledge obtained from this is important and valuable, because it provides information about what we may expect at our sealing site shortly after installation and up to reaching operational conditions (pressure and temperature). It is thus possible to avoid an obviously incorrect selection of a sealing material.

Page 24

Figure 16 Minimum and maximum gasket pressure.

Page 25

Figure 17 Crash point estimation and the influence of grease on a gasket.

Page 26

In addition, all nonmetallic sealing materials are subject to changes by long-term influences, which may also lead to the failure of a sealing site. This behavior is characteristic for every sealing material and in addition depends on the installation conditions, temperature, and environment. The cause or, better, causes have not hitherto been clarified. Luc Marchand [5] thus starts from a purely thermal or thermaloxidative disintegration. His initial idea is that all nonmetallic sealing materials begin to disintegrate at elevated temperature and/or are subject to oxidative disintegration by environmental air. He believes he can analytically detect this by the loss of mass alone, from which, via various measurements and considerations, he defines a maximum permissible loss of mass of 15% for the functional ability of a gasket. He comes to conclusions about longterm applicability at lower temperatures from measurements at high temperatures (see Chapter 5). This idea is fascinating, since it would provide a simple and inexpensive method for predicting the safe functional ability of a sealing site and thereby make safe preventive maintenance possible. Unfortunately, however, nature and technology are more complicated. Gasket manufacturers believe that extrapolation to lower temperatures is not possible, since this presupposes a constant decomposition rate independent of the temperature. The reason why Marchand's approach cannot be correct is quite simple. This would presuppose that a safe functional life is independent of the material's thickness, which definitely is not the case. But the strong dependence on material thickness indicates a dependence on creep-relaxation behavior. This approach is also

followed in German industrial standard DIN 28090, already repeatedly mentioned (Fig. 18). The long-term decrease in thickness in addition to the initial hot-compression behavior (see Section 4) is thus determined at constant temperature and surface pressure [in Fig. 18, T = 300°C and s = 30 MPa (572°F, 4.35 ksi)]. This step is logical, since this decrease in thickness permits conclusions about the decrease in bolt force and with it a reduction in surface pressure. We could thus theoretically calculate when the value falls below the minimum surface pressure, and we must expect leakage. In connection with test standards BS 7531, DIN 52913 and ASTM F 38B (in Section 4), we have already recognized that flange bolts are also subjected to temperature influences. In addition, flanges are not infinitely rigid, but instead exhibit flange rotation that depends on their type. For the realistic evaluation of the long-term behavior of a sealing connection, these flange effects are taken into consideration in DIN 28090 by so-called rigidity factors; i.e., in contrast to the presentation in Fig. 18, the surface pressure is readjusted according to three defined stiffnesses. The value may not fall below a defined value during a defined period at the end of the experiment, otherwise this experiment must be extended for from 16 hours to up to 100 hours.

Page 27

Figure 18 Long-duration creep relaxation.

As already mentioned, the advantage of this DIN 28090 is its consideration of the relaxation behavior for assertions about the long-term applicability of a sealing material. Because of limited experimental time, however, it is disadvantageous to consider only physical technical influences but not chemical thermal influences, which are usually of a long-term nature and may accelerate in the course of time (see Luc Marchand's approach). We will later go into these relations in a model discussion. Another possibility for long-term assertions was presented at the

Fifth FSA Conference [6]. It is based on leakage tests on a real flange, where a still-

Page 28

acceptable maximum leakage before the onset of a blowout is defined as the time application limit or failure of a sealing connection (Fig. 19). The authors now show that, in a logarithmic plotting of the duration up to the failure point in function of the inverse absolute temperature K, all failure points lie on a straight line (Fig. 20). (The authors originally formulated a

Figure 19 Definition of the life expectancy of bolted flanges.

Figure 20 Life prediction for bolted flanges.

Page 29

dependence on the normal temperature in °C. From theoretical considerations, however, such as Luc Marchand also dealt with, dependence on the inverse absolute temperature appears more logical.) The authors moreover indicate that the time up to the failure point is dependent on the: 1. Thickness of the gasket 2. Surface pressure (in these experiments, actually the assembly surface pressure) 3. Creep relaxation Logically, we must also mention the chemical environment (the medium and the air) here as a supplement. In the installation of a freshly manufactured, untreated fiber gasket, it is possible to formulate a minimum function duration of the sealing site according to this process. However, it is possible to optimize this further (Fig. 21) by certain manipulations (heat treatment before installation of the gasket, slow heating, preheating in the flange, retorquing). The processes indicated here appear to include all currently known parameters. As the authors indicate, the dependencies presented here also apply to asbestos-containing CAF gaskets. Thus, for example, the known highly loadable KLINGERit (It 400 according to DIN 3754) is directly comparable with asbestos-

Figure 21 Life prediction for bolted flanges under optimized conditions.

Page 30

free CSF-G, our glass-fiber model material. We know from the application of KLINGERit or similar qualitative CAF sealing materials over many years, however, that they are absolutely safe to apply at far higher temperatures, which would contradict the conclusions from the high-temperature flange leakage tests. The difference between laboratory testing and practical experience possibly arises because the test flange is heated from outside by a collar heater in the test procedure, whereas this takes place in practice by the transfer of heat from the medium. We therefore also have to consider the effect of a temperature gradient that is dependent on the grade of the pipe and flange insulation. If it is now considered that improved heat insulation and protection of power resources have become customary only recently, and reported temperature data are frequently based on the origin site (reactor or steam producer) but not on the application site of the gaskets, we must be cautious in drawing conclusions from old experience reports. The conclusion remains that the laboratory high-temperature leakage test simulates the worst case, without considering installation errors in practice. In our discussion until now, we have conjectured long-term creep relaxation as the cause of the long-term failure of gasket connections. We have not hitherto discussed the reason for this phenomenon. It is logical to look for it in thermal and thermaloxidative decomposition in the case of nonmetallic sealing materials, as Luc Marchand already took as a basis in his approach. The following features would come into consideration: 1. A thermal decomposition for PTFE gaskets 2. Both thermal and thermal-oxidative decomposition for elastomer

gaskets 3. Thermal-oxidative decomposition for flexible graphite Pressure plays a considerable part in the chemistry of the reaction in chemical syntheses, with the participation of gases and in thermal decompositions with the liberation of gases. This leads to the situation that elastomers under surface pressure decompose only at considerably higher temperature than unpressed materials. It is only for this reason that it is possible to use elastomer-base asbestos-containing and asbestos-free fiber gaskets at higher temperatures. If we transfer these relations to our model of a flange connection (Fig. 22), we observe that the long-term creep relaxation of a gasket depends on the: 1. Assembly pressure 2. Load-bearing characteristic This is also the result of the high-temperature flange-leakage test, but other dependencies also become recognizable from our model (Fig. 22), namely: 1. Hydrostatic end thrust 2. Internal pressure 3. Dimensions of the gasket

Page 31

Figure 22 Long-term pressure model of an idealized gasket joint.

Page 32

We furthermore recognize that thermal decomposition begins at the edges of the gasket due to the pressure drop occurring there, and it progresses into the interior of the gasket. As is indicated by the pressure characteristic curves P1 to Pn+1, this leads to a reduction in effective flange pressure. Since the rate of a decomposition is more rapid at lower pressure than at high pressure, however, the loss of surface pressure is accompanied by an accelerating decomposition. It is also observed in practice that a small leak can escalate to a blowout in a relatively short time. We are also able to recognize this in Fig. 19, which is based on a high-temperature leakage test. According to our model, we must expect leakage in our flange connection as soon as the local maximum surface pressure Pn on the gasket reaches the level of internal pressure Pi. It is to be expected in the case of an ideal flange (high rigidity, no flange rotation) that the loss in surface pressure is more rapid at the outer edge of the gasket than at the inner edge because of the greater reduction in pressure there. This is accelerated even more by oxidative effects of the ambient air. This also corresponds to experiences from laboratory tests. These conditions of course apply only if the enclosed medium is not aggressive chemically and/or is not a good solvent. In the case of strong flange rotation, however, these relations may become reversed because of the predominant surface pressure conditions in that case, so the decomposition of the sealing material first takes place at the inner edge of the gasket. The higher temperature is also predominant here under realistic conditions. Of course, our model does not take any other influences on the flange into consideration. As simple and perhaps as insufficient as

it is, however, it does explain various everyday observations. We now also understand why the old-timers preferred to select highly loadable gaskets, although they had no, or only a few, testing technique methods available. With our model, it also becomes clear why it made sense in the older practice to retighten the bolts of a flange connection after some operating time. As was already mentioned, however, this should take place in the case of nonasbestos fiber gaskets only if controlled and only in consultation with the gasket manufacturers. 6. Conclusions Many investigations have been conducted in the last 15 years that, at least in the United States, have led to proposals for new design rules on the basis of the ROTT test process. One of the discussion participants at the last 1996 FSA Conference in Houston, Texas, raised the objection: Practitioners are not interested in these design rules for determining minimum installation surface pressures, but only the maximum permissible surface pressures. The answer to this objection, which is correct, depends on the standpoint of the user. We have attempted to find at least a partial answer in our discussion. We have also observed how

Page 33

difficult the relationships become as soon as we attempt to give consideration also to the most varied influences, including the human factor. The selection criteria necessary for the applicability of a sealing material, from the manufacturers' standpoint, must include the following two criteria: 1. The minimum surface pressure in the installation state, according to which process it is determined 2. The maximum permissible surface pressure in the operational state This was similarly proposed by Ernst Sauter (see earlier and Fig. 16). In addition, however, it is also necessary to take long-term behavior into consideration, as suggested in Section 5. References 1. Sauter, Ernst, and Demel, O. The performance of static gaskets. A new method of characterization. First International Symposium on Fluid Sealing, Nantes (France), 1986, pp. 159166. 2. ASTM Draft No. 9, Standard Test Method of Gasket Constants for Bolted Joint Design. 3. DIN 28090, Deutsches Institut für Normung e. V., Beuth Verlag GmbH, 10772 Berlin. 4. Latte, Jörg, and Rossi, Claudio. Creep and leakage properties of modified PTFE. Sixth Annual Technical Symposium of the Fluid

Sealing Association 1996, Fourth International Symposium of Fluid Sealing of Static Gasketed Joints 1996, pp. 211255. 5. Marchand, Luc, and Derenne, Michel. Long-term performance of elastomeric sheet gasket materials subjected to temperature exposure. CPVT-96 Conference, Montreal. 6. Latte, Jörg, and Rossi, Claudio. High-temperature behavior of compressed fiber materials and an alternative approach of life expectancy statements. Fifth Annual Technical Symposium of the Fluid Sealing Association, 1994, pp. 171186.

Page 35

2 Internal Combustion Engine Gaskets DANIEL E. CZERNIK Fel-Pro Inc., Skokie, Illinois Gaskets for the internal combustion engine appear simple but in reality are highly engineered products. A gasket design is selected for use only after extensive functional testing. Engine dynamometer testing, usually at overrated power outputs and thermal shock conditions, along with long-term field testing, are almost always conducted before the gaskets are adopted for use. Prior to the engine manufacturer's expending the costs involved in this testing, the gasket manufacturer must show that engineering analysis and testing support the expenditure. What is gasket? Webster's New Collegiate Dictionary defines gasket as Plaited hemp or tallowed rope for packing pistons, making pipe joints, etc.; hence, packing or any other suitable material. The American Heritage Dictionary defines gasket as Any of a wide variety of seals or packings used between matched machine parts or around pipe joints to prevent the escape of a gas or fluid. The American Society for Testing and Materials (ASTM) defines gasket as A material, which may be clamped between faces, and acts as a static seal. Gaskets may be cut, formed or molded to the desired configuration. Another definition, and one this author prefers, is: A gasket is a material or combination of materials clamped between two separable members of a mechanical joint. Its function is to effect a seal between the

members (flanges) and maintain the seal

Page 36

for a prolonged period of time. The gasket must be: capable of sealing the mating surfaces, impervious and resistant to the medium being sealed, and able to withstand the application temperature and pressure. Figure 1 depicts a gasketed joint and presents its associated nomenclature. 1. History of Gaskets Gaskets have been utilized from the advent of the industrial revolution through the steam engine period and from the beginning of the evolution of the internal combustion engine. From the advent of the internal combustion engine, almost every material imaginable, from extremely hard to extremely soft, has been (and still is being) utilized for gaskets, including leather, paper, metal, cork, rubber, sponge, and plastic. An interesting sidenote is what used to happen at the old Brickyardthe Memorial Day Indy 500 races: Throughout most of the 1920s and 1930s, one of the most frequent excuses for the failure of race cars to finish was gasket failure, particularly of head gaskets. In addition, excessive oil on the track due to leaking gaskets caused accidents that led cars not to finish. After a couple of hundred miles, the track became so slippery from the oil leakage that drivers felt they were driving on ice. Sand was applied generously, particularly around the corners. This resulted in a false sense of security, and the accidents continued. Engine gasketing today is extremely high tech, with passenger car engines going more than 150,000 miles without significant leaks. Heavy-duty diesel

Figure 1 Nomenclature of a gasketed joint.

Page 37

engines have been known to go 1 million miles or more before any repair service was required. One of the first materials used for gaskets was cork, which comes from the bark of cork oak trees. Cork granules have been combined with a variety of polymers, resulting in the formerly popular corkglycerin product and various other cork-rubber composites. Throughout the past 90 years, paper has been another product utilized for gasketing. Fibers were generally cellulose, and their product was made on a paper machine. To improve the sealing properties of cellulose papers, it was common to saturate them with glycerin and animal glue and then to treat them with formaldehyde and dry them in an oven. The product, known as treated fiber, is still extensively used for sealing water and fuel. During World War I, the Germans controlled the glycerin market, so it was unavailable in the United States. As a result, some of the U.S. manufacturers saturated the base gasket with honey to improve sealability. Then rubber latexes were developed to saturate the sheets to make them even more impervious. This process was an offshoot of the shoe sole and liner technology that was used especially with military shoes for prevention of mildew. The product widely used in the early stages of the automotive industry was compressed sheet. It has a very dense structure made under calendering rolls. This material were first made by the Klinger Company in the 1890s. Their first product was called Klingerit. When a material's name ends in it, this signifies that it is a compressed product. The original fibers used were asbestos; hence the material was called compressed asbestos fiber, or CAF or CA for short.

Immediately after World War II, the manufacturing of paper gaskets went through a major revolution with the introduction of the beater-add (BA) process. In this process, elastomers, fibers, and fillers are mixed with water in a slurry. The slurry is then deposited on a conveyor belt and the water drawn off. After heating, the BA facing is ready to be made into gaskets. The BA rubbers include nitriles, styrenes, acrylics, fluorocarbons, and others. Blends of the rubbers are common. Subsequently, composite gaskets incorporating beater-add facing materials and either a perforated or a solid steel reinforcing core became the dominant products for sealing various engine gasketing applications. Metal gaskets made of steel, copper, and aluminum have been used for a number of engine applications, including head, intake, and exhaust manifold applications. The gaskets can be embossed or corrugated and coated with various polymeric materials. The original head gaskets were copper-clad constructions with a compressible center layer. Rubber, by itself, is used in various forms, either die cut, molded, or extruded for different gasketing applications. Plastics are used in gaskets either as a carrier or in some cases as a gasket itself.

Page 38

During World War II, the U.S. armed forces had a need for standardization of gasket materials because they were experiencing procurement problems. This resulted in a request that the industry take some action. Early in 1944, a committee was formed under the joint sponsorship of the Society of Automotive Engineers (SAE) and the American Society for Testing and Materials (ASTM). The committee functioned as a section of a technical committee on automotive rubbers. Their first order of business was to develop a numbering system for nonmetallic gasket materials. The first ASTM system had a G prefix, and its various subdivisions identified various physical properties. This later developed into a new nonmetallic gasket materials standard for automotive and aeronautical uses that had a prefix P, which first appeared in 1958. One can still find G&P prefix standards today. Since it appeared that the new standard would be more readily fitted with ASTM, the committee, with the concurrence of SAE and ASTM, established ASTM Committee F3 in 1962. Later it became obvious that the P numbering system was not serving the industry, and a new classification system of ASTM D-2000 and SAE J2000 for rubbers was developed. In addition, ASTM F104 and SAE J90B, the standard classification systems for nonmetallic gasket materials, were established. Revisions to these have occurred over the years, and a number of standards are now in existence. 2. Requirements of Gaskets In theory, if the flanges of a mechanical joint were perfectly smooth, parallel, and infinitely rigid, one could bolt them together and seal without a gasket. But in practice, flanges have rough

surface finishes and limited rigidity. In addition, flange loading is often nonuniform across the flange surface. Gaskets, therefore, are introduced in the joint to maintain sealing by: 1. Compensating for the nonuniform flange loading and flange distortion 2. Conforming to flange surface irregularities Engine head gasketing is particularly indicative of the nonuniform flange loading. In addition to sealing and compensating for the mating flange distortions, there are other requirements of engine gaskets: Possess heat and media resistance Have zero leakage through the gasket Have zero leakage over the gasket Be affordable Be environmentally safe Accommodate surface finish conditions of flanges Reduce and/or control port distortion Reduce and/or control flange distortion

Page 39

Accommodate thermal expansion and contraction Possess adequate recovery Minimize torque loss Not require retorquing Transfer heat as desired Possess close tolerance on compressed thickness (maintain shim thickness control) Meter fluid Provide acoustic or thermal isolation Possess antistick properties Pass customer verification testing Possess manufacturability/handlingrobotics, cleanliness, block, visibility for sensors, fit Accommodate service assembly requirements Be recyclable There are literally dozens of commercially available materials that will seal, more or less efficiently, alone or in combination. It is the task of the gasket engineer to qualify and catalog each material for its sealing efficiency in various operational environments, combining materials where necessary to take maximum advantage of the best properties of each. In order to classify materials, the gasket engineer must determine their physical and mechanical properties and analyze the

environment in which the gasket will be placed. Environmental factors such as mating flange thickness and surface finish, along with bolt size, torque, spacing, and length, all are important to gasket performance. A matching of the gasket to the environment requires extensive bench testing and, in many cases, actual engine testing, both on dynamometers and in the field. The gasket may appear simple; in reality, it is a highly engineered product. Gasket sealing requirements vary widely. An example of a leakage rating chart is given in Table 1. Figure 2 depicts the various applications where gaskets are used on a typical Vee engine. Gaskets for engines are available individually or as part of a gasket set. Figure 3 presents some of the sets associated with a typical four-cylinder inline engine. 3. Standard Classification System for Nonmetallic Gasket Materials The American Society of Testing Materials (ASTM) F104 classification system provides a means for specifying or describing pertinent properties of commercial

Page 40

nonmetallic gasket materials. Materials composed of asbestos, cork, cellulose, and other organic or inorganic materials in combination with various binders or impregnants are included. It should be noted that although asbestos is listed in the ASTM classification system, it has been eliminated from essentially all gasketing materials. Rubber compounds are not included, since they are covered in ASTM Method D2000 (SAE J2000). Also, gasket coatings are not covered, since coating details and specifications are intended to be listed on engineering drawings or in separate documents. The ASTM classification is based on the principle that nonmetallic gasket materials can be described in terms of specific physical and mechanical characteristics. Thus, users of gasket materials can, by selecting different combinations of statements, specify different

combinations of properties desired in their gaskets. Suppliers, likewise, can report properties available in their products in accordance with this classification system. In specifying or describing gasket materials, each line callout must include the number of this system (minus date symbol) followed by the letter F and six

Page 41

Figure 2 Vee engine gasket locations.

numerals, for example, ASTM F104 (F125400). Each numeral of the callout represents a characteristic, and six numerals are always required. The numeral 0 is used when the description of any characteristic is not desired. The numeral 9 is used when the description of any characteristic (or related test) is specified by some supplement to this classification system, such as notes on engineering drawings. Table 2 presents the ASTM F104 guidelines. 4. Important Considerations for Fabricating Nonmetallic Gaskets The following is an overview of what is important in nonmetallic

gaskets from the fabricator's standpoint. Identified are some important sheet properties for

Page 42

Figure 3 Four-cylinder engine gasket locations.

Page 43

gasketing purposes as well as for processing and installation needs. Their significances in gasket applications are identified. Properties associated with the creating and maintaining of the seal are also identified. Important environmental conditions are noted. Presented are necessary sheet characteristics for quality assurance. Due to the nature of some applications, the nonmetallic gasket material by itself is inadequate for various reasons. Therefore, included in the paper are some sealing enhancements that are added to the material and/or gasket by fabricators. 4.1. Important Material Properties for Gasketing The following are properties of the gasket material that are important for sealing performance in the application. Compatibilitybe resistant to the media being sealed. Heat resistancewithstand the temperature of the environment. Compressibilityconform to the distortions and undulations of the mating flanges. Microconformabilityflow into the irregularities of the mating flanges' surface finishes. Recoveryfollow the motions of the flanges caused by thermal or mechanical forces. Creep relaxationretain sufficient stress for continued sealing over an extended period of time. Compressive Strengthresist crush and/or extrusion caused by high stresses.

Erosion resistanceaccommodate fluid impingement in cases where the gasket is required to act as a metering device. Tensile strengthresist blowout due to the pressure of the media. Shear strengthhandle the shear motion of the mating flanges due to thermal and mechanical effects of the mating flanges. Z strengthresult in used gasket removal without internal fracture of the material. Antistickensure gasket removal without sticking. Heat conductivitypermit the desired heat transfer of the application. Acoustic isolationprovide the required noise isolation of the application. 4.2. Important General Sheet Characteristics To ensure that there will be no health problems for the fabricator and/or end user, the material must be nontoxic. This is a most important characteristic of the material. Another important characteristic of the gasket, of course, is sealability. This is not only dependent on the sheet material properties but also a function of environmental conditions, such as clamp load, bolt span, and flange rigidity.

TABLE 2 Basic Physical and Mechanical Characteristics for Nonmetallic Gasket Material Numbering code for basic characteristi Meaning of each numeral in the basic six-digit 0 1 2 3 number First number: type of NSAsbestos Cork Cellulose material Second number: class of Compressed Paper and material when first NS sheeter Beater process millboard number is 1 or 7 process Cork Cork and Cork and cellular when first number is 2 NS composition elastomer rubber Untreated Proteintreated Elastomerictreated when first number is 3 fiber fiber fiber PTFE with PTFE filaments when first number is 4 NSSheet PTFE expanded braided or woven structure Homogeneous Laminated when first number is 5 NS __ sheet sheet

(table continued on next page)

(table continued from previous page) Meaning of each numeral in the basic six-digit number Third number: compressibility-% compression loss (ASTM F36)

Numbering co 0 NS

Fourth number: % thickness increase when immersed in ASTM #3 oil NS (ASTM F146) Fifth number: maximum % weight increase when immersed in ASTM #3 NS oil (ASTM F146) Sixth number: maximum % weight increase when immersed in water NS (ASTM F146) NS-not specified. AS-as specified.

Page 46

Sheet materials made by the beater addition process contain voids. These sheets are made by adding a rubber binder to a water slurry of fibers and materials. The water is then drawn out on the papermaking machine, resulting in voids in the material. These voids must be closed for the gasket to seal. (Some enhancements added to close these voids are identified later.) There is another sheet material characteristic that is not necessarily application important but is required for general sealing performance: uniformity of the material, that is, consistency of formulation thickness, density, and surface finish. Consistency of formulation is vital for compatibility of the material with the media being sealed, as well as for other reasons. Many gaskets have precise compressed-thickness requirements, since they function as shims in addition to being gaskets. Thickness tolerance and density uniformity from lot to lot are important for compressed-thickness control. Uniformity of density is also very critical for some of the fabricators' added enhancements, such as saturating, laminating, and surface coating. Surface finish similarity within a lot and from lot to lot is important for surface sealing. For long-term sealing performance, the gasket fabricator desires the material to have little change in load from initial clamp up to long time operation. The load loss is called relaxation. The reason for this can be seen in Fig. 4. The reduction of the sealing stress on the gasket from point G to point K is due to the relaxation of the gasket.

4.3. Important Material Characteristics for Processing and/or Assembly There are a number of sheet material characteristics important to the fabricator for processing sheets into gaskets and/or for the end user who assembles the gasketed joint: amounts of dust and slivers, tool wear and life, scuff resistance, breaking strength, and handling characteristics. 4.3.1. Dust and Sliver Amounts During blanking of the sheet material, lack of dust is desired, since dust causes many problems during processing. Dust gets into bearings and other machine components, resulting in premature repair, downtime, and the like. In addition, dust on the gasket results in adherence difficulties for subsequent operations such as coating and printing. Since the conversion to nonasbestos, dust has become a major problem with gasket materials. The replacement of asbestos fibers in the sheets with inorganics such as slate and various fibers has resulted in dust problems. Dust is also a major complaint of workers.

Page 47

Figure 4 Gasketing joint diagram.

Page 48

Slivers are also characteristic of nonasbestos materials. Slivers are little pieces of the material that result during blanking. They can cause dents in the gasket, resulting in potential leak paths. Sliverfree gaskets are highly desirable for gasket fabricators because manual sliver removal is costly. 4.3.2. Tool Wear and Life The conversion of the asbestos-based sheets to nonasbestos-based sheets has resulted in significant increase in tool wear. Reports of decreased tool life abound. Some fabricators indicate a reduction of life by factors of 1/3 to 1/20. The exact reasons for the increased wear are not known, but the aramid fibers and/or some of the organics used are thought to be possible contributors. Many gasket manufacturers are conducting investigations of tool wear. In addition, the Gasket Manufacturers Association has a study project under way. 4.3.3. Scuff Resistance The sheet material's ability to resist scuffing is important to the fabricator. The rough handling and transportation of materials during fabrication can result in scuffing of the surfaces of the gaskets, thus rendering their sealing performance questionable. 4.3.4. Breaking Strength The breaking strength of the materials must be sufficient to resist fracture during processing. Some of the processing operations exert considerable tensile pull on the material. 4.3.5. Handling Characteristics The handling characteristics of the gasket materials are important

during processing, and these same characteristics are also important during assembly of the gasket. Rigidity of the gasket, for example, may be important for ease of assembly and ensuring proper installation. In some cases, the gasket may be installed by robots, and this type of handling needs to be taken into account. 4.4. Sealing Enhancements All of the just-discussed gasket fabricators' desired properties and characteristics of the base material may not be possible of being contained in the sheet. Thus, some fabricator additions may be required for various applications. Often, a gasket's sealing requirements are such that the produced gasket sheet material cannot accommodate them. In these cases, the gasket fabricator relies on any number of sealing aids to improve the gasket and meet these requirements, including saturation, coating, combining and/or laminating, embossing, eyeletting, printing, and grommeting.

Page 49

4.4.1. Saturation The voids in the gasket material can be filled with a number of chemical resins. A saturant, in addition to filling the voids, can improve the heat and chemical resistance of the gasket. It also can alter the physical properties of the material. 4.4.2. Coating Coatings are applied to the gasket for a variety of reasons, including the following. Improve surface seal. Improve antifret characteristics. Antifret is reduction of scrubbing of the gasket due to flange shearing motion. Reduce or eliminate sticking upon removal. Provide sticking or tack for ease of assembly. Provide the gasket with a barrier coat for subsequent printing (printing is defined later in Section 4.4.6). Reduce sticking of gasket to gasket during processing and shipping. Provide color to the gasket for identification reasons. 4.4.3. Combining and/or Laminating The gasket material can be mechanically combined to a perforated core or chemically laminated to an unbroken rigid core material such as metal. Such reinforcement improves the mechanical properties of the gasket material. Perforated core and laminated unbroken core are depicted in Figs. 5 and 6 respectively.

Figure 5 Perforated core.

Figure 6 Unbroken metal core.

Page 50

Figure 7 Embossed unbroken and perforated cores.

4.4.4. Embossing After a gasket material has been reinforced by combining and/or laminating, it can be embossed. This results in high stresses at the embossments for improved sealing purposes. Embossed gaskets are depicted in Fig. 7. 4.4.5. Eyeletting Metal eyelets are used at port openings to: (1) protect the gasket material from the sealed media, and/or (2) provide high sealing stress on the eyeleted locations. Eyelets can also be used at bolt holes to reduce distortion of the gasketed joint. A metal eyelet is depicted in Fig. 8. 4.4.6. Printing A common sealing enhancement used today is printing of the gasket. This is usually done by a version of the silk screening process. In this process, elastomeric

Figure 8 Metal eyelet.

Page 51

beads are deposited on the gasket in strategic locations. This results in improved sealing at these locations due to the higher resulting stress of the rubberlike sealing interface with the mating flange. Figure 9 shows a printed gasket. In addition the higher recovery properties, the beads also improve long-term sealing. Various elastomers are used, with silicone being the most popular. Rigid beads, such as those that result when rigid epoxies are deposited, are sometimes incorporated in gaskets for compression limiting purposes or for reduction in joint distortion. Figure 10 depicts a rigid bead used for compression limiting. Use of the printing technique has virtually changed the look of the engine gasketing industry. While the printing patents were in existence, there were more than 30 worldwide licensees of the technology.

Figure 9 Printed elastomeric bead.

Page 52

Figure 10 Printed elastomeric and rigid beads.

One extension of this technique is to deposit the bead or beads using a tracer or robot. This permits deposition of more material and increases the thickness possibilities for the beads. Another extension is to combine it with embossing: An elastomeric bead is deposited into the emboss, generally by means of a robot. This combination can be likened to a trapped O-ring. The end product has extensive recovery characteristics. A variation of the printing process is called MIP, for mold in place. An elastomeric bead is molded to the gasket material or gasket. This provides for precision thickness variation of the bead, resulting in exacting threedimensional gaskets. Liquid injection-molded silicone is sometimes used for this. 4.4.7. Grommeting Grommets in the gasket industry are rubber parts and/or rubber parts reinforced with metal or plastic. They are molded products and are added to the gasket to provide improved sealing at difficultto-seal locations. Their cross sections are virtually unlimited, and therefore permit a large range of design possibilities. Figure 11 shows various grommets. Many materials are used for grommets. Most common are nitrile, neoprene, polyacrylic, silicone, and

fluorocarbon. 4.5. Stress Distribution Testing In order for a gasket to provide sealability, it must distribute stress adequately and properly. There are a number of methods that can and have been used to measure stress distribution in a specific gasketed joint: regular carbon paper, carbonless (NCR) paper, stress-sensitive film, and electrical stress sensors. 4.5.1. Regular Carbon Paper Regular carbon paper once extensively used by secretaries in typing, is a two-piece system. One sheet is the carbon carrier; the other is a clean sheet of paper required for carbon transfer. Stress impressions created by this technique provide

Page 53

Figure 11 Rubber grommets and gasket bodies.

on-off (yes-no) visual effect. That is, sufficient stress either is or is not available to transfer the carbon from the carbon paper to the clean sheet. This provides a very narrow range of stress information. In some instances, multiple layers of carbon paper were used for better measuring of the stresses. In such cases, compressibility of the paper stack had to be taken into account, particularly with gaskets that possessed little inherent compressibility. Because of these shortcomings and drawbacks, this method is rarely used. 4.5.2. Carbonless, or No-Carbon-Required (NCR), Paper Carbonless, or no-carbon-required (NCR), paper is available as a single sheet having pressure-sensitive chemicals within the paper. Stress applied to the paper results in a crushing of the encapsulated chemicals that produces color on the paper. NCR paper is an improvement over regular carbon paper because it involves a single sheet and the impression density or color intensity is proportional

to the stress applied. Therefore, elemental comparative impressions at various stress levels, performed on a load machine such as an Instron or MTS, can be used to calibrate the paper-impression color density versus stress level. The paper is available in only one pressure range, which limits the accuracy of the quantified the data collected. 4.5.3. Stress-Sensitive Film Stress-sensitive film, manufactured by Fuji Photo Film Co., Ltd., is available in one- or two-sheet systems and in four pressure ranges. It functions in a manner

Page 54

similar to carbonless paper but is an improvement over carbonless paper in that it permits fine-tuning of the stress distribution. Also, by means of a commercially available densitometer, impression color density can be converted directly to stress readings. Since the film is affected by time, temperature, and humidity, these conditions must be taken into account when analyzing the color intensities. Again, calibration of these films can be accomplished by conducting elemental stress compression-color intensity testing. 4.5.4. Stress Distribution Testing Technique To evaluate stress distribution within a sealing joint, the regular carbon paper, NCR paper, or Fuji film is precut to the shape of the mating flanges, and the holes for the fasteners are punched out. The paper or film is then placed between the flanges and the fasteners, tightened to the specified torque. When the paper is removed from the actual gasketed joint, it shows the stress distribution pattern on the gasket. Figure 12 depicts this. One deficiency shared by all of these three methods is that they indicate a one-time maximum stress during the course of testing. Any reduction in stress during the torquing sequence or from external forces is not reflected in the impressions. 4.5.5. Electrical Stress Sensors A new technology has been invented for determining stress distribution: electrical stress sensors. This system represents a quantum leap forward. It utilizes electrical sensors that measure the stress distribution in real time. That is, stress can be measured during the loading and unloading modes. In some cases, the unloading due to hydrostatic end forces can be determined.

Figure 12 Stress distribution on a gasket.

Page 55

Figure 13 Makeup of electrical sensor.

The sensor is approximately 0.1 mm thick. It is made up of two flexible, thin, mylarlike sheets that have electrical strip patterns (electrodes) formed on them. The inside surface of one sheet has a row pattern, while the inside surface of the other sheet has a column pattern. Therefore, when the two sheets are placed on top of one another, the intersection of rows and columns forms a grid pattern with electrical point junctions. Before assembly, a thin resistive/conductive coating (ink) is applied as an intermediate layer between the two electrical contacts (rows and columns). This ink provides an electrical resistance/conductance between these contacts. Figure 13 depicts the sensor makeup. Of major importance is that the ink's electrical resistance changes with applied external force. Therefore, when installed on a gasket and clamped within the sealing joint, the sensor provides an array of electrical conductive points.

Figure 14 Component diagram of electrical system.

Page 56

Figure 15 Output of electrical system on monitor.

Each of these points is associated with its own discrete electrical resistance/conductance. This resistance can be then related to force. By observing the minute changes in current flow at each crossing point, the force distribution pattern can be measured and displayed on a computer screen. This pattern shows the location and magnitude of the forces exerted on the surface of the sensor at each specific point, and a force profile is created. Figure 14 shows the component diagram of the system. Software has been developed that permits two- and three-dimensional color force displays projected on a monitor to show the relative force at each intersection. Figure 15 shows this. Newly created hardware and software, along with a specialized electrical contact, are used in conjunction with a personal computer to display the output of the sensor. A color bar chart is displayed that depicts the relative amount of force. Designed into the

software is a windowing feature that provides the capability for evaluating localized area stress. In addition, current technology includes a sum-force scale that gives relative differences in load levels at various points. Various types of universal and customized sensors are utilized with this measuring system. This technology is patented in a number of countries, including the United States. 5. Engine Gasket Applications Let us now investigate some of the various engine gasket applications and examine each in detail.

Page 57

5.1. Cylinder Head Gaskets The cylinder head gasket is the most critical sealing application on any engine. Typically, it must simultaneously seal: (a) high combustion pressures and temperatures; (b) mixture of water and anti freeze, with its high wicking and wetting characteristics; and (c) lubricating oil, with its associated detergents, additives, and variable viscosities either built in or changed with the season. In addition, the head is a structural component of the engine; i.e., the combustion chamber is formed by the head, block, piston, piston ring, and gasket. The gasket shares the same strength requirements as the other combustion chamber components. The head gasket is used many times either to meter or to block coolant flow for proper cooling of the engine. It also seals the block-liner intersection in wet liner engines. Its compressed thickness affects the compression ratio of the engine, and the importance of compression ratio control to emission levels, especially in diesel engines, is well known. Today's engine manufacturers require that the head gasket perform without a retorque operation, seal for extensive periods of time, and come off clean so no scraping of the mating flanges is necessary when the engine is repaired. The head gasket sometimes needs very high thermal conductivity to transfer heat efficiently between the block and the head. It must be constructed so as to permit rough handling and extended storage life. The gasket must also perform in temperature ranges well below freezing at start-up to over 700 degrees Fahrenheit in the combustion seal area during engine operation. It must accept occasional instances of detonation without failure. This is especially true today, when premium fuels are

sometimes unavailable and detonation associated with regular or no-lead gas occurs. The gasket must typically withstand combustion pressures of 1000 psi in naturally aspirated spark-ignition engines and 3000 psi or higher in turbocharged diesel applications. Today's gasket must also accommodate greater motions, both thermal and mechanical, because lighter-weight castings and lighter-weight, less rigid materials are being utilized for cylinder heads and engine blocks. As a result of all these requirements, the head gasket is a very complex product. For a head gasket to seal properly, the head bolts must apply a sufficient clamping force on the gasket. As the bolts are tightened, the gasket is squeezed to provide a seal between the cylinder head and the engine block deck. All gaskets relax a certain amount. This is true even with no-retorque gaskets. The ideal no-retorque gasket design compresses sufficiantly at installation to conform to and seal minor surface imperfections. At the same time, the design minimizes relaxation and maintains adequate clamping force over a long period of time. By contrast, a retorque design will relax excessively. This reduces the tension on the bolts and results in excessive torque loss. If one doesn't retorque this type of gasket, the engine could lose compression, and fluid or combustion gas leaks will result. Leaking combustion gases can damage the gasket surface

Page 58

or cause the blowout of an entire gasket section. Loss of coolant could result in engine overheating, leading to engine damage. In addition, a retorque operation is costly. Any technician knows how time-consuming engine work has become. Tighter engine compartments and emission and electronic controls make many engines extremely difficult to work on. With a retorque head gasket, the head bolts typically need to be retorqued after the engine has been warmed and again after several hundred miles. The time needed to gain access to the head bolts for retorquing will quickly reduce the profit the technician makes on the job. Also, the vehicle's owner is inconvenienced because the vehicle needs to be returned to the facility for retorquing. Gasketing of the internal combustion engine has largely been one of reaction, in that the sophistication of the gaskets has mirrored the sophistication of the engine itself. That is, as engines became more powerful and placed greater demands on the gasketing, the gasket industry responded by designing the required sophistication into the gaskets. Table 3 shows the sophistication of engines and TABLE 3 Head Gasket Designs Versus Engine Sophistication ZoneEngine sophistication Head gaskets Sandwich construction containing Low output engine and low A copper-asbestos-copper and gasket compression ratios shellac; also, resin dipped asbestos Additional cylinders added Steel-asbestos-copper, Steel used for more power in-line B in combustion area for increased eights and twelves slight life increase in compression ratio Embossed steel with pre-applied C New Vee and OHV engines coatings, various sealing aids in sandwich gaskets

D

E

F

High performance, large bore engine and automatic transmissions. No retorque of cylinder head bolts Use of lighter weight alloys and castings and higher specific output CAFE and clean air requirements

Development of bitusmastic and rubber-fiber beater sheets used with tanged core. Improved combustion seal design Development of laminated bodies and use of higher temperature resistant materials. Improved antistick and anti-fret coatings. Use of graphite, molded rubber More of zone E plus double and steel and multi-layer steel, over head cam, more valves rubber coated stainless steel and turbocharging materials

Page 59

Figure 16 Sandwich gasket.

head gaskets versus time. The various time zones shown in this figure are described next. 5.1.1. Head GasketsTime Zones 19101930 Zone A. Initially, the basic head gaskets for engines were of the sandwich type, with asbestos millboard center and either tin-plate or copper on the outer surfaces (Fig. 16). Grommets and/or eyelets were incorporated in these gaskets, depending on the specific engine needs. Numerous versions were designed and manufactured. 19301950 Zone B. As engines gained sophistication, gaskets also gained in sophistication, and a variety of designs were produced. These designs utilized various reinforcements at the combustion chamber seal for improved sealing. Metal shims and reinforced filler materials, for example, were incorporated into many constructions (Fig 17.).

Figure 17 Sandwich gasket with reinforcements.

Page 60

Figure 18 Embossed steel shim.

19501980 Zone C. The embossed steel shim gasket was the next popular gasket to be utilized on passenger car engines (Fig. 18). This gasket had a plastic resin coating applied for microsealing purposes. Because it was all metal, good torque retention was inherent with this gasket. However, as engine displacement increased, the output resulted in motions that normally could not be accommodated by the elastic response of the embossed design. In addition, many times the land areas, especially between cylinders, were reduced to a point where the legs of the emboss would fall inside the ports, making adequate sealing impossible. 19651975 Zone D. With the development of rubber-fiber facing materials by the gasket paper manufacturers came improved designs. The majority of these designs incorporated a tanged or perforated-core steel sheet, with these new facing materials mechanically clinched to either side of the core, thus providing soft surfaces of sealing material for water and oil sealing (Fig. 19). One of the major require-

Figure 19 Perforated-core head gasket.

Page 61

ments of gaskets in this time zone was that they function without need for a retorque operation on the cylinder head bolts. Retorquing still is specified on some foreign-made engines but is essentially nonexistent in today's American-made engines. While this may not appear a major requirement, it is indeed major, since the retorque operation greatly aids gasket performance. 1975Present Zone E. Graphite facings became popular in this time zone. Graphite has good sealability and relaxation properties and high heat resistance. One of the newer constructions eliminates the perforated core and uses an unbroken steel core to which an adhesive is applied for bonding the facings (Fig. 20). This laminated gasket has been adopted on many of the more difficult sealing applications. This type of gasket body can be embossed to achieve higher sealing stress at particular passageways. Another technique used to seal critical passageways on today's engines involves silk screening to print elastomeric beads at these locations. (Printing was discussed earlier in Section 4.4.6) In addition, many improvements in seal, antistick, and antifret coating have been incorporated in the latest gasket constructions. Currently, a new array of lightweight, high-output gasoline and diesel engines, both naturally aspirated and turbocharged is being developed. The gasket industry is involved in extensive R&D programs designed to seal these new families of engines. Time will show even newer concepts being utilized in the head gaskets for these engines.

A new lineup of gaskets incorporating multiple layers of embossed, springtemper stainless steel with rubber coatings has become popular (see Fig. 21). These gaskets exhibit little change during engine operation. Use of spring-temper stainless for embossed layers results in high elastic recoveries.

Figure 20 Unbroken-metal-core head gasket.

Page 62

Figure 21 Multilayer-steel head gasket.

5.1.2 Head GasketsCombustion Sealing In most cases a tin-plate or zinc-plate armor is used for sealing the combustion gases of spark ignition engines (Fig. 22). The thickness of the armor is a function of the thicknesses and type of facing material. The overlap and heel are sized for the specific engine, to establish a proper unit seal load at the combustion chamber. The heel may be sized differently at various positions around the combustion chamber in order to obtain the proper unit loading at these positions (Fig. 23). High-output engines and/or turbocharged engines normally require stainless steel armor for improved hightemperature and fatigue resistance. Types 430, 304, and 321 stainless steel are commonly used. In the case of diesel applications, an armored gasket is not generally adequate. Only armors of a certain thickness can be formed and imbedded into given gasket bodies, and the thicknesses that normally fill these requirements are not structurally sufficient to withstand the high combustion pressures of most

Figure 22 Armored head gasket.

Page 63

Figure 23 Variable-width heel.

diesel engines. As a result, other means for sealing combustion are necessary for these applications. The most popular method incorporates a low-carbon-steel ring. This ring gives a high unit sealing stress at very low loading and is widely used in today's diesel engines. The wire is buttwelded and generally attached to the gasket body by means of a stainless steel armor wrapping. (Fig. 24). In some

Figure 24

Wire-ring head gasket.

Page 64

cases, the wrapping may be tabbed to reduce the load required to imbed the armor into the body, thus increasing the loading on the wire ring. In some cases, stainless steel wires are needed to withstand the heat and fatigue characteristics inherent in particular engines. An example is the case where precombustion chambers experience high thermal and mechanical movements. There are some gaskets where more than one wire may be utilized to achieve the desired sealing requirements of the engine. Armored embossed metal is also used to seal combustion in a number of engines (Fig. 25). Varying the height and/or the width of the emboss results in a wide range of load-compression properties. When the embossing is made from the core of the gasket body, variation in thickness tolerance are minimized, since the emboss and the core are made from the same piece of metal. Stainless or low-carbon steel are used as armors. Diesel engines frequently have wet liners, and the gasket is usually charged with sealing the intersection of the liner and the crankcase (Fig. 26). During engine operation, there is motion between the liner and the block, and the likelihood of erosion of the liner seat is high. As a result, coolant can leak to the top of the deck; the gasket is required to seal at this location. In some cases, the soft surface is used to seal; in other cases, the heel of the armor is extended to cover the intersection for sealing purposes. Engine testing normally dictates which is better. Some manufacturers are using roomtemperature vulcanizing (RTV) silicone to seal this application. There are other engines that have liner designs that incorporate ridges and, in some cases, grooves in the cylinder head. This results

in coining, or imbedding, of the gasket for improved combustion gas sealing. Figure 27 shows one of these types. Another head gasket design used on ridged liner engines is a thick (0.080'') steel plate that is embossed for improved combustion sealing (Fig. 28). A unique gasket, which is used to seal very large diesel engines, uses copper-clad steel that has been etched away at various locations. The etching

Figure 25 Armored embossed metal.

Page 65

Figure 26 Head gasket on wet-liner diesel engine.

removes the copper from specific areas, thereby permitting high unit loading at other locations for improved combustion gas sealing (Fig. 29). One of the items to be considered in the design of combustion sealing is bore distortion. Some of the designs may need supplementary aids to keep bore distortion within acceptable limits. A few of the techniques used for this include extending the combustion armor heel at specific locations (Fig. 30), overlapping of the heel around the gasket body, generally at the ends of the gasket, and depositing beads or areas of rigid materials at preselected points. All of these techniques essentially change the load transmitting characteristics of the gasket and are useful for minimizing head bending as well as reducing bore distortion. In some engines, the back-to-back location of exhaust valves results in high thermal growth in the area between cylinder bores. If excessive, this growth can result in combustion leakage. One

means of improving the sealing in this area is to incorporate a metal shim in the gasket at this location. The shim acts as a stopper, permitting the gasket to resist the thermal growth and enhance sealing. Air-cooled engines have somewhat reduced requirements regarding head gasket sealing. Since there are no cooling-water passageways, slight combustion gas leakage can be permitted as long as: (1) engine performance is not affected, and/or (2) the gasket is not affected by the leakage. Most of the gaskets for these engines consist of metal tanged core on both outer surfaces and a hightemperature-resistant fibrous core material (Fig. 31). Because these engines are made mainly from aluminum, high thermal motions occur. The metal surfaces

Page 66

Figure 27 Ridge-on-liner and groove-and-head head gasket.

Figure 28 Embossed-steel-plate head gasket.

Page 67

Figure 29 Etched cooper-clad head gasket.

Figure 30 Extended heel of armor.

Page 68

Figure 31 Air-cooled-engine head gasket.

of the gasket permit head and block motions to occur without serious effect to the gasket's sealing ability. Embossed metal gaskets are also used on these engines, especially when high heat transfer through the gasket is required. 5.1.3. Head GasketsLiquid Sealing The basic factor involved in the creation and maintaining of the liquid seal is to have sufficient sealing stress on the gasket to ensure conformation of the gasket to the flange surfaces. This results in blocking the passage of media between the gasket and the flanges. In addition, the stress must be high enough to close any voids in the base material. The stress, however, must be low enough not to result in extrusion of the base material. To ensure long-term sealing, the material must, of course, retain adequate stress. Therefore, the selection of the facing material and its thickness is critical. The base materials used for the gasket bodies must be extensively evaluated. The various ASTM test specifications for the materials' physical properties before and after fluid immersion and heat aging are reviewed. In addition, bench test results for sealability, creeprelaxation, crush and extrusion, etc., are analyzed before the

material is accepted for use in a head gasket. Today's popular head gaskets utilize soft surfaces for sealing the engine's liquids. These surfaces are rubber-fiber facings that are attached either mechanically and/or chemically to a metal core. The most popular facings are nitrile, neoprene, or polyacrylic elastomers. These are compounded to resist degradation by the oils and coolants, retain torque, minimize extrusion, and exhibit heat resistance. In addition, they must permit coolant infringement on the surfaces without degradation, since many of the gaskets are used to meter and/or block coolant flow in the engines. The mechanically clinched design uses perforated metal that has tangs on each side to which the facing is mechanically attached. They, in general, give adequate performance. At times, in critical sealing areas, the liquids tend to seep along the core tangs, resulting in some leakage problems. In addition, if the tangs

Page 69

penetrate the outer surfaces of the gasket, erosion and/or corrosion of the mating flanges may occur. This is particularly true in the case of neoprene-bound facing materials. Upon aging, the neoprene may release hydrochloric acid, causing corrosion and etching. The laminated or chemically bonded gasket bodies utilize an unbroken metal core to which the facings are bonded. Since there are no tangs, there is no possibility of leakage around the tangs or etching of the mating surfaces. The bonding adhesives, however, must be carefully compounded to accept the heat of the application and be resistant to extrusion. Because there are no tangs in these constructions, extrusion must be resisted by a high-strength, facingadhesive-core bond. As mentioned previously, these bodies may be embossed to achieve high unit sealing stress at various locations. In some cases metallic eyelets may be utilized at high-pressure openings to improve the sealing efficiency of the gasket at these locations. The laminated type of design lends itself to providing multiplethickness gaskets. The steel core is varied in thickness, while the amount of facing material is kept constant from gasket to gasket. This results in essentially equivalent torque retention properties even though the gaskets differ in thickness. The combustion seal is varied to compensate for the change in the steel core (Fig. 32). There are number of thermosetting seal and antistick coatings used on the gasket bodies. They provide microsealing properties to the gasket and eliminate sticking of the facing to the head or block when the engine is disassembled. Some of the European gasket manufacturers impregnate various gasket bodies with various resins for improved sealing properties. The impregnation is usually

associated with bitumastic bound sheets, which are rarely used in the United States today.

Figure 32 Multiple-thickness head gaskets.

Page 70

One important physical property of a gasket body is good torque retention. In general, torque retention is associated with the amount of compressible material in a gasket. Compression set is a function of the compressible amount, and reduction of the compressible amount results in higher retained torque of the fasteners. However, there must be sufficient compressibility consistent with good seal. Coatings such as Teflon* fill in surface irregularities in cylinder blocks and cylinder heads and permit the reduction of the amount of compressible material in a given gasket. Figure 33 shows the cold-flow properties of one type of Teflon coating. Figure 34 shows some of the motions that occur as an engine is operated. Bi-metal engines with aluminum heads and cast-iron blocks are being used on more and more vehicles because their lighter weight contributes to fuel efficiency. However, the aluminum head and cast-iron block expand and contract at different rates as the engine goes through heating/cooling cycles. This presents a special challenge in designing a head gasket because the different rates of expansion produce a shearing or scrubbing action between the head, the block, and the gasket. This shearing action can eventually cause a gasket to fail. Some gaskets incorporate antifriction coatings for bi-metal engines. Their use allows the head and block to slide over the gasket surface and still maintain a good seal. Teflon, moly disulfide, and graphite are examples of antifriction coatings. In cases of large head-to-block motion that may occur at specific liquid passageways, molded rubber, either with or without metal reinforcement, is sometimes incorporated in

Figure 33 Cold-flow properties of a Teflon coating. *Teflon is the registered trademark of the Du Pont Company.

Page 71

Figure 34 Motions in head gasket environment.

Page 72

the gasket for accommodation of the motion and effective sealing. These grommets are bonded and/or staked in place, depending on the gasket design (Fig. 35). As noted earlier, another technique utilized to improve the liquid sealing of the head gaskets is the printing or silk-screening of elastomeric beads at the liquid ports. When a thin-cored gasket body is used, the beads are usually located on one surface of the gasket, generally on the surface facing the weaker mating flange. The thin core allows transfer of the localized stress through the gasket to the opposite flange. The beads may be positioned on both sides of the gasket. One example is when the steel core is of substantial thickness and the transfer of a high unit load is impossible because of the thick core. There are a variety of materials that can be utilize in this technique, the most popular being silicone. The thickness of the bead is somewhat dependent on the nature, type, and thickness of the facing to which it is applied. Normally, as the facing material becomes thicker, a thicker bead is deposited. Some production gaskets, including head gaskets, incorporating the silk-screened beads are shown in Fig. 36. In-line engines many times present difficult liquid sealing problems. They are inherently unbalanced with regard to distribution of bolt loading. In many cases, no studs or cap screws are used to provide clamping pressure between head and block on the outer periphery of the push rod cavity. Supplementary

Figure 35 Grommets in various gasket bodies.

Page 73

Figure 36 Production gaskets incorporating elastomeric beads.

sealing means must be designed into the gasket; otherwise, oil seepage will occur down the side of the block. Figure 37 depicts two gasket design techniques used to solve this problem. They include: 1. Dipping the push rod area of the gasket to give this area a rubber overcoat 2. Applying a bead of high-temperature synthetic rubber to one or both faces of the gasket around the low-clamp-load area In each of these cases the intent is to obtain adequate clamp load at the push rod area by building up the gasket thickness by means of the rubber overcoat or the silk-screened bead. Open deck or open tank engines, which are becoming more popular because of their light weight, pose additional sealing demands on gasket bodies. These engine designs result in large water and steam impingement sections on the gasket. These areas are unsupported

and are undergoing pressure pulsations as the engine operates. Sealing the fluids, being resistant to degradation by them, and withstanding their corrosive attack are requirements of the gasket bodies on these engines. Providing small vent holes in the gasket for steam where possible is recommended on some applications.

Page 74

Figure 37 Head gasket thickness build-up at push rod area.

5.2. Casting Finishes A cylinder head or block can be resurfaced either too rough or too smooth for best gasket sealing purposes. A good range is between 60 and 125 microinches; 90 to 110 is preferred. Metal-faced gaskets, although excellent performers, are particularly sensitive to casting finishes either above or below this range. There is low resistance to combustion blowout with very smooth finishes, and poor combustion and fluid seal with rough finishes. Surface flatness is also important to good sealing. If the cylinder head is not flat enough, milling or grinding is necessary. 5.2.1. General Specifications for Flatness and Surface Finish The following are general specifications for the flatness and the surface finish covering the cylinder head gasket mating surfaces for both diesel and sparkignition engines. It should be understood that

bolt spacing, location of oil and water holes, pressures, etc., will vary from engine to engine and that there will be exceptions where closer tolerances might well have to be observed to seal an engine properly.

Page 75 Maximum out-of-flat amounts* Surface Flatness Length Width Three-cylinder and V6 engines .003'' .002'' Four-cylinder and V8 engines .004" .002" In-line six-cylinder engines .006" .002" *This is the sum of the values of the cylinder head plus cylinder block combined. Since cylinder blocks usually do not display major out-offlat conditions, out-of-flat conditions are usually associated with the cylinder head; but the sum of the two must be kept in mind and must not exceed the recommended specification.

Surface Roughness for Conventional Gaskets Having Either Steel or Fiber Outer Surfaces Cast Iron Maximum 110 Ra (125 RMS) (all values are in microinches) (rougher surfaces limit gasket conformance)Minimum 30 Ra (33 RMS)(smoother surfaces increase tendency for gasket to flow and reduce the gasket's blowout resistance) Recommended range 60100 Ra (65120 RMS) Aluminum Maximum 60 Ra (65 RMS) Minimum 30 Ra (33 RMS) Recommended range 5060 Ra (5565 RMS) Surface Roughness for Engine Assemblies Utilizing Rubber-Coated Multilayered-Steel Head Gaskets Maximum 30 Ra (33 RMS)

Smoother finishes are desirable Waviness A maximum waviness height of .0004"/.0005" is recommended. This value is usually associated with milling machines. A waviness peak spacing of .100" is satisfactory. It should be no less than .030". Miscellaneous Information There should be no sudden irregularities exceeding .001". A maximum out-of-flat of 0.001" in any 3" diameter should not be exceeded.

Page 76

Cleanliness of the mating surfaces and gaskets is of prime importance. Soft-faced gaskets are preferred for the extremes, since there is great surface conformity under clamping pressure. All gaskets, where metal or soft surfaced, are supplied with a coating to aid in sealing rough surface finishes. In general, gasket manufacturers do not recommend use of supplementary sealers on cylinder head gaskets, with the exception of steel embossed shims. The gaskets are supplied as total sealing systems; if properly installed on properly conditioned engines, they do not require supplementary sealers. 5.3. Manifold Gaskets From a gasketing point of view, there are three types of manifold assemblies in common use: intake, exhaust, and a combination of both. Functional requirements usually demand different materials for each type of service; however, for the combination, a compromise in materials is usually made that is acceptable for both. 5.3.1. Intake Gaskets Intake gaskets must always have excellent vacuum sealing ability. This is especially true in today's computer-controlled engines. Even slight air leakage increases the air-fuel ratio, resulting in a hotrunning engine, poor operation, increase in Nox pollution, and reduction of valve life. Many Vee intake manifolds are lightly loaded; and at times, inadequate clamping pressure is manifest. Fiber constructions such as beater-mix or compressed sheeting many times were not adequately compressed to stop wicking of

coolant and fuel vapors. In addition, they tended to break down and rupture in the port walls, which resulted in vacuum or water leaks. The preferred intake gasket structure consists of a metal core to provide rigidity and prevent wall collapse due to vacuum in fuel intake ports and coolant pressure at water crossover ports. Rubber facings are bonded to the steel core to provide surface seal in contact with casting flange faces. As in the case of head gaskets, bonding may be accomplished either by mechanical clinching of metal tangs or by chemical adhesion. In some cases the gaskets are embossed to achieve higher sealing pressure in those areas that are poorly clamped. Some V-6 and V-8 production engines have incorporated intake manifold gaskets with an integral oil splash plate that stops oil splash on the base of the intake manifold (see Fig. 38). This is done to prevent the caking of oil sludge on the hot section of the manifold that is porting the exhaust gas crossover. These gaskets have been called bathtubs, turkey pans, turtlebacks, and dishpans by various engine and gasket manufacturers. Due to the deep forming required to clear the manifold, a soft-temper steel is required. However, the soft temper is

Page 77

Figure 38 Intake manifold gasket with splash plate.

a disadvantage in the case of the embossments. In some gaskets the embossments tend to collapse under loading and don't provide adequate sealing. As a result, some of the gaskets require rubberfiber gaskets in addition to the embossing. The fiber gasket is sometimes supplied loose; sometimes it is attached to the gasket. Intake manifold leaks can lead to serious engine problems that result in repairs far more costly than replacing an intake manifold gasket. Some intake manifold leaks cause changes in vacuum pressure. This can cause the on-board computer to make adjustments in fuel flow and engine timing, which in turn can result in drivability problems or abnormal combustion (preignition or detonation). Abnormal combustion is the most common cause of damage to the head gasket. It can even result in damage to the engine's hard parts, such as pistons, rings, valves, and spark plugs. Some intake manifold leaks can lead to external coolant leakage.

Or coolant might leak into the engine's oil supply, causing serious and expensive damage. If an exhaust crossover port passes through the intake manifold gasket, some leaks can result in a noisy engine, power loss, or even a blown gasket. Of course, leaks such as these can occur in combinations. 5.3.2. Exhaust Gaskets Exhaust manifold gaskets are becoming more important because of emission and noise and vibration hardness (NVH) requirements. In the past, many Original Equipment Manufacturer (OEM)-built engines had no exhaust manifold gaskets.

Page 78

The most common gasket is a perforated steel core with hightemperature-resistant facings clinched to each side of it. For instances of large expansion and contraction the preferred structure is a two-layer-steel facing gasket, with the flat steel surface contacting the manifold to permit manifold-to-gasket slippage. Generally speaking, the exhaust gasket contains a steel surface, at least on the side facing the manifold, to permit manifold-to-gasket slippage. If the gasket has an overlap, it is located on the stationary side of the joint. To accommodate large thermal motion that occurs on some manifolds, some gaskets have slotted bolt holes and some gaskets may have a crimp incorporated into them (Fig. 39). Often, single-port, embossed steel gaskets are used on heavy-duty diesel engines for better accommodation of the large thermal motion that occurs. In many cases, single-bead embossments are adequate, but some assemblies require a double-bead design (Fig. 40). To aid further in cases where large motion is involved, these gaskets can be coated with high-temperature solid-film lubricants such as graphite, molybdenum disulfide, or other special formulations. Embossed multilayer gaskets are becoming popular on today's high-output engines (see Fig. 41). Some exhaust manifold gaskets also incorporate additional material to act as a heat shield to protect ignition wires and/or other engine rubber parts from the high heat of the engine exhaust. Figure 42 shows one of these.

Figure 39 Exhaust manifold gasket with crimp.

Page 79

Figure 40 Single- and double-embossed exhaust manifold gaskets.

Another type of seal that is external to the engine, but one associated with sealing exhaust gas, is the exhaust-pipe ring seal located between the exhaust header and the tail pipe. These gaskets are normally made from bitumastic or cement-bound fibrous material or graphite with metal reinforcement, where the metal is perforated or is a wire mesh. In some cases, where very high temperatures

Figure 41 Multiple-layer steel exhaust manifold gasket.

Page 80

Figure 42 Exhaust manifold gasket and heat shield.

are experienced, a metal sleeve, often stainless, is incorporated into the seal to protect the material from eroding at high temperatures, which results in long-term sealing. This particular gasket is in an even more demanding environment in transverse-mounted engines. Rotation of the engine during operation causes motion on the sealing surfaces. The gaskets in these engines must accommodate this motion while continuing to seal. Six-cylinder in-line engines pose manifold gasket problems when intake and exhaust ports are intermixed. For best sealing, different materials are desired at alternate ports. In some cases the exhaust and intake ports are not flush with each other, and therefore the gasket thicknesses at these ports must be different. Installation of these manifolds is critical, and strict conformance to installation instructions is recommended. 5.3.3. Oil Pan Gaskets

Oil pan gaskets have two different types of assemblies: cast-metal pans and formed-metal pans. The cast-metal pan is rarely used for passenger cars or light trucks because of its weight and cost. It seldom presents any problems to the gasket engineer, since the flange area is usually an unbroken rectangular shape. A wide variety of fiber-type gasket materials in the one-piece construction are satisfactory. The

Page 81

larger gaskets are frequently broken down into dovetailed segments for reason of material economy and convenience in packaging and handling (Fig. 43). The formed-metal pan assemblies frequently become a challenge to the gasket engineer. The thin metal of the pan gasket flange is easily drawn and distorted around the bolt holes by overtorquing. It is an absolute must that these pan flanges be flat, not only around the bolt holes but also along the length of the pan. Flat surfaces with uniform controlled bolt torque is the answer to reliable sealing. Also, remember that a little more torque does a little more harm in the case of thin flanges. Because of the metal distortion problem, softer and thicker materials have consistently been the rule for formed-metal oil pan applications. Cork and vegetable-fiber packings have a marked tendency for the oil to wick through the body if not fully and uniformly compressed. These types of gasket materials are improved by rubber overcoatings that envelop the gasket and improve interface seal between gasket and flange. The coating also contributes substantially to dimensional stability in storage. 5.3.4. Rocker and Cam Cover Gaskets Many of the production rocker and cam covers are of formed metal or molded plastic. Use of plastic for these covers is becoming very popular. As in the case of oil pans, flat surfaces with controlled torque are extremely important. Rocker covers generally have fewer fasteners per inch than oil pans. In addition, the cylinder head land upon which the gasket seals is not always flat, as is the block in the case of the oil pan gasket. As a result, the rocker cover is even more difficult to seal than the oil pan gasket installation. In

some cases the head land is so narrow that the gasket may slip off the land, permitting oil seepage up and along the fastener. In these applications, backup washers are supplied to stop the leakage at the top surface of the rocker cover flange. There are numerous materials used for rocker cover applications. Each has its inherent advantages and disadvan-

Figure 43 Segmented gasket and various types of dovetails.

Page 82

tages. Rubber gaskets generally give the longest seal life, but present installation problems in some applications. Figure 44 shows some molded-rubber gaskets that are in production. Cork-rubber gaskets ease the installation problems at the sacrifice of seal life and are used when cost is the major criterion for gasket selection. To avoid comebacks from leaks, there is sometimes a tendency to overtorque when installing a cover gasket. Unfortunately, overtorquing can cause covers to distort. This leads to poor clamping force along the length of the cover and could result in leaks. Many engines are designed with a special groove in the rocker or cam cover to hold a molded-or extruded-rubber gasket. The molded gasket may be stretched slightly to make it fit in place so it holds tightly against the cover. When the cover is tightened, the rubber material deforms rather than compresses. Because of its memory, the gasket tries to return to its original shape and size, pushing against the metal surfaces, resulting in a long-term seal. The cover openings and the gasket are both precisely sized in order to ensure a proper seal. 5.3.5. Other Gaskets Gaskets for timing covers, water pumps, and other applications follow essentially the same guidelines given for oil pans, since the cover may be either cast, formed-metal, or plastic. Accessory items such as oil pump and fuel pump assemblies

Figure 44 Various production molded-rubber gaskets.

Page 83

require a different type of gasket material. For these functions, power must be transmitted through the flange joint either by rotating shaft or lever. Thus, the flange joint is subjected to a working load. Gasket materials for these cases, although fiber, have higher densities than the usual materials. The higher densities are specified to keep torque loss at a minimum, thereby reducing the possibility of the clamping bolts vibrating loose and causing eventual leakage. 5.4 Gasket Installation When a gasket is installed between mating surfaces, tightening the bolts compresses the gasket slightly, allowing the gasket material to conform to the small irregularities on the surfaces. This allows the gasket to cold seal so it won't leak before the engine is started. The gasket's ability to achieve a positive cold seal as well as to maintain a long-lasting, leak-free seal depends on two things: its own ability to retain torque over time (which depends on the design of the gasket and the materials used in its construction), and the clamping force applied by the head bolts. But even the best gasket won't hold and maintain a tight seal if the bolts have not been properly torqued. The amount of torque that's applied to the bolts, as well as the order in which the bolts are tightened, determines how the clamping force is distributed across the surface of the gasket. If one area of the gasket is under high clamping force while another area is not, it may allow the gasket to leak at the weakly clamped point. Another consequence of failing to torque the bolts properly can be warpage of the mating surfaces. Uneven loading created by unevenly tightened bolts can distort surfaces. Over a period of time, this may cause the surface to take a

permanent set. For a gasket to seal properly, the mating surfaces must be clean, relatively smooth, and flat. The presence of foreign material (dirt, old gasket facing material, abrasive residue, etc.) on either surface can prevent a good seal by preventing firm contact between the mating surfaces and gasket. Debris can act like a bridge and form pockets that create a path for leakage to occur. Debris can also become embedded in the surface of the gasket and damage the gasket. To avoid these kinds of problems, make sure the surface of both the head and the engine deck are perfectly clean before installing the head gasket. Old gasket material, rust, scale, dirt, and other debris may be removed by applying a gasket-removing compound (such as Fel-Pro Gasket Remover), by using a generalpurpose degreaser (such as Fel-Pro Clean Off), and/or by scraping the surface of the head with a scraper or wirebrush. When cleaning aluminum surfaces, use a nonmetallic (plastic) scraper, so relatively soft metal will not be gouged or scratched. Resurfacing a mating surface (milling, grinding, or belt sanding) usually removes all traces of debris. But milled, ground, and sanded surfaces can sometimes pick up debris from the resurfacing operation itself. So again, don't assume

Page 84

the surface is clean just because it's been resurfaced. Check it, and clean it if necessary to remove any metal chips or residual traces of abrasive. Ordinary soap and water works best for a final cleaning. Clean hands are a must when installing the gasket. Greasy fingers can pick up a lot of dirt that may end up on the gasket or head surface. Grease and oil can also interfere with a good seal, so keep the gasket as well as the mating surfaces as clean and dry as possible during installation. 5.4.1. Installation Specifics All the bolts need to be in perfect condition, with clean, undamaged threads. Dirty or damaged threads can give false torque readings as well as decrease a bolt's clamping force by as much as 50%! Wirebrush all bolt threads, carefully inspect each one, and replace any that are nicked, deformed or worn. Dirty or deformed hole threads in the mating surface can reduce clamping force, the same as can dirty or damaged threads on the bolts. Unless specified instructions counterrecommend it, run a bottoming tap down each bolt hole. The tops of the holes should also be chamfered so the uppermost threads won't pull above the surface when the bolts are tightened. Clean all holes to remove any debris. For bolts that screw into blind holes, lightly lubricate the bolt threads as well as the underside of the bolt heads with engine oil. For bolts that extend into liquid openings, coat the threads with a flexible sealer such as Fel-Pro Bold Prep (GRA2). Failure to coat the threads may allow liquid to leak past the bolt.

Check bolt lengths. Make sure you have the correct-length bolts for the application and for each hole location (some locations require longer or shorter bolts than others). Bolts should also be measured or compared to one another to check for stretch. Any bolt found to be stretched should be replaced because (1) it may be dangerously weak, and (2) it may bottom out when installed in a blind hole. When installing bolts in aluminum-surface applications, hardenedsteel washers should be used under the bolt heads to prevent galling of the soft aluminum and to help distribute the force. Make sure the washers are positioned with their rounded or chamfered side up and that there is no debris or burrs under their surfaces. Resurfacing a mating surface decreases its overall height, so be sure to check bolt lengths to make sure they won't bottom out in blind holes. If a bolt bottoms out, it will apply little or no clamping force and, therefore, cause the gasket to leak. Always use the specified tightening sequence and recommended bolt torque values. If no tightening specification exists, use the sequences shown on Figs. 45 and 46. Use an accurate torque wrench to tighten the bolts in three to five

Page 85

Figure 45 Circular-gasket criss-cross fastening sequence.

Figure 46 Spiral fastening sequence for noncircular gasket.

incremental steps, following the recommended sequence and torque specs for the application. Tightening the bolts down gradually creates an even clamping force on the gasket and reduces distortion. Double-check the final torque readings on each bolt to make sure none have been missed and that the bolts are retaining torque normally.

Page 87

3 Industrial Gaskets JÖRG LATTE Istag AG, Egliswil, Switzerland DEREK COOMBER Thermoseal Inc., Sidney, Ohio 1. Introduction to Gaskets A gasket is the interface between two imperfect surfaces. Its function is to prevent leakage of liquids or gases under constantly changing conditions of mechanically and thermally induced stress. Gaskets are used in static sealing applications that range from a simple rubber seal on a garden hose to sophisticated combinations of metallic and nonmetallic elements for high-pressure, hightemperature industrial applications, often sealing toxic or flammable liquids or gases. Gaskets may be specially designed to meet a specific application, such as an automotive cylinder head gasket, or they may be manufactured to American national standards, such as American Society of Mechanical Engineers (ASME) B 16.21 [1] for nonmetallic gaskets or ASME B 16.20 [2] for spiral-wound and jacketed gaskets. These standards specify dimensions, materials of construction, pressure/temperature classes, and, in some instances, methods of construction and methods of identification marking. Nonmetallic gaskets may be specified as ring gaskets, which fit inside the bolts of a circular flange, or as full-face gaskets, which locate onto the bolts. There are many other standards commonly in

use,

Page 88

including the German national standard (DIN) and the British national standard (BS). The importance of good gasketing has been brought into focus by our increasing recognition of the need to protect our environment, demonstrated by the Federal Clean Air Act and by Title V requirements [3] and by our own recognition of how fragile is Spaceship Earth. Further, the safety of personnel operating in the area where gasketing is used and the safety of those living in the area can be dependent upon the integrity of the bolted flange assemblies. More pragmatically, we also recognize the cost benefits of good gasketing and the energy and lost-product costs that can be incurred with leaking flanges. When we refer to leaking flanges or leaking gaskets, we really mean unacceptable levels of leakage. All bolted-joint assemblies leak, but the best ones leak so slowly that it doesn't matter and/or may even be difficult to detect [4]. Gasket applications fall into three major market groups: Automotive Original equipment (OE) Maintenance repair and operations (MRO) Original equipment and MRO gaskets can be further grouped into: Pressure vessel applications Non-pressure vessel applications Very often, the lines between these various groupings get blurred. For example, a manufacturer of heat exchangers would be

considered an OE user of pressure vessel gaskets, but the maintenance of the heat exchangers would be considered an MRO pressure vessel application, and in many instances the type and quality of gaskets used in the maintenance application will be different from the gaskets used by the manufacturer. One big difference is that while the OE manufacturer will probably purchase gaskets that are ready for installation, maintenance personnel will often cut gaskets from sheets of gasket materials, on site, as the gaskets are needed. Automotive gaskets are basically non-pressure vessel applications and are discussed elsewhere in this book (Chapter 1). Therefore, our major focus here is on industrial gasketing applications. The term pressure vessel refers to a specific group of applications. ASME defines pressure vessels as containers for the containment of pressure, either internal or external. This pressure may be obtained from an external source, or by the application of heat from a direct or indirect source, or any combination thereof [5]. Design and construction requirements for these applications are detailed in the ASME Boiler and Pressure Vessel Code, and professional engineers are required to follow the mandatory components of the code [6]. Most follow the nonmandatory (advisory) sections. Gasketing requirements fall into the nonmandatory classification.

Page 89

Section VIII, Division 1, of the code specifies Rules for Construction of Pressure Vessels, and Appendix 2 specifies Rules for Bolted Flange Connections with Ring Type Gaskets. Further information is supplied in Appendix S, Design Considerations for Bolted Flange Connections; nuclear applications are addressed in Section III, Division 1. Included in these sections are rules for the design of flanges and methods for determining the appropriate gasket stress for defined applications and specific gasket types. The amount of stress required to be applied to a gasket (and hence the design of flanges and bolt selection) is currently determined using gasket factors y and m. The American Society for Testing Materials (ASTM) and ASME are introducing new gasket constants Gb, a, Gs, which allow the degree of tightness required to be considered in designing the flange assembly and calculating the required gasket stress. These new gasket constants are addressed in Chapter 8 of this book. Some examples of OE and MRO non-pressure vessel applications are: Refrigeration compressors Stationary and off-road combustion engines Sewer and waste disposal applications Potable water supply Gasoline storage and distribution Some examples of OE and MRO applications that may be considered pressure vessel applications are:

Boilers Heat exchangers Process steam systems and pipe work It is interesting to note that personnel who weld pressure vessels or weld components to pressure vessels are required to be certified. However, personnel who bolt components to pressure vessels are not required to be certified, nor are they required to have any training in the tasks they are performing. This anomaly is now being addressed by ASME. To appreciate the range of gaskets that is available, it is necessary to understand the range of basic materials and manufacturing procedures. One group of gaskets is produced by cutting simple or complex shapes from flat sheets of gasket material. The cutting tools range from scissors through steel blade/plywood tools to sophisticated compound steel tools. Recently, water jet and laser cutting methods have been introduced. With this type of production, very complex shapes can be produced. Typical examples of the base gasket materials are: Beater addition (paper making process) Rubber-bonded compressed asbestos

Page 90

Rubber-bonded compressed nonasbestos (also known as compressed fiber material or compressed synthetic fiber material) Polytetrafluroethylene (PTFE) Flexible graphite A second group of gaskets is manufactured by assembling or forming components to the finished gasket form. Here the manufacturer may use metallic components or combine metallic and nonmetallic components to form sophisticated gasket configurations. While the gasket configuration may be sophisticated, the outline of the gasket is limited to regular shapes such as rings, ovals, squares, and rectangles, with maybe some cross members. Typical examples of semi-metallic gaskets are: Spiral wound Single and double jacketed Filled corrugated Rigid laminated graphite/steel Typical examples of solid metal gaskets are: Plain corrugated Flat ring Lens ring Wedge seals 2. Function of Gaskets

To understand better the parameters that influence the sealing capabilities of a gasketing material, it is necessary to consider how a static gasket functions. If the faces of flanges could be made to mate perfectly, if they were perfectly flat and parallel to each other and stayed that way in operation, then no gasket would be necessary. However, in practice the flange surfaces are always rough, out of parallel to some degree, and deformed. Further, the relationship of the flanges, one to the other, changes during assembly and again while in operation. This unevenness must be compensated by a compressible and recoverable material, a characteristic of the materials discussed. In addition to the sealing material's acting as a compensator for flange imperfections, it must also act as a barrier against the medium inside the pipe and therefore be chemically resistant to that medium and the temperatures encountered. The gasket must also have a load-bearing capacity sufficient to sustain the stress coming from the bolt forces and the internal pressure (Fig. 1). Based on the foregoing, we can list the requirements for gaskets:

Page 91

Figure 1 Forces acting on a gasket.

1. They must be conformable to flange surfaces. 2. They must exhibit resistance to high temperatures, high surface pressure, and chemical attack. 3. They must remain tight at all service conditions. But to achieve a safe flange connection, we also have to consider the flange itself and the bolts (Fig. 2). In the past, process plant and component design did not allow for these requirements to be considered. The gasket was not considered a critical element in the plant design-it just had to work. When a leak did occur, it was always assumed to be the fault of the gasket and never of incorrect flange design or incorrect gasket installation.

Very basic reasons often lead to leakage in a bolted flange connection, and in most cases it is not the gasket itself that causes the problem. The following list summarizes some reasons for bolted joint assembly failures. The list is by no means complete, but it does illustrate the major mistakes that are made. Every

Page 92

Figure 2 Components of a safe flange connection.

manufacturer has a wide experience of many unbelievable misapplications, but when failure occurs it is blamed on the gasket. Assembly Lack of initial compression Uneven bolting load Bad threads on bolts/studs/set bolts Incorrect-quality bolts Misalignment of the gasket Metal Faces Uneven, dirty, damaged, corroded Rotation of flanges Gasket Material

Reuse of old gasket Incompatibility between gasket material and medium Operating conditions that exceed the limitations of the gasket material Excessive use of release compounds on the gasket Incorrect dimensions Permeability of material

Page 93

Design Factors Insufficient bolt load (low gasket stress) Excessive bolt load (high gasket stress) Weak flange/poor bolting arrangement 3. Types of Static Gaskets Fundamentally, all static gaskets belong to one of the following categories: Nonmetallic gaskets Semimetallic gaskets Metallic gaskets These can be grouped into eight categories, by manufacturing process: 1. Beater-addition process 2. Calendar process (compressed asbestos fibers and compressed synthetic fibers) 3. Graphite laminates 4. Rubber-coated metals 5. PTFE-based products 6. Special sheet products, such as mica laminates and sandwiched composites sheets 7. Spiral-wound gaskets

8. Specialty gaskets such as eyeleted gaskets, metal envelope, lined metal gaskets, PTFE-enveloped gaskets, steel/graphite gaskets, corrugated gaskets, solid-metal flat ring gaskets, solid-metal ring joints, and many others Within the scope of this chapter it is not possible to address all types of gaskets and gasket materials in detail. This is because between groups there exist many intermediate materials, and each group can be split into innumerable subgroups and subgroup materials. Items 1 through 6 in the categorization by manufacturing process are the basic sheet materials from which gaskets are cut, and items 7 and 8 are gaskets manufactured directly in their final form. The materials in the first three categoriesbeater addition, calendered sheet, and graphite laminatesare very cost effective. The first two, beater addition and calendered sheet, are fiber reinforced and very user friendly. This means they can be cut and handled easily by maintenance people on site and are the materials of choice for general maintenance applications where gasket cutting is performed by the maintenance people. Compressed nonasbestos sheet can also be further reinforced with woven or expanded steel inserts.

Page 94

Compressed asbestos gasket material is undoubtedly the most versatile gasket material available. However, the health concerns related to the use of asbestos and the legal and insurance problems for this type of material have severely limited its use. Compressed asbestos gasket material is still available, but the majority of compressed gaskets are the nonasbestos type manufactured with synthetic fibers. Beater-addition material, which is primarily used for OE applications, was manufactured for many years with asbestos. Today, very little beater-addition gasket material is manufactured in this country with asbestos; the great majority is manufactured with synthetic fiber. Flexible graphite, expanded graphite mechanically locked into a matrix, has become a more popular material of choice since the decline in the use of asbestos-containing materials. This material is versatile, with excellent thermal and chemical resistance. However it is not particularly user friendly, and gaskets are generally produced by gasket cutters (also known as fabricators) rather than by the maintenance people on site. Rubber-coated metal gaskets, produced with sophisticated tooling from sheets of metal-reinforced rubber, are becoming more popular, especially in industrial OE and automotive applications. In the manufacture of the gaskets, the sealing properties can be enhanced by including beading or corrugating in the cutting process, instead of the conventional flat gasket design. There have been moves during recent years to promote this material for multiple-layer cylinder head gaskets. This approach has recently gained momentum in replacing conventional cylinder

head gaskets based on beater-addition materials. This material is also very suitable for high-volume OE applications, where the high cost of tooling can be justified and where only very low leakage is allowed. PTFE is a highly chemical-resistant resin polymer; therefore, it should have been a superior gasket material. Unfortunately, in its unmodified form it has little resilience against surface stress and temperature, thus leading to unacceptable creep resistance. Gaskets made of PTFE were used only in more exotic applications or as envelopes for standard compressed-fiber gaskets. Starting in the early 1980s, PTFE was modified in two different procedures, resulting in two different new grades: Expanded PTFE Filled fibrillated PTFE In spite of the high price, both types were well accepted in the market; yet neither can be used in fire-safe applications, which essentially prohibits their use in the refinery and petrochemical industry. Recently, a new material of the second type (filled fibrillated PTFE) has been developed and launched on the market by a major gasket manufacturer.

Page 95

This new gasket material is designed to be suitable for highly demanding and critical applications, such as high temperature, high internal pressure, and high surface stress together with excellent resistance in severe chemical environments. This material has successfully passed the API 6FA fire-safety test. While the sheet gasket materials we have discussed so far represent the major part of the market, there are some specialty materials that should also be addressed. Mica foil gaskets are manufactured from pure mica laminated with tanged metal. These gaskets are produced by a number of gasket manufacturers, but the basic mica foil is produced by only one manufacturer. These gaskets are excellent in hot-gas applications, such as industrial and automotive exhaust systems, and have also been successful in molten-lead applications and in the manufacture of lead/acid batteries. The material can also be used in some strong-acid applications. Other special sheet materials include combinations of base sheets already discussed. For example, compressed-fiber sheet with a layer of flexible graphite on both sides uses the good properties of both basic products to get a material that can be easily handled and cut and has the positive properties of flexible graphite and compressed-fiber material. This combining of materials carries over from the sheet material into the specialty gasket field. There are a number of gaskets available that combine steel and a nonmetallic component. The most widely used is the spiral-wound gasket (Fig. 3). This type of gasket is generally circular, formed by winding together layers of thin steel and a nonmetallic filler, such as flexible graphite or PTFE, onto a form, with the cross section of the metal and filler being formed to a bell shape to give resilience

to the gasket. The spiral-wound gasket normally has an outer retaining ring, and in many instances an inner retaining ring, to prevent the windings from buckling. Graphite-filled spiral-wound gaskets are fire safe and are widely used in hydrocarbon applications. An alternative to the spiral-wound gasket is the steel/flexible graphite gasket shown in Fig. 4. This type of gasket is manufactured from steel sheet, normally with some corrugating of the steel to get a spring effect to increase gasket recovery and maintain gasket stress, with either rings of flexible graphite glued to the steel or full coverage of the steel with flexible graphite. This type of gasket is also suitable for fire-safe applications. Another specialty semimetallic gasket is the single- or doublejacketed gasket (Fig. 5). Here a thin metal covering is wrapped around a resilient nonmetallic filling. All of these gaskets are limited to standard flange joints and some basic heat exchanger applications. There are also combinations of steel and expanded PTFE, molded rubber-coated steel gaskets (not cut from sheet), and gaskets with eyelets, fire rings, silicone beads, inserts, and other modifications custom designed to fit specific needs.

Page 96

Figure 3 Spiral-wound gasket.

Figure 4 Steel/flexible graphite gasket.

Page 97

Figure 5 Metal jacketed gaskets.

The last group of gaskets to consider is metallic gaskets. While flat or corrugated metal gaskets are used with flat flanges, the most popular metal gaskets are ring-type gaskets (Fig. 6). These are generally circular gaskets of specific cross section, and must be used with special flange assemblies. These types of gaskets are sometimes referred to as soft-metal gaskets, and while a wide range of metals can be used, including copper, low-carbon steel, stainless steel, and Monel metal, it is a requirement that the gasket be softer than the flange. In most flange designs for ring gaskets, the gasket is fully contained. However, the lens ring gasket is located to the inner part of the flange and is not fully contained (Fig. 7).

Figure 6 Ring gaskets.

Page 98

Figure 7 Lens ring gasket.

For a very high-pressure applications, the wedge seal (Fig.8) probably offers the most reliable seal. Here an annular metal wedge is locked between two flanges, which have angled contact faces that utilize the mechanical advantage of the inclined plane to apply a very high surface stress to the wedge seal. 4. Production Processes In order to understand the properties of sealing materials, it is necessary to understand their manufacturing processes.

Figure 8 Wedge seal.

Page 99

4.1. Beater-Addition Process Figure 9 shows the beater-addition process in outline. Water, latex, fibers, fillers, and chemicals are mixed to form a slurry. The manufacturers' know-how manifests itself not only in the intelligent selection of the type and quantity of the various raw materials, but also in planning the sequence in which they are added and the sequence of time they are exposed to mixing. Consequently, all manufacturers keep the details of their process secret. However, the main fiber basis in all high-quality gaskets of this kind is aramid pulp. For less demanding applications, beater add is also manufactured with cellulose as the primary fiber ingredient. These types of materials have a lower temperature limitation than the aramid-reinforced materials.

Figure 9 Beater-addition production.

Page 100

On a circulating wire sieve, the slurry is largely drained of its water content, dried in a heating tunnel, and vulcanized on heated calendar rolls. At this point, the maximum thickness is around 2.0 mm (0.080''). At this thickness, however, only low-density material can be manufactured, but the material can be compressed in a hydraulic press or two-roll calender, which increases the density and improves sealability. For further improvement, the material must be impregnated or needs additional high-pressure hot compression to be applied. Because of the method of manufacture, beater-add materials generally do not have as good sealability as equivalent compressedfiber materials and are best suited for moderate-pressure temperature applications. Since the process is basically continuous, the end product is relatively cheap compared to compressed-fiber material made on a calender. Beater-addition material is widely used in the making of auxiliary engine gaskets or, after further processing, cylinder head gaskets. For this purpose, the semifinished product is laminated onto both sides of a spiked metal sheet and is physically fixed in place by the spikes. Properties of Beater Addition Materials Advantages Disadvantages Inexpensive Poor sealability Acceptable Temperature limitation to 280°C (535°F) (may go to basic 400°C (750°F) in limited circumstances) properties

4.2. Calender Process Unlike the beater-addition process, which is continuous, the

calendering process is carried out in discontinuous steps. As Fig. 10 shows, the dry ingredients are thoroughly mixed together and then blended with a rubber solution prepared using an appropriate solvent. Where available, powdered rubber is included with the other dry ingredients, which are mixed together until a very fluffy consistency is reached. The manufacturer has to be aware of the danger of static electricity, and the selection of raw materials and good equipment design and work practices can help prevent this. When the dry materials are fully blended, the required solvents will be added, and this will cause the elastomer to liquefy, allowing it to encapsulate the other ingredients. A skilled manufacturer will vary the mixing process depending on the materials being used. The final choice is often influenced by the type of fiber used in the product. After mixing, the compound is removed from the machine and stored temporarily in a container, from where it is conveyed to the calender in batches. The calender consists of a small roll that is cooled and a large roll that is heated.

Page 101

Figure 10 Compressed-fiber production.

In calenders of the conventional type, these rolls are arranged vertically, as shown in Fig. 11. The smaller, top roll, which is cooled, presses against the lower, heated roll with a force of between 6 and 20 tons, depending on the material involved. The blended compound is fed into the nip between these two rolls and is drawn in by the rotary movement. If the mixing process has been properly carried out, the compound will wrap itself around the hot lower roll in layers about 0.02 mm

(0.0008'') thick, depending on pressure, gradually forming a cylindrical lining. In the process, the solvent evaporates and vulcanization commences. The solvent vapors are extracted either for recycling or for combustion or possibly catalytic combustion,

Page 102

Figure 11 Two-roll calender.

as the case may be. Recycled solvents are reused, while the energy produced by combustion is used to heat the calender. The speed of rotation of the lower roll depends on the evaporation rate of the solvent. If the roll rotates too fast, the solvent content cannot evaporate completely before the next layer of compound is applied, so the entrapped liquid creates blisters and the product will have to be scrapped. If the roll rotates too slowly, the material may be too dry to form a homogenous bond to the previous layer, and delamination may occur. As soon as the material has reached its specified thickness [which

in the case of 1.6 mm (1/16") would be after approximately 80 revolutions of the heated roll], compound feeding ends. Material up to and even above 6.4 mm (1/4") thick can be manufactured by this method, but few gaskets are used above 3.2 mm (1/8").

Page 103

There are two options for cutting open the cylinder liner: 1. Across the machine direction, in which instance one gets a sheet as wide as the lower roll and as long as its circumference. 2. In spirals, in which instance one gets strips of material ranging in width between 25 mm (1") and 150 mm (6"), which is the maximum that makes sense. While wider or narrower strips may be produced, they are not normally economical. The sheet thus obtained would then be further processed by trimming, printing, cutting to size, and finally gasket cutting or punching. No additional pressing or heating is required, and the material is ready to perform as a gasket. The calendering process leads to special properties in the sealing material. One of these is tensile strength, which is very different depending on whether it is measured with or across the calendering direction. This effect is a result of the tendency of the fibers to become oriented in direction in the calender nip with the movement of the calender roll. The relationship of the tensile strength of asbestos-containing compressed materials measured in line and across was approximately 2:1. The same relationship in the new, nonasbestos compressed materials is typically 2.5:1 to 3:1. However, the tensile strength is normally of little importance to the quality of a gasket, since a gasket very rarely is subjected to tensile forces. In very rare instances, such as radiator gaskets, the flanges are screwed together, and the resulting torsion leads to a tensile stress on the gasket. In order to keep that stress as low as possible, it is necessary in these cases to use lubricants on the surfaces such as

graphite or PTFE coatings. Generally, however, it is not necessary to apply lubricants to a compressed gasket material, and in fact the addition of lubricants, particularly metallic based, can have a significant negative impact on the gasket and can cause premature gasket failure. Since calender sheets are formed in layers, it is theoretically possible to introduce a different compound at each revolution of the heated roll or to apply two or more compounds alternatively. So far, no one has gone to this extreme. It is current practice, however, to apply special compounds in the first and last layers to ensure specific properties. Here are a few typical instances: 1. Special surface layers featuring enhanced antistick and anticorrosion properties 2. The application of highly resilient surface layers to ensure high compressibility 3. Surface layers with defined chemical properties 4. The application of coatings (PTFE, antistick, adhesives, etc.) Another technique is to introduce a layer of wire mesh or expanded metal into the center of the sheet for enhanced strength. The reinforcement with expanded metal results in an inversion of the previously mentioned relation in tensile strength.

Page 104

Because expanded metal is stronger in the across-machine direction, this dominates the tensile strength in this type of gasket material, up to a thickness of about 2.4 mm (3/32"). Over this thickness the basic material becomes dominant. The advantages and disadvantages of gaskets reinforced by expanded metal are as follows: Properties of Expanded Metal Reinforced Materials Advantages Disadvantages Blowout protection Increased gas leakage Improved load-bearing capacity Unfavorable cutability Stability against bending Improved tensile strength Change of anisotropic effects Decreased brittleness

In all of these instances, however, the compressed-fiber materials are so highly compressed that they form an organic entity that is no longer a true laminate. Another peculiar feature of calendered gaskets is that because of the exposure to the time and temperature gradients involved, the first layers of a gasket sheet, i.e., the side that was in direct contact with the heated roll, is more highly vulcanized than the finish layers, where both the temperature and the duration of vulcanization are much reduced because of the heat loss and the shorter exposure times involved. This effect may also be useful in certain circumstances, because the finished side is softer. On the other hand, it is possible to neutralize these effects through postheat treatment. Among the advantages of the calender process is its versatility,

since it allows the design of composites having specific properties. Because it involves compression at high pressure, it makes the end product a material ideally suited for gaskets and more favorable than beater-addition products. On the other hand, the relatively high expense involved is a drawback. Properties of Compressed Gasket Materials Advantages Disadvantages High loadbearing capacity Hardening due to increased temperature (depending on grade) Good or excellent crush resistance Generally limited to maximum of 350°C (660°F) (depending on due to deterioration of the rubber binder; in specific grade) instances temperatures as high as 425°C (800°F) Good sealability can be accommodated at ambient temperatures

(table continued on next page)

Page 105

(table continued from previous page) Advantages Disadvantages Significantly improved sealability at elevated temperatures Good conformity to flange surfaces Good handlability Large dimensions For oversize dimensions, pieces can be joined together Wide variety of applications Economical

4.3. Flexible Graphite Graphite is one of the two crystalline forms of the element carbon, the other being diamond. The differences between these two forms is explained by their crystal structure, as shown in Fig. 12. Graphite consists of layer planes of hexagonally arranged carbon atoms, which make up a so-called plane structure. The circumstance that these planes slide against each other explains the softness and excellent lubricity and the mechanical properties of this material. The hardness and stability of diamond is due to its threedimensional bond. Both forms, graphite and diamond, are found in nature and can be produced synthetically. To form flexible graphite, natural graphite flake is subjected to a special chemical and thermal procedure (Fig. 13). By this, it expands over 400-fold in sizesimilar to popcorn. This expanded graphite is so light that a volume of 1 liter holds only 2 to 3 grams. Through repeated calendering and heating, the expanded graphite particles are compacted into a sheet or foil, which is accomplished without the use of any organic or inorganic binders. The wormlike

expanded graphite particles (Fig. 14) are interlocked mechanically. By this method, rolls of graphite foils are produced with thicknesses ranging from 0.20 mm to 1.00 mm (0.008" to 0.040") and with a maximum width of 1.5 m. (60"). The standard density of graphite foil is 1.0 g/cc, although densities of 0.7 g/cc to 1.4 g/cc are also available. The various densities are achieved by varying the degree of compaction during the calendering operation to form the foil. Because the density of flexible graphite is a major criterion for its quality, the user should be cautious when cheap material is offered. Often it proves to be inferior material of low density. Manufacturers of flexible graphite always emphasize its excellent load-bearing characteristics. While this is basically correct, the user should take into account that material of standard 1.0-g/cc density has a compressibility of about 40%. For the nominal thickness of 2.0 mm (0.080"), the effective thickness after

Page 106

Figure 12 Crystal structures of graphite and diamond.

Page 107

Figure 13 Process for manufacturing flexible graphite.

Figure 14 Expanded graphite flakes, 50×.

Page 108

fitting is about 1.2 mm (0.047"). High-grade compressed-fiber materials of the same effective thickness also have excellent loadbearing capacities, very similar to flexible graphite, when the nominal thickness is about 1.3 mm (0.050") (depending on actual compressibility). The conclusion is as follows: Comparable Thicknesses Flexible graphite High-grade compressed fiber (density 1.0 g/cc) (compressibility 8%) 3.2 mm (1/8") 2.1 mm 2.0 mm 1.3 mm 1.6 mm (1/16") 1.05 mm 1.0 mm 0.65 mm 0.8 mm (1/32") 0.5 mm

Because of its high compressibility, the conformity of flexible graphite to flange surfaces is excellent. This leads to good sealability even when flanges are badly corroded, damaged by chisel, or distorted. During the early stages of the use of flexible graphite, manufacturers stated a maximum temperature in application higher than 1,000°C (1832°F). This was definitely overstated and has been withdrawn. It is now common among manufacturers to recommend normally in nonoxidizing atmosphere up to 450°C (840°F). But even this may be too optimistic. In long-duration sealability tests, the leak rate increases unacceptably, even at a continuous 400°C (750°F) application. This is most probably due to oxidizing effects of the environmental atmosphere. When the test rig is placed into an inert environmental atmosphere, the leak rate increase is suppressed. In December 1995, TTRLEcole Polytechnique of

Montreal published a report [7] to the Pressure Vessel Research Council (PVRC) indicating in part a limitation of 315°C to 340°C (600°F to 650°F) for a five-year exposure. Most flexible-graphite gasket material is not offered as pure graphite but as laminate. There are mainly two reasons for lamination: 1. Because of its low strength, the graphite sheet, and more particularly the cut gasket, might be damaged through handling. Therefore, it is reinforced by bonding it to both sides of a metal sheet. 2. Because it is difficult and uneconomical to produce graphite foil above 1-mm thickness, thicker sheets are achieved by bonding two or more sheets on top of each other. We call such sheets homogeneous laminate. Bonding the graphite to a plain metal insert or to another graphite sheet requires a good adhesive, applied under controlled conditions. A special process is needed that allows a very even distribution of an adhesive quantity of less than 5 g/m2.

Page 109

Only by this can the content of organic compounds in the graphite laminate, which might act as a lubricant, be kept below 1% of the total sheet. Otherwise the stress relaxation would be affected negatively. The only purpose of the adhesive is to provide a temporary bond between the individual layers until the gasket is fitted between the flanges. Graphite foil can also be bonded onto both sides of a pegged steel insert with or without the use of adhesive. The standard insert is stainless steel, but other inserts are available as well. This physically bonded laminate can stand a 10% higher stress compared with an adhesive-bonded plain insert. Also, laminates with two inserts of plain stainless steel are offered that provide much higher stiffness and are recommended for larger gaskets [above 500 mm (20")]. Flexible graphite tends to stick to flange surfaces. Therefore, it is advisable to take advantage of the antistick surface treatment offered by some manufacturers. Properties of Flexible Graphite Advantages Excellent load-bearing capacity Good sealability Excellent conformity to flange surfaces Good general chemical resistance

Disadvantages Poor handlability Relatively high expense

4.4. Rubber-Coated Metal Sheet Continuous lengths of surfaced-treated metal, normally carbon steel, are coated with rubber latex or rubber solution by means of a roller device or by spraying. Because the water of the latex or the

solvent of the solution has to be evaporated and the rubber has to be vulcanized, long heating tunnels are needed. These can be either horizontal or vertical. If the heating tunnel is horizontal, a lot of floor space is used and the metal sheet can be coated on one side only during each run. When the heating channel is placed vertically, it is very high and needs larger-than-normal building height. The vertical configuration does offer the opportunity to coat the metal sheet continuously on both sides in one run. Additional surface treatments become possible, such as the application of adhesives and protective paper. Typical configurations are acrylic or NBR rubber on cold-rolled steel, with an overall thickness from 0.4 mm to 1.6 mm (1/64" to 1/16"). The thickness of rubber is from 0.03 mm to .20 mm (0.001" to 0.008") per side.

Page 110

Rubber-coated metal sheet is also known as soft metal or silent metal, which also indicates its major applications. It is a basic material for gaskets, for instance, beaded single- or multilayer gaskets. Another popular application is for damping or absorption of sound, for example, in disc brake pad assemblies. Properties of Rubber-Coated Metal Sheet Advantages Disadvantages Continuous process Complexity of gasket design Reasonable cost High cost of gasket-cutting tools Good handlability

4.5. PTFE-(Polytetrafluoroethylene-) Based Materials 4.5.1. Overview PTFE is formed in an expensive process by polymerization of its monomer tetrafluoroethylene(TFE). The molecular structure (Fig. 15) shows that the carbon backbone of PTFE is totally covered by fluorine, thus hindering the access of other aggressive media that might destroy the backbone bonds. Additionally, the bond between carbon (C) and fluorine (F) is extremely strong compared with others, so the replacement of the fluorine is practically excluded. This is why PTFE has such good chemical stability. There are only a few ways to destroy PTFE: Chemically by: Molecular fluorine (F2) or other fluorinating media Molten alkalies Physically by:

Heating over 400°C (750°F)

Figure 15 PTFE molecular structure.

Page 111

High-energy radiation (i.e., nuclear radiation) Both of these methods attack the chain length of PTFE, which is shortened, and monomers TFE are released. These examples show the limitations of PTFE-based gaskets. Users should be aware that in case of a fire, PTFE (which does not burn) can release fumes based on TFE or its reaction products with the atmosphere. These fumes may cause health problems similar to severe influenza. Because even minor contamination with PTFE can cause this problem, workers handling PTFE must not smoke during work and should not have cigarettes or other smoking gear in the working area to avoid contamination. It is strongly recommended that users carefully study the safety data sheets supplied by the manufacturer. PTFE has the following properties: It does not burn. It is stable to light. It does not absorb water; it is hydrophobic. It is physiologically safe. It is insoluble in all solvents, even at increased temperature. It is inert to all media except those mentioned earlier. It has excellent electrical insulating capacity, even at high humidity. Its thermal conductivity is low at about 6 × 10-4 cal/cm × sec × degrees

Due to the highly symmetric molecular structure, as shown in Fig. 16, it has high crystallinity of 9095%. All properties mentioned so far should make PTFE a very good gasket material, but unfortunately there are two unfavorable properties that are very much to its disadvantage: Low strength High creep, caused by a transition point at 19°C (65°F) Because the number of ambient- or low-temperature gasket applications are limited, the usage of this material for gaskets is limited. For this reason, there have been countless trials to improve strength and creep by incorporating into the material inorganic fillers like glass fiber, carbon fiber, precipitated silica, graphite, carbon black, and bronze powder. Success was rather limited, because nothing really changed the basic properties of PTFE, and the slightly improved creep resistance was obtained at the price of reduced chemical compatibility. This material thickness ranges from 0.8 mm to 6.4 mm (1/32'' to 1/4''). During the past 20 years, two processes have been developed to improve the strength of PTFE and reduce its creep. These are discussed in the next two subsections.

Page 112

Figure 16 Molecular conformation of PTFE.

Page 113

4.5.2. Expanded PTFE As indicated in Section 4.5.3, biaxial-oriented PTFE might have too low a compressibility, depending on the grade, which might cause a problem in low-temperature applications. Therefore, ways were found to produce a soft PTFE, so-called expanded PTFE. The production process takes place in five steps: 1. PTFE powder containing a defined quantity of hydrocarbon as lubricant is ram-extruded into a solid band. 2. The hydrocarbon lubricant is evaporated. 3. The PTFE band is stretched (i.e., expanded). 4. The PTFE band is minimally sintered (steps 3 and 4 at the same time). 5. Slow cooling and settling. By stretching, the PTFE gets oriented in the stretch direction and becomes soft, basically because its porosity increases. The dimension of these bands depend on the size of the nozzle of the ram extruder. Usually an adhesive tape is attached on these bands so that they can be easily attached to flange surfaces. They offer a simple solution to many sealing problems, especially when maintenance personnel are pressed for time and the correct dimension of a sheet-based gasket is not available. (Figure 17 shows expanded PTFE bands being applied.) With this type of gasket application, no accurate dimensioning is necessary. Sizes range from a rectangular cross section of 1 mm × 3.2 mm (0.040" × 0.125") to 8 mm × 25 mm (0.312" × 1"). Basically the same procedure is applied to produce sheet material

[8]. In this case, the nozzle used during extrusion produces a wide band or sheet that

Figure 17 Applying PTFE bands.

Page 114

is cut to size and then, in an additional step, calendered alternately in each direction until the desired thickness of 0.40.6 mm (0.015"0.024") is achieved. By this calendering process, the tensile strength increases strongly due to fibrillation of PTFE. The sheet is then expanded and kept so during sintering at approximately 340°C (645°F). Properties of Expanded PTFE Advantages Disadvantages Excellent chemical resistance Expensive Excellent sealability High compression and creep High compressibility Low recoverability after compression High tensile strength

4.5.3. HS10 Process for the Manufacture of Biaxial-Oriented PTFE The HS10 process was developed and patented in 1960 by Du Pont. Unlike the expanded PTFE described in Section 4.5.2, biaxial-oriented PTFE is generally reinforced with an inert filler. In 1986, David G. Lingard [9] reported on gasket material produced by this process and its favorable properties compared with conventional PTFE (see Section 4.5.1). The process consists of seven steps: 1. Deagglomeration 2. Filtering 3. Calendering 4. Drying 5. Compacting

6. Sintering (curing) 7. Cooling The deagglomeration process involves dispersing the resin in a hydrocarbon liquid, using a high-speed mixer. The first step provides an excellent way to disperse the filler material evenly. The resultant slurry is then filtered to remove most of the hydrocarbon liquid. The filter cake is then passed through a set of calender rolls. The filter cake is rotated 90° each time it is passed through the calender, as shown in Fig. 18. The small amount of the hydrocarbon liquid still present in the filter cake acts as a lubricant and helps the PTFE resin to become oriented as the filter cake is calendered. At this point in the process the unsintered sheet has excellent strength in both the transverse and longitudinal directions. The calendering process has caused the PTFE resin to shear under the applied force. The resin has fibrillated, producing a green sheet that not only is biaxially

Page 115

Figure 18 Filter cake is rotated 90° each pass through the calender.

oriented but in which the fillers are homogeneously dispersed. This sheet is then dried, sintered, and cooled. Gasket materials produced according to the HS 10 process exhibit rather favorable and superior properties. As development work continues, the outlook for the next few years is rather promising. In this respect the properties of biaxial-oriented PTFE might change to its advantage even more. Current Properties of Biaxial-Oriented PTFE Advantages Good tensile strength (similar to compressed fiber

Disadvantages Very high

material)

(table continued on next page)

cost

Page 116

(table continued from previous page) Advantages Good creep resistance (depending on the grade, it might be better than high-grade compressed asbestos) Excellent sealability

Excellent chemical resistance

Disadvantages Rather low compressibility (depending on grade) Temperature limitation 250°C (480°F) Not fire-safe (exceptions, see Section 3)

Excellent insulationability (electrical and thermal) Potential for being fire-safe (depending on manufacturing and grade) Good handlability

4.6. Special Sheet Products 4.6.1. Mica Laminates When manipulating a piece of mica, we immediately notice a special characteristic: Mica mineral is easily split into thin layers with a shiny glitter. Chemically, mica is primarily aluminosilicate; but there are different types that, though differing only slightly in chemical composition, have distinct differences in properties, such as heat resistance. Mica laminate is produced using a process developed during the 1950s. According to manufacturers' literature [10] this is done in two basic steps.

1. Splitting mica mineral by high-pressure water jet into small, thin layers. 2. The resulting slurry (called pulp) is dewatered on a papermaking machine, and mica laminate is formed without the addition of any kind of binder. The advantages of this process is that neither the physical nor the chemical structure of mica is changed, and all mica minerals can be processed in this same way. For gasket application, mica laminate is usually physically attached to pegged or plain metal sheet, similar to flexible graphite. It can also be impregnated with resin, but then it is used mostly in the electric industry as an insulator. Properties of Mica Laminate Advantages Disadvantages Excellent heat resistance Low sealability Absolute incombustibility Practical incompressibility

(table continued on next page)

Page 117

(table continued from previous page) Advantages Disadvantages Good chemical resistance, including to acids, solvents, oils Good tensile strength Good flexibility Exceptional dielectric strength

Due to these properties, mica laminate is the best choice for lowpressure, high-temperature applications, such as exhaust pipe and flue vent sealing, strong acids, and molten metal. 4.6.2. Sandwiched Composite Sheets There are countless sandwiched composite sheet materials and gaskets. They are designed to combine the positive properties of the basic materials, either to achieve synergetic effects or to cover weaknesses of one material. One possible combination that should make sense was offered years ago by a major manufacturer but has not achieved the expected market acceptance. The material in question is basically a compressed-fiber material with a thin foil of flexible graphite on both sides. With this, the excellent surface sealability of flexible graphite, due to its high compressibility, is combined with the good core sealability of compressed-fiber sheet, especially at higher temperatures. Also, the susceptibility of flexible graphite to crush is covered by the high-grade compressedfiber material. 4.7. Special Gaskets 4.7.1. Gaskets Made from Segments

There are two basical reasons why gaskets are manufactured in sections: The gasket dimensions are too large to be produced with sheet material. It is not economical, because off-cuts and centers are too large and can't be used for other purposes. When making gaskets from smaller pieces or segments, the critical component is the joint. Manufacturers have developed two safe ways to make such a joint.

Figure 19 Ends of gaskets are tapered for better bonding.

Page 118

By adhesive bonding By mechanical bonding or dovetail coupling For effective adhesive bonding, the ends of gasket segments are tapered, as indicated in Fig. 19. The slope areas obtained by this operation are coated with a thin layer of adhesive and pressed together by force. It is not uncommon to manufacture gaskets of 3.5-m to 4.5-m (12-ft to 15ft) diameter by this procedure. In these instances, one interface is left open for easy transport, because it is possible to roll up the gasket. The last joint is struck by the assemblers on site. For mechanical coupling, the ends of the gasket segments are cut as shown in Fig. 20. Both methods are technically equivalent, but not all sheet materials are suited for this. Not suited Flexible graphite All reinforced materials (with wire mesh, expanded metal, plain metal sheet)

Suited Compressed-fiber materials (asbestos and nonasbestos) Beater-addition materials (asbestos and non-asbestos) All types of PTFE

4.7.2. Eyeleted Gaskets In some countries, authorities demand special protection against blowout for the sealing of critical or dangerous media. There are three possibilities to achieve this: 1. Tongue-and-groove flanges

Figure 20 Ends of gaskets are cut in segments.

Page 119

2. Spiral-wound or other semimetallic gaskets 3. Eyeleted gaskets First-priority choices are the first two cases, but very often the flange connection does not allow these two solutions. Therefore, gasket manufacturers offer the additional third possibility. In this case, a normal ring gasket made of compressed-fiber materials (asbestos or nonasbestos) or flexible graphite is reinforced by a Ushaped inner ring of metal (the eyelet) that reinforces the contact surface of the gasket. (Fig. 21). There may be a thin PTFE tape between the eyelet and the gasket. Eyeleted gaskets are basically unlimited in dimension. The only limitation is the size of the basic gasket, for instance, 1.5 m (60") for flexible graphite (maximum sheet size). Gaskets made in segments may not be strong enough to support the eyelet. Advantages Disadvantages Blowout protection Price Improved sealability due to local high stress Limitation to under the eyelet ring form Protection against chemical attack

4.7.3. PTFE-Enveloped Gaskets For noncritical temperatures but critical media, it may not be economically realistic to use expensive expanded PTFE (see Section 4.5.2) or high-grade

Figure 21 Eyelet gaskets: (1) Metal eyelet; (2) PTFE foil 0.2 mm (not essential); (3) flexible graphite, beater-addition materials, compressed synthetic fiber, or compressed asbestos fiber, of thickness 23 mm.

Page 120

biaxial-oriented PTFE (see Section 4.5.3). In these instances, PTFE-enveloped gaskets may be used. There are two ways to manufacture these: The gasket is wrapped by suitable PTFE tape with an adhesive face. The inner diameter and the two gasket surfaces are protected by a U or Y-shaped PTFE envelope. This type of gasket is often used in electrolysis plants, in which the media is strongly alkaline and the gasket must be totally insulated. Advantages Good chemical resistance Good sealability

Disadvantages Reduced load-bearing capacity Temperature limitation to about 250°C (480°F)

Nonconductivity

4.8. Semimetallic Gaskets 4.8.1. Spiral-Wound Gaskets Conventional spiral-wound gaskets are manufactured by spirally winding, under pressure, a preformed V-shaped metal strip and a soft nonmetallic filler on the outer periphery of metal winding mandrels. The outside diameter of the mandrel forms the inner diameter (ID) of the gasket, and the laminations are continually wound until the required outer diameter (OD) is attained (see Fig. 3). Normal practice is to reinforce the inner and outer diameters with

several plies of metal with no filler. This gives greater stability to the gasket and provides better compression and sealing characteristics. Compressibility can be controlled by varying the ratio of nonmetallic filler to metal plies for a given width and density, and by varying the pressure on the winding mandrel. A wide range of metallic winding strips is used, based on service conditions. These include, but are not limited to: carbon steel, most grades of stainless steel, Monel metal, nickel, titanium, hastalloy, and Inconel. The most common nonmetallic fillers are flexible graphite and PTFE. Ceramic, compressed asbestos, and compressed nonasbestos fillers are also used. The majority of spiral-wound gaskets are manufactured to national size/pressure rating standards. The most common standard is ASME B16.20. This standard also includes a color-coding system for material. For low-seating-stress applications (i.e., ASME 150-lb flanges), the required gasket stress can be lowered by reducing the width of the metal strip,

Page 121

allowing part of the nonmetallic filler to be unrestrained by the metal strip. In many instances, outer and inner retaining rings are required. The sizes are determined in accordance with the flange and gasket specifications. Normally, retaining rings are machined from sheet carbon steel, either by conventional methods, such as trepanning, or by newer technology, such as water-jet or laser-jet cutting. For large diameters, they are formed from flat strip, a process known as ring rolling. The ends of the strip are joined by welding. The outer retaining ring is fitted manually using a softimpact tool. For diameters up to 200 mm (8'') the sealing element can be manually fitted to the inner rings. For larger diameters, winding is carried out directly onto the inner ring to ensure stability. 4.8.2. Rigid Laminated-Graphite Gaskets This type of gasket consists of exfoliated (expanded) graphite layers bonded to both faces of a steel core (see Fig. 4). The standard exfoliated graphite thickness is 0.020'' (0.5 mm), and this can be bonded to a steel core ranging from 1/16" (1.6 mm) to 1/8" (3.2 mm). The graphite components are cut by conventional blade/plywood tool cutting methods, and the steel by trepanning, stamping, or laser-jet or water-jet cutting. The graphite is bonded to the steel using a high-quality elastomer-based adhesive, evenly distributed and thinly applied. Additional processing may be carried out by the manufacturer, including crimping and densifying. A modification to this design is to bond flexible graphite rings to the steel core. Generally with this type of design, the steel will be corrugated. 4.8.3. Single- or Double-Jacketed Gaskets

These gaskets are formed by wrapping a core of nonmetallic gasket material with a soft metal cover (see Fig. 5). Nonmetallic rings are cut from sheet material using conventional steel/plywood cutting tools, and metal rings are cut or cut and formed using simple cutting blank-and-form tools. Assembly of the nonmetallic element into the steel sleeve and press-closing of the metal completes the assembly. The nonmetallic component is normally a compressed-fiber or beater-add material. Aluminium, copper, soft steel, stainless steel, and Monel can all be used for the metallic component. Selection is by chemical compatibility of the metal to the medium being sealed, because the nonmetallic component is protected from the medium. This type of gasket is highly resistant to blowout. 4.9. Metallic Ring Gaskets The manufacture of metallic ring gaskets is unremarkable, using conventional metalworking equipment (see Fig. 6).

Page 122

References 1. Nonmetallic Flat Gaskets for Pipe Flanges. An American National Standard. ASME B 16.21. The American Society of Mechanical Engineers, New York, 1992. 2. Metallic Gaskets for Pipe FlangesRing Joint, Spiral Wound, and Jacketed. An American National Standard. ASME B 16.20. The American Society of Mechanical Engineers, New York, 1993. 3. 1990 Federal Clean Air Amendment for HON (Hazardous Organic Air Pollutants). 4. Bickford, John H. An Introduction to the Design and Behavior of Bolted Joints, 3 ed. Dekker, New York, 1990. 5. 1995 ASME Boiler and Pressure Vessel Code, Section VIII, Division 1, p. 1. American Society of Mechanical Engineers, New York, 1995. 6. 1995 ASME Boiler and Pressure Vessel Code, Section VIII, Division 1, Rules for Construction of Pressure Vessels. American Society of Mechanical Engineers, New York, 1995. 7. Tightness Testing and Research Lab. Ecole Polytechnique of Montreal Final Report of Combined PVRC Projects 91-8 and 93-3. 8. U.S. Patent 4, 187, 390. 9. David G. Lingard. PTFE-Based Gasketing Materials. Valve Manufacturers Association of America. (Presented in the Asbestos Substitute Gasket and Packing Materials Seminar in Houston, Texas, August 67, 1986.)

10. Compagnie Royal Asturienne des Mines Devision Cogebi Huysmanslaan 65 Lot BEERSEL (Belgium).

Page 123

4 Chemical Gaskets JOHN COCCO Loctite Corporation, Rocky Hill, Connecticut 1. Introduction Anaerobic and room-temperature vulcanizing silicone gasketing materialsboth the formed-in-place kind and the cured-in-place kindhave produced leakproof seals that have lasted the life of millions of machines and vehicles. Gaskets prevent leakage of liquid or gases by forming impervious barriers between mating flanges. Fluid seals are divided into static and dynamic systems, depending on whether or not the parts move in relationship to one another. Flanges are classified as static systems, although they move because of vibration, temperature and/or pressure changes, shock, impact, etc. There are three types of flange gaskets: Conventional compression gaskets of cork, paper, rubber, metal, and other asbestos-free materials Cured-in-place liquid compression gaskets cured in seconds with ultraviolet light prior to assembly (CIP) Formed-in-place liquid gaskets cured after the parts are assembled (FIP)

Page 124

They all must perform four functions: Create seals Maintain seals Remain impervious to fluid flow Remain compatible with the machinery Many factors influence gasket choice. FIP silicone gaskets are well suited for low-pressure joints with large gap potential, such as stamped metal cover plates; FIP anaerobic gaskets seal highpressure joints when both surfaces are rigid. Cured-in-place gaskets are ideal for sealing joints that may be frequently serviced. 1.1. Typical Applications of Chemical Gaskets Chemical gaskets have found widespread applications in all sectors: consumer, automotive repair, and industry. Table 1 lists examples of applications where various chemical gasketing materials have been specified successfully. Many others, too numerous to detail, can be discovered when disassembling parts. 2. Formed-in-Place Gaskets Formed-in-place (FIP) gaskets begin as liquids applied to one of the flange joint surfaces. When parts are assembled, the FIP material flows into voids, gaps, and scratch marks, forming a durable seal after cure. The concept offers a convenient way of manually or automatically dispensing complex seals. Two common types of FIP materials are room-temperature vulcanizing (RTV) silicone and anaerobic compounds. TABLE 1 Successful Applications of Chemical Gaskets

Application Oil pan gasket

Gasket type Silicone formed in place Silicone cured in place, Removable conduit cover gasket compression Compressor housing flange Anaerobic formed in place gasket Gearbox, engine casings Anaerobic formed in place Stamped metal housings Silicone formed in place Die-cast housing and casings Anaerobic formed in place

Page 125

3. Advantages of FIP Gaskets Over Precut Compression Gaskets Formed-in-place gaskets offer the following four advantages. Improved Reliability Seal all surface imperfections Allow true metal-to-metal designs Reduce compression set and fastener loosening Anaerobic type can add structural strength to assemblies Unitize assembly to improve torque transmission between bolted flange joints Reduced Costs Allow for relaxed machining tolerances Eliminate gasket inventories Reduce labor costs with automatic application Eliminate bolt retorquing needed with conventional gaskets Allow use of small fasteners and lighter flanges Easier Application Single-component type requires no mixing Applied semi- or fully automatically Vertical and horizontal applications possible Easier Service

Offer easy disassembly and cleanup Service packages offer multiple applications on flanges of any size The decision to use anaerobic versus silicone FIP gasket material involves several issues. Anaerobics are generally used on rigid or stiff joints, such as cast aluminum or iron. Typical applications include pumps, engines, and transmissions. These joints generally have little movement, in comparison with a joint using a stamped steel or plastic molded cover. Typically, a silicone RTV is used for such high-movement joints. 3.1. Part Design/Performance Issues Compression gaskets require an initial compressive load to deform the gasket into the irregularities of the flange surfaces. They therefore must carry the bolt

Page 126

load. The major causes of failure and leakage of compression gaskets are as follows. Compression Set Over time, the gasket loses its elastic properties and becomes less resilient. The load on the gasket and relative motion causes a general decrease in the thickness of the gasket, with subsequent leakage. Flange Bowing The area where the joint is most likely to leak is at the center of the flange, where the smallest compression is produced by the bolts. This is where the maximum separation occurs from internal pressure. This phenomenon is often called flange rotation. Blowout Gaskets can be blown out of flanges by a combination of low compression (less than the minimum sealing stress) on the gasket and internal pressure of the system. Bolt Hole Distortion High stresses are transferred to the gasket material under the bolt head, causing the gasket to crack, tear, rupture, or extrude. 4. Anaerobic Formed-in-Place Gaskets Anaerobic formed-in-place materials cure in the absence of air and in the presence of metal or other active surfaces. Cure rates at room temperature range from a few minutes to several hours. Since there are no solvents, the conversion from liquid to solid is virtually

100%, so the voids surface imperfections, and tool marks are completely filled, thus eliminating potential leak paths. 4.1. Advantages of Anaerobic Gaskets Anaerobic gaskets offer numerous benefits over traditional sealing systems. No Gasket Relaxation Metal-to-metal contact ensures proper bolt tension throughout the life of the assembly. No retorquing is required. Nonshimming Anaerobic gaskets allow flanges to join with metal-to-metal contact. No allowance is needed for gasket thickness, so tolerances can be more accurately maintained.

Page 127

Structural Strength Anaerobic gaskets offer high shear strength that can be used to stop movement due to side loading. This eliminates bolt loosening and fretting between flanges and increases assembly strength. Excess Material Remains Liquid Unlike other liquid sealants, anaerobic gaskets cure only between flange faces. Excess material is wiped away from exterior surfaces or flushed away from interior surfaces (liquid anaerobics are miscible in most fluids). Passages and channels will not be blocked. Relaxed Surface Finishes Anaerobic gaskets allow relaxation of surface finish and flatness tolerances. Scratches and scored surfaces can be sealed without rematching. No Cure Before Assembly Because anaerobic gasket materials cure in the absence of air, they offer limited on-part life when exposed to air. This enables multiple application methods and reduces the housekeeping problems found with the use of evaporation- and/or moisture-cured materials. Reduced Inventory Costs Precut gaskets can be used only on specific flanges. They require careful storage and handling. Large stocks of precut gaskets can create significant purchase and inventory costs. Chemical Compatibility

Cured anaerobic gaskets demonstrate excellent solvent resistance to petroleumbased fuels, lubricating oil, water/glycol mixtures, and most other industrial chemicals. 4.2. Disadvantages of Anaerobic Gaskets While anaerobic gasketing offers many design advantages over conventional alternatives, they do have the following constraints. Flange Movements Anaerobic FIP gaskets will help to restrain flange movement from differential thermal expansion or applied loads. However, if lateral forces exceed the shear strength of the material or the separating force exceeds the tensile strength of the product, the long-term sealability of the assembly will be jeopardized, as it

Page 128

would be with most conventional gaskets. This relative motion must be controlled through design. Temperature Anaerobic gaskets are thermosetting plastics, and the typical operating temperature range is -50°C (-58°F) to +150°C (+302°F); some materials can sustain temperatures up to +204°C. 5. Anaerobic FIP Gasketing Design Considerations Review of successful designs utilizing anaerobic FIP gasketing indicates that critical joint design parameters fall within the following practical rules (depending on design, materials, etc.): 5.1. Bolt Span In flanged joints, bolts should typically not be spaced more than 10 bolt diameters apart. For example, for a nominal bolt diameter of 10 mm, the maximum bolt span should be no more than 100 mm. The practical rules for relating bolt span to flange width are shown in Table 2. 5.2. Surface Finish In addition, reliable sealing of an FIP joint depends on adhesion to the flange surface. Therefore, surface finish is critical to joint integrity. Thermal shock and/or severe structural loading may shear anaerobic FIP gasketing material from very smooth surfaces ( 0.5 is suggested as a guide for the acceptance and consideration of candidate asbestos-replacement gasket materials for predicting their long-term service temperature, Ts, and for further evaluation of their long-term service tightness suitability. The MTI scheme and proposed fiber-reinforced elastomer sheet specifications require that the average Qp > 1, with the minimum Qp > 0.5. From Fig. 37 it is seen that most nonasbestos materials might be considered for applications where Ae lies between 25 and 50. For 25 < Ae < 50, the calculated 10-yr-life temperature Ts is between 350 and 400°F. This performance is well below that of a true asbestos equivalent. 10.4.2. Acceptance for PTFE-based Products As usually accepted, there is no time effect on these materials, and service temperature predictions are made from plots of Qp vs. ATRS test temperatures

Page 260

Figure 37 Qp-Ae plot for elastomeric fiber-reinforced sheet gasket products.

alone. Therefore, the scatter introduced in the quality plots results only from the experimental error attached to each Qp point. For these materials, the use of the lower Ts' at Qp = 1 for the recommended service temperature Ts may seem too stringent for PTFE for joints with confined gaskets, based on experience. Because applications with a confined PTFE gasket are not rare, it is reasonable to consider that for these materials, Ts is given at Ts" for QP = for confined applications and Ts = Ts' for Qp = 1 for unconfined applications. Figure 38 shows a plot of Qp vs. temperature for various PTFEbased materials. This plot indicates that service temperatures Ts' for unconfined applications (based on a Qplimit of 1) and Ts" for confined applications (based on a Qplimit of 0.5) are significantly higher for a new PTFE-based products now on the market (Product

C shown respectively as solid circles and squares) than for a traditional PTFE sheet (Product A). 10.4.3 Acceptance for Flexible-Graphite Sheet Products For flexible-graphite sheet materials, the minimum residual HATR tensile strength, TSX, is close to 1000 psi (6.89 MPa) for unreinforced specimens, and it has been established that 60% stress relaxation is an acceptable threshold for a safe leakage performance of these materials [34]. For a 60% stress relaxation

Page 261

Figure 38 Qp-T (temperature) plot for PTFE-based gasket products.

and a TSX of 1000 psi (6.89 MPa), the value of Qp is close to 0.3 (see Eq. 14). Therefore, in a screening application for flexiblegraphite sheet candidate, the mechanical quality, Qp should remain greater than 0.3 for this candidate to be considered acceptable. Figure 39 plots the data of HATR tests performed on flexiblegraphite sheet materials in the form of Qp vs. Ae. This figure shows that the metal-foil- or tangreinforced materials perform better than their unreinforced counterparts, up to values of Ae around 90110. These materials might be considered safe for applications where Ae is less than about 100. The corresponding 5-year life extrapolates to 600°F. For materials with passivating corrosion and oxidation inhibitors, the recommended Ae is close to 130, corresponding 5year life at 630°F (from Eq. 19). Since these calculated temperatures are significantly below commonly accepted values for these materials, it is important they be confirmed with more

specific tests if they have to be recommended as long-term service life temperatures. 10.5. Recommended Versus Calculated Service Life: a Word of Caution The selection of a 5- or 10-year life is illustrative only, and it is not intended to sanction a particular service life. Plant owners and operating companies may

Page 262

Figure 39 Qp-Ae plot for flexible-graphite-based sheet gasket products.

have different requirements for a probable service life. The calculated life from Qp and Ae does not intend to impose any particular life requirement. Rather, it hopes to provide methods that permit the user to make more rational judgments to estimate the long-term service behavior of a gasketing product. The main point is to show that a basis for performance prediction is available. It follows directly from correlations with ATRS/HATR results with an equivalent aging parameter, Ae, and carries the assumption that similar damage is inflicted from the equivalent exposure. Ae relationships are possible mathematical combinations of temperature and exposure time found by best fit of test results of ATRS/HATR stress relaxation and residual tensile strength obtained in a screen test program performed to qualify a particular gasket product. The long-term gasket service temperature predictions that are extrapolated from the use of equivalent aging parameters have been

derived from laboratory test data that was obtained under realistic flange loads and flexibility with accelerated thermal aging for oxidizing services. They are not fluid or process or equipment design temperatures. It is recognized that the gasket temperature in an application may differ substantially, depending on items such as heat loss considerations, the fluid type and velocity attacks the gasket material, the size of the flange, whether the flange is insulated or not, and whether the installation is indoors or outdoors. Therefore, the true service life of a joint will depend on the average cumulative effect of time at temperature determined by the joint working conditions at the gasket location. These conditions are not always well known, and it

Page 263

is often required that the service temperatures predicted by the mechanical screening tests be confirmed and/or supplemented by elevated-temperature leakage tests. 10.6. Recent Developments in Long-Term Performance Predictions The main purpose of gasket aging tests performed with simple mechanical devices such as ATRS or ARLA fixtures (Section 5) and aging qualification tools developed in the present section is to estimate the maximum safe operating temperature that a gasket material can sustain for a long period of time, typically 310 years. As a result, a first gasket qualification test scheme based, in part, on short-duration mechanical screen tests was published in 1990 [15]. This scheme, presented in Section 12.4, is the basis of the technology used today to predict long-term operating temperatures of many types of sheet gasket materials. Since then, research effort has been continuously pursued to improve the confidence level of the long-term temperature predictions. 10.6.1 Thermal Endurance Graphs for Elastomeric FiberReinforced Gaskets Recently, a major PVRC test program [49] was completed to study the change in the properties of compressed elastomeric fiberreinforced sheet gasket materials subjected to temperature exposure for periods of up to 1 year. This work had two objectives: 1. Verify the validity and the accuracy of the long-term gasket life predictions made for these products and based on mechanical screening tests and qualification tools like Qp and Ae. 2. Improve and refine the elastomeric sheet gasket qualification protocol, especially for new gasket materials that will likely appear

on the market in the future. To achieve these two goals, long-duration (up to 1 year) air and low-pressure dry steam screening tests were performed on six different gasket elastomeric fiber-reinforced sheet materials (three aramid, one carbon, one graphite, and one asbestos fiber sheets) at temperatures ranging from 420°F (216°C) up to 700°F (371°C). Through the results of these tests, a better understanding of the effect of thermal degradation on the properties of elastomeric sheet gasket materials was gained. It was also found that the change of gasket properties such as weight loss and leak rate can be accurately represented and predicted by the Arrhenius model. With this new knowledge, it became possible to define guidelines that should be used to improve the qualification protocol for this class of gasket

Page 264

materials. In Fig. 40, test results are plotted for one typical compressed-aramid-fiber sheet. Similar plots were obtained for the other tested materials. From these plots, the important findings [50] can be summarized as follows: Continuous and interrupted exposure tests result in similar gasket property changes. This allows for the cost-effective technique of interrupted monitoring of the gasket property changes at regular time intervals during a mechanical screening test. Temperature and aging atmosphere do not influence significantly the load relaxation behavior (Fig. 40a). On the basis of similar gasket weight losses (Fig. 40d), short-duration tests performed at high temperature can be used to predict the load relaxation for long-term exposure at lower temperatures. The effect of temperature on gasket weight loss varies with the material, and for the same material it can also vary with the exposure time (Fig.

Figure 40 Long-duration ARLA test results for a typical compressed-aramid-fiber sheet material.

Page 265

40b). However, for a given test temperature, the effect of time can be extrapolated with reasonable accuracy up to a certain extent. Mass leak rates measured at room temperature at regular intervals of time are plotted as a function of exposure time (Fig. 40c). The early stage of exposure is generally characterized by a decrease in the gasket leak rate, as compared to its leak rate immediately after bolt-up. Subsequent exposures lead to a continuous increase in the leak rate, and the final leak rates were above those measured before aging. As was discovered in previous work [38], a strong relationship exists between gasket weight loss and leak rate (Fig. 40e). Gasket weight losses and leak rates resulting from steam-exposure tests are much lower then those obtained from air-exposure tests (Fig. 40c), but the gasket degradation can be highly intensified in the presence of a small amount of air. Since the outside edge of a gasket is usually exposed to the surrounding air, the use of airaging tests in a qualification protocol is more appropriate, because it will ensure safe long-term predictions. The thermal degradation of sheet gasket materials in air atmosphere obeys the Arrhenius model generally used to predict the long-term thermal endurance of numerous organic materials. The Arrhenius equation is: t = AeB/T (20) where: t

= time to a fixed level of degradation (or time

to failure) T=absolute temperature in °K A=frequency factor, h-1 B=energy of activation e =base of natural logarithms Eq. (20) may be expressed as a linear function by taking the logarithms as follows: (21) According to Eq. (21), data from tests performed at different temperatures may be plotted on a log (time) vs. 1/T graph, and a straight line should join the points. Such a graph is called a thermal endurance graph of a material, to plot it, a fixed level of degradation must be selected. The energy of activation, B, is the slope of the endurance lines on this graph.

Page 266

Thermal endurance graphs were obtained for several fixed levels of both gasket weight loss and leak rate (see Figs. 41 and 42). They demonstrate that the degradation process for most elastomeric fiber-reinforced sheet gasket materials obey the Arrhenius model, since, over a temperature range, a linear relationship exists between log (time) and 1/T at different weight loss or leak rate levels. A striking similarity exists between the two sets of thermal endurance graphs. Leak rate is a better candidate than weight loss for generating precise thermal endurance graphs for elastomeric fiber-reinforced sheet materials because, of these two properties, leak rate variations during an ARLA aging test are more significant than weight loss changes. To plot a thermal endurance graph based on gasket leak rate, an acceptable leak rate value for long-term operation must be selected. A tightness criterion has been proposed for sheet gasket materials based on the maximum leak rates obtained with reference asbestos gaskets submitted to elevatedtemperature tightness tests (see Section 10.2). This criterion is equivalent to a leak rate of 3 mg/sec for a 150-mm-OD gasket pressurized with helium gas at 800 psig (5.5 MPa), or 1.2 mg/sec for the for the smaller ARLA gasket specimen (2.31 in., or 58.7 mm, OD). There are other possible tightness criteria that could be used to meet, for instance, fugitive emission requirements that would result in different allowable gasket leak rates. To demonstrate that the thermal endurance graph of a gasket can be obtained for various leak rates, three different levels of leak rate (0.15, 0.5, and 1.2 mg/sec) were selected to plot the thermal endurance graphs in Fig. 42.

Accurate long-term temperature extrapolations can be made on the basis of a thermal endurance graph obtained for a fixed value of gasket leak rate (tightness criterion). This is done by simply extending the endurance line in the lower temperature range. Longterm temperature extrapolations for 3- and 5-year lives were made from the 1.2-mg/sec thermal endurance lines of Fig. 42. The calculated service temperatures are given in Table 9. The 3- or 5year temperatures of the nonasbestos materials are found to vary between 330°F (165°C) and 473°F (245°C), and their 5-year temperatures are only 2040°F (1020°C) lower. For the asbestos sheet, the 3-year temperature is greater than 630°F (330°C). This value is relatively close to 750°F (400°C), the temperature generally considered the upper limit for a premium-quality compressed-asbestos sheet material. Tests performed at temperatures of 750°F and above would be necessary to establish more accurately the asbestos sheet long-term temperature, but this data is not available for the time being. The validity of long-term temperature extrapolations based on a thermal endurance graph depends mainly on two factors: (1) the repeatability of

Page 267

Figure 41 Thermal endurance graphs for fixed levels of gasket weight loss. (a)-(c) Compressed aramid fiber sheets, (d) compressed carbon fiber sheet, (e) compressed graphite fiber sheet, (f) compressed asbestos sheet.

Page 268

the test results, and (2) the lowest test temperature. A good repeatability of the test results was demonstrated when comparing the effect of continuous vs. interrupted exposures. To extrapolate temperatures in the 20,00040,000-hour range (approximately 35 yr), methods based on the Arrhenius equation suggest to limit the temperature extrapolation at a maximum of 45°F (25°C) below the lowest test temperature, if this test lasted a minimum of 5000 hours. Most of the extrapolated temperatures in Table 9 meet that condition, since the lowesttemperature test was performed at 420°F (216°C) for 1 year. For the compressed carbon and graphite fiber materials, the extrapolation limit is extended to 90°F (50°C), and their calculated long-term temperatures are still valid, although they could be less precise due to the extent of the temperature extrapolation. Some materials exhibit a break in their thermal endurance lines (see Figs. 41 and 42). This could be the result of a new degradation reaction, like pyrolysis of the binder, that starts above a certain temperature. Data above the temperature of the endurance line break should not be used for predictions, since this could lead to an unrealistically high long-term temperature for the material. 10.6.2. Guidelines for Improving the Screen Test Scheme From the preceding findings, guidelines to develop an improved qualification test scheme based on the Arrhenius model and the thermal endurance graph can be proposed. The MTI/PVRC qualification test scheme for elastomeric sheet gasket materials presented in Section 12.4 relies mainly on mechanical screening ATRS tests performed in air atmosphere at several temperatures.

These screening tests are of the continuous exposure type; this means that gasket property changes are measured after aging, at the end of the tests. Test durations vary between 1 and 42 days (see Plate 3). To evaluate the long-term temperature, two parameters are used: the equivalent aged exposure parameter, Ae, and the quality parameter, Qp, both presented in the present section. The long-term service temperature of a gasket sheet candidate is computed with Eq. (18), using the critical equivalent aging exposure Ae value that brings the quality Qp below that of the reference asbestos-based gaskets in the Qp vs. Ae plot (see Section 10.4). A comparison between the long-term temperatures computed with Ae and the thermal endurance graphs has shown that similar temperature predictions are made for compressed-aramid sheets. For the other materials (compressed-carbon, graphite, and asbestosfiber sheets), these predictions differed significantly, and the differences were found to be a function of the energy of activation, B, of the materials (see Eq. 20 or 21). When the energy of activation of a material is above or below some threshold value, the temperatures computed with Ae are respectively lower and higher those found with thermal endurance graphs. Thus,

Page 269

Figure 42 Thermal endurance graphs for fixed levels of gasket mass leak rate. (a)-(c) Compressed aramid fiber sheets, (d) compressed carbon fiber sheet, (e) compressed graphite sheet, (f) compressed asbestos sheet.

Page 270 TABLE 9 Extrapolated Long-Term Temperatures for a 1.2mg/sec Leak Rate Criterion Extrapolated temperature in °F-°Ca (calculated from leak rate thermal endurance lines in Fig. 42) Compressed elastomeric fiber3-year exposure 5-year exposure reinforced sheet gasket (26,280 hr) (43,800 hr) materials Aramid fiber:CAR-1 465/240 446/230 Aramid fiber: CAR-2 473/245 455/235 Aramid fiber: CAR-3 420/215 400/205 Carbon fiber:CCA 347/175b 330/165b Graphite fiber: CGR 383/195 347/175b Asbestos fiber: CAF 626c/330c 620c/327c aExtrapolated temperatures are rounded up to the nearest 5°C. bExtrapolated temperatures more than 45°F (25°C) below the lowest test temperature of 420°F (216°C). cAnother mechanical test [25] performed at 700°F (371°C) for 1000 hr indicates a higher long-term temperature.

it appeared that some improvements to the actual MTI/PVRC qualification protocol for elastomeric fiber-reinforced sheet materials are necessary to adapt it better to the qualification of all types of elastomeric sheet gasket materials. The best way to improve the qualification test scheme would be to establish thermal endurance graphs with the results of the mechanical screening tests. These graphs could be obtained for properties that are considered critical for the gasket performance in service and for which specific qualification values (or failure

criteria) are established. Leak rate, load relaxation, residual tensile strength, and possibly brittleness are gasket properties that appear important for its long-term performance. Extrapolated long-term temperatures could be obtained from the endurance graph of one or several of these properties. To achieve this goal, the test procedure should be modified according to the following general guidelines. 1. Whenever possible, a monitoring of the gasket property change should be performed at regular time intervals during thermal exposure (ARLA tests). This will allow for the precise determination of the exposure time required to reach the qualification value of a gasket property. When monitoring of a gasket property is not possible, like for residual tensile strength or brittleness (with ATRS tests), two or more tests having different durations should be performed at each temperature. In this case, monitoring of weight loss could be used to help establish the necessary test durations.

Page 271

2. Test temperatures and durations should be chosen in order to generate a thermal endurance line that is valid in the range of the extrapolated long-term temperature of the material. To obtain this result, two points must be taken into consideration. First, the extrapolated long-term temperature should not exceed by more than 4590°F (2550°C) the lowest temperature of the test scheme. Second, the different test temperatures must belong to the same segment of the endurance line, the segment that extends in the range of the extrapolated long-term temperature. Compared to the existing qualification screening test scheme (Section 12.4), the improved scheme based on the preceding guidelines will require longer test durations. This requirement is, however, in accordance with the recognized test practices used in the evaluation of the long-term property for polymeric materials. Being based on the determination of the thermal endurance graph of a candidate material, the improved test scheme should lead to accurate long-term predictions for any type of elastomer-based sheet gasket materials. 11. Hot Relaxation Resistance Qualification Tools for PTFE-Based Gaskets 11.1. Background Considerations 11.1.1. A PTFE Gasket Qualification Project In 1995 the TTRL of École Polytechnique of Montreal completed an extensive test program of PTFE gaskets that was privately sponsored by gasket producers and users in an effort to develop a more comprehensive protocol for qualifying the tightness and hot performance of PTFE gaskets based on a direct measure of their

margin of safety against blowout [41]. Program objectives were to develop a protocol for qualifying PTFE gaskets based on hot relaxation resistance, gross leakage resistance (blowout), roomtemperature tightness performance (gasket constants), and resistance to high stress (crush resistance). Twelve gasket user and producer companies participated in the project, and a total of 27 currently available gaskets were investigated. These were selected to cover a wide range of PTFE-based gaskets, from sheet materials (virgin PTFE and filled PTFE) to preformed (metal-reinforced and spiral-wound types) and formed-in-place products. In addition to ROTT with CRUSH test extension, to examine the hot relaxation and blowout resistance of these products, baseline HOBT2 tests (see Plate 10 in Section 7) have been performed at 290, 435, and 750 psig (20, 30, and 52 bar approx.) helium pressure under the simulated rigidity of an NPS 3-

Page 272

in. 150-lb slip-on flanged joint and a 5000-psi (34.5-MPa) initial gasket stress. These tests have pointed out that blowout temperatures are clearly material dependent and are in the range of 300700°F (150370°C), that test pressure has little effect on the blowout temperatures, and that gasket stress relaxation behavior before blowout is indicative of the blowout resistance. Note: HOBT2 tests were performed under the simulated axial rigidity of an NPS 3-in. 150-lb slip-on flanged joint because this type of bolted flanged joint is generally considered to have inadequate bolting and is suspected of giving unacceptable performance in chemical or petrochemical plant piping systems. 11.1.2. PTFE Hot Relaxation Behavior Considerations The hot creep/relaxation behavior observed during blowout tests performed in the project is material dependent. The stresstemperature behavior varies from upward to downward curvatures, and it can be grouped into four families of curves, depending on their shape, in a region ranging from 150°F (66°C) up to the blowout temperature, as shown in Fig. 43. These families are described in the following sections. Family F1 In this family, the gasket stress decreases smoothly and less and less rapidly as the temperature is increased, until blowout conditions are reached (concave

Figure 43 Typical HOBT-family curves.

Page 273

curves). F1 curves characterize gaskets with a relatively high reserve for an increased temperature and a poor reserve for thermal or pressure cycles. Family F2 In this family, the gasket stress decrease is rapid at the beginning of heating and is less and less pronounced as temperature increases, but past some point it accelerates again until the blowout happens (S curves). F2 curves characterize gaskets with an intermediate reserve for increased temperature or pressure cycles. Family F3 In this family, the gasket stress creep/relaxation resistance is good as the temperature increases and then decreases uniformly and rapidly to the blowout (convex curvature). For some materials, the creep/relaxation resistance is excellent up to a point when it suddenly decreases very rapidly to produce the blowout. F3 curves characterize gaskets with a relatively high reserve for pressure surge or thermal or pressure cycles and a low reserve for an increased temperature Family F4 In this family, there is practically no creep/relaxation during the temperature increase and there is no tendency to blowout. This behavior is similar to what would be expected for a flat metal ring. 11.2. PTFE Hot Creep/Relaxation Characterization For a gasket in service in a plant, it is useful to know the reserve, or margin, available between the operating service conditions and those that would cause a gross leak (blowout). This is especially

important since it is known that most gross leaks occur during or just after thermal or pressure events, such as plant start-ups and shutdowns, thermal upset, or rain on exposed joints. A bolted joint subjected to such events could be acceptable or unacceptable depending on its potential reserve against gross leakage, because it would be more or less vulnerable to disturbances that would reduce the gasket load. Fortunately, the hot creep/relaxation behavior of gaskets tested with the HOBT2 test procedure provides a means for characterizing the capability of PTFE-based gaskets to resist gross leakage in the event of certain plant operating fluctuations. 11.2.1. Stress/Relaxation Considerations for an Operating Joint Let us suppose a PTFE-based gasket is installed in a flange operating at some arbitrary temperature in a process plant, and refer to the typical HOBT2 hot

Page 274

stress relaxation curve as represented in Fig. 44. The point Qa (Sga,Ta) represents the assembly state of the gasket; the point Qinit (Sginit,Ta) represents the operating state of the gasket at the start of the heating process (Ta is the ambient temperature). Note that Sginit is always less than the assembly stress, Sga, because Sginit is the residual stress after fluid pressurization and ambient creep/relaxation that immediately follows gasket installation. Assuming that some hot stress relaxation has occurred, the arbitrary point Q (Sg,Tg) on the stress relaxation curve represents the operating conditions for that joint containing a PTFE gasket (assuming no time effect on the degradation of PTFE-based materials). Given these initial definitions, the characterization of the potential reserve against gross leakage of a PTFE-based gasket operating in a joint requires the following considerations. Pressure or temperature surge: It is normal to expect pressure or temperature variations at a piping joint. In fact, short-term excursions are defined and permitted for metallic piping systems by Par. 302 of the Refinery Piping Code B 31.3. Since a pressure or temperature surge sufficient to cause blowout is to be avoided, it is essential to characterize the safety potential (or reserve) that any PTFE gasket has against blowout due to pressure or temperature increase when operating at the conditions Q (Sg,Tg).

Figure 44 Generic representation of HOBT test procedure.

Page 275

Shutdown/Start-up: The possibility of plant start-up or shutdown (normal or abnormal) is always present as an expected part of a plant operation. Since it is known that most gross leaks occur during or just after thermal events such as plant start-ups and shutdowns, it is necessary to consider any additional gasket stress relaxation resulting from thermal cycles (e.g., cool-down, restart) when operating at Q (Sg,Tg) Referring again to the typical HOBT2 stress creep/relaxation curve of Fig. 44, these considerations and characteristics can be expressed as follows. The safety concept requires the definition of safety bounds in terms of a lower-bound stress limit, Sglb, and a upper-bound temperature limit temperature, Tub, that a PTFE-based gasket cannot overstep when tested with the HOBT2 procedure. Sglb and Tub have to be determined from the blowout conditions represented by a point Qbo(Sgbo,Tbo), where Sgbo is the remaining gasket stress at blowout and Tbo the blowout temperature. These limits represent the lowest permitted stress and the maximum permitted temperature for a PTFE gasket that gives an assurance of freedom from catastrophic failure in the form of gross leaks in operating bolted joints. The cool-down consideration permits the determination of a safe cool-down operating point, referred to as Q = Qc(Sgcd,Tcd), on the hot stress relaxation curve so that the potential safety that this PTFE gasket has against blowout is maintained. Operating point Qc can be estimated on the basis of the difference in the thermal expansion characteristics between the gasket and the test fixture materials.

For qualification purposes, the determination of a safe cool-down point Qc(Sgcd,Tcd), together with safety bounds Sglb and Tub, will permit the determination of a safe reserve operating point, Qr(Sgr,Tr), that will be useful to identify the potential reserve against gross leakage for PTFE-based gaskets. 11.2.2. Determination of Safety Bounds Against Gross Leakage Estimation of a Safety Stress Limit, Sglb A safety stress limit, Sglb, being equal to 1.5 × Sgbo (Sgbo is the blowout stress), appeared to be a reasonable choice for qualification purposes when using a HOBT2 test at 750 psig He pressure (53 bar). This limit corresponds to a mean HOBT2 relaxation range of 77.5%. This range is representative of the creep/relaxation behavior that a wide range of real B 16.5 bolted flanged joints (class-150 to -900 from 1 to 24 in.) could exhibit during operation when, for any reason, the bolt load is reduced to the point where the joint is close to a blowout situation (Sg = 1.5 × p). A bolted joint in this situation is unacceptable because

Page 276

it is vulnerable to small thermal disturbances or any input that further reduces the load. Sglb = 1.5 × Sgbo (22) Estimation of the Safety Limit Temperature, Tub It is prudent to have a margin of safety regarding the blowout temperatures. From plant operation practices, temperature increases of 100°F ( 55°C) are occasionally encountered. In addition, shortterm allowable stress excursions are defined and permitted for metallic piping systems by Par. 302 of the Refinery Piping Code B31.3. If system pressure is held constant, the allowable stress increase is construed as permitting a short-term temperature increase. It appears that for class-150 and -300 carbon steel piping, a maximum safe temperature of 100°F ( 55°C) less than the blowout temperature, Tbo, would seem to be a reasonable choice of the maximum permitted temperature for a PTFE-based gasket installed in these types of joints. An absolute limit to the long-term PTFE-based gasket temperature was also considered, because fluoropolymer resins are recognized to be stable at temperatures below 300°C ( 570°F). However, they do exhibit a small amount of degradation (weight loss) at higher processing temperatures. Therefore, a reasonable upper-bound limit temperature Tub combining both temperature increases during operation and absolute limit for long-term continuous use is determined as follows: (23) 11.2.3. Determination of a Safe Cool-Down Point, Qc(Sgcd,Tcd)

The shutdown and restart of plant piping can cause a thickness decrease in PTFE gaskets that results in bolt-load loss. The objective is to provide for an adequate residual stress so that there are no leaks on cool-down. Figure 44 represents a typical HOBT2 stress creep/relaxation curve. Assuming that some hot stress relaxation has occurred in an HOBT2 test, the arbitrary point Q(Sg,Tg) on the stress relaxation curve represents the operating conditions for that joint containing a PTFE gasket. From this point, let us suppose that the test rig is cooled down to room temperature, Ta. The gasket thickness change during this simulated cool-down to Ta (assumed to be equal to 75°F or 24°C) can be estimated on the basis of the difference between the thermal expansion coefficients of the gasket and test rig material. From that value, the corresponding equivalent gasket stress relaxation of the test rig, called X, can be calculated

Page 277

using the specific joint axial rigidity value of the test rig. Then, the safe cool-down point Q = Qc(Sgcd,Tcd) can be determined on the basis that the residual gasket stress at room temperature, defined as Sgcd - X, should always remain greater than the safe stress limit Sglb to ensure that there are no gross leaks on cool-down. Therefore, the safe cool-down point Qc(Sgcd,Tcd) is determined as the intercept of the stress creep/relaxation curve and the line Sgcd = 1.5 × Sgbo + X. The gasket temperature Tcd is the safe cool-down temperature. 11.2.4. Determination of the Reserve Operating Point Qr(Sgr,Tr) The reserve operating point Qr(Sgr,Tr) on the HOBT2 stress creep/relaxation curve of a PTFE gasket material is the more conservative of the two controlling points defined from the considerations presented in sections 11.2.2 and 11.2.3. These points are: The safe upper-bound point, Qub: The safe gross leak temperature, Tub = Tbo - 100 < 600 (°F) (Tub = Tbo - 55 < 315 (°C) and its corresponding gasket stress, Sgub (psi, MPA), determined graphically for each material. The safe cool-down point, Qc: The safe cool-down temperature, Tc (°F or °C), and the cool-down gasket stress, Sgcd (psi, MPA), both being determined graphically such that Sgcd = 1.5 × Sgbo + X (X determined graphically). Tbo and Sgbo are, respectively, the blowout temperature and the residual gasket stress as measured in HOBT2 tests performed on that material. The reserve point, Qr(Sgr,Tr) is the one with the highest stress and the lowest temperature.

Typical values of Qub(Sgub,Tub), Qc(Sgcd,Tcd), and Qr(Sgr,Tr) are shown in Fig. 45 for the three typical family classes. 11.3. Role of Safety Margins The simulation of PTFE flange joint contraction and the safety stress and temperature concepts characterize the susceptibility of a gasketed joint to blowout in case of events such as cool-down and heat-up during shutdown and start-up or pressure or temperature surges. When confronted by a choice of candidate PTFE-based gaskets, the following considerations (among others) may be important in selecting a particular candidate. 1. What is the safe operating temperature for this candidate? 2. At that temperature, what is the reserve against gross leakage for temperature surges, pressure surges, and stress decreases? 3. If one chooses to operate a particular gasket below the safe operating temperature, what margin is available for temperature surges and stress

Page 278

Figure 45 Reserve operating point, Qr, for three typical HOBT family curves. In the figure, the safe cool-down point is designated by Qcd.

decreases at, say, 100°F less than the safe operating temperature? In that case, is there a better choice than might be indicated by the considerations of items 1 and 2? These issues are addressed by considering the margins available at the reserve operating point, Qr(Sgr,Tr), and at a second point, Q(Sg,T), representing a less severe operating condition. 11.3.1. Concept of Safety Margins (or Reserves) It would be useful when selecting a PTFE material to know what its potential reserve against blow-out is for any plant operating conditions represented on the hot creep/relaxation curve by an arbitrary operating point Q(Sg,T) below Qr. It is convenient to define two safety margins: The % bolt-load margin, %BLM, when operating at Q(Sg,T):

(24) Where Sga is the initial assembly stress and Sgbo is the blowout stress The temperature margin, TM, when operating at Q(Sg,T)

Page 279

(25) Evaluating the safety margin at the reserve operating point, Qr(Sgr,Tr), is more than worthwhile because it will give additional means of comparison between materials having a similar reserve temperature, Tr. Also, because plants can operate at temperatures lower than Tr, a safety margin evaluation is also considered at Tr minus some threshold temperature to examine the reserve of the gasket material at that temperature and determine its suitability for the type of application envisioned. Since it would be cumbersome to examine several temperatures, a good choice would be to consider evaluating the safety margin at Tr - 100°F (Tr - 55°C). 11.3.2. Application to the Typical PTFE-Based Materials Figure 46 visualizes the concept of safety indicators (reserve temperature, Tr, % bolt-load margin, %BLM, and temperature margin, TM) for a typical tested material. To highlight the usefulness of these indicators for PTFE gasket selection, Fig. 47 plots the % bolt-load margin versus gasket operating temperatures for Ta (ambient), Tr - 100°F (Tr, - 55°C), Tr, and Tbo temperatures for tested materials that are characterized by a similar reserve temperature Tr (within a band of about 40°F or 22°C).

Figure 46 HOBT safety indicators.

Page 280

Figure 47 Comparison of safety indicators for three PTFE-based materials with similar reserve operating point, Qr.

11.4. Conclusion and Recommendations Characterization of the blowout resistance of PTFE-based products is based on safety indicators introduced to ensure freedom from failure in terms of gross leakage when these gasketing products are used in bolted flanged piping joints, essentially Class 150 and 300 lb. These indicators are: A safe operating reserve temperature, Tr A percent bolt-load margin, %BLM A temperature margin, TM These three safety indicators characterize the susceptibility of a gasketed joint to blowout in case of events such as cool-down and heat-up during shutdown and start-up or pressure or temperature

surges during operation. Selection and comparison of PTFE-based materials based on these indicators has become possible. A word of caution: Events such as cool-down and heat-up or pressure or temperature surges have not been tentatively reproduced using the HB or UG rigs. At the time of this writing, a modest exploratory HOBT test program was

Page 281

under way at TTRL under conditions that would mimic these possible events to introduce another version of the HOBT procedure. Additional tests using this modified procedure should be planned on at least one material of each PTFE generic class to confirm the qualification tools developed for these materials. 12. Gasket Qualification Guides or Protocols 12.1. General Approach The test procedures and qualification tools outlined in Sections 411 are being applied to the qualification of promising gaskets for process plant applications. Although the materials and gaskets vary widely, selection mechanisms are in place for accommodating the differences, with variations that appropriately apply the test procedures and qualification tools just summarized. The general approach to qualification is to examine the roomtemperature behavior of a gasket product candidate and to determine whether or not it is the ideal asbestos replacement by direct comparison of its long-term elevated-temperature properties with those of traditional asbestos counterparts. Material classification is important because any test scheme for estimating long-term gasket performance must consider distinctly the different characteristics of the various gasket types and materials that are available. Depending on the class of material (whether elastomer-bound, flexible-graphite, PTFE, or composite gaskets), the test scheme of the qualification guide will recommend ROTT tests, a different mix of FIRS, ATRS, ARLA, and HOTT/AHOT tests as well as EHOT, HOBT, ROMT, and HOMT tests. In terms of applying these procedures, this boils down to the

following: Room-temperature tightness tests (ROTT tests) to determine gasket parameters, including gasket constants Gb, a, and Gs. Screening tests to evaluate fire integrity (FIRS and/or FITT tests) and to evaluate the fire-survival temperature of a sheet gasket product. Screening tests to evaluate mechanical stability, tightness, weight loss, and relaxation of the gasket candidate. Depending on the type of gasket product and on whether the objective is to qualify the product for a specific temperature or to evaluate its long-term service limits, different numbers of ARTS/HATR/ARLA and/or HOTT/AHOT tests are proposed at various combinations of exposure times and test temperatures. Emissions tests (EHOT) to evaluate the effect of a moderate exposure at the maximum gasket service temperature on the emission (leakage) performance of a gasket. The EHOT test should tell us whether the room-temperature gasket constant Gs is still applicable when the gasket

Page 282

is exposed to a high service temperature for an exposure time that represents 2 or 3 years of service (about one-third the maximum service life). Hot relaxation and blowout tests (HORT and HOBT) to evaluate the relaxation and gross leakage resistance of a gasket, as recommended for the process and material being considered. Pure mechanical tests (ROMT or HOMT) to evaluate the gasket stiffness, as needed by a gasket manufacturer or requested by a gasket user or a gasket supplier. To judge the relative performance of a gasket, qualification tools for ambient- and elevated-temperature applications have been established from comparative tests with asbestos sheet products. Through the use of aging parameter Ae, test conditions (temperature and time) are selected in order to qualify a material for a specific long-term service temperature or to establish a recommended temperature limit. It is now possible to specify and qualify virtually all process plant gaskets in terms of their tightness, indicated long-term service temperature, and potential fire integrity. This new technology has emerged in the last dozen years, coincident with new gasket materials and regulation challenges. It provides process plant operators with the opportunity to improve the reliability and emissions performance of bolted flanged joints. These improvements are realized through a better understanding of the mechanism of flanged joint leaks and the availability of a series of standardized tests for the characterization and qualification of gaskets.

Plant experience has confirmed the adequacy of the proposed qualification guides developed essentially for elastomeric fiberreinforced materials. For PTFE-and flexible-graphite-based materials and for composite gaskets, these guides will benefit from further improvements to account for the specific behavior of these various materials. Nevertheless, questions about gasket performance and improvements of these new qualification tools remain, because of limited funding and time pressures in spite of the substantial efforts described. 12.2. Room-Temperature Qualification A gasket user may specify ROTT test requirements for any particular type and rating of gasket in purchase document by asking that the supplier certify the results of two or more tests conducted in accordance with the ROTT test procedure. A minimum of two ROTT tests are required to determine the gasket parameters, including PVRC gasket constants Gb, a, and Gs, with conditions per the current proposed ASTM Draft 9 and Addendum. A third test may be needed, depending on the statistical agreement of the first two tests. Interested users and producers may also specify a CRUSH test as a supplement to the ROTT tests. The maximum CRUSH test loads are limited to the test

Page 283

rig available (40,000 psi gasket stress with the TTRL HR rig). ROTT and CRUSH test procedures and gasket parameter determination are presented, respectively, in Sections 6 and 9. The following ROTT and CRUSH test results shall be available for evaluation and comparison: Gasket constants (Gb, a, Gs), seating values S100, S1000, and S10,000, and Tps with Ss (if tightness hardening is applicable) Smallest ROTT unload-reload cycle Tpmin or Tpmin(B), largest ROTT Tpmax Tpmin(C), and Sc, the Tp and max gasket stress defined by the lowest crush-cycle max gasket stress, where Tpmin(C) > Tpmin(B) Acceptance may be considered if the following shall be within the values specified: Gasket constants Gb, a, and Gs Seating values S100, S1000, S10,000 Crush strength Sc: lowest crush-cycle max gasket stress where Tpmin(C) > Tpmin(B) Acceptance values for gasket constants and seating stress values have to be discussed and specified by judgments formed through comparison of the ROTT results for a proposed material against the typical results available for similar materials and by interpretation of the design objectives of the specifier. Suggested values of acceptable crush strength could be based on ANSI/ASME B 16.5 joints in the NPS 324 size range when an overly aggressive bolting situation is assumed. For example, in the

cases of Class 300 and Class 150 with the bolts stressed to 80,000 psi on a gasket having sheet gasket dimensions, we can specify a minimum crush strength, Sc, as follows: Class 300: Sc > 24,500 psi Class 150: Sc > 12,700 psi 12.3. Fire Resistance Qualification A minimum of three FIRS tests are requested to evaluate the firesurvival temperature, Tf, of a sheet gasket product. The FITT test is optional, but recommended if proof of tightness is required. For composite preformed gaskets (such as spiral wound), the FITT test is preferred to the FIRS. 12.3.1. Elastomeric Fiber-Reinforced (EF) Sheet Products Three FIRS tests at 1200, 1100, and 1000°F (minimum) to determine Tf One FITT test at Tf if proof of fire-survival tightness is recommended (optional)

Page 284

12.3.2. PTFE-Based Products Three FIRS tests at 1000, 900, and 800°F (minimum) to determine Tf One FITT test at Tf if proof of fire-survival tightness is recommended (optional) 12.3.3. Flexible-Graphite-(FG) Based Products Screening tests to evaluate fire integrity (FIRS and/or FITT tests) would be of little use, because it has been proven that flexiblegraphite-based gaskets are usually fire-safe [30]. However, if confirmation is required for a specific product, it is recommended. One FIRS tests at 1200°F One FITT at 1200°F test if proof of fire-survival tightness is recommended 12.3.4. Acceptance Requirements FIRS Acceptance Specimen can be handled with residual TSX > 100 psi. Lack of excessive spread, extrusion or lateral flow. FITT Acceptance Qt'' > 1 (Tpmin > 32 with He medium) Qt'' > 1 (Tpmin > 12 with air medium) 12.4. Screening for Hot Mechanical Performance Screening tests to evaluate mechanical stability, tightness, weight loss, and relaxation of the gasket candidate exposed to thermal aging are recommended. Depending on the type of gasket product

and on whether the objective is to qualify the product for a specific temperature or to evaluate its long-term service limits, a test program can be tailored to specific needs, on the basis of different numbers of ATRS/HATR/ARLA and/or HOTT/AHOT tests proposed at various combinations of exposure times and test temperatures. To judge the relative performance of a gasket, qualification tools developed in Sections 10.1, 10.2, and 10.3 are used. See also Section 10.5 for recommendations regarding the choice of screening test conditions to improve predictions of gasket life. 12.4.1. Qualification Tools for Screening Gasket Properties The reference formulas developed in Section 10 are summarized in the following sections

Page 285

Quality Parameter, Qp

= postaging residual ATRS/HATR specimen TSX tensile strength % = postaging % stress retained by the retained ATRS/HATR test fixture Qp is applicable to all types of gasket products, with the following specifics corresponding to the gasket material class. a. For metal unreinforced products or for products where the gasket material delaminates from the metal foil insert:

b. For metal-reinforced products where the metal reinforcement is uniformly distributed over the thickness (screen, corrugated, or tang metal inserts) TSX = 2500 psi (17.2 MPa) Acceptance is considered when the following applies:

Tightness Quality,

Tpmin is the minimum tightness measured during an unload-reload cycle associated with the standard HOTT or AHOT test thermal disturbance sequence. Acceptance is considered if

1.

Equivalent Aging Parameters, Ae For EF-based sheet products:

For FG-based sheet products:

Page 286

where: T

= exposure temperature in °F

H

= exposure duration in hours

For PTFE-based products: Since there is no significant age effect on known PTFE-based products, there is no equivalent aging parameters for these materials. Only temperature is considered for screening PTFE-based gasket properties. 12.4.2. Objective 1: Qualification for a Specific Temperature Application With this objective in mind, a gasket user wishes to be assured that the material he or she purchases will be serviceable after an expected length of time at some specific gasket temperature (for example: a 5-year life at 600°F). He or she might specify screen and hot leakage tests (usually three ATRS/HATR tests and two HOTT/AHOT tests), with test conditions calculated from the equivalent Ae exposure corresponding to the service life requirements for that particular class of gasket products. For Elastomeric Fiber-Reinforced (EF) Sheet Products Three ATRS tests are recommended at test conditions corresponding to the Ae exposure to be verified with a maximum of 1000 hr exposure. Some ARLA long-term tests (45000 hr) could be

recommended to supplement and/or confirm ATRS test results. As an example, suppose a user wishes to qualify an aramid fiberreinforced sheet for a 5-year service at a 420°F (Ae 50 for EF materials). The user might specify screen and hot leakage tests for, say, the following. Screen tests requirements: 42 days @ 550°F 16 days @ 600°F 7 days @ 650°F with the following minimum requirement: Qp > 0.5 with the average Qp > 1.0. Note: A 6-month ARLA test at 480°F could be recommended. HOTT/AHOT test requirements: 1. The HOTT test shall be conducted with air-media aging: Typical exposure: 5 days (120 hr) or less at 680°F Required posttest tightness: Qt" > 1 (Tpmin > 12)

Page 287

2. AHOT tests shall be conducted with nitrogen-media aging: Typical Exposure: 30 days (720 hr) at 560°F Required posttest tightness: Qt" > 1 (Tpmin > 12) For Flexible-Graphite (FG) Sheet Products As an example, suppose a user wishes to be assured that the tanged-inserted flexible-graphite sheet material purchased will be serviceable after 5 years at a 585°F gasket temperature (Ae = 85 for FG materials). The user might specify screen and hot leakage tests as follows. Screen test requirements: Three ATRS/HATR tests are recommended. Typical exposures are: 42 days @ 770°F 16 days @ 835°F 4 days @ 950°F Typical minimum requirements are: Qp greater than 0.3 minimum Weight loss less than 15% Density Change less than 5% HOTT/AHOT test requirements: The following approach might be used to specify HOTT and AHOT tests for tang-inserted flexible-graphite sheets. 1. The HOTT test shall be conducted with air-media aging:

Typical exposure: 5 days (120 hr) or less at 925°F Required posttest tightness: Qt" > 1 (Tpmin > 12) (typical: Tpmin > 200) 2. AHOT tests shall be conducted with nitrogen-media aging: Typical exposure: 42 days (1008 hr) at 770°F Required posttest tightness: Qt" > 1 (Tpmin > 12) (typical: Tpmin > 200) 12.4.3. Objective 2: Seeking a Recommended Service-Temperature Limit To solve this more difficult problem and depending on the type of gasket product, six to eight screening tests (ATRS/HATR), in air medium, are recommended. While most of the ATRS/HATR tests are short-term tests (up to 1000 hr), at least one long-term ARLA test (6 months) at a temperature that is considered moderate for the specific product under consideration may be needed to supplement and confirm relaxation behavior trends.

Page 288

To judge the relative performance of a gasket candidate, plots of quality parameters, Qp, versus equivalent aging parameters, Ae, are used. From these plots, by using polynomial regressions a range Ae', Ae" is determined for the gasket candidate at the minimum quality, Qpmin, acceptable for its class. A range of temperature limits, Ts' and Ts", are calculated from the Ae formula. Picking a final calculated recommended Ae, referred to as Aer, for calculating a long-term service temperature may require a tightness test (HOTT/AHOT) performed under aging conditions that will represent an equivalent aging exposure between Ae' and Ae". Using the mathematical relationship for Ae, long-term service temperatures can be found for an expected service life of 5 years, for example. Hot leakage tests may be specified if certifiable proof of tightness is required by the client. In that case, one 5-day HOTT test and/or one up-to-a-30-day AHOT test are recommended to confirm and supplement ATRS or HATR results and trends. When applicable, the AHOT test follows the HOTT test. In some instances, tests in both air and nitrogen media may be specified, which may necessitate the use of additional HOTT and AHOT tests. The possibility of additional testing is usually discussed with the client and decided in mutual agreement. Based on ATRS test results, some ARLA tests (up to two) could be recommended to check tightness for one or two exposures in order either to confirm the trends predicted by mechanical screening if full gasket qualification is not expected or to fine-tune the tightness test temperatures (5-day HOTT or 30-day AHOT) required for complete qualification.

Test Conditions for Elastomeric Fiber-Reinforced (EF) Sheet Products Screen tests requirements: Most of the ATRS tests are short-term tests, from 96 hr up to 1000 hr, at temperatures ranging from 400°F to 700°F, typically, six ATRS tests, in air medium, as follows: 6

days @ 450°F (232°C) (Ae = 20)

42 days @ 450°F (232°C) (Ae = 30) 5.5 days @ 600°F (316°C) (Ae = 40) 16 days @ 600°F (316°C) (Ae = 50) 4

days @ 700°F (371°C) (Ae = 50)

10 days @ 700°F (371°C) (Ae = 60) Test results: Determination of Ae' and Ae" at Qpmin = 0.5. Note: A 6-month ARLA test at a temperature 5080°F greater than that predicted from the ATRS test scheme is recommended to supplement and/or confirm the predictions. Elevated-temperature tightness confirmation (optional): One 5-day HOTT (air medium) and/or One 30-day AHOT (nitrogen purge during aging)

Page 289

at a test temperature corresponding to Ae such that Ae' < Ae < Ae". Tightness requirement: Qt" > 1 Test Conditions for Flexible-Graphite (FG) Sheet Materials Screen test requirements: Six to eight HATR tests, in air medium, are recommended. While most of the HATR tests are short-term tests (up to 1000 hr), at least one long-term test (2200 hr at 700°F) may be needed to supplement and confirm relaxation behavior trends, typically: 48 hr@

950°F(Ae = 60)

96 hr@

900°F(Ae = 70)

384 hr@

850°F(Ae = 90)

240 hr@

780°F(Ae = 50)

2160 hr@

720°F(Ae = 80)

96 hr@ 1000°F(Ae = 110) Test Results: Determination of Ae' and Ae" at Qpmin = 0.3. Elevated-temperature tightness confirmation (optional): Tightness confirmation is optional for flexible graphite below 750°F. If tightness confirmation is requested: One 5-day HOTT (air medium) One 30-day AHOT (air or nitrogen purge during aging) at a test temperature corresponding to Ae such that Ae' < Ae < Ae"

Tightness requirement: Qt" > 1 12.5. Emissions Characterization The EHOT test should tell us whether the room-temperature gasket constant Gs is still applicable when the gasket is exposed to a high service temperature for an exposure time that represents 2 or 3 years' service (about one-third the maximum service life). For elastomeric fiber (EF) based products, a total 8-day emission hot tightness test (EHOT) at 450°F (5-day aging; Ae = 20) with helium medium is recommended. For PTFE-based products, a total 5-day emission hot tightness test (EHOT) at 450°F with helium medium is recommended. For flexible-graphite-(FG) based products, a total 14-day emission hot tightness (EHOT) test at 850°F (10-day aging; Ae = 75) with helium medium is recommended.

Page 290

EHOT Acceptance Requirement The acceptance criterion of an EHOT test is based on the validity of Gs at temperature. If calculated Gs remains practically unchanged, it can be used up to the equivalent gasket service temperature that is verified from the EHOT test conditions. In special circumstances where Gs exhibits a significant change in the tightness response to stress cycles, two possibilities are considered: 1. The EHOT test is repeated at less severe test conditions until the room-temperature constant Gs is unchanged. An upper servicetemperature limit for the use of Gs is determined. 2. The EHOT leakage test results and change in constants due to temperature will be used to determine new initial bolt-load calculation that accounts for the temperature effect. 12.6. Blowout and Relaxation Characterization for PTFE 12.6.1 Application Blowout and relaxation characterization is recommended for all foreseeable types of PTFE-based gasket products intended for standard metallic ANSI/ASME B 16.5 flanges. This includes preformed gaskets, such as metal-reinforced, spiral-wound, and envelope types, as well as sheet and formed-in-place products. The bolting and pressure-temperature parameters of ANSI/ASME Class-150 and -300 piping joints is the primary focus. Nevertheless it is possible to extend the protocol to higher ANSI/ASME ratings by appropriate adjustment of the protocol parameters and, to allow the necessary higher test pressures, modification of the HOBT fixture. It is anticipated that most manufacturers and users will opt to qualify materials for the Class-300 rating.

12.6.2. Aging Aging (in terms of days and months) is not now considered an important aspect of PTFE performance. The HOBT test procedure and fixture are for short-term tests. Nevertheless, if it is important for a particular product, there are two ways to consider aging. First, specimens may be preaged within the present HOBT fixture at a specified temperature for a specified number of days, followed by a cool-down and the usual HOBT test. Second, the MTI-developed ATRS test (Section 5.1) may be specified as an option to determine the effects of various exposures in the relaxation performance of a product. 12.6.3. Minimum Required Tests Three blow-out tests (HOBT2). Optional:

Page 291

Include a shutdown cycle in the second and third HOBT tests. This test cycle has to be implemented at the reserve temperature, Tr, indicated by the first HOBT test. Second, third, and fourth HOBT tests at various other pressures so that a temperature vs. pressure curve is established for the material in question. Aged HOBT tests as proposed in Section 12.6.2. ATRS tests as proposed in Section 12.6.2. 12.6.4. Test Conditions The following results (developed in Section 11) shall be available for evaluation and comparison of the gross leakage and blowout resistance of PTFE-based products tested with HOBT2 tests: Hot relaxation stress curve (Sg vs. temperature) Gasket stress just before pressurization, Sga Gasket stress after pressurization and just before heating, Sginit Temperature and gasket stress at the point of gross leak (blowout), Tbo, Sgbo Safety bounds Sglb and Tub, evaluated as follows: Sglb = 1.5 × Sgbo Tub = =Tbo - 100°F < 600°F Safe cool-down temperature and gasket stress, Tcd, Sgcd, Reserve temperature, Tr, and reserve stress, Sgr, where Tr is the lesser of: safe gross leak temperature, Tub and safe cool-down temperature,

Tcd Percent bolt-load margin, %BLM, evaluated at Tr as follows:

Temperature margin, TM, evaluated at Tr as follows: TM = Tbo - Tr 12.6.5. Acceptance Considerations Acceptance may be considered if the following apply: Reserve temperature, Tr, is greater than specified value % bolt-load margin, %BLM, is greater than specified value Temperature margin, TM, is greater than specified value Acceptance values have to be specified by judgments formed through comparison of the HOBT results for a proposed material against the typical results available

Page 292

for similar materials and by interpretation of the design objectives of the specifier. For example, one could specify one or more of the following three conditions: Tr = 425°F %BLM = 20% TM = 125°F In addition, two absolute criteria for acceptance are: Class 300: Gross leak temperature, Tbo, greater than 500°F (avg. of three tests, 480°F min.) Class 150: Gross leak temperature, Tbo, greater than 420°F (avg. of three tests, 400°F min.) 12.7. HORT Characterization The purpose of the HORT test is to evaluate the relaxation resistance of gaskets under simulated rigidities when subjected to excessive loads and temperature cycling. 12.7.1. Specifying HORT Test Requirements A gasket user may specify HORT test requirements in purchase documents for a particular type of gasket by asking that the supplier certify the results of HORT tests on NPS-4 specimens. The test specifications for typical exposure (time and temperature) and typical joint rigidities are given according to specific requirements and to the type of gasket product to be evaluated. As an example, based on service life predictions made for flexiblegraphite-based sheet gaskets (see Section 10.4) and on a gasket

user field experience data, test conditions for a HORT test performed on these products can be a 4-day relaxation phase with three thermal cycles from ambient to 650°F under 800-psig He pressure. For this class of gasketing products, these test conditions are, for the time being, the best suggestion that can be made to meet the approval of the gasket community. However, they may be susceptible to some changes, depending on future improvement, suggestions, and any field experience data contributions from gasket users and producers. 12.7.2. HORT Acceptance At the time of this writing there is no acceptance criterion that has been developed for the HORT test. Its use is primarily as a reference for comparison among various products. 12.8. Mechanical Characterization Mechanical tests such as LCMT are used to evaluate the stiffness of gaskets under loading at room temperature, whereas the ROMT and HOMT are more

Page 293

specifically designed to evaluate their pure creep behavior at room and elevated temperature. 12.8.1. Specifying ROMT/HOMT or LCMT Requirements A gasket user may specify ROMT or HOMT or LCMT test requirements for a particular type and rating of gasket in a purchase document by asking that the supplier certify the results of two or more tests conducted in accordance with the ROMT/HOMT or LCMT procedures. 12.8.2. Mechanical Acceptance There is no specific criteria on acceptable creep resistance of gaskets for the ROMT/HOMT. For the LCMT, specifically developed for spiral-wound gaskets, specifications on gasket compressed thickness, gasket thickness recovery, and gasket visual aspect are usually specified. For example, for a spiral-wound gasket CG style with graphite filler, 4-in. NPS, Class 300 (API 601) having the following characteristics: Gasket OD

5.875 in.

Gasket ID

5.000 in.

Gasket area

7.474 in.2

Gasket initial thickness

0.180 in. nominal

Gasket gage ring thickness

0.120 in. nominal

Acceptance requirements: 1. The final compressed thickness shall be between 0.125 and 0.145 in. 2. Gasket recovery shall be 0.01 in. minimum within 1 hr following complete unloading. 3. There shall be no excessive buckling to the extent that metal piles are separated or that the spiral wound has separated from the guide ring. 4. There shall be no failure of any spot welds. 13. Standardization Status in North America A growing number of process plant owners have applied the new qualification approach to identifying promising products, to specifying them, and to discovering the limitations of others. Some gasket producers are now setting equipment to qualify their products through the use of this new technology. Because resources are limited, a few plant owners or gasket producers or other research organizations cannot test each and every gasket possibility. A more organized and standard approach makes a lot more sense.

Page 294

The adoption of standardized performance testing by national standards bodies such as the ASME and ASTM is seen as the most important step in promoting consistent qualification procedures for functional elevated-temperature gasket performance tests within industry. Because the mitigation of leaks and emissions become ever more important and as gasket producers continue their quest for improved products and introduce tomorrow's advanced technology, the use of national qualification standards is necessary. 13.1. ASME/ASTM Code Revision An ASME Special Working Group (SWG/BFJ) is working to implement gasket constants derived from hundreds of PVRCsponsored gasket tests. A new nonmandatory appendix will appear in Section VIII, Div. 1 of the ASME Code, which will parallel the present Appendix 2, and eventually replace it. This appendix will contain tables with new gasket constants (called Gb, a, and Gs). There will be corresponding changes in the formulas that obtain the design bolt loads. The constants will be derived and condensed from a conservative interpretation of the now-existing PVRCROTT test data (see also Chapter 9). The ROTT test is under examination by the ASTM F3 Committee on Gasket Testing and should become an ASTM standard test following the completion of round-robin testing. This will permit manufacturers to get the new constants for their new and improved gaskets. While no one can say exactly what the final form of these documents will be, it is clear that the changes they represent will give the designer of gasketed bolted joints the opportunity of using gasket constants that have been certified by the gasket

manufacturer/supplier in accord with meaningful new standards. It is anticipated that the ASME table of gasket constants will be updated over the years as a more comprehensive array of gasketsupplier-developed constants becomes available. 13.2. Standardization of Mechanical Screening Tests Drafts of ASTM standards for the ATRS, FIRS tests have been prepared and are under study by the ASTM Committee F3 on Gaskets. The drafts are presently in revision and should be submitted for round-robin testing in a near future. All these tests will be used as screening tests, and test results will be interpreted in accordance with proposed specification schemes. 14. Continuing Developmental and Research Efforts The research effort in North America is aimed at improving the capability to predict and therefore improve the behavior of bolted flanged joints. In these environmen-

Page 295

tally sensitive times there are advantages and benefits for an approach to bolted flanged joint design that considers leakage and makes the tightness of the joint a design criterion. The introduction in the ASME Code of a bolt-load calculation procedure based on the new PVRC gasket constants will be a first step toward solving this problem. However, for joints operating at elevated temperature, more work needs to be done [38]. New design guidelines based on standardized qualification specification and test methods for gaskets are still to be finalized before further Code improvements can be considered. Nevertheless, the diligent designer can find much to consider from the findings and information now available. Table 10 lists the PVRC research programs that will help to achieve these goals. These programs either are currently under way or are to be undertaken soon, in accord with the PVRC 5-year Plan objectives [48]. The experimental investigations will be conducted by UND, JPAC, CETIM, EDF, and TTRL under PVRC and TTRL coordination to provide a common basis for the use of testing procedures and test equipment. Experimental test results will be reviewed, compared, and reconciled and further tests will be conducted if needed. Reporting and recommendation will be produced on the basis of European and North America needs for bolted joints. List of Abbreviations Organizations ANSI American National Standards Institute

ASME American Society of Mechanical Engineers API

American Petroleum Institute

ASTM American Society for Testing and Materials BHRG British Hydromechanics Research Group (UK) CETIM

Centre Technique des Industries Mécaniques (France)

EDF

Électricité de France (France)

EPA

Environmental Protection Agency

ICPVT

International Conference on Pressure Vessel Technology

JPAC Jim Payne Associates Company (USA) MPA

Staatliche Materialprüfungsanstalt of Stuttgart (Germany)

MTI

Materials Testing Institute of the Chemical Process Industries

PVP

Pressure Vessel and Piping Conference

PVRC Pressure Vessel Research Council SAE

Society for Automotive Engineering

TTRL

Tightness Testing and Research Laboratory of École Polytechnique de Montréal

UND University of North Dakota (USA) WRC Welding Research Council

Page 296 TABLE 10 Current cooperative PVRC Research Test Programs as of the End of 1996 Program Test Current experimental programs status (end facilities of 1996) Long-duration ARLA tests at moderate Completed temperatures on EF sheet materials (up to 12 WRC months) to confirm and establish the precision of TTRL Bulletin in long-term predictions based on short-term screen preparation tests and aging parameter Ae. Effect of oxidation inhibitor and fluid type and pressure and gasket stress condition on the elevated-temperature behavior of flexibleCompleted graphite sheet materials. This work is proposed as Expected TTRL an extension of the completed PVRC work on the as a WRC characterization of FG sheet gaskets to confirm Bulletin long-term gasket-life predictions for oxidizing atmospheres. Long-term performance of flexible-graphitebased gaskets under steam exposure. This work is TTRL To be proposed as an extension of the completed PVRC CETIM started in work on the characterization of FG sheet gaskets EDF 1997 to confirm long-term gasket-life predictions for steam. Experimental study to develop a qualification test scheme for confined FG-based products Under way TTRL (jacketed, spiral-wound, and metal corrugated) on Completion CETIM the basis of a series of HATR, HOTT, and AHOT for the end EDF tests at specific conditions that ensure a specific of 1997 service life. To be Plant survey to determine good/bad life exposure JPAC started in for sheet FG gaskets. CETIM 1997

Effect of thermal cycles on the hot stress relaxation and blowout resistance of PTFE gasket materials. This work is proposed as an extension TTRL of the completed PTFE Gasket Qualification Project, Industry Sponsored, to confirm the validity of the qualification tools that were proposed for PTFE products. Characterization of NPS- and NPS-8, Class-300 and - 1500, NPS-4 Class- 150 and NPS- 16 Class300 bolted flanged joints subjected to pressure TTRL CETIM and external bending loads and equipped with four different types of gaskets (PTFE sheet, PTFE UND joint sealant, F.G. sheet, and spiral-wound gaskets). Effect of flange surface finish and gasket width TTRL on room-temperature gasket tightness and CETIM emissions of gasketed joints. Correlation between volume part per million (VPPM) and mass leakage rate of gasketing products, especially for very low mass flow rates TTRL and to verify existing empirical formulas CETIM proposed in the CMA (Chemical Manufacturers Association) literature.

Under way Completion for the end of 1997

Under way Completion for end of 1998

Partially under way

To be started in 1997

Page 297 TABLE 10 Continued Current experimental programs

Program Test status facilities(end of 1996)

Safe load limits for gaskets related to tightness and mechanical integrity. This work will extend the use of the CRUSH test procedure to different gasket To be styles and will assist the ASME SWG on BFG by TTRL started providing safe maximum gasket stress limits to in 1997 supplement the PVRC gasket constants Gb, a, and Gs. Test facilities TTRL: Tightness Testing and Research Laboratory (Canada) CETIM: Centre Technique des Industries Mécaniques (France) EDF: Électricité de France (France) JPAC: Jim Payne Associates Company (USA) UND: University of North Dakota Tested materials EF: Elastomeric fiber-reinforced-based gasket products PTFE: PTFE-based products FG: Flexible-graphite-based gasket products

Test Procedures AHOT aged hot operational tightness test procedure ARLA aged relaxation leakage adhesion test procedure ATRS aged tensile relaxation screen test procedure CRUSHCRUSH test procedure EHOT emission hot tightness test procedure

FIRS

fire simulation screen test procedure

FITT

fire simulation tightness test

HALR

high-temperature aged leakage relaxation screen test procedure

HATR

high-temperature tensile relaxation screen test procedure

HOBT hot blowout test procedure HOMT hot mechanical test procedure HORT hot relaxation tightness test procedure HOTT hot operational tightness test procedure LCMT load-compression mechanical test procedure ROMT room-temperature mechanical test procedure ROTT room-temperature tightness test procedure

Page 298

Test Equipment HB rig simple fixture hot blowout test rig RB rig room-temperature bolted-up test rig RH rig room-temperature hydraulic test rig SG rig elevated-temperature single-gasket test rig UG rig elevated-temperature universal-gasket test rig CR gasket creep/relaxation control mode mode DC gasket deflection control mode mode FS gasket creep/relaxation mode under axial flange mode rigidity simulation SC gasket stress control mode mode Other AARH arithmetic average roughness height EF

elastomeric-fiber-reinforced-based gasket products

FG

flexible-graphite-based gasket products

ID

internal diameter of a gasket

LVDT linear voltage differential transducer NBR acrylonitrile butadiene rubber NPS

nominal pipe size

OD

outside diameter of a gasket

PTFE

polytetrafluoroethylene-based gasket products

RTJ gkts

ring-type joint gaskets

SBR

styrene butadiene rubber

SS

stainless steel

VHAP volatile hazardous air pollutants VOC volatile organic compound Nomenclature Most important symbols are presented here. Other symbols are defined in the text when necessary. Ae

equivalent exposure parameter

Aer

recommended equivalent exposure parameter

Dg

gasket deflection (in., mm)

DDg

change of gasket thickness (in., mm)

DFg

change of total gasket load (lb, N)

Fg

total gasket load (lb, N)

Gb, a, Gs PVRC gasket constants k

global joint axial rigidity factor (lb/in., N/mm)

Page 299

Lrm

total mass leak rate through the gasket (mg/sec)

P

Fluid pressure (psig, MPa)

Q(Sg, Tg)

Operating point at a given stress, Sg, and temperature, Tg

Qp

Quality parameter

Qpr

load retention quality parameter

Qpx

tensile quality parameter

Q''t

tightness quality parameter

Sc

lowest CRUSH-cycle maximum gasket stress, where Tpmin(C) > Tpmin(b) (psi, MPa)

Sg

generic gasket stress (psi, MPa)

Sglb

Safe stress limit against gross leakage (psi, MPa)

Ss

gasket tightness-hardening stress limit (psi, MPa)

S100

gasket stress value corresponding to a Tp of 100 (psi, MPa)

S1000

gasket stress value corresponding to a Tp of 1000 (psi, MPa)

S10,000

gasket stress value corresponding to a Tp of 10,000 (psi, MPa)

Tcd

safe cool-down temperature (°F, °C)

TM

temperature margin against gross leakage evaluated at Tr (°F, °C)

Tp

tightness parameter

Tpmin(B)gasket ROTT lower tightness bound Tpmin(C)

highest CRUSH-test minimum tightness that is less than Tpmin(B)

Tpmax

gasket ROTT upper tightness bound

Tps

gasket tightness-hardening limit

Tr

reserve temperature against gross leakage (°F, °C)

Ts

service temperature (°F, °C)

TSX

residual tensile strength of a gasket specimen (psi, MPa)

Tub

Safe limit temperature against gross leakage (°F, °C)

%BLM

percent bolt load margin evaluated at Tr (%)

Acknowledgments Special thanks are expressed to Professor Amhad Chaaban, Dr. Hakim Bouzid, and Olivier Sakr, professional researchers at the Tightness Testing and Research Laboratory (TTRL) of École Polytechnique de Montréal, for their experienced help and valuable comments in the writing and preparation of this document. The authors are also grateful to École Polytechnique for making the time and facilities available during the last 15 years for the development of pioneering experimental work in the field of static fluid sealing for gasketed bolted flanged joints and other pressurized equipment.

The developments and advancements outlined in this chapter would not have been possible without broad industrial support. The authors are grateful to the PVRC Gasket Program sponsors for their support, including financial, of the various gasket test programs over the years; to the PVRC Committee on Bolted

Page 300

Flanged Connections for its directions; and to the MTI for the support and direction through its Project No. 47. References 1. ASME Boiler and Pressure Vessel Code, Section VIII, Div. 1. New York: American Society of Mechanical Engineers, 1989. 2. Payne, J. R. PVRC flanged joint user experience survey. WRC Bulletin 306 (July 1985). 3. Bazergui, A. Short-term creep and relaxation behavior of gaskets. WRC Bulletin 294 (May 1984). 4. Jones, W. F., and Seth, B. B. Evaluation of asbestos-free gasket materials. ASME/IEE Power Generation Conference, Boston, Oct. 1990, 90-JPGC/PWR-58. 5. Winter, J. R. Gasket selectionA flowchart approach. Presented at the 2nd Intl. Symp. on Fluid Sealing of Static Gasketed Joints, La Baule, France, Sept. 1820, 1990. 6. Design Division Problem No. XIII. Re-evaluation of gasket factors used in flange design. WRC Bulletin 298 (Sept. 1984). 7. Raut, H. D., and Leon, G. F. Report of gasket factor tests. WRC Bulletin 233 (Dec. 1977). 8. Raut, H. D., Bazergui, A., and Marchand, L. Gasket leakage behavior trends. WRC Bulletin 271 (Oct. 1981). 9. Bazergui, A., Marchand, L., and Raut, H. D. Further gasket leakage behavior trends. WRC Bulletin 325 (July 1987). 10. Payne, J. R., and Bazergui, A. More progress in gasket

testingThe PVRC program. In 1981 Proc. Refining Dept. Vol. 60. Amer. Pet. Inst., 1981, pp. 271290. 11. Leon, G. F., and Payne, J. R. An overview of the U.S. PVRC research program on bolted flanged connections. Proceedings of the ICPVT-6, Pressure Vessel Technology, C. Liu and R. W. Nichols, eds. New York: Pergamon Press, 1988. 12. Bickford, J. H., Hsu, K. H., and Winter, J. R. A progress report on U.S. PVRC Joint Task Group on Elevated-Temperature Behavior of Bolted Flanges. Proceedings of the ICPVT-6, Pressure Vessel Technology. C. Liu and R. W. Nichols, eds. Vol. 1, Design and Analysis. New York: Pergamon Press, 1988, pp. 249266. 13. Hsu, K. H., Payne, J. R., Bickford, J. B., and Leon, G. F. The U.S. PVRC Elevated-Temperature Bolted Flange Research Program. Presented at the 2nd Intl. Symp. on Fluid Sealing of Static Gasketed Joints, La Baule, France, Sept. 1820, 1990. 14. Chao, R. C. Behavior of bolted flanges at elevated temperatureProgram overview. Proc. 1985 Pressure Vessel and Piping Conf. ASME PVP-Vol. 98. 15. Payne, J. R., and Bazergui, A. Evaluation of test methods for asbestos replacement gasket materials. MTI Publication 36 (1990). 16. Payne, J. R., Mueller, R. T., and Bazergui, A. A gasket qualification test scheme for petrochemical plants. Part I & II. Presented at the 1989 PVP Conference, Honolulu, Hawaii, July 2327, 1989. 17. Rossheim, D. B., and Markl, A. R. C. Gasket loading constants. Mech. Eng. 65 (1943): 647.

Page 301

18. Bazergui, A., Payne, J. R., and Marchand, L. Effect of fluid on sealing behavior of gaskets. Proc. 10th International Conference on Fluid Sealing, BHRA, Innsbruck, Austria, April 1984. 19. Bazergui, A., and Marchand, L. PVRC milestone gasket testsFirst results. Welding Research Council Bulletin 292 (Feb. 1984). 20. Bazergui, A., Marchand, L., and Raut, H. D. Development of production test procedure for gaskets. Welding Research Council Bulletin 309 (Nov. 1985). 21. Bazergui, A., and Louis, G., Predicting leakage for various gases in gasketed joints. Society for Experimental Mechanics, 1987 Spring Conf. on Experimental Mechanics, Houston, Texas, June 1987. 22. Payne, J. R. Standard test method for gasket constant for bolted joint design. Presented to the ASTM Committee F3, February 1992. 23. Payne, J. R., Bazergui, A., and Leon, G. New gasket factorsA proposed procedure. Proceeding of 1985 Pressure Vessels and Piping Conference, 1985. 24. Payne, J. R., Leon, G., and Bazergui, A. Getting new gasket design constants from gasket tightness data. Special Supplement, Experimental Techniques, Society of Experimental Mechanics, Nov. 1988. 25. Marchand, L., Derenne, M., and Bazergui, A. Weight loss correlation for sheet gasket materials. Journal of Pressure Vessel Technology, Vol. 114, Feb. 1992.

26. Marchand, L., Bazergui, A., and Derenne, M., Recent developments in elevated-temperature gasket evaluation. Presented at the 13th International Conference on Fluid Sealing, Bruges, Belgium, 79 April 1992. 27. Bazergui, A., and Payne, J. R. On the elevated-temperature behavior of gaskets. Proceeding of the ICPVT-6, Pressure Vessel Technology. Vol. I, Design and Analysis. New York Pergamon Press, 1988. 28. Payne, J. R., Derenne, M., and Bazergui, A. A device for screening gasket materials at elevated temperature. Proc. 11th Fluid Sealing Conf., Science Publisher, Cannes, France, April 1987. 29. Marchand, L., Bazergui, A., and Derenne, M. The influence of thermal degradation on sealing performance of compressed sheet gaskets with elastomer binder. Part I: Experimental methods. 2nd International Symposium on Fluid Sealing of Static Gasketed Joints, La Baule, France, Sept. 1990. 30. Derenne, M., Payne, J. R., Marchand, L., and Bazergui, A. Development of test procedures for fire resistance qualification of gaskets. Welding Research Council Bulletin 377 (Dec. 1992). 31. Bazergui, A., Marchand, L., and Payne, J. R. Development of a hot tightness test for gaskets. Proc. 11th Fluid Sealing Conf. Cannes, France, April 1987. 32. Bazergui, A., Marchand, L., and Payne, J. R. Development of tightness test procedures for gaskets in elevated-temperature service. Welding Research Council Bulletin 339 (Dec. 1988). 33. Tightness Testing and Research Laboratory of École Polytechnique of Montréal. Gasket performance characterization for bolted flanged connections (October 1994).

34. Derenne, M., Payne, J. R., Marchand, L., and Muzzo, U. Elevated-temperature characterization of flexible graphite sheet materials for bolted flanged joints. Welding Research Council Bulletin 419 (Feb. 1997).

Page 302

35. Payne, J. R. Bolted joint improvement through gasket performance tests. 1992 NPRA Maintenance Conference, San Antonio, Texas, Bulletin MC-92-76. 36. Payne, J. R. Improved bolted joints through gasket performance tests. MTI/NACE First International Symposium on Industry Piping, Orlando, Florida, 1993. 37. Hsu, K. H., Payne, J. R., and Derenne, M. Recent developments in PVRC elevated-temperature gasket testing. CETIM 3rd International Symposium on Fluid Sealing, Biarritz, France, September 1993. 38. Derenne, M., Marchand, L., Payne, J. R., and Bazergui, A. Elevated-temperature testing of gaskets for bolted flanged connections. WRC Bulletin 391, May 1994. 39. Reid, R. C., Prausnitz, J. M., and Poling, B. E. The Properties of Gases and Liquids, 4th ed. New York. McGraw-Hill, 1987. 40. Chemical Rubber Co. CRC Handbook, 74th ed. 1994. 41. TTRL École Polytechnique. PTFE Gasket Qualification Project: Final Report. October 1995. 42. Batista, P. B., Marchand, L., and Derenne, M. A. proposed model for predicting leakage through porous gaskets. Proceedings of the ASME-PVP Conf. Vol. 305, Hawaii, July 1995. 43. Deshaies, F., Derenne, M., and Marchand, L. Effect of time on the ambient leakage behavior of gaskets. Final Report, PVRC Project 9294. 44. Vignaud, J. C., Nowak, H., and Digat, P. Mechanical and

sealability characteristics of expanded graphite gaskets. 2nd International Symposium of Fluid Sealing of Static Gasketed Joints, La Baule, France, Sept. 1990. 45. Draft 9 of the proposed ASTM Method for the Standard Test Method for Gasket Constants for Bolted Joint Design and Addendum to Draft 9, Section 9. 46. Draft 1 of the proposed ASTM Method for the Practice for Measuring Gas Leakage Through Gaskets. 47. Marchand, L., and Derenne, M. Fugitive emission characteristics of gaskets. Final Report, PVRC Project 92.25, Feb. 1997. 48. PVRC 5-year plan, rev. version, March 1, 1995. 49. Marchand, L., and Derenne, M. Long-duration air and steam screening tests on elastomeric sheet gasket materials. Final Report submitted to the Subcommittee on Gasket Testing and Elevated Temperature Joint Behavior, PVRC, New York, October 1995. 50. Marchand, L., and Derenne, M. Long-term performance of elastomeric sheet gasket materials subjected to temperature exposure. Proceedings of the ICPVT-8, ASME International, Vol. 1, Montreal, Canada, July 1996.

PART III: SELECTING A GASKET

Page 303

6 Gasket SelectionA Flowchart Approach J. RONALD WINTER Eastman Chemical Company, Kingsport, Tennessee 1. Introduction Due to increasing environmental concerns, especially fugitive emissions, it has become necessary to control and predict leak rates from valve stems, agitator shaft seals, pump shaft seals, and gasketed flanges. In general, the leak rate from a gasketed flanged joint is far less than that from the other three sources [30]. However, the number of flanges versus the number of valves, pumps, and agitators makes them just as big a total contributor to fugitive emissions as the other sources. Another factor that should not be overlooked is the cost of leaks associated with lost product. For example, ethylene gas at a price of $0.58/L would cost the user $905 per year at a leak rate of 3 cc/min. If this leak rate were reduced to 0.25 cc/min, the cost would be $75.50 per year, a cost reduction of $829.50 per year. Thus tighter joints, while protecting the environment, will also reduce operating costs. Furthermore, the savings will generally more than justify the use of a more expensive gasket and/or flange. With fugitive emissions looming as a major environmental problem in the future, engineers must reconsider their basis for selecting gaskets and the associ-

Page 304

ated flange design procedures. It should be fairly obvious that the design criteria will have to be based on allowable leakage rates. This is where we, as engineers and as general practitioners of the profession, must make some, perhaps far-reaching, decisions that will have a significant impact on our ability to meet future emission control limits. In this decision process many parameters come into play. If any one of these parameters is not properly considered, the result could be a flange design and/or gasket selection that will not yield an acceptable joint from a leakage point of view. Establishing the key parameters is vital to achieving the desired goal of minimizing leakage at reasonable costs. To augment the technical considerations, one must, to some extent, consider other internal and external effects. These are shown in a fishbone (Ishikawa) diagram in Appendix L. The first parameter, Lrm, is the mass leak rate per unit diameter of the gasketed joint. This sounds rather simple but is not. First ask, What will be my basis for establishing allowable mass leak rates? Then determine if different leak rates will be permissible for different classes of liquids and gases depending on their toxicity, carcinogenic properties, flammability, etc. So the picture becomes somewhat cloudy. Assuming we have established allowable leak rates for several classes of fluids, then we must consider the types of gaskets available, the effects of temperature, pressure, and diameter, chemical compatibility, the flange assembly methods to be used, etc. Therefore, considerable thought will be needed in the initial stage of establishing new criteria for the design of gasketed joints. Certain parameters, once established, will have a long-term effect on the design process. Many of the others will need to be considered each time a new gasketed joint is designed. This fact

will be especially true for nonstandard flanged joints and, more specifically, for those operating in severe environments. Once this technology has been incorporated into the national codes, such as the ASME Unfired Pressure Vessel Code [29], then the present ASME/ANSI flange standard B16.5 [32] can be reevaluated. It is important to realize that the intent of this chapter is to assist process designers in the selection of gaskets for flanged joints [50]. The actual flange design process is not within the scope of this chapter. The special ASME Code Working Group on Flanges is responsible for incorporating this new gasket technology into the ASME Code flange design process and enhancing the design procedure in the future as advances in technology dictate. The tentative new ASME Code flange design procedure is discussed in Chapter 8. The final flange design procedure may be quite different or have additional restrictions [78]. 2. Detailed Discussion of Gasket Behavior The PVRC Gasket Testing Task Group, in an effort to establish reasonable leak rates, obtained leakage data from numerous ASME/ANSI standard flanged joints that are typically used in major petrochemical plants. This data was averaged to

Page 305

establish a proposed nominal leak rate. A corresponding joint tightness parameter, Tp, was also established [2,4,6,16,17,18,23]. The tightness parameter, Tp, is defined by Eq. (1). See Chapter 8 for more details on the development of this equation. (1) where internal pressure, MPa (psi) P= P* =atmospheric pressure, MPa (psi) mass leak rate per unit diameter, mg/sec-mm (lb/hrLrm = in.) reference mass leak rate per unit diameter [ = mg/sec], mg/sec-mm (lb-hr-in.)

=1

A standard tightness class called T2 has been defined. It represents a nominal leak rate per unit diameter of 0.002 mg/sec-mm (0.0004 lbm/hr-in.). To reflect the need for joints that are required to have a lower leak rate (a tighter joint) as well as those where more leakage is permissible, additional classes were added: T1, T3, and T4 [23]. They are presently referred to as the Economy Class, Tight Class, and Very Tight Class. Each class represents two orders of magnitude change in leak rate. To make the differentiation between different classes of chemicals simpler, additional tightness classes could be established, one between T1 and T2, one between T2 and T3, and one between T3 and T4. Each company or code would need to decide on the need for such subdivisions. These tightness

classes are shown in the Table 1. The equation in English units that defines the minimum tightness Tpmin based on the data presented in Table 1 is:

Page 306

Tpmin = minimum tightness = 1.82574(C)(Pr) (2) or Tpmin = 1.83(C)(Pr) = 0.124(C)(P) (English units) (3) where C = a constant from Table 1 P

= internal pressure, psi (MPa)

P*

= atmospheric pressure, psi (MPa)

Some typical classes of materials and/or operating conditions that need to be assigned specific tightness values are listed next. These are discussed in more detail later in this chapter. Spontaneously ignites when released [A] Combustible gases/liquids

Requires an ignition source Will ignite only in the presence of an existing flame Carcinogenic Noncarcinogenic but toxic

[B] Toxic

Toxic to the environment but not

gases/liquids

to man Lethal

[C] Liquids/gases at high temperature [D] Liquids/gases at high pressure [E] Nonharmful liquids/gases Before considering further the gasket selection process, it is necessary to consider briefly the new gasket technology. First consider the typical graph of mass leak rate versus gasket stress shown in Fig. 1. This particular graph shows leak rate variations at three different internal pressures: 400 psi (2.8 MPa), 800 psi (5.5 MPa), and 1200 psi (8.3 MPa) [1,4,18]. The curves shown in Figure 1 were condensed to a single curve by use of the previously defined nondimensional tightness parameter, Tp. The resulting graph is referred to as a tightness curve [17]. A typical tightness curve is shown in Fig. 2. Note that a minimum tightness value Tpmin is shown in this figure. The associated gasket stress is called the minimum assembly gasket stress, SAmin [17,23]. The SAmin value can be used for the initial flange design when using classical flange design relationships [35,37,38,39,40]. As the process is pressurized, the gasket stress usually decreases, following an unloading curve like that shown in Fig. 2. In cases where the gasket stress increases due to differential thermal expansion, one either proceeds up the initial

Page 307

Figure 1 Typical curve of leak rate vs. gasket stress.

loading or seating curve if no unloading had occurred due to pressure or proceeds up from the unloading curve as shown in Fig. 2. In any event, the new gasket stress is often referred to as the operating gasket stress, SGmin or SGo. If the operating tightness value, Tpo, associated with SGo is less than Tpmin, then the assembly stress will need to be increased by some amount DSg such that Tpo is equal to or greater than Tpmin at operating conditions. The adjusted assembly tightness value, shown in Fig. 3, is referred to as TpA. This is the value used in the final design of the flanged joint. Note: In Ref. 17, TpA is designated Tpn. The tightness curve of Figs. 2 and 3 can also be defined by three constants called Gb, a, and Gs, formerly called B, d, and S*. These constants, which are referred to as the PVRC gasket factors, will be used in the new

ASME code flange design procedure. The basic relationship* between gasket stress and tightness based on the new gasket constants Gb, a, and Gs is: (4) *Note: This equation linearizes the data on a log-log plot.

Page 308

Figure 2 Typical gasket stress vs. tightness curve.

Figure 3 Gasket stress vs. tightness curve.

Page 309

where Gb =

intercept of the gasket loading cure (psi)

Tp =tightness a =slope of the loading curve They are further discussed in Chapter 8. But how do we systematically select from the vast number of available gaskets the ones that will meet the increasingly stringent leakage requirements? A flowchart is perhaps the simplest technique to assist designers in selecting gaskets. This will be discussed in detail in the following paragraphs. 3. Flowchart Discussion The flowchart shown just before the appendices to this chapter will be discussed on the ensuing pages. As we proceed through the various parts of the flowchart, support information from test data and in-the-field experiences will be used to expand on the particular block or section to give the reader a wider understanding of the problems and constraints with which he or she must deal. The easiest way to reduce emissions is to eliminate the sources of such leaks, i.e., to eliminate as many flanges as possible. If flanges are needed, they should be strategically located such that they are easy to assemble and maintain. It is important to realize that all joints leak. The leak rate may be parts per billion, but each joint does leak. There is no such thing as zero emissions in any pressurized system that has joints, be it the coolant system in your car, the air in your car tires, a chemical process, or a natural gas

pipe line. 3.1. Establishing Acceptable Leak Rates After eliminating unnecessary joints, one must establish an allowable leak rate for those remaining. The basis for doing this is quite complicated, as shown on the first page of the flowchart. Of course, economics is the basic reason we do not design all flanges for the minimum possible leak rate. The first step is to establish the classification of the confined materials (M1 through M7) and the associated tightness class (T1 thru T4). The material classifications, M1 thru M7, involve chemical hazard ratings, medical (carcinogenic) constraints, pressure effects, thermal effects, and special EPA and/or other government requirements. A typical chemical hazard rating table is shown in Fig. 4. Material safety data sheets (MSDS) sheets can be most useful during this process [79,80]. As shown in the flowchart starting on page 346, there is a tightness class associated with each material classification. The minimum tightness factor, Tpmin, is defined by Eq. (2). Note: Tpmin is associated with the maximum allowable leak rate for a particular material classification. Since the ASME Pressure Vessel Code requires a hydro test at a minimum of 1.5 times the design pressure, it is

Page 310

Figure 4 Typical chemical hazard ratings table.

Page 311

recommended that the tightness value be multiplied by 1.5 as the basis for selecting gaskets that will also maintain a seal during the hydro test; i.e., select gaskets that are capable of reaching a tightness value of 1.5Tpmin. 3.2. Temperature Considerations Once we have selected the gaskets that meet the 1.5Tpmin requirement, we must determine which of these gaskets will meet the process or design temperature requirement. First establish the maximum process operating temperature, Tmax, and determine which of the gaskets are rated for this value. Of the remaining acceptable gaskets, determine which ones cannot maintain the desired tightness value of 1.5Tpmin as the temperature increases to Tmax [8,12,13,15,19,21,57]. Be aware that a different tightness curve may be needed at elevated temperatures since the material properties of the gasket may have changed, resulting in a change in leakage behavior [13,21]. For example, consider the tightness curves shown in Fig. 5 and 6 for two spiral-wound gaskets, one with mica-graphite fill and one with asbestos fill. Each gasket was tested at two temperatures, ambient and 800°F. At ambient temperature, both gaskets perform quite well, i.e., reasonable loading slopes and fairly steep unloading curves. From a tightness point of view, the gasket with

Figure 5 Tightness curve at two temperatures for a spiral-wound gasket with mica-graphite fill.

Page 312

Figure 6 Tightness curve at two temperatures for a spiral-wound gasket with asbestos fill.

mica-graphite is the better performer at ambient temperature. But that changes dramatically at 800°F. The unloading slope of the mica-graphite-filled gasket has become very shallow, showing extreme sensitivity to any change in gasket load or deflection. However, the asbestos-filled gasket tightness curve is virtually unchanged at 800°F. Thus at 800°F the asbestos-filled spiralwound gasket would be the gasket of choice, which is the reverse of the selection at ambient conditions. In this case the PVRC gasket factor Gs has changed dramatically (from 25 to 1300) for the micagraphite-filled gasket. The problem with thermal degradation has become increasingly important. As a result, this topic and its affect on gasket selection will be discussed in great detail. This expanded discussion will also cover the various reasons that have led to increased interest in thermal degradation.

Due to lung problems associated with asbestos fibers* and the resulting overreactive government regulations that were not based on sound research plus the resulting overzealous legal environment, nonasbestos substitute gaskets were rushed to the market without sufficient long-term testing at temperature. Due to the ensuing problems encountered in the field, PVRC and MTI were both requested to *Associated mainly with asbestos insulation and to a lesser degree with manufacturing processes that use asbestos.

Page 313

investigate the degradation of these asbestos substitutes, primarily aramid fibers, as a function of temperature. Asbestos-fiber gaskets were used as the reference material. Due to the extreme importance of this matter, the next several paragraphs will deal with the many facets of thermal degradation of various gasket materials, such as aramid sheet, asbestos sheet, flexible graphite, and PTFE. Studies of compressed-asbestos and nonasbestos sheet gasket at École Polytechnique de Montréal revealed a correlation between gasket thermal degradation, principally weight loss, and leak rate. It was found that aramid-fiber gasket began serious degradation at 350°F (see Figs. 7 and 8.) MTI gasket tests showed similar thermal degradation (see Appendix O as well as Appendix N). But even at temperatures between 200°F and 350°F those materials begin to lose resilience with time. It is important to note in Fig. 7 and 8 that the although asbestos gaskets also deteriorate, they do not continue to deteriorate as the aramid gaskets do. Unfortunately, this limited degradation above a certain temperature level may render this reliable gasket unable to meet the new PPM fugitive emissions limits even though it was more than adequate for past definitions of a leak, i.e., drops per minute. Subsequent work explored the effect of time at temperature and led to the development of correlations whereby one can estimate the useful life of fibrous sheet gaskets at a given temperature or temperatures [42,43,51,53,57]. This is further discussed in Chapter 4. Subsequently, another potential major replacement for asbestos was rapidly introduced, with considerable acclaim. It was flexible graphite. However, time

Figure 7 Thermal degradation of compressed-fiber gaskets: leak rate vs. temperature.

Page 314

Figure 8 Thermal degradation of compressed-fiber gaskets: fraction of weight loss vs. temperature.

at temperature again dispelled initial claims due to oxidation. The Westinghouse Corporation was the first to experience major problems with the flexible-graphite-sheet gasket. This occurred with steam turbine gaskets. They were attempting to replace the old standard-asbestos-sheet gaskets with flexible graphite. They had gaskets failing with time at temperatures well below the commonly advertised 850°F oxidation limit for flexible graphite. As a result, they performed a series of tests at their facilities. The results were later shared with PVRC and published in a major journal [53]. The basic results are shown in Fig. 9. This data was subsequently duplicated by PVRC at the École Polytechnique Tightness Testing Research Laboratory (TTRL); see Fig. 10. It is obvious that at temperatures as low as 600°F this material, after a dwell time between 1000 and 2000 hr, will begin to degrade. Additional curves from the TTRL tests [58] are shown in Appendix M. As experience continued, problems also began to occur with spiral-

wound gaskets filled with flexible graphite. For this type of gasket, the degradation process is slower, since the flow of oxygen is reduced by the metal spirals. In any event, numerous failures have been encountered, as shown in Fig. 11 and 12. As one would conclude from the previous discussions, time at temperature has proven to be an important parameter in evaluating gaskets, especially sheet gaskets, for long-term in-the-field use. This led to the development of new elevated-temperature PVRC gasket tests to shed additional light on the short-term effect of elevated temperature on gasket performance, i.e., creep, relaxation, and leak rate. Some typical tests are ROTT, ATRS, HOTT, ARLA, AHOT, HOBT,

Page 315

Figure 9 Westinghouse flexible-graphite-sheet test data.

EHOT, FITT, HATR, FIRS, and ROMT. The results of a typical HOTT test are shown in Fig. 13. Longer-term elevated-temperature tests have also been developed. These tests, of course, are more expensive. Another test developed specifically for Teflon (PTFE) materials is the HOBT test, i.e., the HOt Blowout Test. The purpose of this test is to determine the temperature at which a PTFE gasket will blow out at a given pressure. See Chapter 4 for a further discussion of these gasket tests.

Figure 10

PVRC duplicate flexible-graphite-sheet test data.

Page 316

Figure 11 Photos of a graphite-filled spiral-wound gasket that failed in 3 days after exposure to 1400°F-1500°F temperatures. (The actual gasket temperature was in excess of 1300°F. In addition, the gasket load may have been low and not uniform.)

Page 317

Figure 12 Photos of a graphite-filled spiral-wound gasket showing oxidation from the OD inward. The process temperature was between 1100°F and 1200°F.

Page 318

Figure 13 Results of a typical HOTT test.

All of these tests, plus field experience related to meeting the EPA's volatile hazardous air pollutants (VHAPs) emission constraints has proven that virtually all gaskets deteriorate with time. At one time, certain gaskets were thought to last indefinitely, but the low emission limits has dispelled this myth. The higher the temperature, the faster the deterioration. Some suggested long-term temperature limits for several generic gasket types and/or materials is shown in Appendix N. Note that these are actual gasket temperatures, not process temperatures. See Ref. 52 for more detailed information. 3.3. Gasket Temperature Rating Limits Temperature limits for gaskets have caused a lot of concern. Vendor claims versus field experience, as well as plant-to-plant and process-to-process discrepancies, have been the major causes for the confusion. For instance, one company may claim to have

successfully used a gasket at 500°F while another says they had the same gasket fail at 350°F. This has been a quite common situation, which has led to a great deal of confusion as well as many accusations. An example of such a confusing situation involved the use of flexible-graphite-filled spiral-wound gaskets in a hightemperature process (1200°F to 1500°F) without severe deterioration in a typical 9-month run cycle. In an effort to explain this situation, infrared temperature measurements were taken of the uninsulated flanges. Typical

Page 319

results are shown in Fig. 14 and 15. It is obvious from these figures that the flange and gasket temperatures are far below the process temperature. Thus the reason for the limited degradation was understood. This situation resulted in a large investigation entitled Flange Thermal Parameter Study [52]. The study consisted of a large finite-element heat-transfer analysis of uninsulated flanges. Typical graphs of gasket temperatures versus the heat transfer film coefficient, h, are shown in Fig. 16 and 17. Typical transient heattransfer analyses results [52] (thermal profiles) for two time slices (1 min and 12 min) are shown in Fig. 18 and 19. It is obvious from these results that the film coefficient has a dramatic effect on the actual gasket temperature, provided the flanges are not insulated. It also explains why one person can successfully use a given gasket at a quite high process temperature while another person may have the same gasket fail at a much lower process temperature. In some cases, gasket failures have been blamed on temperature, poor assembly, or the gasket itself when none of these were actually at fault. For instance, extrusion failures of various types of PTFE gaskets were initially blamed on the PTFE material itself, temperature, or poor assembly. Careful review of the process procedures as well as the maximum possible temperatures indicated

Figure 14 Temperature measurements on a 4-in. lap joint nozzle flange on a condenser.

Page 320

Figure 15 Temperature measurement on a 5-in. stainless steel weld neck (integral) flange.

Figure 16 Gasket OD temperature vs.h for 24-in. Class 150 and Class 1500 lap joint flanges.

Page 321

Figure 17 Gasket OD temperature vs. h for 3-in. Class 150 and Class 1500 lap joint flanges.

that some other major force was involved. Tests [60] also showed that assembly was not the real cause, even though it could have been a contributor. HOBT tests of PTFE gaskets at the Tightness Testing Research Laboratory (TTRL) at Ecole Polytechnique in Montreal also indicated that temperature was not the problem. In these situations, externally heated pipe lines were being valved off while full of liquid but without removing the heat source. The liquid, being an incompressible fluid, expanded as it was heated, producing pressures that could have exceeded 1000 psi. In fact, pressures in the 2000- to 3000-psi range have occurred in Class 150 flanged pipe systems. Needless to say, this resulted in extrusion failure of the gaskets [60,61]. Two typical examples are shown in Fig. 20 and 21. Obviously, one must be very careful when reviewing gasket failures.

If elevated-temperature tightness curves or other high-temperature leakage/weight loss test data is not available, you can get a good indication of changes in gaskets performance with increasing temperatures by noting changes in the gasket stress deflection curve at different temperatures. As an example, consider Fig. 22, which contains stress deflection curves for a flexible-graphite-filled spiral-wound gasket at four temperatures: ambient, 350°F, 850°F, and 1200°F. It is quite obvious that a major change took place between 850°F and 1200°F. Before using that particular gasket at a temperature above 850°F, you should

Page 322

Figure 18a Thermal profile at time = 1 min and h = 8 (24-in. Class 150 weld neck flange).

obtain an elevated-temperature tightness curve to assess its change in leakage behavior properly. You also must determine the gasket life as dictated by oxidation. From the previous comments and examples it is obvious that temperature alone can pose a rather difficult problem but that this situation can be greatly aggravated by thermal gradients associated with cyclic operation: start-up/shut-down cycles, poor (especially nonsymmetric) heat transfer, etc. These situations are discussed in more detail in the next paragraph as well as later in this chapter. The next step (block in the flowchart), which involves diameter effects, is difficult, because it requires both judgment and experience. There is little data to substantiate this decision process

other than simple logic. It is fairly easy to understand that as the diameter of a vessel increases, the design process becomes more difficult. Many equations that work well for small-diameter vessels will not yield an adequate design for a large-diameter vessel under the same operating conditions. In a sense, gaskets are the same way. For instance, at a given temperature and pressure, a gasket that works well in a small-diameter flange may not

Page 323

Figure 18b Thermal profile at time = 1 min and h = 128 (24-in. Class 150 weld neck flange).

be satisfactory in a large-diameter flange. For example, a compressed-asbestos-sheet gasket is generally acceptable for 600psi steam for diameters less than 6 in., but becomes more susceptible to blowout as the flange diameter increases [7,14,15,19]. Such conclusions may not be due to the gasket alone, but may also be due to the inadequacy of the gasketed joint design and/or the assembly process. A general rule of thumb is that sheet-type gaskets, such as compressed asbestos, compressed nonasbestos fiber, PTFE, and flexible graphite, can be successfully used at smaller diameters, lower pressures, and/or lower temperatures. As diameter, temperature, and/or pressure increases, sheet gaskets become less

acceptable due to the increasing possibility of blowout. In the latter case, composite gaskets, such as spiral-wound gaskets [28], metal O-rings, double-jacketed gaskets, and metal corrugated gaskets [19],* and solid metal gaskets, *At process temperatures above 600ºF, it is generally preferable not to insulate flanges to keep the bolts and flange as cool as possible [52]. This prevents or reduces the effect of stress relaxation and creep of all flange components. See Appendix J.

Page 324

Figure 19a Thermal profile at 12 min for h = 8 (24-in. Class 150 weld neck flange).

such as oval rings, delta rings, and flat metal rings, are needed. A series of curves showing this selection philosophy is presented in Fig. 23. 3.4. Fire Integrity Next in the flowchart is the term fire integrity. It refers to the ability of a gasket in a flange to withstand flame impingement for a maximum of 30 min without failing and allowing additional combustible material to feed the fire. In this case, an asbestos gasket, a flexible-graphite gasket, a composite gasket with a noncombustible fill material, or a solid metal gasket would be needed [8,13,15,16,21]. The PVRC tests FIRS and FITT developed at Ecole Polytechnique are used to simulate a fire environment. See

Chapter 4 for a discussion of these tests. 3.5. Cyclic Service Considerations Is the gasket sensitive to cyclic service? Actually all gaskets are somewhat sensitive to cyclic service, i.e., a service where relatively large changes in thermal

Page 325

Figure 19b Thermal profile at 12 min for h = 128 (24-in. Class 150 weld neck flange).

and/or mechanical loads are normally expected during a typical process run, a service that is batch operated, or a service that has relatively short run times. Gaskets that are used in processes that operate at temperatures below ambient can also pose some difficult sealing problems. In this situation, the materials of the joint contract. If the gasket material, say, PTFE, has a greater coefficient of thermal expansion (contraction) than the other materials in the joint, then the gasket load will decrease, resulting in a leak. The use of a different material may be required. The addition of a corrugated metal core to an expanded PTFE material can help this situation considerably. If you have cyclic service, you will want the gasket that is the most tolerant to such service. This requires a gasket that is quite resilient.

The best indicator is the change in tightness, Tp, that occurs with changes in either load or deflection. For cyclic service you would ideally like a gasket that, once seated, can experience relatively large changes in load or deflection with only small changes in tightness or leak rate [16]. As an example, consider the tightness curves for two metal ring gaskets, one made of soft aluminum and the other of annealed copper. From the curves shown in Fig. 24 and 25, it is obvious that the unloading curve for

Page 326

Figure 20 Extrusion failure of a 4-in. filled PTFE gasket.

the copper gasket is considerably steeper than the one for aluminum, clearly indicating that the copper gasket is the better of the two for cyclic service. In other words, the PVRC gasket factor Gs for the copper gasket is considerably smaller than that for the aluminum gasket. In some cases, a gasket may be more sensitive to thermal cycles than to pressure cycles, or vice versa. A gasket that has a large hysteresis effect is not desirable. Likewise, a gasket that undergoes considerable creep or stress relaxation is undesirable in cyclic service [1,2,8,13,16,18,21]. A more subtle type of cyclic service is that involving vibration and/or pressure pulses associated with reciprocating equipment, such as positive displacement pumps. Such cyclic loads, which are usually of a low frequency, can lead to a loss of bolt load and increased leakage. Of course, high-frequency, low-amplitude loads

can also lead to a loss of bolt load and to leakage. Bolt preload is very important in dealing with this problem. Another area of concern is large-amplitude cyclic or noncyclic loads associated with the thermal expansion of pipe systems. Thermal growth of vessels can also induce large loads in nozzle and pipe flanges that can lead to seal problems.

Page 327

Figure 21 Extrusion failure of a 2-in. filled PTFE gasket.

3.6. Creep and Stress Relaxation Creep and/or stress relaxation can often be a severe problem. As previously mentioned, gasket creep and/or relaxation is detrimental in cyclic service. It can, likewise, be a problem for noncyclic service. A gasket that creeps excessively will often seal very well initially, but the seal will deteriorate with time. PTFE is an example of this type of behavior. The addition of specific fill materials to PTFE has served to reduce significantly the effect of creep and stress relaxation. The addition of special metal inserts or corrugated metal inserts with expanded PTFE has also helped the inherent PTFE creep problem. In any event, materials that creep excessively have a greater tendency to begin leaking during a process transient or shutdown. Many gaskets are in this group. Of course, increasing temperature has a detrimental effect on this problem, as shown in

Fig. 26 and 27 [5,18,19]. These figures represent the creep behavior of two spiral-wound gaskets with different fill materials, chlorite-graphite and flexible graphite, at two temperatures, ambient and 650°C (1202°F). It is obvious that both gaskets suffer a considerable increase in creep at 650°C (1202°F), but at both temperatures the graphite-filled gasket has the lower creep rate. Thus, it would be the preferred

Page 328

Figure 22 Stress deflection graphs for a graphite-filled spiral-wound gasket at 70°F, 350°F, 850°F, and 1200°F.

gasket if creep was an important selection parameter.* However, it is important to realize that due to flexible-graphite oxidation, the graphite-filled gasket would have to be replaced periodically to maintain the desired seal. One should also be aware that at elevated temperatures, other flange components may undergo creep and/or stress relaxation, thus adding to long-term sealing problems. To contain highly toxic materials or highly combustible materials, a special gasket is needed to ensure maximum safety. This is referred to as an intrinsically safe gasket. This means the gasket will not blow out and, thus, will minimize leakage under adverse

conditions and allow the particular system to be shut down *Note: Flexible graphite suffers long-term oxidation at these temperatures. See pages 311 and 312.

Page 329

Figure 23 Gasket diameter, pressure, temperature selection graph.

Figure 24 Tightness curve for a soft-aluminum ring gasket at room temperature.

Page 330

Figure 25 Tightness curve for an annealed-copper ring gasket at room temperature.

safely without a major release. In general, this will require either a metal composite gasket or solid metal gasket. At pressures above roughly 2500 psi, solid metal gaskets are generally preferred. Of course, this transition pressure is dependent on the diameter of the particular flange. Some typical solid metal gaskets are oval rings [33], delta rings, lens rings [34], and coated flat metal rings [15]. Typical flanges associated with such pressures are ring joints [33] and lens ring joints [31]. Some examples of proprietary highpressure joints are TAPER-LOK and GRAYLOC [14]. 3.7. Tolerance to Process Chemicals The gaskets that remain in the selection list now must meet another critical test: Will they tolerate the chemicals with which they will be in contact? That is, are they compatible with the contained

chemicals? If test data or vendor data is not available, then special laboratory testing may be required. Beware of assuming either that: (1) if the gasket is good at one concentration and/or temperature or temperature range, it will be good at other concentrations and/or temperatures; or (2) if it is good for two or more chemicals alone, it will be good for a mixture of these chemicals. It is always wise to compare your selection with the type of

Page 331

Figure 26 Creep curves for a chlorite-filled spiral-wound gasket at 72ºF and 1202ºF.

Figure 27 Creep curves for a graphite-filled spiral-wound gaskets at 72ºF and 1202ºF.

Page 332

gasket presently being used in the process or in a chemically similar process. Industrial experience is very valuable here and serves as a good check on your basic logic. 3.8. Other Selection Considerations Different types of gaskets may have unique problems due to their formulation or construction. For instance, spiral-wound gaskets may buckle radially inward due to differential thermal expansion, thermal gradients, and/or the incompressible nature of certain fill materials [3,83]. See Appendix C for typical examples of buckled spiral-wound gaskets. The buckling problem can usually be corrected by using an internal gage ring of the proper size. Other gaskets may have similar unique problems. Other requirements might involve FDA approval for use in food processing, photosensitivity requirements for the photographic industry, or radiation tolerance for nuclear applications. Next, determine the minimum gasket seating stress, SAmin, based on Tpmin and 1.5Tpmin. This can be established either from the gasket stress-tightness curve, through an equation based on regression analysis of the actual leakage test data, or from the PVRC gasket factors Gb, a, and Gs (see Eq. 4). These new gasket factors are discussed in Chapter 8. An ASTM Standard* should be available in the near future for gasket testing and for the conversion of the test data into tightness curves and the associated PVRC gasket factors Gb, a, and Gs [16,17,23]. If the flange is to be designed by a vendor, one option is to have the design seating stress listed in the specification to be no less than the gasket stress associated with 1.5Tpmin. If test data or field experience with a similar environment indicates that a more

conservative flange design is needed, then increase the gasket seating stress to some value you feel will be adequate. Another option would be to use a higher tightness value or tightness class, which automatically results in a higher gasket seating stress. Note: The 1.5 factor is already included in the code-oriented procedure discussed in Chapter 8. If the flange already exists or if it is a standard flange,** then, from the remaining list of gaskets, select one that will not result in overloading the flange and produce high stress levels or large flange ring rotations. If no gasket meets this criteria, then the flange should be replaced by either a higher-rated standard flange or by a customized design. If the flange (gasketed joint) is to be custom designed, then, from the remaining gaskets, select one that is the most economical based on a combination *If the ASTM standard has not completed the approval process, a draft version can be obtained from the ASTM Committee on Gaskets. **A special program is being proposed by the PVRC Subcommittee on Flange Parameter Studies to rerate the standard flange based on tightness.

Page 333

of the gasket cost, the flange cost, and assembly (maintenance) costs. Once a gasket has been selected, one should specify the gasket surface finish required. Generally, the gasket manufacturers will recommend specific surface finishes for their gaskets. For example, a surface finish range of 125 to 250 AARH is usually recommended for spiral-wound gaskets, while a value of 500 AARH is more common for sheet gaskets. Some designers prefer a serrated finish (5001000 AARH) for sheet gaskets. However, manufacturers of some of the newer sheet gaskets are recommending the 125250 AARH range. The present ANSI Flange Pipe Standard B16.5 specifies a surface finish range of 125500 AARH [32]. For solid metal gaskets, the surface finish should be less than 125 AARH, with the final value depending on the type of metal used to manufacture the gasket. See Appendix A [45]. 3.9. Selecting an Assembly Procedure Next, an assembly procedure should be selected. The cross or star assembly patterns are commonly used. But just as important as the torque pattern is the method used to control the bolt preload. Four basic alternatives are listed in Table 2 [7,912,20,22,23,25,26,36,46,77]. Some of the older methods to control bolt load, such as turn of the nut, have been shown to be quite inadequate. See Appendix I. Recent in-the-field flange assembly experiences with the use of an ultrasonic extensometer indicate that the thinner the flange, the more sensitive the gasketed joint is to variations in bolt load. Likewise, the stiffer the flange, the less sensitive it is to variations in bolt load. In other words, for the same variations in bolt load, a thinner flange is more likely to develop a major leak than is a

thicker (stiffer) flange of the same diameter using the same gasket [7,17]. TABLE 2 Flange Assembly Alternatives and Assembly Efficiencies Method to Stud-to stud load Assembly control bolt Tools required variations from the Efficiency preload mean (percent) (e) 1. No Power impact, lever, torque/stretch Over ±50 0.75 or hammer wrench control Calibrated torque 2. Torque wrench or hydraulic ±30 to ±50 0.85 control wrench 3. Tensioner Multiple stud ±10 to ±30 0.95 load control tensioners 4. Direct Calipers, ultrasonic measurement of extensometer, strain ±10 or less 1.0 stress or strain gages, etc.

Page 334

Consider the bolt load/stretch variations on a large 600 psi steam heat exchanger body flange [7]. Figure 28 shows bolt stretch variations associated with a normal assembly using the star pattern and torque control. With this assembly procedure, this in.diameter unit never maintained a seal for more than 3 months. However, by controlling the bolt stretch through the use of an ultrasonic extensometer, the unit was sealed for several years without a leak. The bolt stretch results are shown in Fig. 29. This indicated the sensitivity of this flange to variations in bolt load. Subsequent classical analyses confirmed the relative flexibility of this flange. At a later date, the tube bundle in the unit had to be replaced. At this time, the head was redesigned and the flange thickness changed from in. to in. The unit was then assembled using the old procedure, without any of the previous seal problems. The type and size of spiral-wound gasket was not changed. On two similar units the flange sensitivity to bolt load was corrected by increasing the flange thickness from in. to in. for a 37-in.-diameter unit and from

Figure 28 Bolt stretch variability after the third and normally final pass.

Page 335

Figure 29 Bolt stretch variability after using a bolt extensometer.

in. to in. for a 50-in.-diameter unit. This flange stiffness sensitivity is reflected in the term assembly efficiency, e, shown in Table 2. Division by e at the appropriate location in the ASME Code flange design process or division of the final flange thickness by e will result in a thicker flange for the bolt preload control methods that produce large variations in bolt preload. The same end result can also be achieved by limiting flange ring rotation, since this will lead to an increase in flange ring thickness, i.e, a stiffer flange. The ASME Code now has a rigidity index, J, to check the flange flexibility or rotation. This check may or may not be sufficient in all cases, since it is based on a fixed rotation, which may not be adequate for some flange designs and/or gaskets.

Which of the methods in Table 2 you select will, again, depend basically on economics. If the flange will be carbon steel or carbon steel with an overlay of a more exotic metal, then method 1 or 2 would most likely be desired, since it is less costly to purchase a thicker flange than to pay the higher cost associated with assembly method 3 or 4. For medium-cost metals, such as the stainless

Page 336

steels, methods 2 or 3 and possibly 4 would be preferred. However, method 4 would be desirable if the size of the flange is such that it is difficult to locate a vendor with adequate forging capability for thick stainless steel flanges. This fact could also come into play in the preceding cases. If the flange will be made of some exotic material, e.g., titanium, hastelloy, or zirconium, then method 4 may be desirable, to minimize capital equipment costs since it will result in a thinner flange. All of the preceding cases assume the flange is to be an integral type.* The selections would be modified for a loose ring or lap joint flange [3]. For this type of flange, methods 1, 2, and possibly 3 would be appropriate. Of course, even if method 1 or 2 is selected, you still may want to use method 4 in the field if the material being contained is highly toxic or flammable. This increases the design safety factor and, thus, further reduces the risk of a leak. At this point, you are ready to do the initial flange design using a modified version of the ASME Code flange design procedure [2,20,25,29]. A tentative new procedure for determining the ASME Code design bolt load is presented in Chapter 9. A more sophisticated flange design method such as the finite element method can be used for flanges made of exotic metals, and/or if the flange operates in a very hostile environment or must contain a highly toxic or flammable materials [2,12,35,37,38,39,40]. 3.10. Flange Rotation Considerations Another problem that is inherent to bolted joints is flange ring rotation. This leads to a variable compression across the gasket from the ID to the OD. In most cases, variable compression across a gasket is associated with flange rotation caused by the bolt load,

i.e., the assembly bolt load plus any other loads reacted through the bolts. However, if you have a tube sheet sandwiched between the flanges, similar to that shown in Appendix H, the amount of variable gasket compression across the gasket can be either increased or decreased by deformation of the tube sheet caused by temperature gradients. A pressure drop across the tube sheet can also contribute to tube sheet deformation (bending). In the case shown in (Fig. 30), the temperature difference across the tube sheet ranged from 34°F to 265°F. This resulted in the tube sheet's bending or rotating in the same direction as the head flange, thus decreasing the variable compression across the gasket [12]. The slopes or angles q1 and q2 of the flange and tube sheet are indicative of the amount of rotation. The difference q1 - q2 represents the variable compression that the gasket actually encounters during steady-state operation. This explained why measurements of the gasket after disassembly of the unit did not reveal a very large angle of rotation, as had been predicted by classical *Flange costs can also be reduced by using a carbon steel flange with a overlay of an exotic metal when feasible.

Page 337

Figure 30 Rotation of gasket contact surface during start-up.

analyses that did not account for thermally induced tube sheet deformations. Only the estimated effect of differential thermal expansion was included. In-the-field experience, plus controlled tests of larger flanges, has shown that rotation itself does not cause a leak. The facts indicate that a flange that encounters considerable rotation due to assembly loads or operational loads is very sensitive to variations in bolt load, since such variations lead to large variations in gasket deflection or stress. The opposite is true for a stiff flange. As mentioned previously, the ASME Code has, in an attempt to correct this problem, added a rigidity constraint, J, to the revised

flange design procedure. This change, plus the use of efficiency factors for various assembly procedures, will help future flange designers avoid this problem. The present rigidity factor for integral flanges is roughly equivalent to a flange ring rotation limit of 0.3 degrees. For loose ring (lap joint flanges), it is about a 0.2degree limit.

Page 338

3.11. Anticipating and Minimizing Start-Up Problems As the process is brought on-stream, the pressure and/or the temperature change. As pressure increases, the bolts tend to stretch, thus reducing the load on the gasket; i.e., for stiffer flanges, the bolt load increases while the gasket load decreases. In the case of flexible flanges such as the ASME/ANSI B165 24 Class 150 flange, the bolt load actually decreases as the pressure increases due the vessel wall dilation. As the temperature changes, several things can happen in addition to the creep and relaxation problems previously discussed. If the flange and bolts are made of the same materials, then as the temperature increases the flange will become hotter than the bolts, and thus, due to differential thermal expansion, the gasket will be further compressed. The amount of additional compression will reach some maximum value during the transient heat-up phase and then decrease somewhat as the bolts reach steady-state temperatures [7,12]. If the bolts and flanges are made of different materials, rather dramatic effects on the gasket compression can be encountered. As indicated previously, in some cases the joint may become tighter, whereas in others it may loosen; i.e., the gasket may be further compressed or may lose compression. The latter situation can lead to a significant increase in leak rate (decrease in tightness, Tp), which may result in the joint's not meeting government (EPA, OSHA, etc.) emission regulations. In the case of joints operating in a cold process environment, decompression is likely to occur. Gasket contraction can be a problem in these situations. As previously discussed, PTFE gaskets contract a large amount as they are cooled, thus tending to be poor performers. However, the

insertion of a corrugated metal core retards this effect. In any event, the change in gasket stress associated with a change in gasket compression can be determined from a gasket stress-deflection graph. Then, by considering the unloading part of the tightness graph shown in Fig. 3, you can determine if the joint tightness, Tp, drops below the minimum value Tpmin [7,12]. Substantial axial, radial, and/or circumferential thermal gradients can make a difficult thermal problem considerably more difficult relative to maintaining a seal. See Appendix H for some typical flange thermal gradients. Also see Figs. 18 and 19. In most cases thermal gradients are most severe during process transients, such as start-up/shutdown cycles and process upsets. However, poor heat transfer can result in a flange's being distorted during steady-state operation as well as during process transients. A severe situation can be encountered with jacketed vessels or pipes. A flanged head on a jacketed vessel can pose a very difficult problem, especially if the head contains nozzles, lifting lugs, etc. A body flange on a jacketed vessel likewise poses a difficult problem. See Appendix H, p. 366. A thick, jacketed flat head can be particularly troublesome in elevated-temperature service. If the jacketed head is noncircular, then an even more difficult

Page 339

problem can be encountered [25,47,48]. In these situations thermal gradients can cause flange rotation, which will lead to a reduction (or an increase) in bolt load. This usually is worst during the first part of a start-up (or shutdown) cycle. As the flange temperature increases, differential thermal expansion will generally cause the bolt load to increase and thus compensate for some or all of the loss in bolt load due to the thermally induced flange rotation. The variation in bolt load measured ultrasonically during a heat-up and cool-down cycle of the flange shown on the second page of Appendix H is shown on page H3. Note that the bolt stress actually went to zero. At this time, the nuts could be turned by hand. They were loose. In any event, the selected gasket must be able to tolerate the radial, axial, and thermal variations in load and displacement without leaking. In many cases the flange and/or vessel will need to be redesigned. For jacketed vessels operating at temperatures above 400°F and pressures greater than 50 psi, a special heat transfer analysis is highly recommended. Such an analysis will help in establishing a start-up (heat-up) and shutdown (cool-down) procedure that will not lead to excessive flange distortion. Every effort should be made to keep thermal gradients axially symmetric. Insulating flanges can also lead to seal problems by increasing joint creep or relaxation, reducing the residual bolt load at steady-state conditions, etc. Beware! See Appendix J. In the previously discussed situations, if the gasket will tolerate the change in gasket stress or deflection without developing a major leak, yet the tightness or leak rate exceeds company or government requirements, then the design tightness value, Tp, should be adjusted upward by an amount that ensure that, upon unloading, the

tightness will not fall below Tpmin. This adjusted or revised value is referred to as TpA. See Fig. 3. If the change in tightness or leak rate after adjusting the gasket load is still such that the gasket loses its ability to maintain a seal, or fails to meet the applicable government regulations, then another gasket will need to be selected and/or the flange redesigned. Once TpA has been established, determine the required gasket seating stress SA from the tightness graph for the particular gasket or use Eq. (4). Now recheck the flange design to determine if it still meets code requirements or other, more stringent constraints. A redesign of the flange may be needed. Note: The new ASME Code Flange Procedure has a built-in factor to deal with unloading effects. This may be adequate for many cases, but not for all of them. 3.12. Assembling the Joint Once the flange design is completed, then every effort must be made to reduce the possibility of a poor assembly. Needless to say, there are numerous factors that affect assembly. Many of these are not recognized until assembly problems develop after installation. Strategically locating flanges was mentioned at the beginning of the discussion and needs to be reiterated at this time. In pipe lines,

Page 340

flanges should be located insofar as possible at locations where they can be easily assembled, e.g., accessible from a platform, not from a ladder, etc. External loads acting on pipe flanges and vessel nozzle flanges due to thermal expansion, gravity loads, misalignment, cold springing, etc., can also cause seal problems. A thorough pipe stress analysis is highly recommended. On vessels, especially heat exchangers, you should have easy access to the bolts if you want to achieve the designed tightness; i.e., equipment layout is important. The clearance between the body flange and a nozzle should not obstruct access to the bolts. The same is true for lifting lugs. I have encountered numerous situations where a torque wrench could be used on 7585% of the bolts, while a hammer wrench was needed on the remaining bolts. Through ultrasonic bolt stretch measurements it has been shown that this leads to extremely large variations in bolt load (over ±50%) as indicated by method 1 in Table 2. This, of course, greatly increases the probability of developing leaks. This has been borne out in the field numerous times. Some typical clearance recommendations for proper equipment layout are shown in Appendix D. It is important to realize that the arduous gasket selection process and the meticulous flange design work will have been done in vain if you cannot properly assemble the joint. Overtightening bolts is also a source of many gasket failures. Each type of gasket has an optimum gasket seating stress range. To exceed this range can be just as detrimental as not having sufficient gasket loading. The revised ASME Code Flange Design Procedure will have gasket factor tables that give the stress limits for the various gaskets. Another problem that comes into play with time is bolt deterioration. A good bolt inspection procedure and a good

acceptance/rejection criterion is a necessity. To supplement this, a good bolt specification sheet, as shown in Appendix E, is desirable. This will ensure receipt of the proper bolts for the assembly [22,26,44,29]. The next step in the flowchart is to specify a lubricant, emphasizing that it should be placed on the bolt threads and under the head or nuts [7,22]. Placing lubricant under the head is just as important as placing it on the threads. For larger-diameter bolts, say, .gif in. or larger, one should also specify hardened washers in an effort to further reduce friction. AISI 4140 steel hardened to a Rockwell C hardness range of 40 to 45 is about an optimum washer material. See Appendix F [7,22,26,36]. Using the revised gasket seating stress, SA, determine the in-thefield bolt load, stretch, and torque that is needed to achieve the desired tightness value. Due to joint elastic interaction, the in-thefield torque value needed to achieve the desired average bolt load is generally 1030% greater than the calculated single-bolt torque value. (Note: This increase applies only when using torque control.) The percentage increase is related to the joint stiffness. Note that the calculated bolt load is not associated with the ASME Code design bolt load. The minimum assembly bolt load, Fb, is based on the full gasket width. The applicable

Page 341

equation where SA has been determined from a finite element analysis or some other analytical procedure is [7,912,26,44,49]. (5)

ODg outside diameter of gasket, raised face, or lap = ring, mm (in.) IDg inside diameter of gasket, mm (in.) = SA =

adjusted or final gasket seating stress based on a fine element analysis, MPa (psi)

Nb number of bolts = Note: If the revised seating stress, SA, is only estimated, then Eq. (5) would need to be modified to include the hydrostatic end-force effect. In this case, we will call the modified or estimated seating stress SAE. The revised equation is: (6) where SAE =

estimated new gasket seating stress, MPa (psi) Pinternal pressure, MPa (psi)

A quick check on the calculated bolt load is to use the old rule of thumb that the assembly bolt load should always be greater than 1.5 times pressure-induced bolt load. If the calculated value of Fb is less than this value, then the value of Fb should be set equal to 1.5 times the pressure-induced bolt load. Once the bolt load is known, then the assembly torque for the specific bolt and lubricant can be determined. Usually tables or graphs of bolt torque versus bolt load for different lubricants (coefficients of friction) are available to assist in rapidly estimating the assembly bolt torque. A typical example is presented in Appendix G. A simple equation using a nut factor is another option. See pp. 226-233 of Ref. 59 or Chapter 11. To determine the required bolt stretch, DL, for use in conjunction with an ultrasonic extensometer, use the following relationship. (7) where bolt load, N (lbf) Fb = effective length of Bolt, i.e., from middle of nut to Le = middle of nut or middle of bolt head, mm (in.)

Page 342

bolt diameter, mm (in.) [Use either the minor db diameter or one-half the sum of the minor and = root diameters.] Eb modules of elasticity of the bolt material, MPa = (psi) Once all of this information has been determined, it should be placed on the applicable drawings. A typical format for specifying such information is shown in Appendix B. 4. Closing Comments The gasket selection flowchart gives all of the major parameters one should routinely consider in selecting a gasket for a given environment. However, the designer should always be aware of other unique or unusual situations or constraints that can further affect the gasket selection process. Many specialized industries will have additional unique requirements. Also see Appendix L. Nomenclature Ab

cross-sectional area of the bolt, mm2 (in.2)

Ag

gasket seating area, mm2 (in.2)

Gb

loading intercept on a tightness curve, MPa (psi) [Formerly called B.]

C

Constant associated with the Tightness Class

a

slope of loading curve in a tightness graph, MPa (psi)

[Formerly called d.] db

bolt diameter, mm (in.)

Eb

modulus of elasticity of the bolt material, MPa (psi).

Fb

bolt load, N (lbf)

IDg

gasket inside diameter, mm (in.)

DL

bolt stretch or elongation, mm (in.)

Le

effective length of the bolt, mm (in.)

Lrm

mass leak rate per unit diameter, mg/sec-mm (lbm/hrin.)

.gif Nb

reference mass leak rate per unit diameter [ mg/sec], mg/sec-mm (lbm/hr-in.)

.gif = 1

number of bolts in the flange

ODg gasket outside diameter, mm (in.) P

internal pressure, MPa (psi)

P*

atmospheric pressure, MPa (psi)

Pr Gs

unloading intercept on a tightness curve, MPa (psi) [Formerly called S*.]

SA

adjusted assembly gasket seating stress, MPa (psi)

SAmin Minimum Assembly Gasket Seating Stress, MPa (psi) SGmin SGo = minimum operating gasket stress, MPa (psi)

Page 343

Tp

tightness parameter of factor (nondimensional)

TpA Tpn = adjusted assembly tightness Tpminminimum tightness Tpo operating tightness value Ti

tightness class, i.e., T1, T2, T3, etc.

Acknowledgments I would like to acknowledge the helpful suggestions of the members of the PVRC Bolted Flange Committee and its subcommittees. The excellent effort put forth in preparing the flowchart and graphs by the Eastman Chemical Company Art Department, as well as the typing by the Information Processing Center, is also acknowledged. References 1. Raut, H. D., and Leon, G. F. Report of gasket factor tests. Welding Research Council Bulletin 233 (December 1977). 2. Welding Research Council Bulletin 271: (1) Methods of analysis of bolted flanged connectionsA review, by A. E. Blach and A. Bazergui; (2) Gasket leakage behavior trends, by H. D. Raut, A. Bazergui, and L. Marchand (October 1981). 3. Winter, J. R., and Leon, G. F., Radially inward buckling of spiral-wound gaskets. Presented at the 1985 ASME PVP Conference, New Orleans, LA.

4. Bazergui, A., and Marchand, L. PVRC Milestone gasket testsFirst results. Welding Research Control Bulletin 292 (February 1984). 5. Welding Research Council Bulletin 294: (2) Short-term creep and relaxation behavior of gaskets, by Bazergui, A. (1) Creep of bolted flanged connections, by Krause, H., and Rosenkrans, W. (May 1984). 6. Bazergui, A., Marchand, L., and Payne, J. R. Effect of fluid on sealing behavior of gaskets. Proceedings of 10th International Conference on Fluid Sealing, BHRA, Innsbruck, Austria, 1984, Paper H2, pp. 365385. 7. Winter, J. R. Use of an ultrasonic extensometer to determine the variations in the assembly bolt loads of a problem industrial flange. Presented at the 1987 Society of Experimental Mechanics Spring Conference, Houston, Texas, June 1419, 1987. 8. Bazergui, A., and Payne, J. R. On the elevated temperature behavior of gaskets. Presented at the 6th International Conference on Pressure Vessel Technology, Beijing, China, 1988. 9. Bickford, John H. That initial preloadWhat happens to it? Mechanical Engineering (October 1983): 54ff. 10. Bickford, John H. New twists in bolting. Mechanical Engineering (May 1988): 31ff. 11. Bickford, John H. Ultrasonic control of bolt preload. Preprint 3481, 1981 ProceedingsRefining Department, 46th Midyear Meeting, American Petroleum Institute, Chicago, May 1114, 1981.

Page 344

12. Hayashi, K., Chang, A., and Winter, J. R. A preliminary evaluation of the elevated temperature behavior of a bolted flanged connection. Welding Research Council Bulletin 341 (February 1989). 13. Marchand, L., Derenne, M., and Bazergui, A. Elevated temperature gasket behavior trends. Presented at the CETIM Symposium on Fluid Sealing, Nantes, France, June 1986. 14. Webjorn, Jan. The bolted joint: A series of problems. Ph.D. dissertation no. 130, Institute of Technology, Linkoping, Sweden, 1985. 15. Birembaut, Y., and Bravo F. Influence of high temperature on mechanical and sealing properties of flat gaskets. Proceedings of 11th International Conference on Fluid Sealing, BHRA, pp. 684697, 1987. 16. Payne, J. R., Mueller, R. T., and Bazergui, A. A gasket qualification test scheme for petrochemical plants, part I: Test methods and application results; Part II: quality criteria and evaluation scheme, Proc. ASME/JSME Pressure Vessel & Piping Conf., Hawaii, July 1989. 17. Payne, J. R., Bazergui, A., and Leon, G. F. Getting new gasket design constants from gasket tightness data. Experimental Techniques, Society for Experimental Mechanics 12. 22s27s (1988). 18. Bazergui, A., Marchand, L., and Raut, H. D., Further gasket leakage behavior trends. Welding Research Council Bulletin 325 (July 1987): 110.

19. Winter, J. R., and Bazergui, A. Room temperature and elevated temperature tests of a metal corrugated gasket with flexible graphite fill. Presented at the 1989 ASME/JSME PVP Conference in Honolulu, Hawaii, July 2427, 1989. 20. Blach, A. E., Bazergui, A., and Baldur, R. Bolted flanged connections with full face gaskets, Welding Research Council Bulletin 314 (May 1986). 21. Bazergui, A., and Marchand, L. Development of tightness test procedures for gaskets in elevated temperature service. Welding Research Council Bulletin 339 (December 1988). 22. Bazergui, A., and Clement, B. Classification of variables which affect bolt joint assembly results and/or behavior. Draft presented at the April 21, 1989, meeting of the Bolting Technology Council, United Engineering Center, New York. 23. Payne, J. R., Bazergui, A., and Leon, G. F. New gasket factorsA proposed procedure. Proc. of the 1985 PVP Conference, ASME, PVP, pp. 8593. 24. Bazergui, A., and Louis, G. Predicting leakage for various gases in gasketed joints. 1987 Spring Conference, Society of Experimental Mechanics, June 1519, 1987. 25. Blach, A. E. Bolted flanged connections for non-circular pressure vessels. ASME/JSME PVP Conference, Honolulu, Hawaii, July 2427, 1989. 26. Bickford, J. H. The bolting technology council and the search for more accurate preload. ASME/JSME PVP Conference, Honolulu, Hawaii, July 2427, 1989. 27. Payne, J. R. PVRC flanged joint user survey. Welding Research Council Bulletin 306 (July 1985).

28. Crago, W. A., and Stevens-Guille, P. D. Application of spiral wound gaskets for leak-tight joints. ASME Winter Annual Meeting, New York, November 1722, 1974. 29. American Society of Mech. Engrs. ASME Boiler and Pressure Vessel Code, Sec. VIII Div. 1.

Page 345

30. Protocols for Generating Unit-Specific Emission Estimates for Equipment Leaks of VOC and VHAC. Doc. No. EPA-450/3-88010, October 1988, Table 2-1. 31. MSS Standard High Pressure Chemical Industry Flanges and Threaded Stubs for Use with Lens Gaskets, MSS SP-65, 1988. 32. ANSI Standard Pipe Flanges and Flanged Fittings, ANSI B16.5, 1981, P.5, ¶ 6.3.4. 33. API Specification for Wellhead and Christmas Tree Equipment, API Spec 6A, 15th ed., April 1, 1986. 34. USA Standard Unfired Pressure Vessel Flange Dimensions, USAS B16.30-1969, Section 2, ASME, United Engineering Center. 35. Murray, N. W., and Stuart, D. G. Behavior of large taper hub flanges. Proceedings of the Symposium on Pressure Vessel Research Towards Better Design, Jan. 18, 19, 1961, London, England, 1962. 36. Bickford, John H. An Introduction to the Design and Behavior of Bolted Joints. Dekker, New York, 1981. 37. Harvey, J. F. Theory and Design of Modern Pressure Vessels. 2nd ed. Van Nostrand Reinhold, New York, 1974. 38. Timoshenko, S. Strength of Materials. Part II. Van Nostrand, New York, 1956. 39. Deen, P. J., van Campen, D. H., and Latzko, D. G. H. Deformation of large-diameter, high-pressure flanges, PhD dissertation, University of Technology, Delft, The Netherlands.

40. Menkem, C. M. Influence of bolt loading on deformation of pressure vessel flanges. PhD dissertation, Eindhoven University of Technology, The Netherlands. 41. Marchand, L., Derenne, M., and Bazergui, A. Weight loss correlation for sheet gasket materials. 1990 ASME PVP Conference in Nashville, Tennessee, June 1721, 1990. 42. Marchand, L., Bazergui, A., and Derenne, M. Influence of thermal degradation on sealing performance of compressed sheet gasket materials with elastomer binder. Part I: Experimental methods. 2nd International Symposium on Fluid Sealing of Static Gasketed Joints, La Baule, France, September 1820, 1990. 43. Marchand, L., Bazergui, A., and Derenne, M. Influence of thermal degradation on sealing performance of compressed sheet gasket materials with elastomer binder. Part II: Analysis. 2nd International Symposium on Fluid Sealing of Static Gasketed Joints, La Baule, France, September 1820, 1990. 44. Bickford, J. H. Bolt torqueGetting it right. Machine Design (June 21, 1990). 45. Payne, J. R. Effect of flange surface finish on spiral wound gasket constants. 2nd International Symposium on Fluid Sealing of Static Gasketed Joints, La Baule, France, September 1820, 1990. 46. Brown, S. J., and Brown, T. J. A computer simulation of a bolted flange with a spiral wound gasket in a TEE shell. 1990 ASME PVP Conference in Nashville, Tennessee, June 1721, 1990. 47. Blach, A. E., and Lihua X. Rectangular bolted flanged connections: Finite element analysis and test results. 1990 ASME PVP Conference in Nashville, Tennessee, June 1721, 1990. 48. Blach, A. E., Bolted flange connections for rectangular

pressure vessels: Exploratory investigation. 2nd International Symposium on Fluid Sealing of Static Gasketed Joints, La Baule, France, September 1820, 1990.

Page 346

49. Bickford, J. H. Improving bolted joint performance. Plant Engineering (February 8, 1990). 50. Czernik, D. E. Selecting and testing industrial gaskets for bolted joints. V. P. Technologies and Technical Planning (Fel-Pro Inc., Skokie, Illinois). 51. Payne, J. R., and Bazergui, A. Evaluation of test methods for asbestos replacement gasket materials. MTI Publ. No. 36, Materials Technology Institute of the Chemical Process Industries, St. Louis, 1990. 52. Winter, J. R., and Coppari, L. A. Flange thermal parameter study and gasket selection. Presented at the 1996 ICPVT/ASME PVP Conference in Montreal, July 2128, 1996. 53. Jones, W. F., and Seth, B. B. Evaluation of asbestos-free gasket materials. Presented at the ASME/EEEE Power Generation Conference, October 1990. 54. Winter, J. R. Gasket selection flowchart. ASME PVP-185, 1990 ASME PVD Conference, June 1721, 1990, Nashville, Tenn. 55. Winter, J. R. Gasket selection: A flowchart approach. Presented at the 2nd International Symposium on Fluid Sealing of Static Gasketed Joints, LaBaule, France, September 1820, 1990. 56. Derenne, M., Batista, P. P., and Muzzo, U. Effect of oxidation inhibitor, fluid type, fluid pressure, and gasket stress on the elevated temperature behavior of metal reinforced flexible graphite sheet materials. Draft PVRC Report to PVRC Subcommittee on Gasket Testing and Elevated Temperature Joint Behavior, June, 1995.

57. Derenne, M., and Marchand, L. Long duration air and steam screening of elastimeric sheet gasket materials. Final Draft Report to PVRC Committee on Gasket Testing and Elevated Temperature Joint Behavior, October, 1995. 58. Derenne, M., Payne, J. R., Marchand, L., and Muzzo, U. Elevated temperature characterization of flexible graphite sheet materials for bolted flanged joints. Final Report, June 1995. 59. Bickford, John H. An Introduction to the Design and Behavior of Bolted Joints. 3rd ed. Dekker, New York. 60. Keywood, S. Testing and evaluation of PTFE-based gaskets for chemical plant service. 1992 ASME Pressure Vessels and Piping Conference, New Orleans, LA, June 2125, 1992. 61. Winter, J. R., and Keywood, S. Investigation of extrusion-type gasket failures of PTFE-based gaskets in pipe-line flanges. Presented at the 1996 ICPVT/ASME PVP Conference in Montreal, July 2128, 1996. 62. Derenne, M., Marchand, L., Payne, J., and Bazergui, A. Elevated temperature testing of gaskets for bolted flanged connections. WRC Bulletin 391, New York, (May, 1994). 63. Marchand, L., Derenne, M., and Bazergui, A. The influence of thermal degradation on sealing performance of compressed sheet gasket materials with elastomer binder. Part I: Experimental methods. 2nd International Symposium on Fluid Sealing of Static Joints, CETIM, La Baule, France, Sept., 1990, pp. 225266. 64. Marchand, L., Derenne, M., and Bazergui, A. Weight loss correlation for sheet gasket materials. ASME Journal of Pressure Vessel Technology 114 (February 1992): 17. 65. Martens D., and Porter, M. A. Investigation and repair of heat

exchanger flange leak. 1994 ASME Pressure Vessels and Piping Conference Minneapolis, Minn. June 1923, 1994.

Page 347

66. Sawa, T., Hirose, T., and Kumano, H. The behavior of pipe flange connection in transient temperature field. 1992 ASME Pressure Vessels and Piping Conference in New Orleans, June 2125, 1992. 67. Payne, J. R., Bazergui, A. Evaluation of test methods for asbestos replacement gasket materials. MTI Publ. No. 36, Materials Technology Institute of the Chemical Process Industries, 1990. 68. Derenne, M., Marchand, L., Payne, J., and Bazergui, A. Elevated temperature testing of gaskets for bolted flanged connections. Welding Research Council Bulletin (May 1994). 69. Bazergui, A., and Marchand, L. Development of tightness test procedures for gaskets in elevated temperature service. Welding Research Council Bulletin (December 1988): 120. 70. Payne, J. R., Derenne, M., and Bazergui, A. Estimating elevated temperature gasket performance with ATRS tests. 1989 SAE International Congress and Exhibition, Detroit, February, 1989. 71. Bazergui, A., Marchand, L., and Payne, J. R. The aged hot tightness test for gaskets. 1989 SAE International Congress and Exhibition, Detroit, February 1989. 72. Payne, J. R., Mueller, R. T., and Bazergui, A. A gasket qualification test scheme for petrochemical plants. Part I: Test methods and application results. PVP-ASME/JSME, Honolulu, Vol. 158, pp. 5368, July 1989. 73. Payne, J. R., Mueller, R. T., and Bazergui, A. A gasket qualification test scheme for petrochemical plants. Part II: Quality

criteria and evaluation scheme. PVP-ASME/JSME, Honolulu, Vol. 158, pp. 6979, July 1989. 74. Derenne, M., Payne, J. R., Marchand, L., and Bazergui, A. Development of test procedures for fire resistance qualification of gaskets. Welding Research Council Bulletin, (December 1992). 75. Marchand, L., Bazergui, A., and Kockelmann, R. H. The influence of the stiffness of flanges and bolts on the creeprelaxation behavior of gaskets. 3rd International Symposium on Fluid Sealing of Gasketed Joints, Biarritz, September 1993. 76. Marchand, L., and Derenne, M. Fugitive emission characteristics of gaskets. PVRC Project No. 9295, June 1995. 79. Bibel, G., and Ezell, R. Bolted flange assembly: Preliminary elastic interaction data and improved bolt-up procedures. Welding Research Council Bulletin 408 (January 1996). 80. Payne, J. R., and Schneider, R. W. Comparison of proposed new ASME rules and gasket constants for bolted flanged joints. 81. Payne, J. R., and Winter, J. R. Fugitive emission estimates from gasket tightness test data. ASME PVP Conference Presentation, New Orleans, 1992. 82. U.S. Dept. of Health and Human Services. NIOSH Pocket Guide to Chemical Hazards. DHHS (NIOSH) Publ. No. 90117. 83. Meuller, R. T. Recent buckling experiences with spiral wound flexible graphite filled gaskets. Presented at the 1996 ICPVT/ASME PVP Conference in Montreal, July 2128, 1996.

Page 348

Gasket Selection Flowchart

Page 349

Page 350

Page 351

Page 352

Page 353

Page 354

Page 355

Appendix A: Typical Gasket Surface Finishes Note: These surface finishes are examples and may not be applicable to your particular flange, gasket material, process conditions, or contained fluid. Most of the new gaskets work well in the 125250 AARH range. This range appears slowly to be becoming the industry standard finish. Type of Gasket

Surface finish (AARH) 125 Serrated* (500+) 200 200

Surface finish range (AARH) 100125 5001000 125250 125250

Spiral wound Teflon Filled Teflons Expanded Teflons Compressed sheet Serrated* (500+) 5001000 (old) Compressed sheet 250 125250 (new) Flexible graphite 250 125250 O-ring (metal) 63 3263 Oval ring/delta ring 63 0.06 in. Note: AARH = average arithmetic roughness height; units are microinches.

A typical flange gasket surface finish specification is shown next. Typical Flange Gasket Surface Finish Specification TECHNICAL NOTE

Gasket Seating Surface Finish Based on PVRC gasket tests, vendor recommendations for the newer gasket materials and/or fills, and industrial experience, one gasket surface finish range is recommended for all pipe and equipment flanges for the commonly used gaskets1. The Gasket Seating Surface Finish Range1 for all Pipe and Equipment Shall Be a: 125 to 250 Ra2 Serrated Finish with no Radial Tool Marks Notes: 1. Specialty equipment gaskets for severe environments (P > 600 psi, T > 700°F) may require special finishes. Some typical examples are solid metal rings, metal

Page 356

o-rings, API oval/octagonal rings, delta rings, double jacketed, etc. Flanges made of softer metals such as copper may also require special finishes. 2. Ra means Roughness average and is expressed in microinches (min). See MSS SP-6 and ANSI/ASME B46.1. Terminology used in previous years was AARH (arithmetic average roughness height), AA (arithmetic average), and RMS (root mean square roughness). These older terminologies may be found in many older textbooks and standards and on many drawings.

Page 357

Appendix B: Assembly Information for Body Flanges and Large Nonstandard Nozzle Flanges

Page 358

Page 359

Appendix C: Typical Examples of Buckled Spiral-Wound Gaskets Figure C1, C2, and C3 show a 72-in.-diameter asbestos-filled 316 stainless steel spiral-wound gasket with a carbon steel external gage ring but no internal gage ring. The radially inward buckling was due to differential thermal expansion [3].

Figure C1

Figure C2

Page 360

Figure C3

Figures C4 through C9 show Teflon-filled Hastelloy spiral-wound gaskets with a carbon steel external gage ring but no internal gage ring for a standard ANSI/ASME B16.5 24-in. flange. The gage ring had separated from the gasket, as shown in Fig. C4. The radially inward buckling was due to the assembly load only. The new gaskets were removed before process start-up. The incompressible nature of Teflon (PTFE) is the major culprit. A properly sized internal gage ring corrected the problem. Similar buckling is encountered with flexible graphite-filled spiral-wound gaskets [3,83].

Figure C4

Page 361

Figure C5

Figure C6

Figure C7

Page 362

Figure C8

Figure C9

Page 363

Figures C10 and C11 show a 40-in.-diameter asbestos-filled 304 stainless steel spiral-wound gasket with a carbon steel external gage ring but no internal gage ring. The radially inward buckling was attributed primarily to differential expansion, but it was believed that the extremely flexible flange contributed to the problem [3].

Figure C10

Page 364

Figure C11

Page 365

Figure C12 shows a 2-in.-diameter class 2500 spiral-wound gasket with 316 stainless steel windings, asbestos fill with a flexiblegraphite tape overlay, carbon steel external gage ring, and a small 316 stainless steel internal gage ring. This type of radially inward buckling can generally be attributed to one or both of the following reasons: (1) thermal shock or a large thermal gradient between the gasket ID and the outer part of the gasket and gage ring, and/or (2) large assembly loads (more prevalent with flexible-graphite-filled or Teflon-filled spiral-wound gaskets) [3,83].

Figure C12

Page 366

Appendix D: Lifting-Lug and Nozzle-to-Flange Clearance Recommendations

Figure D1

Page 367

Appendix E: Bolt-Stud/Nut Specification Sheet

Page 368

Bolt Specification Explanation Type of Machine Bolt, Stud, Cap Screw, etc. Bolt Carbon Steel, 304 St. St., Sa-193 B7, 410 St. St., Material SA-307 Gr. B, Hastelloy C-276, SA-325, SA-193 B8M, etc. For flange bolting, 7 or 8 threads per inch is generally used. For smaller flange bolts, 12 threads Threads per inch is often used. Bolts used in machine design, per Inch especially for precision equipment, generally have a greater number of threads per inch. Thread Series (Form 1)

UNC (unified national coarse) is usually specified for pipe and vessel flange bolts. UNF (unified national fine) is usually used in machine design functions. Note: Threads may be rolled or machined. Rolled threads are preferred in situations involving high loads, cyclic loads, and/or brittle materials.

2A is used for all pipe and vessel flange bolting as well as structural bolting. It is also used in most Class of machine design applications. However, 3A should Fit

be specified for precision equipment. The equivalent designations for nuts are 2B and 3B respectively. Hex Head, heavy Hex Head, Socket Head, Square Head, Hex Socket, etc. Heavy Hex Heads and Nuts Type of are generally used with pipe and vessel flanges. In Head addition you can specify semifinished, unfinished, or finished in conjunction with the previous items. Semifinished is preferred for flange bolts.

Page 369

Length

Should always include thickness of all clamped parts, including the gasket and washers plus 1.5 times the bolt diameter. Same applies for studs except that the 1.5 should be changed to 3.0.

Special This might involve some ASTM, ANSI, or Specification(s) ASME Code specifications. Some typical special requirements are: Lefthanded threads, chrome-plated, nickelSpecial plated, cad-plated, Teflon-coated, luberized, Requirements rolled threads, machined threads, heat-treat after machining, etc. Example

Standard Bolt and Corresponding Nut Designations: Hev. Hex Head × 21 long. Hev. Hex. Nut Materials to be SA-193 B7; Rolled Threads are required.

Page 370

Appendix F: Hardened Washer Specification Revision No.: _______ Date: _______ Item No. _______ Due to galling/excessive friction, hardened washers should be used on piping and vessel flanges that contain bolts with a diameter greater than 1. This becomes more important as the bolt diameter increases and is especially critical when relatively high bolt loads (stresses) are required to seat the gasket. Typical gaskets that require large assembly bolt loads are spiral-wound, doublejacketed, oval ring, or other metallic gaskets. In most cases heavy hex bolts and nuts should be used in conjunction with hardened washers.

Material: AISI 4140 steel hardened to a Rockwell C hardness of 40 to 45. Note: Washers should be polished on one side. Alternate Specification:

In some situations, depending on the bolt spacing, ASTM specification F 436-86 (Standard Specification for Hardened Washers) can be used. This ASTM specification shall apply only if the flange bolt spacing is such that these hardened washers will not overlap. If the washers will overlap, the preceding specification shall apply.

Page 371

Appendix G: Typical Bolt Load-Torque Table and Graph

Figure G1

Page 372

Figure G2

Page 373

Appendix H: Typical Flange Thermal Profiles Note: These thermal profiles are based on in-the-field temperature measurements.

Figure H1 Circumferential flange thermal profile for a two-pass 600 psi super-heated steam reboiler.

Figure H2

Axial flange thermal profile for a 600 psi steam reboiler.

Page 374 Temperature Variations Vs. Time Across the Face of the Shell Flange of a Jacketed Vessel Temperature at locations X, Y, Z(F) Time X Y Z Z' DTZ-X DTZ'-X 0 100 108 140 40 __ __ 13 117 123 156 39 __ __ 21 126 135 175 49 204 78 39 145 157 202 57 250 105 51 160 173 220 60 275 115 113 245 255 292 46 310 64

Figure H3

Flange temperature versus time during start-up of a jacketed conical mixer.

Page 375

Figure H4 Bolt Stretch vs. Time for Bolted Joint Shown in Figure H3.

Page 376

Appendix I: Inherent Inaccuracies of the Turn-of-the-Nut Method of Bolt Load Control for Gasketed Bolted Flanged Joints Theoretically if you control the amount a nut turns as it clamps two pieces of metal together, you should get a certain amount of bolt stretch (load); e.g., for an 8 TPI bolt, one full turn should give you an elongation of in. This would be true if the bolt and nut threads did not distort and the two pieces of metal were perfectly rigid (infinitely stiff relative to the bolts). This idealized behavior is shown by the solid line in Fig. I1. In real-life situations, however, the bolt threads do deform elastically, with perhaps some plastic deformation. Thus, just from consideration of the bolt and nut you can see that the end result will be less than the theoretically predicted value. This is shown by the dashed line in Fig. 11. If you now consider the fact that a flange rotates as it is loaded,* then you get an additional decrease in the theoretically predicted value, as indicated by the dotted line in Fig. I1. If the joint contains a gasket, then a substantial reduction in actual bolt load will occur due to the nonlinear behavior of most gaskets.** This is shown in Fig. I2. There are other facts that lead to further discrepancies, but they are generally of a minor nature compared to those previously mentioned. This is another case where the methods to predict bolt load resulted in loads considerably less than persons thought were present. For instance, if you used the turn-of-the-nut method to reach a specified load on a gasket, your actual load was probably 3050% below the value you thought you had, without even considering the effects of elastic interactions in the joint. This may explain why such high assembly bolt loads were specified in the past; i.e., by specifying a

very high assembly bolt load you actually reached the value that was truly needed. Thus we must be careful in selecting an assembly bolt load for use in the field, since the end result will ultimately depend on the type of assembly tools used and the method for controlling the bolt load. *This assumes that the flanges do not have full face contact, i.e., flanges with full-faced gaskets. **This will not be true for solid metal ring gaskets as long as they are not loaded beyond the elastic limit of the material.

Page 377

Figure I1

Figure I2

Page 378

Appendix J: Insulation of Flanged Closures Insulation of flanged closures has become desirable due to the increasing cost of energy. However, wholesale insulation of all closures without regard to temperature, pressure, materials of construction, process requirements, or stress levels can lead to serious problems. The situations that usually must be approached with caution involve high temperatures and moderate to high pressures. In these situations bolt relaxation (creep) can lead to a decrease in gasket compression, resulting in process leaks. As discussed at the end of this appendix, thermally induced gasket deterioration can also limit the allowable flange insulation temperature. For the most commonly used high-tensile bolt material (SA-193 Gr. B-7) one can completely insulate flanges, provided the temperatures do not exceed roughly 550°F. At temperatures above this value one can still insulate portions of a flange, but one should attempt to keep the bolts open in an effort to keep them as cool as possible. In these cases, easily removable insulation assemblies should be used. See the accompanying figures. The allowable bolt temperature is also affected by the pressure. In general, low pressure and vacuum service can be insulated when temperatures exceed the previously stated values simply because the operating bolt stress required to maintain a seal is lower; thus you do not encounter as much relaxation. In some cases, such as vacuum distillation, it is often a necessity to insulate body flanges to prevent the development of internal condensation that can seriously impair the operation of the column.

For temperatures above 550°F, the following guide might be helpful: As the flange diameter, pressure, and/or temperature increase, the need to keep the bolts cool becomes increasingly important. In these situations, the units should not be insulated until an analysis has been performed. Their are many techniques (material and/or design modifications) that can be employed to increase the acceptable temperature level at which a given bolt assembly can be totally or partially insulated.

Figure J1

Page 379

Due to the mandated low-fugitive-emission requirements, thermally induced gasket deterioration has become a major concern, since it leads to increased leak rates. As a result, the temperature at which one would consider using insulation may in the future be controlled by the allowable gasket temperature rather than the temperature at which creep and/or stress relaxation becomes a problem with the other flange components. In general, the lower the gasket temperature, the slower the rate of gasket deterioration and thus the longer a given gasket can remain in service, i.e., before the gasket has to be replaced to meet the emissions regulation. In those cases where flanges need to be insulated, it would be wise to use removable insulation that will allow periodic inspection of the flanged joint. Thermal deterioration of gaskets as verified by PVRC elevated-temperature gasket tests and substantiated by in-the-field experience indicate that periodic flange/gasket inspection/replacement programs will be needed in the future. Periodic gasket replacement is becoming a routine maintenance requirement. A recommended configuration for insulating near flanges is shown. In these cases a shroud may be placed around the hot flange to protect workers. There should be at least a 1-inch gap between the flange and the shroud.

Figure J2

Page 380

Appendix K: Typical Gasket Types Flat--Non-Metallic Flat--Metallic Metal jacketed gaskets

Hollow metal Spiral wound

Solid metal

Page 381

Appendix L: Gasket Selection Impact (Ishikawa/FishBone) DiagramOther Considerations

Page 382

Appendix M: Additional TTRL Test Results on the Degradation of Flexible Graphite Sheet Gaskets

Page 383

Appendix N: Suggested Long-Term Temperature Limits for Gaskets Notes: (1) As the gasket temperature approaches these limits, the pressure and/or diameter at which the gasket can be safely used decreases. (2) All data is for long-term use at the listed temperature (>12 months). (3) This information is based on the latest elevated-temperature gasket test data as well as in-the-field experiences. The data listed in this table may change as more test data becomes available or if results from the field so indicate. (4) In certain situations a heat transfer analysis may be needed to verify the gasket temperature. (5) These are gasket temperature, not process temperatures. See Note 2.

Gasket type/ material Asbestos sheet

Aramid-fiber

Maximum Environmentrecommended usage oxidizing/ temperature reducing (°F)5 Comments 1200°F (P < May vary somewhat with Both 30 psi) binder. 1000°F (P < May vary somewhat with 100 psi) binder. 900°F (P > May vary somewhat with 100 psi) binder. Binder and aramid deteriorate with time.

sheet (nonasbestos)

Both

Graphite-fiber sheet Both (nonasbestos) Carbon-fiber sheet Both (nonasbestos) Flexible-graphite sheet Oxidizing (uninhibited) Reducing

200°F*

May be somewhat conservative.

200°F

Binder deteriorates.

200°F

Binder deteriorates.

500°F at ID

May be somewhat conservative.

May be somewhat conservative. This may 500°F at OD need to be the ID Limit too.

(table continued on next page)

Page 384

(table continued from previous page) Maximum Environmentrecommended usage Gasket type/ oxidizing/ temperature material reducing (°F)5 Comments Flexible-graphite May be somewhat sheet (oxidation Oxidizing 600°F at ID conservative. inhibited) May be somewhat Reducing 600°F at OD conservative. This may need to be the ID limit too. Corrugated May be somewhat metal/flex. Oxidizing 500°F at ID conservative. This may graphite need to be the ID limit too. May be somewhat Reducing 500°F at OD conservative. This may need to be the ID limit too. Severe creep at Teflon sheet temperatures above 100°F. Both 100°F (virgin PTFE) May be somewhat conservative. Have had some problems Filled Teflon Both 350°F at 350370°F. May be somewhat conservative. A softer material that creeps much more than Filled Teflon Both 1.5(Tpmin). Accordingly, if in the iterative process Sm1 = Sm2 but 1.5(Sa) (Am)/Ag is less than the value of Sya corresponding to Tpa = 1.5(Tpmin), the determination of Wmo and Am is completed using Tpa = 1.5(Tpmin). And in this case, as for consideration (f), Smo = greater of Sm1, Sm2, 2P, and SL Am is now Am required; Am in (c) and (d) may be looked upon as a trial value. In order for a joint to develop the desired tightness, each bolt must be tightened at assembly to a minimum total load of Sya(Ag + Ap). So far, the ideal situation in the joint was assumed; that is: 1. The actual total bolt load at assembly equals the theoretical or required value. 2. There is no variation in the load from bolt to bolt. 3. A uniform gasket stress exists around the circumstance of the gasket.

Page 439

4.6. Gasket Constants and Data Table 2 provides a summary of typical PVRC gasket constants. It also gives the constraints that apply on gasket stress and tightness: SL, Sc; and Tpmax. Table 3 presents the gasket ID codes referred to in Table 2. A key aspect of Appendix BFC of the ASME B&PV Code is to require that the gasket constants Gb, a, and Gs be certified by a test capable of producing the constants, such as the PVRC ROTT procedure. (See chapter 5 for a discussion of this procedure) Constants taken from a table to be provided by the new appendix will be considered acceptable. Other constants may be used provided the source test document certifying the constants for the specified gasket material and type is cited. Effectively this means that gasket constants provided by gasket producers may be used. Certain notes and restrictions apply to Table 2: The stress Smo shall be greater than the value SL unless a lower stress is verified by the ROTT test. Tpmax is the lesser of Tps (tightness hardening limit) and Tpu (Max value associated with 110% of the maximum test stress) unless higher values are verified by the ROTT test. Credit for additional tightness is not permitted for Tpa > Tpmax. The gasket stress Sc represents a maximum acceptable value that may not be exceeded on assembly unless it is confirmed by test that tightness performance will not be impaired by a greater stress. It is up to the designer to establish an appropriate value of Sc where none is given. Also: Sc is limited by the yield strength of respective compression stops,

and Sc is the compressive yield strength of solid metal gaskets. For PTFE-based gaskets, the value of Sc is established by a tightness crush test as described in chapter 5. Otherwise the value of Sc is indicated by the ROTT test data and the judgment of the authors and has not been determined by a tightness crush test. In the case of spiral-wound gaskets, the constants are valid only for designs that employ a compression limiting feature such as a groove or gage rings. The limit Tpmax is set as the lesser of the tightness hardening limit Tps and the stress-based tightness limit Tpu, which is related to the stress limit SU, as in SU = Gb(Tpu)a. The limits SU and SL recognize the upper and lower extremes of the test load range used to determine the constants. Their use prevents excessive extrapolation beyond this range. The possibility of gasket crushing or collapse with corresponding changed unloading performance or impared tightness performance with increased gasket stress is a factor that limits the useful upper stress of the gasket constants. The validity of the constants is typically to within 1020% of the maximum ROTT test stress level. The new rules have adopted 110% of the maximum ROTT test stress level. Beyond this, a special ROTT test is needed to prove constants for a different range of stress levels.

TABLE 2 Gasket Materials and Data Gasket Material Compressed elastomers reinforced with: Asbestos fibers (high wt%) 1/8 in. thick 1/16 in. thick Aramid fibers (less than wt50%) 1/8 in. thick 3/32 in. thick 1/16 in. thick Glass fibers (less than wt50%) 1/16 in. thick 1/32 in. and 1 mm thick Beater process elastomers with Aramid fibers 1/16 in. thick Unfilled PTFE sheet (virgin) 1/8 in. thick 1/16 in. thick Filled skived cut PTFE sheet 1/8 in. thick (glass) 3/32 in. thick 1/16 in. thick (glass) Restructured filled PTFE sheet 1/8 in. thick (barium sulfate) 1/16 in. thick (silica)

(table continued on next page)

Gasket ID code Gb a

CAU4 400 0.380 CAU2 25000.150 CFU4 19000.210 CFU3 19000.210 CFU2 560 0.334 CGU2 11500.300 CGU1 285 0.450 BFU2 900 0.450 PVS2 PVS4

6 0.710 30 0.520

PFG4 PFF3 PFG2

430 0.270 220 0.400 520 0.256

PRB4 500 0.339 PRS2 15000.227

(table continued from previous page) Gasket Material 1/16 in. thick (glass) 1/16 in. thick (barium sulfate) Expanded PTFE Joint sealant chord, 3/8 in. Sheet, 1/8 in. thick Sheet, 1/16 in. thick Laminated flexible-graphite sheet reinforced with: Tanged st. stl. sheet Chemically bonded steel sheet Steel screen Bonded plasic film Unreinforced Spiral-wound stainless steel Asbestos filled Flexible-graphite filled (Consistent w/c1 150) PTFE filled (Consistent w/c1 150) Mica filled Spiral-wound stainless steel with inner ring Flexible-graphite filled PTFE filled (consistent w/c1 150) Corrugated metal jacketed with soft insert Soft copper or brass Soft steel or iron Monel or 46% Cr Stainless steel or 12% Cr

Gasket ID code Gb a PRG2 200 0.364 PRB2 320 0.279 PEJ3 PES2 PES4

1000 0.250 1700 0.200 1400 0.222

LGSM LGSB LGSS LGPB LGUB

1400 0.324 816 0.377 1700 0.260 970 0.384 970 0.384

SSAE SSGE SSG1 SSPE SSP1 SSFE

3400 0.300 2300 0.237 600 0.390 4500 0.140 6720 0.100 2600 0.230

SSGI SSP2

2530 0.241 2280 0.190

KCNU KFNU 8500 0.134 KMNU 8500 0.134 KSNU 8500 0.134

TABLE 2 Continued Gasket Material Corrugated metal sheet (0.0150.025 in.) Soft copper or brass Soft steel or iron Stainless steel or 12% Cr Flat metal jacketed with soft insert Soft copper or brass Soft steel or iron Stainless steel or 12% Cr Solid flat metal (3/64-3/32 in.) Soft aluminium Soft copper or brass Solid flat metal in. nubbin facing) Soft copper or brass Soft steel or iron *Indicates that tightness hardening sets Tpmax.

Gasket ID code

Gb

WCN1 WFN1 WSN1

1500 3000 4700

JCNU JFNU JSNU

1800 2900 2900

FAN FCN

1525 5000

NCN4 NFN4

2400 12000

Page 443 TABLE 3 Gasket Identification Codes A code for gasket identification consisting of four digits is presented, e.g., SSG6. Digit 1 describes the product form and type of gasket, (i.e., compressed sheet, spiral wound, etc.). Digits 2, 3, and 4 describe material, class, grade, thickness, or other attributes keyed to the type of gasket. Digit 1: Type of Gasket Sheet: C = compressed-fiber reinforced. B = beater-process-fiber reinforced. E = elastomer, fabric, or other soft material. L = laminated (as in flexible-graphite or PTFE layers) P = PTFE sheet or formable sealant rope Solid metal: F = flat ring, with or without soft facing G = grooved or profiled ring, with or without soft envelope or facing N = nubbin facing against ring gasket R = ring joint, oval, hexagonal, or API J = flat envenlope with soft filler, with or without soft Jacketed: enveloe or facing K = corrugated metal envelope with soft filler, with or without soft envelope or facing O-ring O = hollow metal or solid elastomeric Spiral wound S = various metal windings and fillers Corrugated W = thin corrugated ring, with or without soft metal envelope or facing Digits 2, 3, 4, Depend on Digit 1. BXXX or CXXXfiber-reinforced sheet Digit 2 is fiber material: A (asbestos), F (nonasbestos), G (glass), M (mineral), etc. Digit 3 is grade: P (premium), S (service), E (economy) Alternatively for binder: N (NBR, nitrile), B (SBR), C (CRNeoprene), P (EPDM), etc.

Digit 4 is thickness; E.g., 2, 3, or 4 for .gif .gif or .gif in. thick, 1 for 1 mm, etc. FXXX, GXXX, NXXX, or WXXXsolid metal ring Digit 2 is core ring material: F (iron or low-carbon steel), A (aluminum), C (soft copper), M (Monel), N (nickel), S (stainless steel), etc. Digit 3 is envelope or facing material: G (graphite), P(PTFE), N (none), etc. Digit 4 is thickness: E.g., 2.3, or 4 for .gif .gif .gif in. thick, 1 for .gif or 1 mm or less, etc. JXXX, KXXXenvelope with soft filler Digit 2 is jacket material: F (iron or low-carbon steel), A (aluminum), C (soft copper), M (Monel), N (nickel), P (PTFE), S (stainless steel), etc. Digit 3 is envelope or facing material: G (graphite), P (PTFE), N (none), etc. Digit 4 is filler material: A (asbestos milboard), M (mica), G (flexible graphite), F (iron or low-carbon steel), etc.

Page 444 TABLE 3 Gasket Identification Codesmndash;Continued LXXX Digit 2 is laminate material: G (flexible graphite), P (PTFE), etc. Digit 3 is core reinforcement sheet material: S (stainless steel) P (Polymer), M (Monel), N (nickel), etc. Digit 4 is method of bond: M (mechanical tang), B (chemical adhesive), N (none), O (other) PXXXPTFE sheet or formable sealant rope Digit 2 is type: F (skived cut), B (biaxially oriented filled PTFE) E (expanded or microcellular or low density), R (restructured filled PTFE), etc. Digit 3 if filler: B (barium), C (composite), F (filler unknown), G (glass), S (silica), V (not filled, virgin), etc. If digit 2 is E (PEXX), then digit 3 is product form: S (sheet), R (rope sealant), M (metal reinforced.) Digit 4 is thickness: Eg., 2, 3, or 4 for .gif .gif or .gif in. thick, 1 for 1mm, etc. SXXX Digit 2 is winding material: S (stainless steel), M (Monel), N (nickel), etc. Digit 3 is filler material: G (graphite), P (PTFE), A (asbestos), etc. Digit 4 is for rating class and gage ring configuration: N (no rings such as for a groove) No inner ring: 1 (C1 150), 3 (C1 300), 6 (C1 600), 8 (C1 1500), C (C1 2500), E (not rated) With inner ring: 2 (C1 150), 4 (C1 300), 7 (C1 600), 9 (C1 900), B (C1 1500), D (C1 2500), I (not rated) RXXX Digit 2 is ring material: D (iron or low-carbon steel), F (46 Cr), M (Monel), N (nickel), S (stainless steel), etc. Digits 3 and 4 are ring profile type: OV (oval), OX (octagonal), BX (API BX type), RX (API RX type)

In the case of tightness hardening (see Chapter 5), Tps and Ss are defined by a tightness hardening finding from an ROTT test. Ss is related to the tightness limit Tps, as in Ss = Gb(Tps)a Gaskets with a compression limiting device are assumed to perform as though there is tightness hardening. In the case of a compression stop, or for tightness hardening gaskets, no credit is permitted for tightness beyond the tightness hardening limit or after the stop is activated. 4.7. Gasket Service Temperature Limitations The constants Gb and a are associated with the initial assembly and seating of the gasket, which is assumed to take place at room temperature and for a new gasket. Any in-service changes in Gb and a do not affect the design. In effect,

Page 445

Gb and a are valid for any gasket service temperature. However, constant Gs reflects the operating condition, and its validity depends on the long-term service temperature of the material. The new rules provide that the constant Gs is valid for the following service temperature conditions: 250°F for elastomer or reinforced-elastomer sheet materials. 600°F for flexible-graphite sheet materials. 900°F for flexible-graphite steel or high-alloy composite gaskets where the graphite is encapsulated by the metal component after seating. This includes spiral-wound, metal-jacketed, profiled metal, and corrugated metal components. In addition, the designer is required to be satisfied that the gasket materials are suitable for the design conditions over the intended length of service. Corrosion, chemical attack, creep, and thermal degradation of gasket materials over time shall be considered. On the question of Gs validity for other materials, the authors are aware that certain gasket manufacturers have performed a special test called the EHOT test (see Chapter 5) that determines Gs before and after an exposure sufficient to realistically age the material for 35 years. These manufacturers should be consulted for further information. In addition, HOTT test results (see Chapter 5) may be used to estimate postexposure values of Gs. 4.8 Assembly Efficiency and Assembly Procedures To ensure reasonably that sufficient bolting is being provided, the rules incorporate an assembly efficiency, a factor related to gasket stress variation as a function of assembly bolt load control, which,

therefore, relates to the method of tightening. As the variability of a tightening technique increases, h decreases from 1 to 0.75 and Am increases. When the strain of each bolt is measured and carefully controlled throughout the tightening process, h = 1; h = 0.75 pertains to tightening by ordinary manual wrenching methods. Refer to Table 4 for the recommended values of h. Refer to Appendix A for additional information. The proposed new rules also require that bolted flanged joints designed in accordance with ASME B&PV Code Appendix BFJ be assembled by qualified assemblers using qualified procedures. Bolted joints designed for assembly efficiencies greater than 0.75 will be assembled and bolted up in accordance with a written procedure that has been qualified by test to achieve the specified assembly efficiency. Further, the joint assemblers will need to be qualified by test prototype assembly to demonstrate that they are able to apply the qualified procedure and achieve the specified assembly efficiency.

Page 446 TABLE 4 Assembly Efficiencies Bolt load variation from efficiency Bolt preload control method mean h percent ±50 or 0.75 Power impact, lever or striker (manual or power) more 0.85 Accurately applied torque (approx. ±3%) ±30±50 Simultaneous application of direct tension to three 0.95 ±10±30 or more bolts Direct measurement of stud stress or strain or the ±10 or 1.00 simultaneous hydrostatic tensioning of all bolts less Assembly

4.9. An Example Calculation In simplest terms, the design process for arriving at Wmo, which is illustrated in Fig. 1c, may be summarized as follows: 1. Determine the minimum tightness required (Tpmin) for a given Tc and P from Eqs. (2). 2. For a given h Sa/Sb, flange geometry, gasket, and an associated trial assembly tightness Tpa determine a trial seating stress Sya to ensure that Tpmin will be achieved. There are several methods available for this, including the possible initial assumption of a Tpa that is:

3. Determine the minimum design gasket stress Smo that is required to maintain the specified minimum level of tightness

(Tpmin) during operation. This refers to steps (3) through (7) of Fig. 1c. As described in the foregoing, Smo is determined, from both seating and operating tightness requirements, as the greatest of Sm1, Sm2, 2P, and SL by an iterative or approximate process. Checks must be made to ensure that stress and tightness are within specified limits. 4. Determine the minimum bolt are load required for operating conditions Am = Wmo/Sb, where Wmo = Smo(Ag) + P(Ai) + He. The equation for Wmo means that the bolts are designed for the sum of the pressure load (also called the hydrostatic end force), as represented by P(Ai), plus any external loads, as represented by He, plus a gasket load that is sufficient both to maintain a seal and to seat the gasket adequately. The preceding process also considers external loads essentially by adding them to the term P(Ai) wherever it appears.

Page 447

Specific Example This process is illustrated by EXAMPLE SHEET No. 1, Case 1. (See Fig. C2, Appendix C of this Chapter) The calculation format of the EXAMPLE SHEET intentionally resembles the format of the examples of Taylor-Bonney Bulletin 502 [12]. Also, the flange design problem selected for this example is exactly the same as the welding neck flange design EXAMPLE 1 found in Bulletin 502, namely, for a 2-in.-thick 32-in.-ID flange designed for 400 psi and 500°F with a 34.5-in.-OD compressed-asbestos sheet gasket. In this example the solution of Sm1 = Sm2 is by the iterated flexible method, where the factor X = Tpal Tpmin is increased from 1.5 to 13.05 until Sm2 = Sm1. Figure 1d illustrates the way Sm1 approaches Sm2 until they intersect at X = 13. 5. Comparing the New Rules (NR) With Asme Boiler Code Appendix 2 Rules (CR) Clearly, the ratio of the amount of bolting required based on the new and the current rules is simply Am (NR)/Am (CR), where the numerator and denominator are figured per the applicable rules. As will be seen later, this ratio is 1.08 for the preceding example, where it is assumed that Ab = Am, Tc = 1.0, and h =

Figure 1d This chart shows how Sya, Sm2, Sm1, and Smo vary with the ratio X = Tpa/Tpmin. The curves are drawn from the solution for Example 1, Case 1, where the optimum bolt load by the flexible method is found at a value of X = 13. Smo is the greater of Sm1 and Sm2.

Page 448

0.75 and the thickness is left the same. However, since there is a small overstress (3%), the designer would need to increase the flange thickness in this case. Also, since Am = Ab was used for this example, which is unlikely, Ab would need to be increased, which in turn would require the flange thickness to be increased somewhat further by the new rules unless some other action to reduce the required thickness were taken. Possible other actions might be to use tightness class T1 (Tc = 0.1), a higher h, or a lower pressure or to choose a tighter gasket. If a higher h is chosen, then the new rules would require a qualified assembly method and qualified assemblers to ensure the design assembly efficiency is actually achieved. If a reduced tightness class were proposed, the designer and the user would need to recognize that tightness is a design requirement and that the proposed tightness reduction (and resulting leak increase) may be unacceptable. Similarly, a reduced pressure may be unacceptable. 5.1 Simplified Comparison Methodology In general, for the purpose of broad comparisons, figuring the ratio of the results is not quite as obvious as the preceding example because there is the question of how to compare them. For example, should it be by weight, dimensions, or some other measure? Therefore the approach taken in this chapter is to present both broad comparisons for a range of flange diameters and pressures and more detailed comparisons for a few specific flanges. For broad parametric comparisons a simplified methodology has been developed that is explained by the following, where the ratio Am (NR)/Am (CR) for simple rings provides a consistent basis. a. Current Appendix 2: Based on ASME B&PV Code Appendix 2

(with Appendix 2 notation), the seating moment is Ms = 0.5 (Am + Ab)Sa(hG), and the operating moment is Mo = HG × hG + HT + hT + HD × hD. Although it is geometrically improbable, it is convenient to assume that hG = hT = hD = 1.0. This simplifying assumption has a negligible effect on the result, especially since it is used for both ASME B&PV Code Appendix 2 and the new rules. It follows from this assumption that Ms = (Ab × Sa)hG, if Ab = Am, and Mo = (H + Hp) × hG. If Sfo = Sfa, the larger of Ms and Mo governs the design of the flange. In general terms, the larger of Ms and Mo(Sfa/Sfo) controls (or establishes) the size of the flange, where Mo(Sfa/Sfo) is the operating moment adjusted to the room temperature allowable flange stress. In other words, M (Adj.) = Mo(Sfa/Sfo), whereas the controlling moment of the current rules is M (CR) which is the greater of Ms and Mo (Adj.). Finally, the dimensions of a flange that satisfies ASME B&PV Code Appendix 2 are determined using M (CR) along with Sfa (the allowable room temperature flange stress). Many factors will determine the final dimensions of the flange; however, the best index of size appears to be the moment upon which flange dimensions are to be based, i.e., M (CR). Accordingly, we have selected the controlling moment as our index

Page 449

of size; the reader may then optimize the dimensions according to his or her definitions of optimum, using, of course, M (CR). b. New rules based on PVRC gasket constants: When the new rules are used, after Wmo and Am are determined, a flange design proceeds essentially per the rules in ASME B&PV Code Appendix 2; there are two differences, however. In Appendix 2, Ms = 0.5(Ab + Am)(Sa)(hG) and Mg = HG × hG where HG for the operating condition is 2b(3.14)GmP. This assumes that the operating load on the gasket is exactly equal to the required load of 2b(3.14)GmP. Theoretically, if the total load used to figure Ms is 0.5(Ab + Am)Sa, then the load on the gasket in operation, discounting any relaxation of the bolt load, should be 0.5(Ab + Am)Sa - H, and not 2b(3.14)Gmp. In the new rules, the seating (assembly) moment is based on the total bolt load of Ab × Sa; therefore HG for the operating condition becomes Ab × Sa - H. Again, using the simplifying assumption hG = hT = hD = 1.0, under the new rules Ms = (Ab × Sa)hG and Mo = (Ab × Sa)hG. The reasons for the simplifying assumption and for making Mo = (Ab × Sa)hGshould now be apparent. Adjusting Mo to a room temperature allowable stress gives Mo (Adj.) = (Sfa/Sfo) × (Ab × Sa)hG. Mo (Adj.) is always equal to or greater than Ms; therefore, the controlling moment for the new rules is M (NR) = (Sfa/Sfo)(Ab × Sa)hG, where Am (NR) = Wmo/Sb (Note: Reverse flanges are an exception.) c. Comparing the new rules with the current rules: Two ratios are used to compare the new rules with the current rules (ASME B&PV Code Appendix 2), specifically, Am (NR)/Am (CR) and M (NR)/M (CR). The ratio of the controlling moments is a good

measure of the ratio of size. If the flanges being compared are considered as simple loose-type rings without hubs with the same OD and ID, it follows that the ratio of flange thicknesses is:

For example, if the ratio of the controlling moments is 2.0, a loosetype hubless flange based on the new rules will be 2½, or 1.414, times the thickness of the flange based on ASME B&PV Code Appendix 2 (i.e., the current rules). When a hub is introduced, the moment-to-thickness relationship is more complex. Nevertheless, for broad comparisons, the ratio of controlling moments is always a valid and meaningful indicator of relative size, which is ideal for our purposes. 5.2 Finding the Maximum Value of Am (NR)/Am (CR) For a given type and size of gasket, the minimum bolt load per ASME B&PV Code Appendix 2 is Wm2 = (3.14)bGy, and the minimum value of Am (CR) is Am2 = Wm2Sa. Am remains constant with increasing pressure until Am1 = (H + Hp)/Sb = Am2 at pressure P = P(A). When P > P(A), the amount of bolting

Page 450

required increases continuously with increasing pressure; for any pressure P > P(A), Am = (H + Hp)/Sb. As the pressure is increased from 0 to P(A), per ASME B&PV Code Appendix 2, Am is constant and equal to Am2 = Wm2/Sa, whereas, using the new rules, Wmo and Am increase continuously with increasing pressure. As a result, the maximum value of Am (NR)/Am (CR) occurs when P = P(A). P(A) is found as follows: P(A) is the pressure at which Am1(CR) = Am2(CR). Am1 = [(3.14)G2P/4 + (3.14)2b(GmP)]/Sb Am2 = [(3.14)(bGy)]/Sa Equating Am1 and Am2 and solving for P, now defined as P(A), yields

where all terms are as defined in Appendix 2 of the ASME B&PV Code. 5.3. The Maximum Value of M (NR)/M (CR)

For pressure 0 to P = P(A), the seating moment per ASME B&PV Code Appendix 2 is a constant, Ms = (3.14)(bGy)hG. When P > P(A), Ms = (Ab × Sa)hG. The first step is to find the pressure P(M) at which Mo(Sfa/Sfo) = Ms = (3.14)(bGy)hG. When P = P(M), the design of the Appendix 2 flange may be based on Ms = (3.14)(bGy)hG or on Mo × (Sfa/Sfo). In either case, the room temperature flange design stress must be used. When P > P(M), Mo × (Sfa/Sfo) governs the design of the ASME B&PV Code Appendix 2 flange. For Wm1 = (H + Hp), Am1 = Wm1/Sb, Wm2 = 3.14(bGy), and Am2 = Wm2/Sa. When Am1 < Am2, Ms = (3.14)(bGy)hG and Mo (Adj.) = (Sfa/Sfo)(H + Hp)hG, or Mo (Adj) = (Sfa/Sfo)[(3.14)G2/P4 + (3.14)2b(GmP)]hG by definition, at P = P(M), Mo (Adj.) = Ms = (3.14)(bGy)hG; therefore (Sfa/Sfo)[(3.14)G2P/4 + (3.14)2b(GmP)]hG = (3.14)(bGy)hG. Solving the expression for P, which is now P(M), P(M) = 4(by)(Sfo/Sfa)/(G + 8bm) The expressions for P(A) and P(M) are similar. The allowable bolt stress multiplier (Sb/Sa) appears in the equation for P(A), whereas the allowable flange stress multiplier (Sfo/Sfa) is in the expression for P(M). As a result, P(M) may be equal to, greater than, or less than P(A) depending on these multipliers. When P = P(A), Ms = (3.14)(bGy)hG; but when P > P(A), M (CR) = M (adj) > (3.14bGy)hG. Therefore, the maximum value of M (NR)/M (CR) occurs at a pressure equal to the smaller of P(A) and P(M). This is an important relationship! 5.4. Useful Relationships Simple relationships exist between M (NR)/M (CR) and Am (NR)/Am (CR), denoted, respectively, in these paragraphs as M/M and A/A for convenience. In

Page 451

a-d following, which give M/M as a function of A/A for various situations, A/A applies to a specific pressure P for the specific value of the ratio Sb/Sa used in the determination of A/A. In the following, Sfa/Sfo is the variable. a.

For P(A) > P < M/M = (Sfa/Sfo) × (A/A) P(M):

b.

For P(A) > P > M/M = [(P(A)/P) × (Sa/Sb)] × P(M): (A/A)

c.

For P(A) < P < M/M = (Sfa/Sfo) × (A/A) P(M):

d.

For P(A) < P > If P(A) < P(M), then M/M = P(M): (Sfa/Sfo) × (A/A) If P(A) > P(M), then M/M = (Sa/Sb) × (A/A)

Note that when Sfa/Sfo = Sa/Sb, then P(A) = P(M). 5.5. Predicting the Variation of M/M and A/A It is now possible to predict what a plot of M/M and A/A should look like as a function of pressure. To do this, all other parameters that affect M/M and A/A, including Sfa/Sfo and Sa/Sb, must be held constant. Although it is not essential, it is convenient at this point to assume that Sfa/Sfo = Sa/Sb. As a result, P(M) = P(A). For pressures P = 0 to P = P(A), Am per ASME B&PV Code Appendix 2 is a constant equal to 3.14(bGy)Sa; per the new rules, Am increases continuously with increasing pressure. Accordingly,

the ratio A/A increases from P = 0 to P = P(A). The value of A/A peaks at P = P(A), owing to the fact that Am (NR) always increases with increasing pressure, whereas Am (CR) starts to increase at P = P(A) when Am1 = Am2. As the pressure increases beyond P = P(A), the ratio A/A decreases quite abruptly at first with the curve of A/A vs. P flattening off in the higher range of pressures as Am (NR) catches up with Am (CR). The M/M curve may be deduced from the A/A curve and the relationships given in Section 5.4. For Sfa/Sfo = Sa/Sb, P(M) = P(A) and relationships (a) and (d) apply. Accordingly, the entire M/M curve becomes M/M = (Sfa/Sfo)(A/A) = (Sa/Sb)(A/A). Figures 2 show the M/M curve for Sb/Sa = 0.84. In summary it is observed that the maximum value of Am (NR)/Am (CR) occurs when P = P(A). The maximum value of M (NR)/M (CR) occurs at a pressure P equal to the lesser of P(A) and P(M). 5.6. Example CalculationsBroad Comparison Figures 2a and 2b illustrate the plot of A/A vs. pressure visualized earlier. These are the result of example calculations illustrating the broad comparison methodology just described. The calculations are for a joint containing a 375-mm OD × 23.8-mm (14.75-in. OD × 15/16-in.) wide (stainless steel) spiral-wound gasket, designed for tightness class T2 (corresponding to a mass leak rate of 0.002 mg/sec/mm of gasket diameter) and the following:

Page 452

Figure 2a A/A represents the ratio between the total cross-sectional (root) area of the bolting suggested by the new rules to the total root area suggested by the traditional ASME B&PV Code rules. M/M is the ratio between the controlling moment suggested by the new rules to the controlling moment suggested by the traditional rules. See Section 5 for a further discussion of controlling moment. This plot shows how those ratios are affected by variations in the design pressure for a spiral-wound asbestos-filled gasket.

1. Bolts are to be tightened using a method that allows the use of an assembly efficiency equal to 1.0. 2. The iterative (flexible) method in the new rules is to be used to compute Wmo. 3. Design pressure = 3.79 MPa (550 psi) and design temperatures of ambient and 427°C (800°F). 4. Elevated temperature flange: Sfa = 121 MPa (17,500 psi), Sfo = 82.8 MPa (12,000 psi); bolts: Sa = 172 MPa (25,000 psi), Sb = 145 MPa (21,000 psi).

5. Gasket constants: m 3.00 and y = 10,000 (per Appendix 2) for spiral wound = Gb23.45 MPa (3400 psi), a = 0.300, and Gs = 0.64 MPa =(93 psi) for SW asbestos filler per the new rules Gb15.86 MPa (2300 psi), a = 0.237, and Gs = 0.0896 MPa =(13 psi) for SW graphite filler per the new rules Calculation Sheet SW 1 (Appendix C of this chapter) shows typical calculations for the ratios A/A and M/M for the asbestos-filled spiral-wound gasket at 427°C

Page 453

Figure 2b A plot similar to that of Fig. 2a, but this time for a spiral-wound graphite-filled gasket.

(800°F) and the calculations for the bolt load (Wmo) for this case according to the new rules. These calculations are illustrated in Fig. 3. Figure 2a summarizes A/A and M/M for the same spiral-wound gasket with asbestos filler for various pressures at ambient and elevated temperature conditions. Calculation sheets SW 2 and Fig. 2b do the same for the flexible-graphite-filled gasket. Figure 4 summarizes A/A and M/M for the same spiral-wound gasket with flexible-graphite-filled gasket at 3.45 MPa (500 psi) for ambient and elevated temperature conditions over a range of diameters. Comment: It is seen from Sheet SW1 (Appendix C) that A/A = 2.62 and M/M = 3.12 for the hot case. These figures can be put into perspective by assuming that the flanges being compared are loosetype hubless and dimensionally identical except for thickness. An M/M = 3.12 means the flange based on the new rules will be 3.12½,

or 1.767, times the thickness of the flange designed using ASME B&PV Code Appendix 2. By comparison (from Sheet SW 2), the hot case for the graphite-filled gasket (which is tighter) resulted in an M/M of 1.59. This means a ring flange that is 1.59½, or 1.26, times the required thickness of the Appendix 2 flange. It is seen from Sheet SW1 that the maximum value of A/A is found at P = P(A) = 516 psi(3.56 MPa). Therefore, repeating the calculations with the pressure changed from P = 550 psi(3.79 MPa) to P = P(A) = 516 psi(3.56

Page 454

Figure 3 Plot of gasket stress vs. the tightness parameter used to illustrate the results of the new rules example tabulated on Sheet SW 1. The gasket here is spiral-wound asbestos-filled Class 600 with stainless steel windings.

MPa) yields A/A = 2.723. This is the maximum value of the ratio for the subject gasket and Sb/Sa = 21,000/25,000 = 0.84. Similarly, the maximum value of M/M occurs when P = P(M) = 421.1 psi (2.90 MPa), and at this pressure M/M = 3.66 is the maximum value of the ratio for the subject gasket and the given stress ratios (Sb/Sa = 0.84, Sfo/Sfa = 0.69). The A/A and M/M ratios found in this case may seem to be too large considering the good service experience that industry has had with flanges designed in accordance with ASME B&PV Code Appendix 2. It should be remembered that in Appendix 2, mass leak rate (i.e., leakage) is not a design parameter and that, generally, tightness was evaluated during a routine hydrostatic test. In contrast, the preceding calculations were based on tightness

class T2, which corresponds to a mass leak rate of 0.002 mg/sec/mm. Changing the tightness criteria from T2 to T1 (0.2 mg/sec/mm), which increases the permitted leak rate 100 times, yields the following results for the Case SW1 otherwise described earlier: A/A = 1.28; M/M = 1.53 (The corresponding T1 maximums are A/A = 1.32 and M/M = 1.75.) Thus, using tightness criterion T1, instead of T2, reduces A/A and M/M to about half the T2 values in this case. On this basis, the new rules require at most

Page 455

Figure 4 A plot showing the influence of gasket diameter on the two ratios A/A and M/M (which are defined in the caption for Fig. 2a).

only 29% more bolting than does ASME B&PV Code Appendix 2 and a flange thickness approximately (1.53)½, or 1.24, times the flange thickness based on Appendix 2. This is a result where the new rules are only slightly more conservative than Appendix 2. The term more conservative is a misnomer, in that a specific value of a maximum acceptable leak rate is used when designing per the new rules. When using Appendix 2, it is implicit that one is designing for a tight joint (a word that is undefined), which often may not be realized. 5.7. Specific Flange Example The example weld neck flange of TaylorBonney Bulletin 502, 1978, was analyzed by the new rules for several cases, as tabulated in Table 5. Case 1 was referred to earlier. The 32-in. (813-mm) ID flange is to be designed for 400 psi (27.6 MPa) and 500°F (260°C)

with a 34.5-in. (876-mm) OD compressed-asbestos sheet gasket. In Case 1 the solution of Sm1 = Sm2 is by the iterated flexible method, where the factor X = Tpa/Tpmin is increased from 1.5 to 13.05 until Sm2 = Sm1. Appendix C gives a detailed comparison between the new rules and the Code's present Appendix 2 rules for this Case 1. Case 2 increased Ab, which in turn increased the maximum stress to 1.08 times the allowable stress and A/A to 1.12. Case 3 is the same as Case 2, except that external loads were added as a design condition. The stress results were

TABLE 5 Summary of Results for 813-mm (32-in.) flange example. Case Flange thickness, in. Ab > Am? Value of Ae Value of Tc External loads Value of X for Sm1 = Sm2 Longitudinal hub, sh, MPa Longitudinal hub, sh, psi Radial flange, sr, MPa Radial flange, sr, psi Tangential flange, st, MPa Tangential flange, st, psi Average of sh and sr, MPa Average of sh and sr, psi Average of sh and st, MPa Average of sh and st, psi Ratio stress (NR)/stress (CR) Max stress (NR)/allowable

1 2 No 0.75 1 No 13.05 168.028 24,364 80.703 11,702 49.966 7,245 124.366 18,033 108.993 15,804 1.07 1.03

2 2 Yes 0.75 1 No 13.05 176.807 25,637 84,917 12,313 52,579 7,624 130.862 18,975 114.69 16,630 1.12 1.08

3 2 Yes 0.75 1 Yes 15.87 176.807 25,637 84,917 12,313 52,579 7,624 130.862 18,975 114.69 16,630 1.12 1.08

4 2 No 1.00 1 Yes 72.03 163.662 23,731 78,607 11,398 48.669 7,067 121.138 17,565 106.166 15,394 1.04 1.00

Page 457

exactly the same as in Case 2, because for the new rules W = Sa × Ab, which was the same for both cases and in turn results in the same flange moments. Of course, Am did increase for Case 3. Case 4 is the same as Case 3, except the assembly efficiency (h or Ae) is increased to 1.00 and Am = Ab. In this case the stress ratio NR/CR is 1.04. Case 5 increases the required tightness to class T3 (Tc = 10) and has no external loads. The Case 5 stress ratio increased to 1.21 for the 2-in.-thick flange. In Case 6 the flange of Case 5 was increased to a thickness of 2.27 in. from 2.00 in. to reduce the ratio maximum stress (NR) allowable to 1.00. It is observed from this particular flange example that a tightness class T2 results in loads and stresses that are quite close to an ASME B&PV Code Appendix 2 design. Cases 1, 3, and 5 are tabulated in Appendix C. 5.8. The Effect of Certain Parameters on A/A and M/M When the new rules are issued, people representing various segments of the pressure vessel and piping industries, including regulatory and environmental, will want to see how the new rules may affect them. To accomplish this, example problems will be devised and solved using both sets of rules; conclusions will then be drawn based, probably, on too few samples. In any case, knowing, even in a qualitative way, how changing various parameters affects A/A and M/M should make it possible to devise fewer and more meaningful sample problems. Parameters that affect A/A and M/M are listed next, each with a brief commentary. 1. Tightness class: Increasing the tightness class, e.g., from T1 to

T2, is equivalent to reducing the allowable mass leak rate; the effect is to increase A/A and M/M. Figure 5 illustrates the effect of tightness class from Tc = 0.1 to Tc = 100 (class T1 through class T4) on the ratio A/A. In this example the PVRC gasket constants are for the compressed-asbestos gasket of the earlier example, and the bolt areas are calculated and compared for a 500-mm (19.8-in.)diameter by 12.7-mm ( )-wide gasket. A tightness limit of 50,000 was applied. At the pressure P(A) = 1.03 MPa (148.9 psi) in this example it is seen that tightness class has a large effect. In the case of class T4, the bolts are 4.5 times larger than required for ASME B&PV Code Appendix 2. However, at pressures above 5.5 MPa (800 psi) the effect is much less, because limitations of Smo = 2P, or Tpmax, are reached. 2. Assembly efficiency: Increasing the assembly efficiency, e.g., from h = .75 to h = 1.0, requires more control in tightening the bolts at assembly. As a result, less extra bolting is needed to compensate for bolting and gasket stress variations. When h = 1, Sya falls on the loading curve; when h < 1, an average gasket assembly stress (or seating stress) of Sya/h is required, so, theoretically,

Page 458

Figure 5 Plot showing how the ratio A/A (vertical axis) is affected by a combination of tightness class and service pressure. Tightness class T1 allows more leakage than tightness T2; T3 and T4 allow progressively less still. See Table 1 for tightness class data.

the effective gasket assembly stress equals Sya (on the loading curve) corresponding to Tpa. 3. Pressure: The ratio A/A increases with increasing pressure until P = P(A); thereafter, the value of the ratio decreases. The ratio M/M increases with increasing pressure to P = P(M); thereafter, the value of the ratio decreases. 4. Gasket width (N): In the new rules, the width of the gasket in contact with the flange surface, i.e., N, is used in all calculations involving the width of the gasket (for example, Ag). On the other hand, ASME B&PV Code Appendix 2 uses the concept of an effective gasket seating width b, where b = N/2 and b = bo when bo < in. (6.3 mm); b = 0.5 (bo)0.5 when bo > in. (6.3 mm). bo

is called the basic seating width. For a given design problem, if everything is held constant except N, and N is increased from N1 = 23.8 mm (0.938 in.) to N2 = 35.7 mm (1.406 in.), A/A will be higher when N = 35.7 mm (1.406 in.) The question is approximately how much higher it will be. For pressures below P(A), when N is increased by a factor of 1.5, Am(NR) will increase by slightly less than N2/N1 (= 1.5). Using ASME B&PV Code Appendix 2, Am (CR) increases by the ratio of effective gasket widths b2/b1. Therefore, in this case, b2/b1 = (1.406/0.938)½ = 1.225. This means that if A/A1 is known for Case 1 when N = N1 = 23.8 mm (0.938 in.), then A/A2 for Case 2 when N = N2 = 35.7 mm (1.406 in.) becomes

Page 459

The factor 1.22 is somewhat high, because it assumes that Am (NR) increases by a factor equal to N2/N1 = 1.5. The actual increase is slightly less than 1.5, because the change in Ag results in a change in Sm2, which in turn causes a shift in the iterated solution toward a lower Am. 5. Gasket diameter: The diameter of the gasket and its width N are two additional parameters that affect A/A and M/M. As the diameter of a gasket increases with N = constant, the ratio of the area acted on by pressure (Ai) to the gasket area (Ag) increases, and this is reflected in the results. For example, for a spiral-wound asbestosfilled gasket, h = 1, N = 25.4 mm (1.0 in.), tightness class = T2, Sb/Sa = 0 84, Sfo/Sfa = 0.686, and using the iterative method (flexible), Table 6 shows how the maximum values of A/A and M/M vary as the gasket size increases from 152 to 915 mm (6 to 36 in.). 6. Ratio Sb/Sa: A plot of A/A vs. P for a given value of Sb/Sa shows that as pressure is increased, A/A increases and reaches a maximum value at P = P(A); A/A then decreases with increasing pressure. If Sb/Sa is reduced, the curve reaches a maximum value at a lower pressure P = P(A) corresponding to Sb/Sa. For a series of Sb/Sa ratios, the plot consists of a series of wavelike curves, each with its peak displaced to correspond with its value of P(A). The lower the value of Sb/Sa, the lower the pressure at which A/A reaches its maximum value. TABLE 6 Effect of Diameter on Max A/A and M/M Gasket P(A) (A/A) P(M)

(M/M)

outside diameter 152 mm 6 in. 203 mm 8 in. 305 mm 12 in. 447 mm 18 in. 610 mm 24 in. 915 mm 36 in.

MPa psi 5.945 862 5.193 753 4.145 601 3.172 460 2.579 374 1.869 271

max 3.074 3.074 3.019 3.019 2.892 2.892 2.726 2.726 2.59 2.59 2.386 2.386

MPa psi 4.855 704 4.241 615 3.379 490 2.593 376 2.103 305 1.524 221

max 4.166 4.166 4.087 4.087 3.902 3.902 3.662 3.662 3.457 3.457 3.188 3.188

Page 460

When h = 1 and the flexible method can be used without X = 1.5Tpmin. taking over so iteration cannot be completed or Smo being controlled by its minimum of 2P, then A/A max. over a range of Sb/Sa ratios is constant. In other words, the maximum value of A/A does not depend upon the value of Sb/Sa. The ratio of Sfo/Sfa does not appear here, since it has no effect on A/A vs. P over a range of Sb/Sa ratios. 7. Ratio Sfo/Sfa: When the new rules are used, the operating moment Mo adjusted to the room temperature flange design stress Sfa is always the controlling moment. [However, when Sfo/Sfa = 1, Mo = Ms = M (NR)]. This is based on the assumptions that (1) hG = hT = hD, (2) Ms = (Ab × Sa)hG, and (3) HG, for the operating condition, equals (Ab × Sa - H)hG. When ASME Boiler Code Appendix 2 is used, the ratio Sfo/Sfa has no effect on the controlling moment when P(A) > P < P(M); for higher pressure, the effect of Sfa/Sfo is seen by referring to the relationships in Section 5.4. These relationships are most helpful when interpreting a mass of comparative results involving (1) different gaskets; (2) a range of pressures; (3) a range of the ratio Sb/Sa; etc. For a single case involving one set of conditions, it is probably easier to use the new rules and Code Appendix 2 and to compare A/A and M/M directly. 6. The Minimum Assembly Bolt Stress In the new rules, the gasket constants Gb, a, and Ga define the behavioral characteristics of a gasket during the initial loading and repetitive unloading cycles. Accordingly, the desired joint tightness at the design pressure is a design parameter. If the required amount

of bolting (Am) for a given joint tightness class is not provided, it should be assumed that the selected tightness is not achievable. On the other hand, if Am is adequate, the desired tightness is achievable only if the bolts are stressed adequately and properly at assembly. The term stressed adequately means that the calculated gasket stress Sya must be developed at assembly and that the gasket operating stress at the design pressure must be at least equal to Smo. In the case of manual (uncontrolled) joint assembly there is no way to ensure that the gasket is stressed adequately except to provide that skilled craftspeople perform the assembly. Further, the use of h = 0.75 provides an additional margin of bolt load in this case. For h = 0.85 or more, chances that the gasket is stressed adequately are greatly improved when assembly is in the hands of skilled craftspeople. The design rules are structured so that when iteration is used, Sm1 = Sm2 and no gasket stress constraints are involved (i.e., Smo = Sm1 = Sm2), stressing the bolts at assembly to develops Sya, and (theoretically) the required gasket operating stress Smo will be achieved if the bolts are unloaded during operation to as low as Sb (the allowable stress at the design temperature). This unloading occurs as a result of the application of operating loads plus relaxation and possibly

Page 461

thermal effects. As a result, the assembly bolt stress conforms with the intent of Appendix S of the Code; and the design value of the operating bolt stress conforms with the intent of Code Appendix 2. For the conditions just described above (i.e., Sm1 = Sm2 = Smo), the minimum assembly bolt stress (or load) will be established by Sya. When , or if constraints take over, for example, Smo = 2P or Smo = SL, and/or the assembly efficiency is less than unity, then the minimum assembly bolt stress may depend on Sya or it may depend on Smo. If the minimum assembly bolt stress, based on considering both Sya and Smo is developed at assembly, the desired joint tightness is considered to be achievable, at least theoretically. This is provided there is no loss of bolt load upon pressurization due to flange rotation, creep, relaxation, interactive effects, etc. While the new rules provide a small margin to compensate for loss of bolt load, ultimately it is the designer's responsibility to account adequately for these effects if the desired joint tightness is to be considered achievable. Additionally, it may be necessary to increase the assembly bolt stress to account for the extra end force produced by the test pressure, which is usually at least 1.5 times the design pressure. The minimum assembly bolt load and bolt stress are taken as the greater of Wma, Wma', and Sma, Sma', where: (A1) (A2)

The minimum assembly bolt stress (MABS) is the greater of Sma and Sma'. 6.1. Assembly Load Example The example that follows is intended to illustrate some of the points just discussed and show the effect of varying Sb/Sa on the required minimum assembly bolt stress. For this example: The test pressure equals 1.5 times the design pressure; i.e., Pt = 1.5P. The iterative method is used to make Smo = Sm1 = Sm2. The problem is to calculate the minimum assembly bolt stress for: A stainless steel flexible-graphite-filled gasket, ID = SSG6. OD = 20 in. and N = 0.50 in. Gb = 2300, a = 0.237, and Gs = 13.0 (from Table 2) Design pressure = 400 psi, test pressure = 600 psi h=1

Page 462

Tightness class T2 Constraints Tpmax = 4268 (from Table 2) Sa = Sb = 25,000 psi The result of this calculation is summarized as follows: Tpmin = 49.71 and, after iteration, Tpa =

200.6 (X = Tpa/Tpmin = 4.036 at Sm1 = Sm2)

Sya = 8080 psi Smo 1487 psi = Wmo 164,995 lb = Am

6.60 in.2

Wma 224,725 lb, Sma = 34,050 psi = Wma 247,492 lb, Sma = 37,500 psi (MABS = = 37,500 psi) As mentioned, MABS represents a minimum value. The Code leaves it up to designers to provide an ample margin to cover the

loss of bolt load for all reasons and to be sure that maximum operating bolt stress values (considering yield, or stress corrosion cracking limitations, for example) are not exceeded. The reason for this should be obvious considering the complex nature of the problem. Further, the consequences of not achieving the desired tightness class may have a profound effect on how the loss of bolt load is to be computed. In other words, in some cases an estimate may suffice, whereas, in other instances a sophisticated finite element analysis may be called for. Nevertheless, the new Code rules will likely require a check on the maximum gasket stress to ensure that tightness performance is not impaired by an excessive gasket load. Most likely this check will take the following form: AgSc > (greater of Wma and Wma')(2 - h) where Sc (gasket crush limit) is as tabulated in Table 2 and the (2 h) factor accounts for the high-side extreme of gasket stress variation that results from assembly bolt load scatter. For our example, Ag = 30.63 in.2 for the gasket. But because there is a compression stop, Sc is defined and controlled by the area and yield strength of the compression stop. For purposes of illustration here we will assume that the compression stop area (Astop) is Ag/2 = 15.3 in.2 and that the yield strength of the stop is 36,000 psi. Therefore, since h = 1:

Page 463

Astop (36,000) =

15.3 (36,000) =551,340 lb >Wma (2 - 1)247,492 lb = OK!

6.2. Significance of 1.5Sa The preceding example shows that the minimum assembly bolt stress (MABS) is determined by the need to develop a gasket stress at assembly equal to Sya. It turns out that for this geometry the equation for Sma' governs for elevated-temperature designs with values of Sb/Sa of 1 to less than 0.5. For Sb/Sa of 0.75 or more, MABS is equal to about 1.5Sa. The new rules offer the best frame of reference for evaluating the bolt-up requirements of Appendixes 2 and S of the ASME B&PV Code. Accordingly, when Am per Code is compared with Am per the new rules, it is observed that, for a range or gasket types, Appendix 2 designs are nominally equivalent to tightness class T1 of the new rules in many cases, particularly where the more traditional (less tight) gaskets are involved. (T1 constitutes a generous leak equivalent to 0.20 mg/sec/mm at the design pressure.) This observation is based on calculations where h = 1 with iteration, whereby Am is minimized. Now, if Am (NR)/Am (CR) is about equal to unity, then whatever must be done to bolting at assembly to make T1 achievable under the new rules must also

be done to Appendix 2 flange bolting. This means that to achieve a nominal tightness of T1 at the design pressure, the Appendix 2 bolting must be assembled to at least 1.5Sa when Sb/Sa = 1 or thereabouts. ASME B&PV Code Appendix S addresses the matter of the required assembly bolt stress almost exclusively from the standpoint of passing a hydrostatic test that is 1.5 or more times the design pressure. Appendix S explains why an initial bolt stress greater than 1.5 times the design value (i.e., greater than 1.5Sa) is needed and states in part, in any event, it is evident that an initial bolt stress higher than the design value may, and in some cases must, be developed in the tightening operation and it is the intent of this Division that such a practice is permissible provided it includes necessary and appropriate provisions to insure against excessive flange distortion and gross crushing of the gasket. Appendix S refers to assembly bolt stresses of 1.5Sa and greater from the standpoint of passing a pressure test. In the case of the new rules, a bolt stress equal to 1.5Sa at assembly is required for the case of Sb/Sa close to unity in order to develop the desired tightness at the design pressure irrespective of the pressure test. How can these two statements be reconciled? It would be reasonable to conclude that the good service history of Code flanges is attributable, to a large

Page 464

degree, to stressing the bolts at assembly to 1.5Sa and above to handle the hydrostatic pressure test. For example, the higher bolt load at assembly increases the gasket operating stress at the design pressure, and the higher gasket assembly stress reduces the gasket operating stress required for a given joint tightness at the design pressure. It can be shown that if bolts are assembled to Sa instead of 1.5Sa (and above) by controlled tightening and the system is then pressurized to the design pressure, then leakage may be unacceptable even if zero loss of bolt load is postulated. To illustrate consider the following joint: 14.75-in.-OD by N = 0.9375-in. SSA6 gasket Gb = 3400, a = 0.300, and Gs = 93 (from Table 2) h=1 Tightness class T1 Sa = Sb = 25,000 psi Tpmax = 500 (from Table 2) SL = 900 psi P = 325 psi. Using the new rules for this joint, Am (NR) = 6.203 sq. in., whereas per ASME B&PV Code Appendix 2 Am (CR) = 6.051 sq. in. Therefore, A/A = 6.203/6.051 = 1.025, or a nominal value of essentially unity. Based on the new rules, if the bolts are assembled to 1.5Sa and then unloaded to Sb, the joint tightness will be tightness class T1 at the design pressure with a mass leak rate of 0.20 mg/sec/mm.

Since A/A = 1.025, the same tightness should apply to the Code Appendix 2 flange bolting, On the other hand, suppose the designer of the Appendix 2 flanged assembly decided to tighten the bolts with a controlled method at assembly to Sa on the basis that (we suppose) no pressure test in excess of the design pressure was planned and, therefore the guidelines of Code Appendix S were not needed. On this basis, the estimated leak rate at the design pressure would be an unacceptable 2.9 mg/sec/mm even without unloading to Sb. This is 15 times the T1 leakage value of the new rules even when a zero loss of bolt load is postulated. There have been situations where well-intended designers have attempted to limit the bolt stress at assembly to Sa, perhaps because Sa is defined as the allowable design stress or perhaps because Appendix S refers to 1.5Sa and greater in terms of a pressure test. They then discover they have a leakage problem at the design pressure. In general, the extra bolt stress at assembly (i.e., 1.5Sa and greater) referred to in Appendix S is required to handle the extra force generated by the test pressure and to reduce the required gasket operating stress (Smo) for a given joint tightness at the design pressure. We are left with the following tantalizing questions. Was it by design or is it a fortuitive situation that Code flanges per Appendix 2 and the guidelines

Page 465

of Appendix S have a good service history? Is the good service history simply a result of having to tighten the bolts at assembly to pass a 1.5P pressure test? Was it known that assembling the bolts to 1.5Sa and above would decrease the required gasket operating stress? Was it surmised at the time that assembling the bolts to 1.5Sa and above might be essential in order to achieve a tight joint at the design pressure? The answer may be by design, since the logic may have been that if the flanged assembly passes a pressure test at 1.5P, then the assembly will be tight at the design pressure. Since the word tight in ASME B&PV Code Appendix 2 and Appendix S is without definition (contrary to the new rules), this seems to be one valid explanation. Further, this explanation accounts for some effects that tend to relax the joint, for example, flange rotation and interaction of the elements comprising the assembly. It does not account for thermal effects as a result of an elevated design temperature. In the foregoing, the new rules were used to demonstrate that the assembly bolt stresses of 1.5Sa and greater discussed in Code Appendix S is essential not only because of a pressure test but to improve leakage performance by reducing the gasket operating stress required for the design pressure. 7. Concerning Elevated Temperature 7.1. Effect of High Temperature on M/M At the time this chapter was being completed, the draft of the new rules defined the flange design bolt load as W = Ab × Sa; therefore, Ms = (Ab × Sa)hG and the load HG for the operating condition was (Ab × Sa) - H. As a result, when the simplifying assumption hD =

hT = hG is used, MS = MO, and multiplying Mo by Sfa/Sfo adjusts the operating moment to the room temperature allowable stress, and as a result it is seen that MO adjusted will always equal or be greater than MS. As a result, MS never governs a design and the preceding comparisons are exactly valid (reverse flanges excluded). Using (Ab × Sa) - H for the operating moment does not account for any reduction in the operating bolt load as a result of an elevated design temperature. Accordingly, when (Ab × Sa) - H is used as the operating load, which we designate HG (Op.), very high temperature designs (over 850°F for low-alloy steels) will be conservative to very conservative compared with Code Appendix 2, which uses HG (Op.) = Wm1 = 2b × 3.14(GmP) For the new rules, consideration is being given to changing HG (Op.) from (Ab × Sa) - H to (Ab × Sb) - H. Philosophically, the new rules would then parallel Code Appendix 2 insofar as HG for the operating condition is concerned.

Page 466

Revising HG(Op.) in the new rules would have an impact on the comparisons of this chapter and especially on the expressions for M/M in Section 5.4. In view of this, the authors suggest that the comparisons of this chapter be considered as applying to design temperatures for which Sb/Sa and Sfo/Sfa = 3/4 or more. In this range, the comparisons and examples are valid enough and quite useful whether HG (Op.) = (Ab × Sa) - H or HG (Op.) = (Ab × Sb) - H. However, designs become increasingly conservative as Sb/Sa or Sfo/Sfa falls below 3/4, i.e., as the design temperature increases. The reader is cautioned, however, that the expressions for M/M in this Section 5.4 in this chapter are exactly valid only for HG (Op.) = (Ab × Sa) H. 7.2. Gs and Tpmin Long-term effects of operating temperatures on the applicability of the constant Gs was briefly discussed in Section 4.7. 8. Closing Observations First, it appears possible to make consistent and meaningful broad comparisons of the new rules with ASME B&PV Code Appendix 2, as illustrated by the foregoing. Further, given a specific set of conditions, flange geometry, and parameters, it is possible to estimate the maximum difference in Am and M by means of the pressures P(A) and P(M). From the limited examples of this chapter it appears that many Code Appendix 2 designs with traditional gaskets fall between tightness classes T1 and T2 in terms of their sealing performance.

Appendix 2 designs using the more modern, very tight gaskets would appear to fall mainly between classes T2 and T3. It is also seen that the new rules permit a greater flexibility of design by recognizing a wider range of gaskets, several tightness classes, and a range of assembly efficiencies. Incidentally, several additional examples can be found in Appendices B and C of this chapter. Depending on how tight the gasket is, the new rules may result in heavier flanges. When this occurs for certain gaskets it does not mean that the new rules are more conservative than Code Appendix 2. The heavier flange is the result of designing for a specific tightness, whereas tightness is not a design parameter in Appendix 2. This chapter indicates trends, i.e., how certain parameters generally effect comparisons and differences between the new rules and ASME B&PV Code Appendix 2. They should be used, however, with a certain amount of caution since, in the larger picture, the many parameters are seen to interact in a complex way to affect the final results.

Page 467

Appendix A: Assembly Efficiency The degree of gasket stress uniformity around the circumference of a joint, relative to the mean stress, is referred to here as the assembly efficiency, h. As shown in the accompanying sketch, Sya is assumed to vary by ±(1 - h)Sya around a mean of Sya for an assembly efficiency of h. In this context, h decreases and the gasket assembly stress increases with increased assembly bolt load variation. That variation, in turn, is affected by, assembly technique, frictional effects, and elastic interaction during the joint assembly process. Elastic interaction depends, in turn, on the stiffness of the flange, bolt, and gasket system that comprises the joint. In fact, wide variations in bolt preload are known to exist in a joint following assembly, but less is known about the variation of stress at the gasket during and after assembly. An exact value of Sya for every situation is complex, if not impossible, to determine. It depends, among other things, on: The degree of stud preload control and related scatter during and after assembly The relative stiffness of the joint components Elastic interaction (cross talk) between flanges and gasket, which in turn is related to the relative deflection or rigidity of all the components of a joint The short-term inelastic response of the joint components Contributing to this is the observation that many types of gaskets exhibit inelastic-elastic load-unload behavior similar to that of mild

steel. Bickford [13] and Bibel [14] have reported Max/Min bolt preload variations of more than 50%, even with the careful application of accurate torque-based stud bolting procedures. The Min/Max variations are greater with less controlled or manual methods. Assembly efficiency as applied here is a variation of gasket stress about a mean value. This is not to be confused with significant total bolt load losses (as high as 50%) from the intended target that have been reported [11] even when using controlled bolt tightening methods. The introduction of h, the assembly efficiency, into the new rules allows the designer to take credit for the use of improved joint assembly techniques such as carefully applied torque, multiple hydraulic tensioners, and the use of stud stretch measurement. The concept recognizes that the least stressed points on the gasket during assembly will govern the design, because it is at these points that the minimum tightness condition (Tpmin) must be met. Hence, if h is low, the mean value of Sya must be increased so that the low points of the gasket are sufficiently loaded. Since seating and operating gasket stresses interact with each other, a more uniform Sya results in a theoretically lower required bolt load because a lower mean value of Sya suffices. Since the advantage of a lower Sm1 is credited to the use of higher Sya, it is fair that there should be means to distinguish and credit the use of the more

Page 468

reliable bolted joint assembly methods that actually achieve the more uniformly applied Sya. Without an assembly efficiency it would be necessary to assume a default degree of gasket load scatter that would likely assume the lowest common denominator of no bolt load control (as is the situation for ASME B&PV Code Appendix 2). Guidance on the selection of h is offered in Table 4. As compared to bolt preload variation, the amount of initial gasket variation is smoothed by gasket and joint flexibility. However, much more information is needed on how to estimate the extent of smoothing. Meanwhile the values for h suggested in Table 4 are recommended until more definitive values (hopefully supported by experiment) can be found. The values found in Table 4 are based on the following formula:

Page 469

Appendix B: A Graphical Solution (R. W. Schneider, August 1996) A graphical solution may be used to find the value of Sya corresponding to Sm1 = Sm2, or for finding Sm1 and Sm2 for a particular value of Sya. Once Sm1 and Sm2 are known, Smo may be calculated and the solution completed in accordance with the equations in this chapter. With a little practice (and care), you will find the method surprisingly fast and accurate. Since it is graphical, it shows the relationship of various terms and permits a designer to see the effects of varying certain terms. There are times when a designer will find a need for this approach. Although the graphical method outlined here covers the important steps, it will become apparent that variations are possible, and in certain circumstances steps may be omitted. In any case, after a solution has been arrived at, it should be examined to see that it satisfies all the constraints for the new rules: for example, 1.5 < X < Tpmax/Tpmin; Smo > 2P and SL. Finally, it is important to remember that for any value of Sya, the corresponding stress on the loading curve is Sg = Ae(Sya) and the tightness parameter Tp is given by the expression Ae(Sya) = Gb(Tp)a. 1. Use log-log paper (suitable scale) and a sharp pencil. The vertical axis is Sg, and the horizontal axis is Tp. 2. Draw the loading curve from Sg = Gb, Tp = 1, with slope equal to a. Better precision will result if the loading curve is drawn between two points, for example, [Sg = Gb, Tp = 1] and [Sg = Ae(Sya) = S, Tp = (S/Gb)(1/a)], where any value such as 15,000 or 30,000 may be used for S provided the two points are not too close to each other.

3. Calculate Tpmin, and draw a vertical line at Tpmin (unloading curves cross the Tpmin axis at values of Sm1). 4. Mark (Gs) at Sg = Gs, Tp = 1; all unloading curves terminate at this point. 5. Calculate (Sya)a and (Sya)b and use the larger as (Sya)1:

where Smo

> (2P or SL). Then the larger is

(Sya)1> [(Sya)a or (Sya)b] 6a. Draw the unloading curve from (Sya)1; it starts at Sg = Ae(Sya)1 on the loading curve and terminates at [Gs, TP = 1].

Page 470

6b. Calculate Sm2 based on (Sya)1 using

6c. Read (Sm1)1 as the value of Sg where the unloading curve of step 6a crosses the Tpmin axis. 6d. Draw a horizontal line on the Sg vs Tp plot at Sg = (Sm2)1; mark the intersection of this line and the unloading curve as point (1). 7a. Make (Sm2)2 = (Sm1)1 by drawing a horizontal line at Sg = (Sm2)2 = (Sm1)1, starting at the Tpmin axis. 7b. Calculate (Sya)2 using

7c. Draw an unloading curve from Sg = Ae[(Sya)2] on the loading curve to Gs; the value of Sg where the unloading curve crosses the Tpmin axis = (Sm1)2. 7d. Mark the intersection of the horizontal line at Sg(Sm2)2, with the unloading curve based on (Sya)2 as point (2). 8. Connect points (1) and (2) by a straight line; where it crosses the Tpmin axis is (Sm1)3 = (Sm2)3. The corresponding value of Ae(Sya)3 is found by drawing a line from GS through Sm1 = Sm2

on the Tpmin axis; the intersection of this line with the loading curve is Ae(Sya)3. If no constraints have been violated, (Sm1)3 = (Sm2)3 represents an iterative solution and

9. This step illustrates the proper way to handle a constraint. If Tp at (Sya)3 is greater than Tpmax, a constraint has been violated. In this case, (Sya)3 exceeds the stress and tightness for which the validity of the loading curve has been established by tests, where Su = Gb Tpmaxa. As a result, Sya must be set equal to Su. To account for this: Draw a new unloading curve from Sg = Ae(Sya) = Ae(Su) at Tpmax on the loading curve to Gs; it crosses the Tpmin axis at Sm1. Sm2 is found using

Page 471

and Smo > (Sm1, Sm2, 2P, or SL). The final (design) value of Wmo and Am are then based on this value of Smo, and the calculations are complete. Other constraints are handled similarly. For example, if a compression stop or a gasket with tightness hardening is employed, the maximum value of (Sya)3 of Step 9 is then based on Tpmax, which is the lesser of Tps and Tpu. Example Using the Graphical Method (Refer to Sheet SW1 in Appendix C) 1 and 2. The loading curve is drawn between the points [Sg = 3400 psi, Tp = 1] and [Sg = Ae(Sya) = 15,000, Tp = (15,000/3400)(1/0.30) = 140.8]. 3. Tpmin = 68.365. 4 and 5.

6a. Draw the unloading curve from (Sya)1; it starts at Sg = Ae(Sya)1 on the loading curve and terminates at [Gs, Tp = 1]. 6b.

6c. Locate (Sm1)1 = 9,000 psi at the intercept with the Tpmin axis. 6d. Draw the horizontal, and mark point (1). 7a. Draw the horizontal at (Sm2)2 = 9,000 psi = (Sm1)1. 7b.

Page 472

7c. Draw the unloading curve from Sg = Ae(Sya)2. 7d. Mark the intersection of the horizontal with the unloading line from (Sya)2 as point (2). 8. Connect points (1) and (2) with a straight line to find: (Sm1)3 =(Sm2)3 =6500 psi Ae(Sya)3 15,500 = psi 15,500 (Sya)3 = psi Assuming that no constraints have been violated: Smo > (6500, 2 × 550 = 1100 or SL = 900 psi) So Smo = 6500 psi, and

Note from the numerical solution that Am = 16.90 in.2. 9. Tp at (Sya)3 < Tpmax at Su; i.e., 15,500 < 17,160 psi. Therefore, no constraints have been violated, and Step 8 gives the correct values for Wmo and Am! Note that Step 5 ensures that Tpa/Tpmin > 1.5.

Page 473

Appendix C: Summarized Examples (see Figs. C1-C4)

Figure C1 Plot of gasket stress vs. the tightness parameter, used to illustrate the graphical method example of the new rules. This graphical result may be compared with the example of sheet SW1 in Fig. 5. Note: Only English units are used in these examples. The following table shows how to convert to metric units. Multiply By in. 26.5 mm lb 0.454 kgf psi 6.895 × 10-3 MPa in-lb 0.0115 m-kgf

Page 474

Figure C2

Page 475

Figure C3

Page 476

Figure C4

Example of Comparison Between New and Old Rules (Sheet SW1) Gasket:

SW CL600 ASB. FILLED (SSA6)

Tightness Class:

STANDARD T2

New Rules

P =550.00 psi Go =14.75 in. = gasket OD N =0.9375 in. = nom. width

Gb =3400 a = 0.3 = Gs =93

He =0 lb ext. load factor

Ae =

Sa =25,000 psi

Tc =1

Sb =21,000 psi

Tpmin 68.36 =

Sfa =17,500 psi

X=

Sfo =12,000 psi

Tr =

2.333 1.2006

n =0.6847 no or (N/2).5

Tpa = 160 =

G =14.065 = (Go - n)

Sya =

Sl =900 psi = 90% min. test stress

Sm1 =

15,571 6,619

Tpmax =221 > Tpa Sc =18,000 psi hd =

1.00 simplifying assumption when hg = hd = ht = 1.0

hg =1.00 ht =1.0 Wmo =354,726 lb Am (NR) =16.892 > Wmo/Sb M (cont/NR) 615,843 = (Ab * Sa) (Sfa/Sfo)hg = Mo = (NR) * Sfa/Sfo

Sm2 =

6,619

Sheet SW1 (Continued) Appendix 2 Rules b =0.3423 in. m =3 y =10,000 psi. Ab =

6.443 = actual bolt area or Am (CR) as a minimum

Wm2 =151,266 = 3.14 * b * G * y Ms =161,080 = hg * Sa(Ab + Am)/2 Hp =49,892 = 3.14 * 2b * G * m * P H =85,415 = 0.785 * G2 * P Wm1 =135,307 = H + Hp P(M) =

421.41 = 4(Sfo/Sfa) * by/(G + 8bm): P(Wm1 = Wm2)

Mo =135,307 = Wm1 * hg Am (CR) =6.443 > Wm2/Sa or Wm1/Sb (req'd area) P(A) =

516.23 = 4(Sb/Sa * by/(G + 8bm): P(Am1 = Am2)

M(adj) =197,323 = Mo * (Sfa/Sfo)

Sb/Sa Sfo/Sfa

M (Cont/CR) =197,323 = greater of Mo(adj) and Ms Conclusions A/A =2.622 = Am (NR)/Am (CR) M/M =3.121 = M (con/NR)/M (cont/CR)

Example of Comparison Between New and Old Rules (Sheet SW2) Gasket:

SW GRAPH FILL CL300+ (SSG6)

Tightness Class:

STANDARD T2

New Rules P =550.00 psi Go =14.75 in. = gasket OD N =0.9375 in. = nom. width

Gb =2300 a =0.237 Gs =13

He =0 lb ext. load factor

Ae =

Sa =25,000 psi

Tc =1

Sb21,000 psi

Tpmin 68.36 =

Sfa17,500 psi

X=

Sfo12,000 psi

Tr =

2.702 1.2353

n0.6847 no or (N/2).5

Tpa = 185 =

G14.065 = (Go - n)

Sya =

Sl900 psi = 90% min. test stress

Sm1 =

7,923 2,336 2,336

Tpmax4816 > Tpa Sc18,000 psi hd

1.00 simplifying assumption when hg = hd = ht = 1.0

hg1.00 ht1.0 Wmo180,489 lb Am (NR)8.595 > Wmo/Sb M (cont/NR)

313,349 = (Ab * Sa)(Sfa/Sfo)hg = Mo (NR) * Sfa/Sfo

Sm2 = Smo =

2,336

Page 480

Sheet SW2 (Continued) Appendix 2 Rules b = 0.3423in.

Sb/Sa 0.84 =

m=

Sfo/Sfa 0.69 =

3

y = 10,000psi. Ab =

= actual bolt area or Am (CR) as a minimum

6.443

Wm2 =151,266= 3.14 * b * G * y Ms =161,080= hg * Sa(Ab + Am)/2 Hp = 49,892= 3.14* 2b * G * m * P H = 85,415= 0.785 * G2 * P Wm1 =135,307= H + Hp = 4(Sfo/Sfa) * by/(G + 8bm): P(Wm1 = Wm2)

P(M) = 421.41

Mo =135,307= Wm1 * hg Am (CR) =

6.443> Wm2/Sa or Wm1/Sb (req'd area) = 4(Sb/Sa) * by/(G + 8bm): P(Am1 = Am2)

P(A) = 516.23

M(adj) =197,323= Mo * (Sfa/Sfo)

M(Cont/CR) =197,323greater of Mo(adj) and Ms Conclusions A/A = M/M =

1.334= Am (NR)/Am (CR) 1.588= M (cont/NR)/M (cont/CR)

Example of Comparison Between New and Old Rules [Refer to Example No. 1 Gasket:

COMP. ASB. SHT. 1/16

Tightness Class:

STANDARD T2

New Rules P =400.00 psi Go =34.5 in. = gasket OD N =0.75 in. = nom. width

Gb =2500 a =0.15 Gs =117

He =0 lb ext. load factor

Ae = 0.75 =

Sa =25,000 psi

Tc =1

Sb =25,000 psi

Tpmin 49.71 =

Sfa =17,500 psi

X=

Sfo =17,500 psi

Tr =

13.054 1.6577

n =0.6124 no or (N/2).5

Tpa = 649 =

G =33.888 = (Go - n)

Sya =

Sl =900 psi = 90% min. test stress

Sm1 =

Tpmax =30,000 > Tpa

Sm2 =

8,805 1,333 1,333

Tpmax =30,000 > Tpa Sc =18,000 psi 1.94 simplifying assumption when hg = hd hd = = ht = 1.0 hg =1.56 ht =2.03 Wmo =466,773 lb Am (NR) =18.671 > Wmo/Sb M (cont/NR)726,386 = (Ab * Sa)(Sfa/Sfo)hg = Mo (NR) =* Sfa/Sfo

Sm2 = Smo =

1,333

Page 482

Comparison Example No. 1 Case 1 (Continued) Appendix 2 Rules b = 0.3062 m= y=

2.75 3,700psi. = actual bolt area or Am (CR) as a minimum

Ab = 18.671

Wm2 =120,609= 3.14 * b *G * y Ms =699,536= hg * Sa(Ab + Am)/2 Hp = 71,677= 3.14 * 2b * G * m * P H =360,859= 0.785 * G2 * P Wm1 =432,265= H + Hp = 4(Sfo/Sfa) * by/G + 8bm): P(Wm1 = Wm2)

P(M) = 111.55

Mo =672,686= Wm1 * hg > Wm2/Sa or Wm1/Sb (req'd area)

Am (CR) = 17.291 Sb/Sa =

1

Sfo/Sfa =

1.00 = 4(Sb/Sa) * by/(G + 8bm):

P(A) = 111.55

P(Am1 = Am2) M(adj) =672,686= Mo * (Sfa/Sfo) M 699,536greater of Mo(adj) and Ms (Cont/CR) = Conclusions A/A = M/M =

1.080= Am (NR)/Am (CR) 1.038= M (cont/NR)/M (cont/CR)

Example of Comparison Between New and Old Rules [Refer to Example No. 1 Gasket:

COMP. ASB. SHT. 1/16

Tightness Class:

TIGHT T3

New Rules P =400.00 psi Go =34.5 in. = gasket OD N =0.75 in. = nom. width

Gb = 2500 gasket con a=

0.15

Gs = 117

He =0 lb ext. load factor

Ae =

Sa =25,000 psi

Tc = 10

Sb =25,000 psi

Tpmin 497.13 0.1243( = =

Sfa =17,500 psi

X=

23.585 X = Tpa/Tp =

Sfo =17,500 psi

Tr =

1.5090 log =

n =0.6124 no or (N/2).5

Tpa =

11,725 XTp =

G =33.888 = (Go - n)

Sya =

10,193 Gb/Ae(Tpa =

Sl =900 psi = 90% min. test stress

Sm1 =

2,259 Gs[Gb/Gs( =

1 =assembly e

tightness c

2,259[

Tpmax =30,000 > Tpa Sc =18,000 psi 1.94 simplifying assumption when hg = hd = hd = ht = 1.0 hg =1.56 ht =2.03 Wmo =540,380 lb Am (NR) =21.615 > Wmo/Sb M840.932 = (Ab * Sa)(Sfa/Sfo)hg = Mo (cont/NR) =(NR) * Sfa/Sfo

Sm2 =

2,259[ =He)/(Ag +

Smo =

2,259 > >

Page 484

Comparison Example 1, Case 5 (Continued) Appendix 2 Rules b = 0.3962 m= y=

2.75 3,700psi. = actual bolt area or Am (CR) as a minimum

Ab = 21.615

Wm =120,699= 3.14 * b* G * y Ms =756,809= hg * Sa(Ab + Am)/2 Hp = 71,677= 3.14 * 2b * G * m * P H =360,589= 0.785 * G2 * P Wm1 =432,265= H + Hp = 4(Sfo/Sfa) * by/(G + 8 bm): P(Wm = Wm2)

P(M) = 111.55

Mo =672,686= Wm1 * hg Am (CR) = 17.291> Wm2/Sa or Wm1/Sb (req'd area) Sb/Sa =

1

Sfo/Sfa =

1.00 = 4 (Sb/Sa) * by/(G + 8bm): P(Am1 = Am2)

P(A) = 111.55

M(adj) =672,686= Mo * (Sfa/Sfo) M(Cont/CR) 756,809greater of Mo(adj) and Ms = Conclusions A/A = M/M =

1.250= Am (NR)/Am (CR) 1.111= M (Cont/NR)/M (cont/CR)

Page 485

References 1. ASME Boiler and Pressure Vessel Code, Section VIII. Div. 1, American Society of Mechanical Engineers, New York, 1995. 2. Blach, A. and E., Bazergui, A. B. Methods of analysis of bolted jointsA review. Welding Research Council Bulletin 271 (October 1981). 3. Rossheim, D. B., and Markl, A. R. C. Gasket loading constants. Mechanical Engineering 65: 647 (1943). 4. Raut, H. D., and Leon, G. F. Report of gasket factor tests. Welding Research Council Bulletin 233 (Dec. 1977). 5. Schneider, R. W., and Rodabaugh, E. C. Article 5.2, flanges, gaskets and closure systems. In: Pressure Vessels and Piping: Design TechnologyA Decade of Progress. ASME PVP, 1982, pp 347357. 6. Design Division Problem No. XIII, Re-evaluation of gasket factors used in flange design. WRC Bulletin 298 (Sept. 1984). 7. Bazergui, A., and Marchand, L. PVRC milestone gasket testsFirst results. Welding Research Council Bulletin No. 292 (Feb. 1984). 8. Bazergui, A. B., and Marchand, L. Development of a production test procedure for gasket. Welding Research Council Bulletin 309 (Nov. 1985). 9. Payne, J. R., Bazergui, A., and Leon, G. New gasket factorsA proposed procedure. Proceedings of 1985 Pressure Vessels and Piping Conference. ASME/PVP, PVP Vol. 98.2, 1985.

10. Payne, J. R., Bazergui A., and Leon G. F. Getting new gasket design constants from gasket tightness data. Experimental Techniques, Special Supplement, (Nov. 1988): 2227. 11. Winter, J. R. Gasket selectionA flowchart approach. Presented at the 2nd Intl. Symp. on Fluid Sealing of Static Gasketed Joints, La Baule, France, Sept. 1820, 1990. 12. Modern Flange Design. Bulletin 502, Edition VII, G+W Taylor-Bonney Division, 1978. 13. Bickford, J.H., Hayashi, K., Chang, A. T., and Winter, J.R.A preliminary evaluation of the elevated temperature behavior of a bolted flanged connection. Welding Research Council Bulletin 341 (Feb. 1989). 14. Bible, G. Bolted flange assembly: Preliminary elastic interaction data and improved bolt-up procedures. Welding Research Council Bulletin 408 (Jan. 1996).

Page 487

10 Bolted Flanged Connections for Noncircular Pressure Vessels A. E. BLACH Consulting Engineer, Montreal, Quebec, Canada 1. Introduction Modern chemical and process industries would not exist in their current form if the bolted flanged connection were not available as a convenient means of joining together the various pieces of equipment that make up such a plant. Flanges constitute a very important part of all pressure-containing apparatus. Due to the favorable circular geometry employed in most pressure vessels or piping systems, the majority of the flanges used are circular. This includes a variety of pressure classes and facing types. There are, however, many applications in which circular pressure containers or conduits cannot be used for various reasons. For example, the inlet nozzles on cyclones are usually of rectangular shape, as are the wood-chip chutes on pulp digesters; the headers on steam boilers are normally square; and the inlet and outlet headers of air-cooled heat exchangers are rectangular. For such noncir-

Page 488

cular cross sections, flanges are often required, be it for access or for connection to other equipment. Bolted flanged connections, being part of pressure-containing systems, are governed by rules in accordance with the various pressure vessel codes that exist in all industrialized countries, for example, the ASME Boiler and Pressure Vessel Code [1]. (see also Chapter 9.) The ASME Code, Section VIII, Division 1, contains rules for the design of pressure vessels and pressure vessel components, including rules for noncircular pressure vessels in Appendix 13, covering side plates, reinforcing ribs, and end plates of such vessels, but no rules for noncircular bolted flanged connections. Two literature searches of publications on bolted flanged connections one by Blach and Bazergui [2], one by Cassidy and Kim [3], have shown that little has been published on this subject. The only type of noncircular bolted flanged connection that has received some attention is the flanged joint in split steam-turbine housings. This type of bolted connection, important for turbine housings, is, however, not representative of the majority of noncircular flanges used. In the following, two approximate design methods are discussed that are presently used by pressure vessel designers, including their limitations. Both methods are compared for accuracy with numerical results from finite element calculations and with experimental data obtained from strain gage measurements of rectangular test pressure vessels. Numerical sample calculations for two approximate methods are included. Results are discussed and suggestions are made as to the accuracy and limitations of both.

2. Approximate Design Methods There are several approximate design methods that can be used in the design of noncircular bolted flanged connections. Two of these methods, sometimes used by pressure vessel designers, are discussed next. One, based on an equivalent circular flange, may be used for square or nearly square flanges [4]. The other uses a combination of frame analysis for the ability of the flange to retain its rectangular shape and of bending of an infinitely long flanged section in a plane perpendicular to the frame [5]. 2.1. Equivalent Circular Flange Method This method is similar in approach to the procedure used in the ASME Code, Section VIII, Division 1, Article UG-34 [1], for the design of noncircular flat covers. In the Code, a factor Z is defined that relates a flat cover of rectangular shape to a circular one. The factor Z is given by

Page 489

(1) with the limitation that Z does not need to be larger than 2.5. The square root of this factor is used as a multiplier of the small side of the rectangular cover to obtain an equivalent diameter to be used in the thickness formula given in the Code. The term added in the Code to account for a bolt edge moment is considered to be included in the flange bending moment subsequently obtained. For a welded flat cover, this approach yields a cover thickness on the safe side, as can easily be seen by comparing square and rectangular cover stress factors with equivalent circular cover factors. In the case of a rectangular cover with little rotational restraint at the rim (thin shell, thick cover), from Roark and Young [6], the factor b in (2) is 0.287 for a square plate and approaches 0.750 for a long rectangular plate with alb > 4. Using the Code formula for rectangular plates, a comparison is made with the factors given in [6]. Equating the stresses from both [1] and [6], (3) it can be seen that the factors b and cZ should have the same values. In fact, using the Code constant c = 0.33 and the appropriate values for Z, the product cZ is always numerically larger than b. Hence, the Code formulas yield results on the safe side.

In this design method, the square root of Z is used as a multiplier for the small side of the rectangular pressure vessel, and thus an equivalent circular shape is obtained, similar to the Code rule of UG-34. Any obround or rectangular flange can then be designed or analyzed as an equivalent circular flange, and all flange design Code rules per Appendix 2 or Appendix Y of the ASME Code [1] are applicable without modification. In the case of full-face gaskets, often used with rectangular flanges (see Fig. 1), no ASME Code rules exist at present. References [7] and [8] describe flange design methods for full-face gasketed bolted flanged connections. An example of the application of this method is shown in Fig. 2, where the equivalent circular flange is superimposed over a rectangular flange of 200 × 300 mm (8 × 12). In this case, the equivalent inside diameter is calculated as follows:

Page 490

Figure 1 Gasket types: (a) Ring gasket (ASME Code, App. 2); (b) O-ring gasket (ASME Code, App. Y); (c) full-face gasket (no ASME Code rules).

Figure 2 Equivalent circular flange.

Page 491

Equivalent bolt circle and outside diameters are then obtained by adding to the equivalent inside diameter twice the distance from the inside edge of the rectangular flange to the line of bolting and to the outside edge, respectively. The equivalent circular flange method has recently been verified. Preliminary test results reported in [9] and [10] indicate that this method yields safely designed rectangular flanges if the length-towidth ratio is close to unity, that is, for almost square rectangular geometries. For long rectangular flanges with length-to-width ratios of over 2, however, there is no longer any resemblance in the behavior of the rectangular flange when compared with a circular flange, and a length-to-width ratio limit of 1.5 is usually applied. The equivalent circular flange method can be used only for flanges of unreinforced pressure vessels. For rib-reinforced rectangular vessels, this method alone cannot account for heavy frame bending stresses, which must also be accommodated by the flanges; hence it is not recommended. A flange designed according to the equivalent circular method must also be checked for adequacy in supporting the frame bending stresses resulting from the thin vessel wall. For unreinforced vessels, frame bending stresses are attenuated in the vessel walls and do not greatly influence the flange stresses. 2.2. Frame Bending Flange Design Method The equivalent circular flange method discussed in the previous section, in addition to length-to-width ratio limitations, is applicable only for unreinforced noncircular pressure vessels, where the frame bending stresses are fully absorbed by the pressure vessel side plates.

A large percentage of noncircular pressure vessels, however, is of the reinforced type, as shown in Fig. 3. In this case, the flange must also act as a stiffener for the vessel side plates, in addition to providing a tight seal between components. Thus, such flanges have to resist frame bending moments, which occur when a rectangular frame is subjected to internal pressure. These moments cause deflections in a plane perpendicular to the vessel axis. In addition, rectangular flanges also have to resist flange bending in planes parallel to the axis of the vessel, caused when a flange is bolted-up about the gasket as fulcrum or when internal pressure effects tend to open up the bolted connection. The two stresses caused by bending in two different planes, of course, are the result of biaxial bending and thus should not really be separated; however, for simplicity these two stresses may be added, in order to compute a safe flange thickness. Preliminary test results reported in [9] show that this procedure yields flange designs on the conservative side. 2.3. Frame Bending Stresses For rectangular flanges of uniform thickness, the frame bending stresses can be found from frame-type structural analysis. The cross section assumed to resist

Page 492

Figure 3 Reinforced noncircular pressure vessel.

bending is usually taken as the flange or reinforcing rib area, plus a length of 16t of the connecting shell (similar to Appendix 13 of the Code). For rectangular frames of uniform cross section (flanges of uniform width), the corner moments are given by (4) and the moments at the center of the long span are given by (5) where the length dimensions l1 and l2 are taken between the centroids of the flange sections, or of the flange shell junction if part of the connecting shell is included in the calculations (see Fig. 4). For flanges of unequal width, a stiffness factor must be included in Eq. (4). The frame load w is the load per unit length of the

frame: w = lp. For an unreinforced noncircular vessel, w = p, the internal pressure; in this case, l = 1. 2.4. Flange Bending Stresses The flange bending stresses due to bolt-up and operating pressure for a long rectangular flange (a/b > 2) can be approximated by considering a unit width

Page 493

Figure 4 Frame bending stresses.

of the flange at the center of the longer side. The pressure distribution in such a flange is not uniform, but the maximum is known to occur at this point. For this purpose it is convenient to introduce a factor b that can be taken from the stress distribution of a rectangular plate with fixed ends, listed in [6] and other texts on structural analysis. This factor b varies numerically from 0.308 for a/b = 1, to 0.500 for a/b > 2. Rectangular flanges with large a/b ratios are not normally used with strip gaskets, due to the inability of a long flange strip to resist rotation. Flanges with small a/b ratios are sometimes used with

strip gaskets, gaskets that are fully inside the flange bolts, as shown in Fig. 5. For such flanges, using the nomenclature similar to that of the ASME Code, Appendix 2 [1], the flange bending moment per unit length is given by Mo = HDhD + HGhG + HThT (6) Mo = bbGPhD + 2bGmPhG + (g - bG)PhT

Page 494

Figure 5 Flange bending with strip gasket.

Figure 6 Flange bending with full face gasket.

For flanges used with full-face gaskets, as in Fig. 6, the gasket compression on the outside of the flanges provides some resistance against rotation. This resistance has been included in the flange bending moment, using the flange design method described in [8]. From Fig. 6:

(7)

Page 495

(8) The moment in Eqs. (6) and (8) must be resisted by a unit width of the flange, assuming that the connecting strip of shell plate is in direct tension only. This assumption is based on the fact that the shell plate attached to a rectangular flange is usually much thinner and thus more flexible than the flange. The same reasoning is made by the ASME Code, Appendix 2 [1], in the case of the loose optional flange, where the contributions of the connecting shell in resisting bending are neglected. The section modulus of a rectangular section of unit width is (9) Thus the stress in the flange, which corresponds to the radial stress in a circular flange, can be calculated as follows: (10) Stresses due to frame bending are combined with stresses due to flange bending. At the inside of the flange, at the center of the long side, frame bending stresses are compressive, as can be seen in Fig. 4, and should thus be added to flange bending stresses occurring at this location for flanges with strip gaskets. For flanges with fullface gaskets, flange bending stresses should be added to frame bending stresses at the inside and at the outside of the flange.

It could be argued that flange and frame bending stresses occur in different planes and need not be combined. However, it is suggested that these stresses be added, for a conservative design, resulting in a thicker flange that will distort less under load and provide easy sealing of the joint. For flanges of nonreinforced rectangular pressure vessels, the connecting shell plate is usually of considerable thickness. In this case the frame bending stresses of the flange are very small and need not be included in the computations. 3. Numerical Examples Several numerical examples are given: a rectangular flange with strip gasket on an unreinforced pressure vessel, using the equivalent circular flange method; a rectangular flange with fullface gasket on a rib-reinforced pressure vessel, plus

Page 496

an obround flange with full-face gasket, both using the frame bending flange design method. 3.1. Example 1 A rectangular pressure vessel is required with flanged covers at both ends for a design pressure of 1000 kPa (145 psi). Dimensions are as shown in Fig. 7. The

Figure 7 Rectangular pressure vessel.

Page 497

material of construction is carbon steel SA-516-70, for which an allowable stress in tension of 120 MPa (17,500 psi) is given in the Code. A strip gasket of compressed asbestos (or other mineral fiber) with gasket factors m = 2 and y = 11 MPa (1600 psi) is chosen, as is a bolt material of SA-193-B7 with an allowable stress of 172 MPa (25,000 psi). As a first step, the equivalent circular flange dimensions are calculated:

Using this equivalent inside diameter, the equivalent bolt circle and outside diameters and the equivalent gasket diameters are obtained, as shown in Fig. 8. From here on, the flange can be analyzed like a round flange, using the method of the ASME Code, Section VIII, Division 1, Appendix 2. These calculations may be performed longhand, as demonstrated in the following, or by using a form sheet as given in the Taylor-Forge Booklet Modern Flange Design [11] or on a spreadsheet program on a personal computer.

Figure 8 Equivalent flange dimensions.

Page 498

Effective Gasket Diameter

G = DG - 2b = 318 - (2)(8.9) = 300.2 mm Flange Loads

HT = H - HD = 70,771 - 56,410 = 14,361 N HG = HP = 2bpGmP = (2)(8.9)p(300.2)(2.0)(1.0) = 33608 N Wm1 = H + HP = 70,771 + 33,608 = 104,380 N Wm2 = bpGy = (8.9)p(300.2)(11) = 92,423 N Bolt Area, 14 Bolts 19-mm Diameter (314-10 UNC), a = 215 mm2

Ab = na - (14)(215) = 3010 mm2 Moment Arms

Flange Moment, Operating MD = HDhD = (56,410)(31.0) = 1,748,724 N·mm MG = HGhG = (33,608)(21.9) = 736,246 N·mm

Page 499

MT = HThT = (14,361)(30.0) = 430,171 N·mm Mo = MD + MG + MT = 219,240 N·mm Flange Moment, Gasket Seating

MGS = WGShG = (311,050)(21.9) = 6,814,960 N·mm Flange Parameters

U = 5.653 T = 1.720 Y = 5.144 Z = 2.690

F = .909 V = .550 f = 1.00

Flange Stresses, Operating

Page 500

Flange Stresses, Gasket Seating

As can be seen from the foregoing, all flange stresses are within the allowable limits: 1.5S for hub stresses and S for radial and tangential stresses, and also for the combinations of hub stress with radial and tangential stress. The governing condition in this case is the gasket seating; for the internal pressure only, stresses are much lower. A change to a softer gasket, such as synthetic rubber, would reduce the flange thickness. The bolt area seems to be too large; however, for a rectangular flange, more bolting is generally needed to keep the joint tight. In addition to the flange, of course, the vessel wall thickness must also be verified. The procedure of Appendix 13 of the ASME Code gives mandatory design formulas for all types of noncircular pressure vessels. For a first check, a simple formula, based on the theorem of three moments, may also be used. Since it is known that the maximum moment occurs at the corners, this corner moment,

for a uniform wall thickness of the four sides, is easily obtained:

Page 501

The maximum bending stress in the vessel is thus less than the Code limit of 1.5S, for combined membrane and bending stresses. 3.2. Example 2 A pressure vessel of the same size as the one used in Example 1 is required in stainless steel, type 316. Due to the high cost of stainless steel, a rib-reinforced design with a minimum shell thickness is suggested. Thus the weight of the stainless steel can be minimized if carbon steel stiffeners, welded to the outside of the vessel, are acceptable. The acceptable material listed in the Code is SA-240316, with an allowable stress at 150°C (300°F) of 127 Mpa (18,400 psi). To reduce the flange thickness, a full-face gasket design is chosen. As a first step, the stiffener spacing must be obtained. This spacing is related directly to the plate thickness used. Assuming a 6.3-mm ( in.) plate, which is less than one-half of the thickness of the unreinforced vessel, the maximum spacing is obtained:

For the vessel length of 800 mm (31.5 in.), six ribs are chosen with a spacing of 114 mm (4.5 in.). A layout of this reinforced vessel is shown in Fig. 9. The required stiffener size can be obtained by using the three-moment equation for a closed frame: w = pl = (1.0)(114) = 114 N/mm For a stiffener size of 12.7 × 44 ( × in.), section modulus and stresses are calculated. These stiffeners may be carbon steel for

most applications.

For a full-face gasketed flange, a thickness of 16 mm ( in.) is suggested. As a first step, the frame bending stresses in the flange are computed. For this, the same formulas as for the stiffeners are used.

Page 502

Figure 9 Rib-reinforced pressure vessel.

Page 503

Flange Bending Stresses

The combined stresses of frame and flange bending are lower than the flange stresses in the previous example, which usually gives results on the safe side. Results of this method were compared with experimental data from strain gage measurements on test pressure vessels, and they compared favorably. Nomenclature a long side of rectangular vessel or plate b short side of rectangular vessel or plate bDeffective width of gasket c bolt line distance

C Code flat head factor d

bolt hole diameter; Code short side of rectangular plate

D Code long side of rectangular plate E modulus of elasticity

Page 504

EG gasket modulus g effective gasket distance hD flange moment arm for HD hG flange moment arm for HG hT flange moment arm for HT HD

hydrostatic end force inside of flange

HG gasket load HT

hydrostatic end force under gasket

l

length

l1 frame long side l2 frame short side m Code gasket factor MA frame corner moment MB frame center moment MOflange operating moment p

pressure, unit load/unit width

P Code pressure t

thickness

tG gasket thickness w frame load, load/unit width W bolt force Z

Code rectangular plate factor

a full-face gasket factor b rectangular plate factor l reinforcing rib spacing s stress References 1. ASME Boiler and Pressure Vessel Code, Section VIII, Division 1, Pressure Vessels. American Society of Mechanical Engineers, New York, 1989. 2. Blach, A. E., and Bazergui, A. Methods of analysis of bolted flanged connectionsA review. WRC Bulletin 271 (Oct 1981): 115. 3. Cassidy, L. M., and Kim, T. J. Literature search and interpretive study on the design of bolted flanges with external loads and noncircular flanges. Unpublished report, PVRC, Sept. 1979. 4. Blach, A. E. Equivalent circular flange method for rectangular pressure vessel flanges. Unpublished lecture notes, 1977. 5. Blach, A. E. Rectangular pressure vessel flanges. Unpublished lecture notes, 1978. 6. Roark, R. J., and Young, W. C. Formulas for Stress and Strain. 5th ed. New York. McGraw-Hill, 1975.

7. Anonymous. Design of Flanges for Full-Face Gaskets. Taylor Forge Inc., Engineering Department, Bulletin 45, Chicago, 1951.

Page 505

8. Blach, A. E., Bazergui, A., and Baldur, R. Bolted flanged connection with full-face gaskets. WRC Bulletin 314 (May 1986). 9. Blach, A. E. Bolted flanged connections for non-circular pressure vessels. Proceedings of the Sixth International Conference on Pressure Vessel Technology, Beijing, China, September 1988, pp.267280. 10. Blach, A. E. Bolted flanged connections for non-circular pressure vessels. Proceedings of the ASME-JSME Pressure Vessel and Piping Conference, Hawaii, July 1989, Vol. 158, pp. 97104. 11. Modern Flange Design. Taylor-Bonney Division of Gulf and Western, Southfield, MI, 1979.

Page 507

11 Bolted Flanged Connections with Full-Face Gaskets A. E. BLACH Consulting Engineer, Montreal, Quebec, Canada 1. Introduction In the early years of the process industries, bolted flanges were usually made of cast iron. To avoid any undue stress on this material, full-face gaskets were always used. No flange design methods as such were in existence, and often flanges broken off the attaching vessel were reported. Even the addition of gussets did not eliminate these problems. When cast-iron process vessels were replaced by vessels made of wrought-iron rolled plates with riveted joints, flanges were then made of rolled angle sections, riveted to the vessel shell. Full-face gaskets were still used in order to accommodate the unevenness of rolled angle flange faces. When pressures to be sealed rose to higher levels, and welding was used as the joining technique in constructing pressure vessels, the forged-steel flange became popular. To reduce the bolt area required to compress a gasket, its area was reduced and the ring gasket, lying fully within the bolt circle of a bolted

Page 508

flanged connection, became the industry standard (see Fig. 1). The full-face gasket did not disappear, however. It has been used for a number of low-pressure applications, also for noncircular flanges, and certainly for flanges of cast-iron valves. For such cases, the accepted practice in industry has been to remove the raised face from a standard forged-steel flange and use this modified flange with a full-face gasket as mating flange on cast-iron valves. Design rules for bolted flanged connections given in the ASME Boiler and Pressure Vessel Code started as a cooperation between industry and research. Waters and Taylor published their first report on the strength of commercially available flanges in 1927 [1]. This was followed by the extensive investigation into the strength of flanges with ring gaskets by Waters et al. [2]. This work, published in 1937, formed the basis for the rules for bolted flanged connections found in the ASME Code, Section VIII, Division 1, Pressure Vessels [3]. The same rules are used in other sections of the ASME Code, in the ASME/ANSI B31.3 Code for Chemical Plant and Petroleum Refinery Piping [4], and in many national pressure-vessel codes of industrialized countries. In addition to design rules for flanges with ring gaskets, given in Appendix 2 of the ASME Code, Section VIII, Division 1, there are also rules for flanges with metal-to-metal contact outside the bolt circle, using self-energizing gaskets such as O-rings (see Fig. 1). These rules are based on the work of Schneider [5] and can be found in Appendix Y of the same Code. Flanges with gaskets covering the full face of the flange, so called full-face gaskets or flat-face gaskets are not covered by Code rules at present. Several design methods have been published and are

used by designers. One method, based on the work of Lonngren [6], was adapted into a systematic computation procedure by the Taylor Forge Company in 1956 [7]. This so-called

Figure 1 Gasket types: (a) Ring gasket (ASME Code, App. 2); (b) O-ring gasket (ASME Code, App. Y); (c) full-face gasket (no ASME Code rules).

Page 509

Taylor-Forge Method is now included in a booklet, Modern Flange Design, published by Gulf & Western, Energy Products Group [8]. It is also recommended in the ASME Code, Section X, FiberReinforced Plastic Pressure Vessels [9], to be used in the analysis of nonstandard FRP flanges. Another method, published by Schwaigerer in 1961 [10], forms the basis of design rules for fullface gasketed flanged connections contained in the German Standard DIN-2505 [11]. A third method was proposed by Blach et al. in 1986 [12]. All of these methods are based on an elastic analysis of the interaction of flange bending and gasket compression. The TaylorForge method uses two equivalent ring gaskets instead of the fullface gasket, one inside and one outside the bolt circle. The Schwaigerer and Blach methods are based on the uneven gasket compression to resist bending of the flange. In comparing test data, obtained from strain gage measurements on test pressure vessels, with calculated values from various design methods, it was observed that the Taylor-Forge method consistently produced much higher stresses than actually are present. The reason for this phenomenon lies in the fact that the loadings for the equivalent inner and outer gaskets are based on their respective widths. Since the bolt circle is usually closer to the outside than to the inside of the flange, a lower compression value is obtained on the outside gasket than on the inside gasket, obviously just the opposite of the actual gasket compression, which is highest on the outside of the flange! This, in turn, produces a higher bending moment in the flange, since the resistance of the gasket against flange rotation is neglected. The Taylor-Forge

method is thus very much on the safe side and predicts stresses much higher than actually are present. The design methods by Schwaigerer and Blach, on the other hand, have proven to be very sensitive in the case of relatively thin flanges, when the flange face after rotation does not remain straight. To alleviate this problem, in a paper in preparation [13], the triangular gasket force distribution used in [10] and [12] was replaced by a parabolic distribution, simulating the deformation of the flange face when subjected to a radial bending moment. This proposed new design method employs nondimensional flange geometry and stress parameters that can be plotted in a design curve. This curve may be used to obtain the maximum tangential stress in the flange, given flange geometry and operating pressure, or, alternatively, to obtain a flange thickness for a flange of given diameter, operating pressure, and allowable stress. Today, flanges with full-face gaskets instead of ring gaskets are often used in order to reduce or minimize flange bending in brittle materials. Examples are flanges in cast iron, glass or porcelain, or other ceramic materials. Due to the large gasket area and the associated heavy bolting required to compress gaskets, full-face gaskets are limited to low-pressure applications, using relatively soft gaskets such as rubber or asbestos compositions.

Page 510

It might be of value to point out that flat-face flanges used with ring gaskets without contact outside the bolt circle behave in the same manner as raised-face flanges and may thus be designed or analyzed in accordance with the rules for flanges with ring gaskets. 2. The Taylor-Forge Method This design method was never published but was disseminated as an Engineering Department Bulletin by the Taylor-Forge and Pipe Works in 1951, in Chicago [7]. The method was subsequently included in a booklet, Modern Flange Design [8], by the same company. No authors are given; however, it appears that it was suggested by D. B. Wesstrom, one of the original co-authors of the Waters method [2]. It seemed to have been an afterthought to the Appendix 2 design method and probably resulted from the early cooperation between E. O. Waters and J. H. Taylor [1]. The analysis is based on the decomposition of the full-face gasket into two separate ring gaskets, one lying inside and one outside the bolt circle. The gasket force is then distributed proportionally to the widths of the two hypothetical gaskets: the higher load to the inner (wider) gasket, the lower load to the outer (narrower) gasket. This assumption, while giving numbers to be used in subsequent ASME Code calculations, is flawed: When the flange faces are rotated, the outer part of the gasket is under higher compressive load than the inner part, which may be fully unloaded before a bolted flanged connection starts to leak. From strain gage measurements on test pressure vessels and from a multitude of successfully operating flanged connections in industry, often using less than one-half of the thickness calculated

using this method, it can be concluded that this design method is very much on the safe side. 2.1 Numerical Example A full-face gasketed flange is required for a vessel of 24-in. inside diameter with a in. wall thickness, for a design pressure of 120 psig (827 kPa) and a temperature of 300°F (149°C). The material of construction is carbon steel SA-516-70 with an allowable stress of 17,500 psi (120 Mpa). It is desired to use a flange cut from plate and attached to the shell by groove-and-fillet welds to satisfy the rules of the ASME Code, Section VIII, Division 1. A synthetic rubber gasket of in. thickness is specified, as are bolts of material SA-193-B7. For a relatively thin shell, a ring flange is advantageously designed as an optional loose flange, per Appendix 24(c). Assuming bolts of in. diameter, a layout is made to accommodate the bolts in such a way that the flange outside diameter is kept to a minimum. A preliminary flange size is shown in Fig. 2.

Page 511

Figure 2 Flange layout.

A in.-thick synthetic rubber gasket is chosen, m = 1.0, y = 200 psi, and moment arms are calculated:

Page 512

Gasket reactions are computed next:

HGY = bpGY = (.750)p(26.028)(200) = 12,266 lb HGP = 2bpGmP = (2)(.750)p(26.028)(1)(120) = 14,719 lb

Then bolt loads are obtained:

HT = H - HD = 63,853 - 56,572 = 7,281 lb

Ab = (20)(.462) = 9.240 in.2 > 3.999 in.2 Wa = AbSa = (9.240)(25.000) = 231,000 lb

Finally, flange moments and stresses are calculated: Mo = HDhD + HThT = (56,572)(1.50) + (7,281)(1.118) = 92998 in.

= lb

Page 513

In the Taylor-Forge Method, a check of the radial bending stress at the bolt centerline is also recommended:

From the preceding it can be seen that a flange thickness of slightly more than in. is required. Comparing this thickness with a flange of the same size, using a ring gasket of in. outside and 24-in. inside diameter, for which a minimum thickness of 1.625 in. would be required, also using an optional loose designation per Appendix 2 of the ASME Code, it appears that no great saving in the flange weight can be achieved using the Taylor-Forge design method. 3. The Blach-Bazergui-Baldur Method This design method is based on the resistance that an elastic gasket provides against rotation of the flange faces under pressure. The method was published in 1986 for similar pairs of flanges [12], and in 1993 for dissimilar pairs such as reducing flanges and blind flanges [14]. As in the design method for flanges with ring gaskets, rotations and displacements at the junction of flange ring and connecting hub are equated and discontinuity forces and moments are obtained. The main difference between the full-face gasket and the ring gasket is the absence of the large meridional bending moment found in the ring gasket. This fact usually leads to thinner, more economical flanges. A disadvantage, however, is the large gasket area required for a full-face gasket. To obtain the necessary gasket compression force for a large area, only relatively soft gaskets can be used within the customary available bolting

arrangements. The original method [12] uses fifteen dimensionless parameters that aid in the calculation of the operating and gasket seating moments used to calculate stresses in flange and connecting hub. It was shown in [12] that certain simplifications do not drastically reduce the accuracy of the method; hence, the original fifteen parameters could be reduced to one if the following assumptions are made: The radial displacement of the flange ring at the junction and the effects of internal pressure on the flange are neglected; also, a uniform pressure distribution is assumed for gasket seating. All three are reasonable assumptions, the first one because the flange stiffness is much greater than the stiffness of the hub, the second one (also used in the analysis of flanges with ring gaskets in the ASME Code) for the same reason, the third one depending only upon the flange geometry and the location of the bolt circle.

Page 514

3.1 Numerical Example A flange with full-face gasket is required for the operating conditions used in the example demonstrating the Taylor-Forge method. All dimensions shown in Fig. 2 are the same, except that B is taken at the flange inside diameter and, in addition to the ASME Code nomenclature, the gasket compressive modulus EG is used. There is little information about numerical values for EG, except for some data given by Schwaigerer [10]. For synthetic rubber of 75 Durometer, a value of 20,000 psi is suggested. Assuming a flange thickness of 1 in., the effective gasket diameter and the moment arms are computed:

Then flange loads and moments are computed:

also the operating moment:

Page 515

The bolt area is verified:

Stresses are then calculated. For a relatively thick flange welded to a thin hub, there is little support against rotation; hence the optional loose designation per Appendix 2 of the ASME Code, Section VIII, Division 1, is used:

As can be seen, the initial assumption of a 1-in.-thick flange is sufficient. This is less than the thickness obtained using the TaylorForge method. There are, however, many existing flanges of similar size and operating conditions that use in. or in.-thick flanges without difficulties or failures. This means that the method described is still safe to use. 4. An Improved Design Method There has been much interest among users of full-face gaskets since this method was published. From communications received, it appears that the method is sensitive if relatively thin flanges are

involved. This is probably due to the original simplifications, which assume that the flanges remain straight after rotation, an assumption not unreasonable when dealing with relatively thick flanges and soft gaskets. However, for low-pressure thin flanges, the rotation and bending deformation cannot be neglected. Thus a new test program was undertaken, using different outside-to-inside-diameter ratios K, as would be the case between small-diameter flanges with large K values and large-diameter flanges with small K values.

Page 516

As a result, an improved design method, based on a parabolic load distribution under the gasket, has been proposed. This method uses nondimensional parameters and permits an explicit solution for the flange thickness, given pressure and a flange geometry. This will eliminate the iterative calculations required in the design of fullface gasketed flanges, normally requiring the assumption of a flange thickness before stresses can be calculated, a procedure that may have to be repeated several times before an acceptable design is obtained. 4.1 Gasket Pressure Distribution The triangular gasket pressure distribution assumes that the flange faces remain straight during rotation. This is true for relatively thick flanges and soft gaskets. For thin flanges, however, the bending deformation of the flange faces will cause a shift in the centroid of pressure. This can be easily approximated using a parabolic gasket pressure distribution, as shown in Fig. 3. In both cases shown in Fig. 3, a limiting gasket load has been assumed to be the case before there is separation between gasket and flange faces. As shown in [12], the moment arm hG for the triangular distribution is

For a parabolic distribution, using the same procedure, the moment arm hG becomes

which is now used in the simplified method published in [12].

Numerical values of stresses versus flange thickness were calculated for a wide variety of flange geometries and plotted in nondimensional form.

Figure 3 Gasket pressure distribution.

Page 517

4.2. Nondimensional Parameters Since full-face gasketed flanges are normally used for low pressure with soft gaskets, attached to relatively thin shells, the preferred design method is similar to the optional loose method of Appendix 2 [3]. In this case, the contributions of the connecting shell are neglected. The flange ring alone is designed to resist rotations and deformations. Longitudinal hub stresses and radial (meridional) flange stresses are assumed to be zero; only tangential flange stresses are determined. Three nondimensional parameters are introduced:

Using these parameters, numerical values were calculated for a range of K from 1.1 to 2, for t/B from .025 to .150, and for resulting P/s values. Calculated values, when plotted on log-log graph paper, yield straight lines with only slight variations. These differences are due mainly to the location of the bolt circle with respect to the width of the flange face. The different number of bolts used in particular bolted flanged connections also may cause some deviations from the straight lines plotted. Figure 4 shows the design curve thus obtained. To verify the curves plotted in Fig. 4, three test pressure vessels with full-face gaskets were strain gage instrumented and tested over a variety of pressures to obtain maximum stress values. To cover a range equivalent to small-diameter and large-diameter

flanges, K values of 1.175, 1.417, and 1.500 were used. After each test sequence, the flanges of the test pressure vessels were machined down to a different thickness and the pressure tests repeated. All tests were performed with two different gasket materials, compressed mineral fiber and synthetic rubber, in several thicknesses. It was observed that the results for all gaskets tested were quite close to each other. This permitted the elimination of the gasket material and gasket thickness in the establishing of the nondimensional parameters. All data from strain gage measurement on the three pressure vessels was also compared with results of a finite element analysis, which will be reported in an upcoming paper. Adjustments to the curves in Fig. 4, in order to obtain straight lines on a log-log grid, were done on the safe side, in order to avoid any overstressing of the bolted connection.

Page 518

Figure 4 Full-face gasketed flange design curves.

4.3. Design Procedure An easy design procedure is available using Fig. 4. For a given set of design conditions and the required flange size, a flange outside diameter is assumed, based on the expected bolt size for the pressure in question. The nondimensional parameters A/B and P/s are calculated. From Fig. 4, the parameter t/B, and thus the required flange thickness, is obtained. For a given flange thickness subjected

Page 519

to a certain pressure, A/B and t/B are calculated, and P/s is obtained from the figure. The maximum flange stress is calculated and compared with the allowable stress. This procedures does not include a verification of the adequacy of the bolt area provided. This can be done by a few simple calculations as outlined in the Code, Appendix 2, except that, since the gasket compression force for the operating condition is not obtained, a simple assumption has to be made. Since in most cases extra bolt area provided does not result in greatly increased costs, it is here suggested that the hydrostatic bolt load HD plus the bolt force under the gasket HT be doubled to account for the unknown force. The bolt load due to gasket seating is calculated per Appendix 2. As can be seen from the foregoing, a full-face gasketed flange can be designed directly from the initial operating and size conditions. The required number and size of bolts may be obtained from the bolt check calculations after the flange size has been determined. 4.4. Numerical Example A stainless steel pressure vessel is required with a full-face gasketed flanged joint. The vessel inside diameter is 10 in., and a design pressure of 200 psig is specified at a temperature of 200°F. Due to size limitations, a flange outside diameter of 11.75 in. was chosen, and 3/8-16 UNC bolts should be used. The allowable stress of the vessel material is 18,800 psi. As a first step, the flange thickness is obtained:

t = (.035)(10) = .350 in. The minimum thickness obtained is rounded off to a .375-in. ( flange thickness, and the maximum stress is calculated again using Fig. 4.

Page 520

The calculated stress of 18,180 psi compares well with the measured value of 18,690 psi for a rubber gasket, or of 17575 psi for a compressed mineral-fiber gasket as obtained on a test pressure vessel of the same size. The number and size of bolts must also be verified. Bolts of lowalloy steel SA-193-B7 are selected, with an allowable stress of 25,000 psi:

Wm1 = 1.5(HD + HT) = (1.5)(15,708 + 1817) = 26,290 lb

To check for gasket seating, a compressed mineral-fiber gasket is assumed with a gasket factor m = 2.0, and a minimum seating stress y = 1600 psi is assumed. Wm2 = bpCy = (.438)p(11.0)(1600) = 24,190 lb

Sb = (16)(.0775) = 1.24 in.2 It can be seen that 16 bolts of in. size are sufficient for the operating and gasket seating conditions as specified in the Code. References

1. Waters, E. O., and Taylor, J. H. The strength of pipe flanges. Mechanical Engineering 49 (5a):531542 (1927). 2. Waters, E. O., Wesstrom, D. B., Rossheim, D. B., and Williams, F. S. G. Formulas for stresses in bolted flanged connections. ASME Trans. 59: 161169 (1937). 3. Pressure vessels. ASME Boiler and Pressure Vessel Code, Section VIII, Division 1. New York: American Society of Mechanical Engineers, 1995. 4. ASME/ANSI B31.3. Chemical Plant and Petroleum Refinery Piping. New York: American Society of Mechanical Engineers, 1993. 5. Schneider, R. W. Flat-face flanges with metal-to-metal contact beyond the bolt circle. ASME Trans. 90: 8288 (1968). 6. Lonngren, H. E. Flange design calculations. Petroleum Refiner 26 (Nov. 1947): 130134; 27 (Jan. 1948): 6373; (Feb. 1948): 116119; (July 1948): 105108; (Aug. 1948): 102103; (Oct. 1948): 145146.

Page 521

7. Taylor Forge Inc. Design of flanges for full-face gasket. Engineering Department Bulletin 45. Chicago: 1951. 8. Taylor Forge Inc. Modern flange design. Chicago: 1938 (fullface flanges only in later reprints of this bulletin). 9. Fiber-reinforced plastic pressure vessels. ASME Boiler and Pressure Vessel Code, Section X. New York: American Society of Mechanical Engineers, 1995. 10. Schwaigerer, S. Festigkeitsbrechnung von Bauelementen des Dampfkessel-, Behälterund Rohrleitungsbaues. Berlin: Springer Verlag, 1961. 11. Berechnung von flanschverbindungen. German Standard DIN2505. Berlin: Beuth Verlag, 1988. 12. Blach, A. E., Bazergui, A., and Baldur, R. Bolted flanged connections with full-face gaskets. WRC Bulletin 314 (May 1986). 13. Blach, A. E. Bolted flanged connections with full-face gasketsProposed ASME Code rules. Proceedings 10th International Conference on Fluid Sealing, April 1984. 14. Blach, A. E. Nonidentical flanges with full-face elastic gaskets. WRC Bulletin 381 (May 1993).

PART V: THE GASKETED JOINT IN SERVICE

Page 523

12 Assembling a Gasketed Joint JOHN H. BICKFORD Consultant, Middletown, Connecticut 1. Introduction Many of the previous chapters have contained suggestions or information about the amount of clamping force or seating stress we'd like to create on a gasket when we assemble a joint and tighten the bolts, information developed during years of experience and research by gasket manufacturers and by such groups as the Pressure Vessel Research Committee (PVRC) of the Welding Research Council (WRC). These chapters and these sources, however, provide little or no information about how to achieve those clamping forces (and seating stresses). There is good reason for this apparent oversight: A large number of variables affects the results when we tighten a group of bolts. As a result, a given amount of clamping force can be achieved, with certainty, only by the use of procedures and equipment that are economically impractical for all but the most critical joints. Standards organizations and gasket manufacturers understandably refuse to mandate the use of such procedures and, instead, leave it up to the user to choose assembly tools and practices that he or she can afford and that produce results good enough for his or her applications. Occasional leakage is combatted

Page 524

by retightening in-service joints or, in extreme cases, by shutting down a system and trying again. Frequent shutdowns, of course, can lead to a search for better assembly techniques even if that raises assembly costs. We'll look at several options in this chapter. Accurate control of clamping force is difficult, so it helps to understand some of the many things that can affect results. Such understanding can lead to modest improvements in bolt-up procedures, and these can make the difference between acceptable and unacceptable joint life and behavior. The main goal of this chapter, in fact, is to review some of the common assembly problems we must be able to cope with. As we proceed you'll learn that bolt tension, or preload, is not always equal and opposite to the clamping force that that bolt exerts on the joint. Sometimes we'll be able to control one of these things, sometimes another, and the distinction is important. The thing we'd always like to be able to control is clamping force. More commonly, however, all we're able to control is preload. The most practical and economical way to do that is to control the torque we apply to the bolts or nuts, so we're going to look at that first. Specifically we're going to consider the accuracy with which we can control preload when using a torque wrench. 2. Torque vs. Preload Accuracy The relationship between torque and preload can be expressed by the familiar short form torque-preload equation: T = KDFP where:T = applied torque (in.-lb, mm-N

(1)

FP = preload developed in the bolt (lb, N) D = nominal diameter of the bolt (in., mm) K = experimentally determined nut factor (dimensionless) Unfortunately, the nut factor, K, is affected by a large number of variables. A study sponsored by Wright-Patterson Air Force Base [1] identified 76 statistically significant ones, including the materials from which bolt, nut, washer, and joint members are made; the hardness of each of those parts; the surface finish of each; whether or not they're plated; whether or not they're lubricated and, if so, by which lubricant, by the quantity used, by where and how it was applied; etc. The fit between bolts and holes, the perpendicularity of holes to the surface of the joint, the speed with which the bolts were tightened, and many other factors have also been found to have an impact [1,2]. The result of all this is that the preloads achieved when we apply the same torque to a large group of as-received steel nuts and bolts, tightened one by one

Page 525

against steel joint members, will vary by ±30%, typically. If the nuts and bolts have been received from a variety of sources and/or if some are new and some are being resused and/or some are clean and some are nicked or contain rust spots, the variation can far exceed ±30%. Best results can be obtained by using new parts all received from the same source at the same time; by making an experiment on an actual joint to determine the actual nut factor for those fasteners (using ultrasonics or strain gages to measure preload); by lubricating nuts and bolts with a good, high-pressure thread lube; and by being as consistent as possible in the sequence and way in which the bolts are tightened. Scatter in the individual preloads created in the bolts, one by one, can sometimes be reduced to ±10% by these means, but that, unfortunately, is not the only variable we face. 3. Joint Members that Resist Closure The torque-preload equation defines the relationship between the torque we apply to the nut and the tension created in the bolt. This tension is usually equal and opposite to the clamping force that that bolt creates between joint members, that all-important clamping force that generates the seating stress or pressure on the gasket that fights leakage. There are times, however, as mentioned earlier, when bolt tension and clamping force are not equal and opposite, making torque a very inaccurate way to control seating stress on the gasket. We're going to look at a couple of these situations in this and the next section. Let's assume that the flanges we're trying to assemble lie in a

horizontal plane and have not yet been pulled together, as in Fig. 1. Let's further assume that we are going to pull them together by tightening the lower nuts. To start with there is no clamping force between joint members, because these aren't even in contact. There will be tension in the bolts, however, thanks to the weight of the blind flange we're trying to pull up into contact with the upper flange. So it will require torque to turn the nuts. And Eq. (1) will define how much torque, but not how much clamping force, which is a far more important factor. The torque we need to apply in this case is that required to lift the blind flange plus that required to create the desired seating stress on the gasket: T = KD(FP + W)

(2)

where: dimensionless nut factor K= D =nominal diameter of the bolts (in., mm) that portion of the tension in the bolt required to FP = create a desired seating force (lb, N) that portion of the tension in the bolt required to pull W= the flanges together (lb, N) FP + total tension in the bolt (lb, N) W=

Page 526

Figure 1 Illustration of the weight effect. It would require torque to turn the nuts and move the blind flange up into contact with the upper flange, and this torque would be related to the tension in the bolts. All of that tension is created by the weight of the blind flange, however, and does not represent a clamping force between joint members (which, at this point, aren't even in contact). This can be a problem if we're controlling the applied torque in order to create a desired seating stress on the gasket.

T = torque required to create that total tension (in.-lb, mm-N) The bolts in most joints, of course, have been selected to create the desired seating force. If selected by ASME Code rules they may be theoretically capable of supporting four times that much tension (because of safety factors built into the Code). But they may not always be strong enough to pull things together and still squeeze the gasket. Most maintenance workers, presumably, will know better than to try to lift a heavy flange by tightening the bolts. Clamps or

something will be used to pull the joint together before the nuts are tightened. But this weight effect can be a factor even if the joints members must be moved only a few thousandths of an inch while compressing a gasket. In any event you should consider this possibility when dealing with a chronic leaker that defies other explanations. 4. Nonparallel Joint Members A closely related problem is shown in Fig. 2. The flanges are in some sort of contact along one side, but a significant gap exists on the other side. This is

Page 527

Figure 2 Nonparallel joint members create the same kind of problem as illustrated by Fig. 1. Torque is required to pull the flanges into full contact, torque that, once again, does not contribute to seating stress on the gasket.

probably a more common situation that discussed in the last section, but it creates the same problem. If the flanges are brought into full contact merely by tightening the nuts, and the same torque is used on all bolts, there will be a big difference between the clamping force created on the closed and open sides. Nonparallelism of this sort can also create high stresses in the pipes and other equipment to which these flanges are connected. A maintenance superintendent in a chemical plant once told me that he wouldn't attempt to bolt up flanges unless they were parallel enough to be within 12 mils of each other all the way around. Bolting nonparallel flanges, he said, was a waste of time; They'll always leak. The U.S. Navy has had similar troubles and, as a result, has developed the flange parallelism criteria tabulated in Table 1. Note that the worst nonparallelism they're prepared to accept is 35 mils

(0.035 in.). Nonparallel or resistant flanges can have a large impact on the seating stress created on a gasket when we assemble a joint, but we can usually detect and avoid such a problem. The next one we're going to address is more insidious because it's always present and cannot be readily detected during normal bolt-up procedures. 5. Elastic Interactions 5.1. What They Are We can think of the joint members as stiff, elastic members, springs if you will. They are clamped together by less stiff, elastic bolt springs. The gasket, whose behavior is elastoplastic, is compressed as we tighten the bolts. Its stiffness

Page 528 TABLE 1 Flange Parallelism Criteriaa Flanges adjacent to rotating equipmentLine flanges 1 in. or through 8 Pipe size All sizes less in. Wall SCH 40 or Over 40 through All All thicknesses thickness below SCH 80 Raised 0.005 in./in. of contact 0.10 in. 0.025 in. 0.020 in. face diameter 0.020 0.005 in./in. of contact Flat face 0.035 in. 0.030 in. in. diameter aIf the pipe diameter is over 10 in. or wall thickness is greater than schedule 80, the specifications say that a special analysis is required.

is typically less than that of either the joint members or the bolts, at least in large pressure-vessel joints. The in-service behavior of the gasket, however, is fairly elastic. (For more detail, see Chapters 5 and 8.) When we tighten the first bolt we partially compress the joint spring in its vicinity. By doing so we achieve preload in that bolt that presumably lies within ±1030% of some norm. When we later tighten the bolts adjacent to that first one, we create preloads in each that probably lie, again, within the same ±1030%. But tightening these bolts further compresses the joint spring in that region of the joint, and allows the first bolt to relax. Chances are its residual preload will now lie below the original ±1030% range. For more than two decades engineers have been measuring residual preload in problem pressure vessel or nuclear power joints using ultrasonic equipment [2,3]. Results of one such set of measurements is illustrated in Fig. 3. Sixteen X 8 ASTM A193

Grade B7 bolts were tightened in a three-pass procedure, using torques of 225, 550, and 825 ft-lb, applied in a conventional crossbolting, or star, sequence. The residual preloads in half of the bolts, after the final pass, are illustrated by arrows above each bolt hole. (I've drawn arrows on only half the holes for clarity. The bolt preloads in the front half of the joint showed a similar pattern.) The residuals shown here vary, not by the conventional wisdom amount of ±1030%, but by exactly 10:1. You'll find another example in Chapter 6, Fig. 28, where the variation in preload achieved by a normal three-pass bolting procedure was about 3.8:1. In other situations the residual preloads have been found to vary as much 20:1. And all of this when a normal three-pass bolting procedure was used during assembly, with well-trained mechanics using properly calibrated torque wrenches. George Bibel of the University of North Dakota has studied elastic interactions in some detail [46]. He has found that a relatively thick spiral-wound gasket can encourage an average preload loss, through interaction, of 2550%

Page 529

Figure 3 Sixteen ASTM A193 B7 bolts were tightened one by one in a conventional three-pass star-pattern sequence. The arrows represent the residual preloads (forces) exerted by the bolts on the joint at the end of the third pass. As you can see, these forces vary in magnitude by 10:1, thanks to elastic interactions between the bolts when they were tightened one by one. The first bolts to be tightened at any point around the joint have later relaxed when adjacent bolts pulled the joint further together at that point. Only the forces along the back edge of the joint are shown here, for clarity, but forces along the front edge show a similar pattern.

in large-diameter pressure-vessel flanges. (His tests involved ANSI Class 150 flanges having nominal diameters of 16 and 24 inches.) Thinner gaskets reduced the interaction loss, as did metal-to-metal contact. In one test involving a 24-inch flange bolted-up metal-tometal, without a gasket, for example, the average preload loss was only 16%. Bibel also discovered that tightening bolts one by one can rock a flange, further altering the preloads in previously tightened bolts [3]. Many engineers challenge interaction data when first exposed to it,

claiming, for example, that every gasketed joint in a petrochemical plant would leak if residual preloads really varied as much as this. Interactions are real, however, and do create the kind of results we've been looking at. There are a couple of reasons, I think, why this scatter in preload doesn't produce more leakage than it does. First of all, safety factors built into the Code lead to grossly overdesigned joints. Fewer and/or smaller bolts could theoretically clamp each joint together with sufficient force to prevent leaks. The fact that the actual, oversized bolts don't have much preload merely makes them, in effect, smaller bolts. Probably more important is the fact that whether or not a gasketed joint leaks excessively depends as much (if not more than) upon the initial seating stress created during assembly as upon the residual in-service maintenance stress, as explained in Chapter 8. The higher the seating stress, the less the maintenance

Page 530

stress required to prevent a leak. As a result, although every other bolt in a joint may have very little residual preload, and therefore be creating very little maintenance stress, these bolts were originally tightened to the same preloads as those that have retained that much preload. The per-bolt seating stress, in other words, may have varied by as little as the ±1030% we expect when we tighten individual bolts. Elastic interactions, however, always increase the probability that problem joints will leak. It's useful, therefore, to learn about ways in which the bolt-to-bolt scatter in residual preload can be reduced. 5.2. Combatting Elastic Interactions Many attempts have been made to identify a torquing procedure that will create more uniform residual preloads than the normal three-pass procedure. One thing that seems to help is to go around the clock several times after the third pass, applying the final, thirdpass torque to each bolt. Each pass reduces the scatter a little, but quite a few passesa half dozen or moreare required for good results. Here's a less time-consuming procedure that often helps. Apply the final torque to each bolt once more, in a star-sequence fourth pass, but this time do it in the reverse order, retightening the last bolt you tightened first, the next to the last next, etc. back down to number one. The results of one such procedure are shown in Fig. 4. Residual preload in these bolts varied by over 5:1 after the third pass, but by only 1.8:1 after a reverse-order fourth pass. Unfortunately, this procedure doesn't always work as well as it did in the example illustrated in Fig. 4, but the reverse pass always improves things.

Bibel has also developed a two-pass bolt-up procedure that can reduce the scatter in residual preload to something like ±2% [5]. Actual preloads must be measured as the joint is bolted-up, however, using strain gages, load washers, ultrasonics, or some equivalent means. The preloads achieved during a first pass are used to develop a matrix of interaction constants. The bolts are then loosened and retightened to new preload values determined by equations based on the interaction constants. At this point (mid1996) the method has not been used in the field, but it is being considered as a means to improve results while reducing radiation exposure in nuclear power plant maintenance operations. Incidentally, Bibel's method has also been tried, successfully, on an automotive cylinder head and engine block joint [6]. One sure way to reduce elastic interactions is to tighten all bolts in a joint simultaneously. This is done in automotive production, for example, where multispindle nut runners, like those shown in Fig. 5, are used to tighten all bolts on an engine head at the same time. Similarly, large carousels of hydraulic tensioners are used in some nuclear plants to tighten all of the bolts in the head

Page 531

Figure 4 The left-hand sketch shows the final preloads in half the bolts in a joint tightened by a three-pass star-sequence procedure. Once again, elastic interactions have caused wide variation in the magnitude of the residual preloads in the bolts (this time an approximately 5:1 ratio between maximum and minimum forces). The right-hand sketch shows preloads in the same half of the same joint after a fourth pass in the reverse order, using the same torque as was used on the third pass. The variation in preload has now been reduced to less than 2:1. See the text for a more complete discussion.

of a nuclear pressure vessel simultaneously. If the tensioners are too big to fit on adjacent bolts, then the carousel will tighten half the bolts in the first pass, the other, in-between ones in a second pass. In other applications, or where money is tight, bolts may only be tensioned four at a time, as also shown in Fig. 5. The more bolts you tighten at once, the less effect subsequent tightenings will have on the preloads in those tightened first. 6. Miscellaneous Factors The weight effect, nonparallelism, and elastic interactions can all affect in-service clamping forces on a gasket. Lesser factors can also have an influence. Thread surfaces can embed themselves in

each other as these heavily loaded surfaces yield and creep in an effort to redistribute contact stresses. Nuts and bolt heads can embed themselves in the joint surfaces. (Hard washers help here.) Although embedment may allow a bolt to relax by only a few thousandths of an inch, tightening them only stretches them by 13 thousandths of an inch per inch of grip length, so embedment will often cause preload loss of 10% or more. If the holes in the mating flanges are not lined up perfectly before the bolts are tightened, there could be interference between the bolts and the walls of their holes. Torque will be required to pull the bolts past the walls as the bolts are

Page 532

Figure 5 Elastic interactions can be reduced if several bolts are tightened simultaneously. Four hydraulic tensioners can be ganged, for example, as shown in the lower right of this sketch, to tighten four at once. Multispindle air tools of the sort shown to the left and in the center, above, are used in mass production operations to tighten all or most of the bolts in a jointan engine head, for examplesimultaneously. The more you tighten at once, the less the interaction.

tightened (i.e., stretched), robbing some of the torque we think is creating clamping force on the gasket. Although it doesn't happen during assembly, differential thermal expansion between bolts and joint members can either increase or decrease in-service preloads when the newly assembled system is turned on. This can happen if bolts and joint members are made of different materials and/or reach different-in-service temperatures. Insulating the joint can help reduce this problem (but can mask leaks caused by other problems). Thermal collars can also be used

to combat differential expansion, as in Fig. 6, because they allow us to use a longerand therefore less stiffbolt in a given hole. The greater the joint-to-bolt stiffness ratio the less a given amount of differential expansion will affect bolt preloads. Belleville washers provide even more forgiving, flexible clamping springs than do bolts. 7. Other Preload Control Options Until now we've assumed that a torque wrench of some sort would be used to tighten the bolts, with torque being used to control preload. As we've seen,

Page 533

Figure 6 Differential thermal expansion between bolts and joint members can cause large changes in the clamping force when the system is put in service. Such changes can be reduced to an acceptable level by using longer bolts contained within thermal sleeves, as on the left, or by using longer bolts and a stack of Belleville washers, as on the right.

however, torque leaves a lot to be desired from a control point of view. Next, therefore, we're going to take a quick look at some other options, most of which are more expensive than pure torque control, but each of which gives us some advantage over torque, usually including more accurate results. 7.1. Multi-Jackbolt Tensioners One factor that can have a major impact on preload accuracy is poor accessibility. The mechanic can hardly get a wrench onto a bolt, much less tighten it properly. It has been said, in fact, that the amount of tension in any bolt in this world is inversely proportional to its accessibility, suggesting that this is a problem with bolts of

any size. It can, however, be especially challenging when we're dealing with large-diameter fasteners. One interesting solution to the problem is the Multi-Jackbolt Tensioner (MJT) provided by Superbolt, Inc., of Carnegie, Penn. [3,7]. Some 30 different versions of this productboth nut type and bolt typehave been designed, one of which is shown in Fig. 7. They have been used to tighten fasteners ranging in size from 3/4 to 32 inches, and to provide preloads of as little as 10,000 lbs to as much as 20,000,000 lbs. The original MJT consists of a cylindrical nut threaded onto the stud or bolt to be tensioned. A circle of small fasteners mounted within this nut are then tightened, one by one, pushing down against a hardened washer placed against

Page 534

Figure 7 Nut-type Multi-Jackbolt Tensioner (MJT) made by Superbolt, Inc., of Carnegie, Penn. A cylindrical nut is run down on a large-diameter bolt or stud. The group of small fasteners circling the nut are then tightened one by one to push the nut away from the joint surface, tensioning the large fastener. As you can see, a small wrench can, therefore, be used to tension a very large fastener. Bolt-type MJTs are also available; in these the circle of bolts is mounted in the cylindrical head of a bolt, rather than in a nut.

the joint member. As compressive stress builds up within the small fasteners, the large one is tensioned. The wrench required to tighten the small bolts is, of course, very much smaller than the wrench or hydraulic tensioner that would be required to tighten a regular nut on the same stud. In one example, a in. torque wrench is used to produce 10,000 tons of tension in a large stud. As a result, studs or bolts located in pockets, or too

closely spaced to accommodate conventional tools, can easily be tightened. Gallinga common problem when large-diameter studs are tightenedis also eliminated, because the cylindrical nut is never turned against the mating threads while loaded. The device is used in many applications, including high-temperature gasketed joints. One of its advantages there is that it adds elasticity to the clamping system, providing the same help as the arrangements shown in Fig. 6. 7.2. Torque-Turn Control Many modern bolting tools measure both the torque applied to the bolt and the turn through which the head of the bolt or the nut is moved. The integral of the

Page 535

torque-turn function is equal to the work or energy required to tighten the fastener. About 90% of that energy is absorbed by friction between nut and joint and between male and female thread surfaces, and these losses can vary quite a bit. A variety of torqueturn strategies has been developed, however, to use the combination of torque/turn/consumed energy information to improve the accuracy with which we can tighten fasteners [2,3]. One, strategy, called the turn-of-nut procedure, is widely used in the structural steel industry and can be performed with hand tools or even with impact wrenches, but it can be used only to tighten certain types of ductile fasteners well past their yield points. Most of the other torque-turn strategies have been developed for use on mass production lines (primarily automotive) and use fixed-station or semiportable air-powered or electric power tools of the sort shown in Fig. 5 to tighten fasteners to some preload below yield. Portable hand tools with some of the same capability are now available as well (Fig.8) [8]. Either the Fig. 5 or Fig. 8 equipment can be used on small gasketed joints. Such equipment is not available if you're dealing with

Figure 8 This Sensor I wrench, manufactured by Ingersoll Rand, measures both the torque applied to a bolt and the resulting turn of the nut. The combination of both torque and turn information provides more accurate control of bolt preload than does torque alone. The tool can also be used to audit residual preload after elastic interactions etc. have done their worst, allowing us to compensate for such interactions [2,3]. This tool, like air-powered tools of the type shown in Fig. 5, can also be used for yield control.

Page 536

bolts requiring torques over 200 ft-lb or so. (Check your tool suppliers for the latest limits.) 7.3. Yield Control The most accurate form of torque-turn control is called yield control, because each bolt is tightened to the threshold of its yield point [9]. The computer that controls tools like those shown in Fig. 5 senses yield by observing the slope of a torque-turn curve. That curve is nearly straight when bolt stretch is fully elastic, much like the stress-strain curve of the bolt material. The slope of the curve becomes less steep when the fastener starts to yield. At this point the computer orders the tool to stop tightening the fastener. A microprocessor in the tool shown in Fig. 8 performs the same task, so that hand tool can also be used for yield control. 7.4. Stretch Control Those who must tighten large bolts accurately have long resorted to something even more accurate than torque-turn control: They control bolt preload by measuring the change in length of the bolt as it's tightened. The relationship between bolt stretch and tension can be predicted using Hooke's Law or can be determined with improved accuracy by making a test in a tensile machine. The traditional way to measure stretch in large bolts is to use a depth micrometer to measure the distance between one end of the bolt and an unstretched gage rod mounted in a hole that has been gun-drilled along the axis of the bolt. As the bolt is tightened, this distance increases. This depth micrometer procedure is illustrated in Fig. 9. Special fasteners with built-in stretch gages are available and two

examples are also shown in Fig. 9. These can be obtained from Valley Forge and Bolt of Phoenix, Ariz. or from RotaBolt, Ltd., Dudley, West Midlands, England [3,10,11]. A V-shaped spring in the Valley Forge product reacts against the curved end of the gage rod and moves across a viewing window in the end of the bolt to indicate stretch. The motion is usually calibrated in terms of bolt tension instead of stretch, but the measurement is stretch. RotaBolt, on the other hand, builds an accurate but inexpensive micrometer into the head of each bolt. This measuring device is then covered with a protective cap. One big advantage of stretch control is that it allows us to measure residual preload as easily as initial preload. This means that we can detect things like elastic interactions and embedment relaxation and to retighten to compensate for them if necessary. Note that we can't compensate for nonparallelism or the weight effect because we're measuring the tension in the bolt, not the clamping force between joint members; but stretch control is still a major step forward.

Page 537

Figure 9 Bolts are stretched a small amount when they're tightened, and this stretch is proportional to the tension, or preload, in the bolt. We can measure such stretch by measuring the distance between one end of the bolt and the end of an unstretched gage rod mounted in a hole along the axis of the bolt. Depth micrometers, such as that shown to the right above, have long been used to make these measurements. Special fasteners with built-in stretch gages are also available. The two shown here are made by Valley Forge and Bolt of Phoenix, Ariz. (to the left above) and RotaBolt, Ltd., of Dudley, West Midlands, England (center). Their operation is described in the text.

7.5 Ultrasonics Figure 10 shows ultrasonic instruments that can be used to measure initial and residual bolt stretch or tension. These instruments have been around in some form or other for 20 years or more, and have found considerable use in gasketed joint applications [12]. They

can be used on most bolts over in diameter. Skill and knowledge are required to use the equipment, but continuing development work is slowly making the technique less intimidating and more accurate. It has been of immeasurable help in studying elastic interactions and other phenomena in the field. One company, with considerable financial assistance from two large European tool manufacturers, has developed low-cost ultrasonic transducers that are permanently mounted on the ends of each bolt, reducing one of the major difficulties in using ultrasonics on bolts [3,13].

Page 538

Figure 10 Ultrasonics can be used to measure either the stretch of or the tension in a bolt. Several different instruments are available to do this. The two shown here are manufactured by the Industrial Tool Division of the Bidwell Industrial Group, Inc., of Middletown, Conn.

References 1. Stewart, Richard. Torque-tension variables. A list prepared by Wright-Patterson Air Force Base, Dayton, OH, April 16, 1973. 2. Bickford, John H. An Introduction to the Design and Behavior of Bolted Joints. 3rd ed. New York: Marcel Dekker, 1995. 3. Bickford, John H., and Nassar, Sayed, eds. The Handbook of

Bolts and Bolted Joints. New York: Marcel Dekker, 1998. 4. Bibel, G. D. Experimental and analytical study of elastic interaction in a pipe flange. The International Conference on Pressure Vessel Technology, Dusseldorf, Germany, 1996. 5. Bibel, G. D., and Ezell, R. Bolted flange assembly: Preliminary elastic interaction data and improved bolt-up procedures. Welding Research Council Bulletin 408 (May 1996).

Page 539

6. Goddard, D. L., and Bibel, G. D. Achieving a selected load distribution in the bolted joint of a cylinder head of variable stiffness and contact geometry. SAE Technical Paper Series 940693. 7. From information received from Rolph E. Steinbock, Chairman and CEO of Superbolt, Inc., Carnegie, Pennsylvania, July 1996. 8. Eshghy, S. The LRM fastening system. Fastener Technology (July 8, 1978): 47. 9. Boys, J. T., and Junker, G. H. Modern methods for the tightening of fasteners with power tools. Design Engineering (January 1975): 21. 10. Corbett, R. F. The importance of high-integrity bolt tension control. Paper presented at the 2nd CETIM International Symposium on Fluid Sealing, Labaoule, France, 1990. 11. Private correspondence with Ronald Clark, President of Valley Forge and Bolt Co., Phoenix, Arizona, October 1995. 12. Meisterling, Jesse. Accuracy of ultrasonic bolt load determination. Paper presented at the CETIM Symposium on Fluid Sealing, Application to Bolted Flange Connections, Nantes, France, June 1820, 1986. 13. McEnroe, Tony. Will impulse wrenches influence the assembly industry? Fastener Technology International (December 1995): 54.

Page 541

13 In-Service Inspection of Gasketed Joints JOHN H. BICKFORD Consultant, Middletown, Connecticut 1. Introduction We will often want to know whether or not a bolted joint has been assembled properly, whether or not, for example, there is enough clamping force on the gasket to prevent leakage or blowout. To do this we must find a way to measure or at least estimate the present tension (usually called preload) in the bolts, or, even better, find a way to measure the distribution and magnitude of the contact pressure between joint members or between a joint member and a mating gasket. Unfortunately, none of this is easy to do. But there are a couple of traditional ways to do it, and a number of new techniques have been developed in the last few decades. None of these techniques are as simple or as economical as we'd like, but they can often save us more than they cost, by preventing premature and/or dangerous failure. In this chapter we're going to examine the basic ways to measure bolt tension or interface clamping force after assembly. Some of these procedures can be used when the joint is in active service. Others, thanks primarily

Page 542

to temperature limitations, can be used if only the joint is taken temporarily out of service, but without disassembling or otherwise disturbing it. First we'll take a brief look at a couple of traditional methods for measuring bolt tension. Then we'll look at a special group of bolts (or studs) that contain tension-indicating mechanisms. Next we'll consider ways to use ultrasonic measurements to measure bolt elongation or tension. Finally we'll look at ways to reveal the distribution and magnitude of clamping pressure between joint members. All of the techniques to be discussed can be and have been used on gasketed joints. Note that each of the procedures we're going to look at can be used to measure the postassembly and/or in-service condition of the joint. Some of the techniques and equipment described in Chapter 12 can be used to measure the tension created in bolts or studs during assembly but cannot be used to estimate that tension if it subsequently changes because of service conditions, self-loosening, embedment, or the like. Our concern in this chapter is only for those techniques that can be used to inspect the joint in some fashion after it has been placed in service. 2. Traditional Methods 2.1 Strain Gages Strain gages have been used for many decades to measure the present tension or preload in bolts. Two basic techniques have been used. In the first the gages are mounted on the bolts themselves; in the second they're mounted in washers placed under the head of the bolt or its nut. Either technique provides measurement accuracies

of ±2% or better [1, p. 341]. The gages have to be mounted on a surface of the bolt or washer that truly reflects the axial tension in the bolt. Traditionally, therefore, bolt gages are mounted within a hole gun-drilled along the axis of the bolt and hopefully at a reasonable distance (one or more bolt diameters) from the head of the bolt or from the thread runout point to avoid the stress concentrations at those points. The gages must obviously be installed and monitored by a knowledgeable technician. At least one company, Strainsert, of Conshohocken, Penn., sells bolts and studs in which strain gages have been installed, plus, I believe, strain-gaged cartridges that you can insert in your own bolts. The gaged parts must, of course, be left in place if they are to be used to monitor in-service bolt loads. Accurate, real-time static or dynamic data can be obtained by use of this well-developed technology. Strain gages, in fact, probably offer the most accurate and reliable way to monitor bolt tension at present. The cost, however, can run to several hundred dollars per fastener.

Page 543

2.2 Restarting Torque Most bolts and studs in this world are tightened using measured torque to control the process. It seems logical, therefore, to use a reapplication of torque to determine whether or not the bolts are still as tight as they were when first installed. In practice a mechanic applies clockwise (tightening) torque to a previously tightened fastener until he or she hears a sharp breakaway noise or until the torque indicated by the gage on the mechanic's wrench peaks and starts to decrease. The mechanic assumes that the torque required to create the noise, or the peak torque if there is no noise, can be used to estimate the present tension in the bolt. There are, however, a number of potential problems here. First of all, service conditions will often change the lubricity of the parts, drying up lubricants or creating rust, for example. This inevitably changes the relationship between the torque applied to the bolt and the tension in it. So even if our restarting torque turns out to be the same as the torque used at assembly, it doesn't necessarily mean that the tension in the bolt equals the initial preload. Another difficulty comes when we try to determine the exact amount of torque required to restart the nut or bolt. Work done by Ralph Shoberg of R.S. Technologies of Farmington Hills, MI, shows that the actual restarting point is very difficult to determine by use of a hand tool [2]. As the torque applied to the tightened fastener is increased, the friction restraint between the face of the nut and the joint surface breaks first. Male and female threads are still locked together by friction forces between those surfaces. If the fastener is fairly well lubricated, the nut usually breaks free when the torque being applied to the nut reaches the point at which

it could create tension in the bolt equal to that already there. This, in fact, is the inspection torque the operator is looking for. But there is no noise at this point to tell the operator that the nut is turning relative to the joint, and the torque being applied to the joint continues to climb. As a result the mechanic is unaware that he or she has gone too far. The bolt now twists, and the tension in the bolt increases past that initial preload created in the bolt at assembly. Finally, relative motion occurs between male and female threads, and the torque starts to drop off. The operator records the peak torque, which will probably be higher than the assembly torque, and decides that everything is OK. In the process, however, the operator has actually tightened the bolt more than it was tightened during assembly. If the fastener is poorly lubricated, the nut will still break loose from the joint before the threads move relative to one another, but this time the break occurs at a torque slightly greater than the assembly torque. There is oftenbut not alwaysa sharp noise at this point. But the torque being applied to the nut continues to climb to first one and then, after a brief drop-off, to a second and

Page 544

higher peak. According to Shoberg, the tension in the poorly lubricated bolt does not start to increase until this second peak is reached. In summary, a clockwise restarting torque, or, as it's often called, break-away torque, appears to be a reasonable way to determine whether or not the tension in the bolts is similar to that created during assembly. Noise and/or peak torque are not perfect ways to inspect the assembled fastener, and we must recognize that they usually suggest that the bolts are tighter than they actually are. If restarting torque equals the assembly torque, then the bolts had less preload in them than desired. But the differences will probably not be great; this test can certainly tell us if the bolts are a lot looser than expected. So much for the traditional breakaway torque test. Shoberg has continued his studies in this area, and he has developed techniques and equipment that use a combination of torque and turn-of-the-nut information to measure axial tension in a bolt. I don't know whether or not his recent work could be applied to in-service inspection of bolted joints, but those interested in better breakaway tests would be well advised to contact him. 3. Using Special Fasteners to Reveal Preload 3.1 The Gage Rod Stud or Bolt For many years, large-diameter studs in critical joints have been supplied with gage rods that sit loosely in holes that have been gundrilled down the axis of the bolt. The far end of the rod is fastened to the stud in some fashion, perhaps by a few threads. When the stud is installed and tightened it is stretched a small amount. For

example, an ASTM A193 B7 stud will stretch about two thousandths of an inch per inch of grip length if tightened to 60% of its yield strength. A depth micrometer is used, as suggested in Fig. 1, to measure the distance between the end of the gage rod and the end of the stud, both before and after the stud is tightened. Using Hooke's law we can relate this stretch to tension or preload in the bolt as follows [1, p. 313]: (1) where: DLc = combined change in length of the threaded and nonthreaded sections of the stud (in., mm) fp = preload in the bolt or stud (lb, N) Lbe = effective length of the unthreaded body of the bolt, this is the length of the unthreaded bolldy itself plus onehalf the height of the head of the bolt; if a stud, this is simply the length of the unthreaded section) (in., mm)

Page 545

Figure 1 A depth micrometer can be used to measure the distance between the end of a gage rod, which is loosely mounted in an axial hole in the bolt or stud, and the end of the fastener. When the fastener is tightened, this distance increases. The change is proportional to the tension created in the fastener when it's tightened, and so it can be used to measure that tension, even after the bolt or stud has been placed in service.

effective length of the threaded portion of the stud; Lse equal to the length of threads within the grip length = plus half the combined thickness of the two nuts (or nut plus tapped hole) (in., mm)

cross-sectional area of the unthreaded body of the = fastener (in.2, mm2) tensile stress area of the threaded section of the = fastener (in.2, mm2) E Young's modulus (psi, MPa) = I've used primed terms for body and thread areas to remind us that, in this case, we must subtract the cross-sectional area of the gage rod hole from the normal body or tensile stress area, as follows: (2) (3)

Page 546

where:

As = tensile stress area of the bolt (in.2, mm2) Ah = cross-sectional area of the gage rod hole (in.2, mm2)

We can get AS from a table of tensile stress areas in ASME Standard B1.1 (e.g., Table 13 in the 1989 edition) [3] or in the screw threads section of any edition of Machinery's Handbook [4] or in Appendix F of ref. [1]. We must, however, subtract the area of the gage rod's hole (Ah) from the tensile stress areas tabulated in these references. Alternatively, we can compute the tensile stress area by using an equation found in Federal Standard FED-STD-H28/2b (e.g., in Table II.B.1 in the 20 August 1991 edition) [5] as follows: (4) where:As = tensile stress area (in.2) D = nominal diameter of the stud or bolt (in.) n = number of threads per inch If we're dealing with metric threads, then we must use, instead, the following equivalent expression for AS [1, p. 33]: AS = 0.7854(D - 0.938p)2

(5)

where:As = tensile stress area (mm2) p = pitch of the threads (mm) D = nominal diameter of the fastener (mm) Once again, of course, we must subtract the area of the gage rod's hole (Ah) from AS, as in Eqs. (2) and (3), before using Eq. (1). If made carefully, by mechanics trained in the proper use of depth micrometers, the gage rod procedure provides a way to measure the current tension in a stud or bolt with an accuracy that is usually acceptable. It's a good idea to have them take and average several measurements rather than relying on the accuracy of a single one. A log must be kept of the initial end distance of the rod, of course, before the fastener is tightened, because we're interested in the change in length of the fastener. Ideally, the before and after assembly measurements are made at the same fastener temperature, since a change in temperature will also cause a change in length of the stud or bolt. This is no problem if the technique is being used to control assembly preloadits normal usebut causes a complication if we want to use it to inspect an in-service joint that is operating at something other than ambient temperature. If that's the case, we must subtract the temperature-induced change in length (DLt) from the change in depth micrometer readings before using Eq. (1). We can estimate the temperature induced change from [1, p. 317]:

Page 547

DLt = (T2 - T1)r where:

(6)

DLt = change in length of the bolt created by a change in temperature (in., mm) T2 = current temperature of the bolt (°F, °C) T1 = temperature of the untightened bolt (°F, °C) r = linear coefficient of expansion of the bolt material (in./in./°F) (mm/mm/°C)

You'll find coefficients of expansion tabulated in a handbook of physics or in Table 4.5 in ref. [1]. Note that temperature can create more than just a measurement accuracy problem for us. It can make an in-service measurement difficult and/or dangerous, and may make it necessary to take the joint out of service, temporarily, to inspect it. This is also true of some, but not all, of the techniques we're going to consider in this chapter. 3.2 The RotaBolt Change in length gives us a much more accurate measurement of preload or tension than does torque or other more common control means. As a result, a number of specially designed gage rod bolts have been placed on the market. We're going to look at two of them, starting with the RotaBolt, which is made in England. A cutaway view of this device is shown in Fig. 2. The gage rod, here called a gage pin, is designed to read the stretch of only a portion of the unthreaded body of the bolt. This pin is threaded into its hole,

but is loose for most of its length. Standard bolts and studs are converted to RotaBolts by installing a micrometer-like load indicator into the head of each bolt [6,7]. Each instrumented bolt is calibrated for tension, not change in length or gap, so we no longer must be concerned about Eq. (1) or its followers. A wide range of bolt sizes can be used. When a RotaBolt is first installed, the load indicator dial in the head of the bolt can be turned freely, thanks to a preset air gap that represents the desired bolt tension. When that tension has been reached, the dialcalled a Rota disccan no longer be turned because the gap, this time, has closed, not opened. If the bolt subsequently looses tension, thanks to embedment or elastic interactions or to temperature change or to any other phenomenon, the Rota disc will turn freely again until the tension has been reestablished. Tests suggest that the technique can control preload within ±5% of a preset value. A second version of the RotaBolt contains two gage pins, one apparently inside the other. This allows the operator to determine whether or not bolt tension is currently within a desired range.

Page 548

Figure 2 The RotaBolt, manufactured in England. This device is a gage rod bolt similar in concept to that shown in Fig. 1. This time, however, a permanent measuring device is built into the head of each fastener and is left there.

Note that the RotaBolt, unlike the gage rod bolt discussed earlier, cannot tell us how much tension exists in a bolt at the present time. It can only tell us whether or not that tension equals a desired preset value (or falls within a desired range). This, however, is all we usually wish to know. As far as temperature is concerned, the manufacturer can accommodate a reasonable range through proper calibration; but, since the indicator must be handled to recheck a bolt, extreme temperatures can once again cause difficulties and/or

require a temporary shutdown. 3.3. The MAXBOLT Another type of load indicating bolt is shown in Fig. 3 and 4. It is manufactured by Applied Bolting Technologies of Phoenix, Ariz., and is an improvement over

Page 549

Figure 3 Drawings showing the spring-loaded lever system used to measure bolt stretch in the MAXBOLT. Tightening (stretching) the bolt causes the gage pin to move away from the end of the bolt, allowing the indicating lever to rotate clockwise. If the bolt relaxes, the lever moves counterclockwise. Lever movement is monitored visually in a calibrated

scale mounted in the end of the fastener. (Drawings courtesy of American Bolting Technologies, Phoenix, Ariz.)

Page 550

Figure 4 A photograph of the MAXBOLT shown in Fig. 3. (Photo courtesy of American Bolting Technologies, Phoenix, Ariz.)

the Valley Forge and Bolt product shown in Fig. 9 of Chapter 12 [8]. This time a proprietary lever mechanism is used to measure the change in distance between the gage rod and the end of the bolt and to display it as bolt tension. A readout indicator permanently mounted in the head of the bolt can display the current tension in the bolt in any one of several ways: as a direct readout in pounds, as a percentage of proof load, or simply as an unidentified band of desired load. These load monitoring mechanisms are inserted into standard studs or bolts and are said to be able to indicate the current tension in the fastener within 3% of the design specification. Once

again each bolt is calibrated separately, and, like RotaBolt, the device can be built into a wide range of bolt sizes and materials. Sizes instrumented so far range from ¾ inch to several inches in diameter. Whenever possible, the lever mechanisms and gage rods are made of materials having the same coefficient of expansion as the fastener in which

Page 551

they're installed, to minimize readout errors caused by a change in temperature. As a result, the device has been used at temperatures ranging from -40°F to over 1000°F. Note that this time it is unnecessary to handle the bolt to determine the current tension, making in-service inspection at elevated temperatures possible. Early applications of this product, to prevent leakage from a gasketed joint, include large-section flange bolts on a ball mill in an Arizona copper mine, a gasketed joint in an autoclave used to process gold ore, gas duct joints working at 400°F in a direct reduction steel processing facility, and flanges on a last blast/ thermal simulator designed to seal explosive charges at White Sands Proving Ground in New Mexico. 4. Ultrasonic Measurements For several decades now it has been possible to monitor the current preload or tension in bolts and studs using ultrasonic techniques. Although other technologies are available, the most popular one uses the so-called time-of-flight (also called transit time) techniques to measure the time it takes an ultrasonic signal to travel from one end of the bolt to the other and to return. When the bolt is tightened, that time of flight increases because the bolt is getting longer and because the velocity of sound decreases when tensile stress is created in the bolt. The resulting increase in the so-called transit time can be interpreted, by a suitably programmed computer or microprocessor, to estimate the tension in the bolt. Once again a log must be kept of the original transit time, before the fastener was tightened; but currently available instruments have built-in memories to store such information.

The technique is capable of measuring the current tension within a fastener with state-of-the-art accuracy, but it is far from problem free. Operators must be skilled in the use of the equipment. The transducers that pump ultrasonic energy into the bolts and that also receive the echo signals must be carefully placed on the end of the bolt in the same place and way that they were placed when reading the untightened bolt. A change in temperature will affect the readings. The computer built into the instrument can accommodate this, but to do so it so it must be fed the correct thermal data, which includes the coefficient of linear expansion, the change in sonic velocity caused by a temperature change, and the current temperature of the bolt. Another temperature problem arises because of the physical limitations of the transducers. At present it's my understanding that these can't be used above about 300°F, although some work has been done on products that can be used at higher temperatures. Ultrasonic measurement of bolt tension is still an emerging technology, however, and, like most such, tends to advance fairly rapidly. You

Page 552

should contact current suppliers for the latest equipment specifications and capabilities [9]. Several different instruments are currently available. Most of these are battery powered and, therefore, portable. Typically they are about as large as a small oscilloscope. Two products that were available a few years ago are shown in Fig. 5. A further discussion of ultrasonic measurements can be found in ref. [1,6]. If you plan to reinspect the tension in bolts periodically, then it's best to use ultrasonics from the start, measuring the acoustic length of each bolt before assembly and storing this data in a log or in the instrument's memory system. You can then measure the present preload in a bolt by telling the computer which bolt you're looking at and placing the transducer back on the bolt. The instrument will immediately display the current tension in the bolt. All is not lost, however, if you decide only after assembly that you want to use ultrasonics to measure

Figure 5 Sketch of typical ultrasonic instruments available from the Bidwell Industrial Group of Middletown, Conn., a few years ago and still available in modified form.

Page 553

the preload in your bolts. To do this you first measure the acoustic length of each bolt before loosening any of them. Next you loosen one of them and use the instrument to measure the changethis time the lossin tension. You now retighten that bolt to its beforeloosening tension. This will presumably reverse any elastic interactions that occurred when you loosened that bolt, restoring the current preloads to the rest of the bolts in that joint. You repeat this procedure on a second bolt, then on a third, etc. [10]. Once you have determined the amount of tension in a large-enough sample of bolts, or in all of them, you can then retighten them, if necessary, using ultrasonics to control the process. One final note: Ultrafast, Inc., of Malvern, Penn., has developed a way to bond a low-cost ultrasonic transducer permanently to the head of a bolt, and has developed a hand-held powered assembly tool that automatically engages with this transducer as it tightens the bolt [11]. Production facilities are, I've been told, being built, leading to mass production that, it is hoped, will lower the cost of these bolts to a competitive level. The main thrust at present is toward smaller bolts, the kind used in automotive joints rather than petrochemical ones, but the basic technique could and may someday be used on larger fasteners as well. 5. Measuring Clamping Pressure Although it is very useful to know how much tension exists in the bolts or studs, we want this information only because it helps us determine the amount of force with which the joint members are being clamped together. This is the factor that determines the life and behavior of the bolted joint. And clamping force is usually

directly related to the combined tension in the bolts, though not always. Interference between the bolts and their holes can reduce the amount of force available to clamp the joint members. The weight effect discussed in Chapter 12 can also create a big difference between bolt tension and interjoint clamping force if joint members are misaligned or if their weight must be carried by some of the bolts. So in the best of all possible worlds we'd be able to measure not bolt tension, but the actual force that clamps the joint members together. Even better, in a gasketed joint we'd be able to see the distribution of contact pressure in order to satisfy ourselves that the entire gasket is properly loaded. A couple of techniques have been developed for doing this. Neither entirely answers our needs, but each can provide very useful information. Let's take a look. 5.1 Fuji Prescale Film Fuji prescale film is a pressure-sensitive film that is hand-cut to fit the joint and then placed between the joint members during assembly. The film turns

Page 554

color when pressure is applied to it: the greater the pressure, the darker the final color. The joint must be disassembled to recover the film and view the results, but the film provides a permanent record of the maximum pressure applied to each portion of its surface. A special hand-held meter, provided with the film, is used to measure the density of color at each point and to convert this information to pressure in kg/cm2. Depending upon which film is used, pressures ranging from a low of 5 kg/mm2 (7.1 ksi) to 1300 kg/mm2 (1846 ksi) can be recorded. There are two main problems with this product. First of all, the films can be used only at temperatures ranging from 41°F (5°C) to 95°F (35°C) [12,13]. This allows them to be used to test various assembly procedures, but rules out their use for many, if not most, in-service applications. The other problem, of course, is that the joint must be disassembled to read the film. This can tell us why an in-service joint was leaking, but only after we have shut the system down. Nevertheless, at present writing (late 1996) this is the only technique I'm aware of that maps the entire contact pressure distribution in a bolted joint. 5.2. UniForce Sensor The UniForce sensor, available from Force Imaging Technologies, Inc., of Chicago or from Camcar Textron in Rockford, Ill., can also be used to monitor contact pressure between joint members, this time on a continuous, in-service basis, without disturbing the joint. The sensor consists of a 0.003-in (0.076-mm) -thick sandwich of conductors. The two conductors are separated by a thin film of material whose resistance is lowered when it is placed under pressure. This change is read by special software and hardware

provided by the manufacturers, and can be displayed in several ways on a computer screen [14]. A single UniForce sensor is shown in Fig. 6. A group of them, plus the computer that reads them and records and/or displays the resulting data, is shown in Fig. 7. The standard interface package available from the manufacturer allows you to monitor up to eight individual sensors at once, at data acquisition rates of up to 1000 Hz. A more expensive, high-speed interface system reaches data rates of up to 20,000 Hz. In either case the accuracy of force measurement is said to be within ±1.5%. More elaborate sensor assemblies, in which several sensors are combined to monitor a number of locations in a single joint, are also available [15]. When this technology was first announced, by Fel-Pro, Force Imaging's parent company, it was made available in large, customdesigned panels able to measure the full distribution of contact pressure in a pressurized joint [16]. Because each panel had to be manufactured separately to fit a customer's

Page 555

Figure 6 A single UniForce sensor. The sensing element, 0.025 in. in diameter, lies near the outer end of the strip. It consists of a 0.003-in.-thick sandwich of conducting elements separated by a thin film of material whose electrical resistance is lowered when it is compressed. One or more of these sensors are placed, during assembly, in the joint to be monitored. The interface clamping pressure between joint members can be monitored by a computer that interprets the change in resistance of the sensors. (Photo courtesy of Force Imaging Technologies, Inc., Chicago, III.)

joint, the technique was not inexpensive, but it provided real-time data on full joint contact pressure. Unfortunately, however, it turned out (as I understand it) that the readout was not stable, but drifted with time. These products were, therefore, at least temporarily removed from the market. Hopefully they will one day return. The present sensor is used to measure contact pressure at any desired location within the joint. A group of sensors will, of course, measure it at many points. Each sensor measures the force created

on a quarter-inch-diameter sensing area. These forces can range from 0.5 to 1000 lb (2.2 to 4400 N), depending upon the sensor used. As a result the equipment can be used to measure contact pressures ranging from about 30 psi to over 20,000 psi (0.02 to 345 MPa).

Page 556

Figure 7 A group of UniForce sensors with the computer that reads them and displays the resulting clamping force information. (Photo courtesy of Force Imaging Technologies, Inc., Chicago, III.)

References 1. Bickford, John H. An Introduction to the Design and Behavior of Bolted Joints. (3rd ed.) New York: Dekker, 1995. 2. From a presentation made by Ralph Shoberg of R. S. Technologies, Inc., of Farmington Hills, Michigan, to the Bolting

Technology Council, Cleveland, Ohio, May 1991. 3. Unified Inch Screw Threads (UN and UNR Thread Forms). ASME Standard B1.1, New York: ASME, 1989.

Page 557

4. Oberg, Erik, and Jones, F. D. Machinery's Handbook. New York: Industrial Press, republished in new editions every few years. 5. Screw Thread Standards for Federal Services (Unified Inch Screw ThreadsUN and UNR Thread Forms). FED-STD-H28/2B; General Services Administration, Washington, DC, 20 August 1991 (for example). 6. John H., Bickford, and Nassar, Sayed, eds. The Handbook of Bolts and Bolted Joints. New York: Dekker, 1998. 7. Information supplied by RotaBolt, Ltd., Peartree Business Park, Dudley, West Midlands, England, 1996. 8. Information supplied by American Bolting Technologies, Phoenix, Ariz. 1996. 9. Information supplied by the Bolting Products Division of the Bidwell Industrial Group, Inc., Middletown, Conn., 1996. 10. Information supplied by Wayne Wallace of Applied Bolting Technologies, Ludlow, Ver., 1996. 11. McEnroe, Tony. Will impulse wrenches influence the assembly industry? Fastener Technology International (Dec. 1995): 54. 12. Brochure published by Fuji Film Co., Ltd., Tokyo, Japan, in 1987, and provided by G. T. Technology Company, Southfield, Mich., 1985. 13. Cavicchiolli, James A. Fuji prescaleA picture of pressure. SAE Technical Paper 850186, presented at the International Congress and Exposition, Detroit, Mich., February 25-March 1, 1985.

14. Information provided by Spike Schonthal of Force Imaging Technologies, Chicago, Ill., 1996. 15. Clamp control, Appliance Manufacturer (June 1995): 64ff. 16. Czernik, Daniel E., and Misczak, Frank L. A new technique to measure real-time static and dynamic gasket stress. SAE Technical Paper 910205, presented at the International Congress and Exposition, Detroit, Mich., February 25-March 1, 1991.

Page 559

14 Stopping Leakage of In-Service Joints PAT KEARNS Team, Inc., Alvin, Texas 1. Introduction In the past 20 years or so, the technology for stopping leakage of in-service joints has been widely developed and utilized to keep processing facilities on line and in production until a scheduled shutdown of the processing unit can be planned and accomplished. In-service leak repairs are generally considered to be temporary, and it is recommended by leak repair service companies that inservice repair materials be removed from the leaking component, such as a set of flanges or a valve bonnet, during the next scheduled unit shutdown. Permanent repairs can then be made, using conventional materials and methods, before the unit is again placed in service. It is not the intent here to cast doubt, however, on in-service repairs. When properly performed by trained technicians using proven designs and materials, these in-service repairs may last for years even though the units may experience

Page 560

several thermal operating cycles in the interim. Personnel safety and plant equipment safety must also be prime concerns when applying in-service repair technology. Many hazards are encountered that would not be obvious to the untrained or illtrained technician performing in-service repairs. Hazardous chemicals, high pressure, high temperature, and severely blowing leaks are just a few of the hazards encountered by in-service repair technicians. Generally speaking, every in-service repair is unique and must be planned accordingly in order to achieve a high degree of success at sealing the leak and to ensure that the leak will remain sealed even though adverse operating conditions may exist. To address effectively the many potential operating conditions encountered in industry it is necessary that varying designs for clamps and enclosures be available. It is also necessary that sealants be available that will perform effectively in a broad range of temperatures, pressures, and chemical environments. Finally, it is necessary that procedures by established and followed that address all technical aspects of the job as well as all safety aspects of the job. For obvious reasons this chapter on stopping in-service leaks will be limited to those gasketed joints that are encountered in pressurized systems in power plants, pulp and paper mills, petrochemical plants, chemical plants, and refineries. These gasketed joints will include line flanges, valve-end flanges, valve bonnet flanges, and inlet and outlet flanges for pumps, turbines, and compressors. 2. Flange Clamp Techniques

Although there are several flange types that exist in plants, only the more commonly encountered ones will be discussed in this chapter. The most common repair for flanges utilizes a flange clamp designed with a seal element in the inner bore of the clamp. The clamp ring is made in two halves, with ears for bolting the two halves together to form a clamp ring completely encircling the flange. The clamp then provides the means for injecting sealant as well as the means for retaining the sealant in place after it is injected. It is then the combination of the clamp ring and the injected sealant that effectively seals the leak and keeps it sealed to prevent an untimely shutdown of the operating unit. There are several clamp designs and also several seal designs that can be utilized by the clamp designer to address specific leaking conditions. Figure 1 depicts a typical flange clamp. Sealant would be injected through several injection ports around the circumference of the clamp. Clamps are designed as unfired pressure vessels to the requirements of Section VIII, Division 1 of the ASME Code [1]. Field technicians provide the designer with the dimensional data and process information needed. Many clamps are designed as unique, one of a kind. Standard designs that are premanufactured and stocked account for less than half of those utilized for inservice leak sealing.

Page 561

Figure 1 Typical flange clamp.

The flange gap dimension is frequently found to vary on a set of flanges. Flanges are sometimes found to be out of round. The clearances in the stud holes between the stud and the hole allow flanges to become misaligned during assembly. These contributing factors frequently require that a clamp be designed and manufactured specifically to fit a set of leaking flanges. To seal the leak successfully, a clamp must fit the flange well with no more than .010'' to .015'' clearance existing at any point through which the sealant may extrude during the injection process. 2.1 Tongue Clamp

As the name implies, the tongue clamp (Fig. 2) is made with a tongue that has specially machined teeth and grooves that is inserted between the flange faces

Page 562

Figure 2 Tongue clamp.

as the two halves of the clamp are bolted together. The teeth on the tongue are sharp pointed, and the points are intended to make metal-to-metal contact with the flange faces to effect a labyrinth that retains the sealant during the injection process. The counterbore on the clamp ring is larger than the diameter of the flanges so that a clearance is maintained between the outside diameter of the flange and the inside diameter of the clamp ring. These clamps are used for sealing flanges that have high-pressure or high-temperature leaks or when the severity of the blow is high. In order for this clamp to be effective, the gap between the flanges must be a minimum of 3/8" in order to accommodate the tongue with the teeth and the sealant injection hole. Note that when the clamp ring is installed, a closed sealant cavity is formed that may

now be injected with sealant to fill the entire flange gap area and all bolt holes as well. When properly injected, compacted, and cured, the sealant seals off the leak. Tongue clamps are a favorite among designers and technicians due to their ease of installation and high success rate.

Page 563

2.2 Packing Clamp The packing clamp, shown in Fig. 3, utilizes square, braided packing installed in a groove cut into the inner diameter (ID) of the clamp ring. The packing is compressed as the clamp halves are drawn tightly together. It seals against the outer diameter (OD) of the flanges as shown. Sealant is then injected and bridges against the packing. Note that the bore of the clamp ring does not contact the OD of the flange. This type of clamp is limited in use to a service pressure of 300 psig and less. It is not used on severe blows, because as the volume of the leak can force the packing out of the packing groove. The packing clamp is preferred on large-diameter flanges, since the packing compresses with less torque on the ear bolts than does any other type of seal. Compression of the packing is necessary to compensate for mismatch and surface irregularities on the flange OD. This clamp can be reused simply by replacing the packing material after each use.

Figure 3 Packing clamp.

Page 564

2.3 Tubing Clamp The tubing clamp (Fig. 4) is similar to the packing clamp except that instead of braided packing, metal tubing is used. The grooves are the same; for example, in. braided packing and in. tubing require a groove with the same dimensions. The tubing clamp is used on pressures above 300 psig; however, greater force is required to compress the tubing than is needed to compress packing. For this reason, softer copper tubing is used as a standard for temperatures up to a maximum of 750°F for the service and 650°F for the skin temperature. Stainless steel tubing is used above 750°F (650°F skin temperature) and is a standard in nuclear plants. Carbon steel tubing is rarely used but can be chosen when chemical compatibility with the process requires it. All tubing is silver-soldered into the clamp and is therefore able to withstand severe blowing leaks. As with packing, the size of tubing used depends on the size of the clamp and the maximum measured by the technician.

Figure 4 Tubing clamp.

Page 565

2.4 Tongue and Packing Clamp When both a primary and a backup sealing system are desired or needed, a tongue with crunch teeth designed to fit in the flange gap can provide the primary seal. Various types of seals can be used in combination with it to provide the secondary seal. In Fig. 5, packing is used to close on the flange OD and provide the secondary seal. 3. Flange Types on Which Clamps Cannot Be Used 3.1. Flat-Face Flanges with Full-Face Gaskets There is one type of flange that cannot be sealed using only a clamp at the flange OD a flat-face flange with a full-face gasket. If a flange clamp were installed on this flange to seal a leak at the gasketed joint, the leak could move to an adjacent bolt hole. The stud leak then could not be sealed by injecting additional sealant into the clamp, because sealant cannot travel from the clamp to the bolt hole due to the full-face gasket.

Figure 5 Tongue and packing clamp.

Page 566

The solution, as shown in Fig. 6, is to install cap nuts and to inject sealant into the stud holes to seal the leak. If sealant extrudes from the gasket area, cable can be installed by using a cable tensioning device to allow the sealant to bridge. 3.2. Lap-Joint Flanges and Total Flange Enclosure Another type of flange that requires special consideration is the lap-joint flange (Fig. 7). Since this flange is not attached to the pipe, when the leak is sealed at

Figure 6 Method for sealing a flat-face flange containing a full-face gasket.

Figure 7 A lap-joint flange.

Page 567

the flange OD with a clamp, it can move to the flange hub. A flange clamp seldom is successful at stopping a leak on this type of flange. The solution is to design a total flange enclosure with closure seals against the pipe (Fig. 8). To seal the leak completely, the entire cavity is injected with sealant. 4. The Wire-Wrap Method This procedure, illustrated in Fig. 9, has been widely used for many years and has proven to be an effective method for repairing flange leaks. It enables the technician to repair these types of leaks immediately and with materials on hand. It is an inexpensive repair and does not damage flanges or studs. Destructive techniques such as peening and drilling of the flange is not required. These techniques require extensive rework at turnarounds. 4.1. Features Cost effective Nondestructive Easily removed during shutdown Immediately repairable with materials stocked on service truck Proven effective as a repair procedure

Figure 8 Sealing a lap-joint flange by use of a total flange enclosure.

Page 568

Figure 9 A wire wrap can be used instead of a flange clamp to seal a joint.

4.2. Standard Applications Performed on flanges with

gaps or less

Performed on service pressures of 650 psig or less Performed on service temperatures of 650°F or less Performed on flanges with 30" diameters or less Steam, water, and air services It should be pointed out that these established limits are for the purpose of guaranteeing work. Many jobs have been successfully performed that vastly exceed the listed limits. For instance, this repair has been successful on flange gaps up to 1" wide and on service pressures as high as 2000 psig. Wire-wrap repairs have been performed on flanges as large as 72 in. in diameter, when the service pressure was very low. When the flange OD is greater than

24 in. It is recommended that a cap nut be installed on every flange stud, instead of every other stud. The additional injection points allow better sealant travel and require lower injection pressures. The wire-wrap repair is not recommended for severely blowing, high-pressure leaks. This type of high-velocity leak can cut through the wire in a very short period of time. In this situation a clamp would be better for sealing the leak than attempting the wire wrap.

Page 569

4.3. How the Cap Nut and Slotted Stud Work The first step is to examine all flange studs. If any of the studs appear to be badly corroded or damaged in any way, they should be replaced with new ones. The stud nuts are uniformly tightened. This is an important step because the studs frequently are loose, and tightening them will often reduce the severity of the leak. Uniform tension on the studs is also important during the sealant injection stage. This prevents the possibility of adding additional stress on the studs with sealant pressure. Uniform tension compresses (loads) the gasket and prevents sealant pressure from pushing sealant into the leaking orifice. If the stud size is or larger, it is advisable to use a torque multiplier or a hydraulic torque tool to tension the studs properly. Next, a nut is removed from one of the studs and is replaced with a cap nut. When installing cap nuts on studs, ensure that there is sufficient thread engagement. A cap nut is as strong as a hex nut only when the thread engagement is greater than the thickness of a standard heavy hex nut. A guideline would be to thread the cap nut onto the stud using all the available cap nut thread. Then install the hex nut on the other end of the stud and tighten. After the cap nut is tightened, another hex nut is removed and replaced with a cap nut. This procedure is repeated until cap nuts have been installed on at least half of the flange studs. In the case of a four-stud flange, however, it is necessary to install cap nuts on all four studs. Caution: When removing a stud from a four-stud flange, C-clamps/ strongbacks should be applied. Sealant enters the cap nut through the injection gun hose and the injection valve. As the sealant enters the capnut, it can travel to the

leak source by one of two ways: 1. Through the slot milled in the stud 2. Through the slot milled in the cap nut then via the space between the stud and the bolt hole ID. Sealant travels around the flange gasket OD and into the other bolt hole cavities, thus sealing the leaks Slotted studs are recommended when the stud gap is small or bridging aids such as extra fiber and steel wool are called for; see Fig. 10. Note: Under no circumstances may more than one-half of the flange studs be replaced with slotted studs. A slotted stud is slightly weaker than a standard flange stud. With the cap nuts in place and tensioned, wire is tightly wrapped in the flange gap, completely filling to the OD of the flange. The standard wrapping wire is .035-in.-diameter type 316 stainless steel. Other types of wrapping wire are used only when compatibility with the leaking process requires them. It is very important that wire tension be maintained during the wrapping procedure. Initially, the wire is tied off to a stud. As the wire is tightly wrapped

Page 570

Figure 10 Slotted studs are sometimes used in wire wrap procedures for reasons given in the text.

into the flange gap, it is kept in tension. Also, during the wrapping procedure the wire is tied off several more times, with a final tie in the wire when the flange gap is filled. If the wire is loose, the sealant pressure may push the wire out of the gap and outside the OD of the flange. Finally, the wire wrap is backed up with either banding or a cable tension device. This is done for two reasons: 1. As a safety feature in the event the wire should break 2. To prevent corrosive atmospheres from attacking the wire Injection of sealant is accomplished by utilizing the cap nuts. Sealant travels through the injection valve, the cap nut, and the stud hold into the flange gap area. The wire retains the sealant and allows the entire gap as well as all stud holes to fill with sealant. As the sealant is compressed, the leak is sealed.

The process of injecting sealant for this repair is the same as that previously described for a flange clamp repair. To recap, initially the gap would be filled with sealant injected at minimum pressure. This is accomplished by venting during injection and allowing the cavity to fill uniformly. To do this, injection begins at a point opposite the leak. Buildup of pressure would cause a shift to the adjacent valve (a first indication of sealant travel) and then a shift to the opposite side of the first valve injected, etc. Compaction of the sealant is accomplished by pressurizing all injection points after the initial curing has taken place.

Page 571

The job is completed when the Job Completion Tag is placed on the injection valve nearest the original leak. 4.4 Disassembly of a Wire-Wrapped Flange During a shutdown or an outage, the wire-wrap repair can easily be removed and the flanged connection restored to API or ANSI specifications by simply removing the wire and sealant and replacing the flange gasket. Disassembly of a wire-wrapped flange is accomplished by removing a nut or cap nut from a stud, doublenutting the other end of the stud, and then wrenching or unscrewing the stud from the flange. In this case the second nut acts as a jam or locknut, and by wrenching on the first nut or the one next to the flange the stud can be twisted out. This disassembly procedure should by fully explained to the customer, otherwise damage to the studs and possibly to the flange could result. 5. Flange Clamp Sealant Injection Procedure This procedure is initiated after the flange clamp has been installed and the ear bolts tightened. At this point all injection valves will have been installed and left open, allowing the sealant cavity to vent. Venting is done to minimize the existing pressure in the sealant cavity. As a result, since existing pressure does not have to be overcome, the injection pressure required to fill the cavity with sealant is minimized. The injection valve furthest from the leak will be the first one to be injected. As an example, if the leak was at 12:00 o'clock, the first injection would be made through the 6:00 o'clock valve (Fig. 11). This will allow the sealant time to cure partially before reaching the leak source. It is desired that the sealant be in a partially cured state

when it reaches the leak source so that possible intrusion or mainlining through the leak will be minimal. It should also be noted that by beginning to inject opposite the leak, the chances of trapping gas or pockets of service in the sealant is minimized. In our example (Figure 11), sealant would be injected at the 6:00 o'clock location until back pressure begins to build up. This is indicated by the gage reading and the increased force necessary to operate the gun handle. At this point the technician would stop injecting, close the injection valve, and move the injection gun to the closet adjacent valve that first gave indication of sealant travel. Sealant would be injected until a buildup of pressure indicates resistance to sealant travel. The injection gun would be moved again. This time to the valve on the opposite side of the first valve that was injected. Then alternate injecting valves until all have been injected and pressured.

Page 572

Figure 11 Plan view showing sealant being injected into a clamped flange.

In this manner the entire sealant channel is uniformly filled while utilizing minimum injection pressure. The sealant is then allowed sufficient time to cure partially and solidify. At this point the injection sequence is repeated, but this time significant gun pressure will be applied at each injection port to achieve compaction and compression of the sealant. Most sealants harden and shrink during the curing process, and sometimes gases are produced. As a result of this gassing off, the sealant may become honeycombed or porous. To combat this condition, pressure applied during the second injection phase consists of the application of significant gun pressure plus the line pressure, never exceeding the calculated maximum injection

pressure. This additional pressure is situational dependant and can vary from a few hundred to several thousand psi. Compaction of the sealant is necessary to keep the leak sealed. Each injection valve is drilled out and repressurized in the same sequence as in the original injection. It should be pointed out that even though we may be applying several thousand psi of gun pressure at each valve, this does not mean that this pressure is being applied over the entire circumference of the

Page 573

flange. Remember that the sealant filling the cavity is now cured or semicured. In this state it may transmit localized fluid pressure. The pressure now being applied is only acting over a small area in the immediate vicinity of the injection port. However, for the purpose of calculating the maximum injection pressure, it is assumed that the pressure will be transmitted to the full circumference of the flange as if the sealant were a hydraulic fluid. Each valve is pressured up, and the pressure is held for 3060 seconds. The injection valve is then closed while the pressure is being held on it. This will ensure the necessary compaction of the sealant cavity. This newly injected sealant is allowed to cure, and the job is checked for leaks. If none are found, the job is then completed by tagging the appropriate valve. 6. Preventing Sealant from Getting Into the Process Inadvertent injection of sealant into the process stream is a potential problem on any job. In the rare case where sealant pressure exceeds line pressure at the leaking orifice, sealant may be forced into the process stream. It must be emphasized, however, that hydraulic injection pressure is not sealant pressure. Because sealant is being pushed through various orifices and openings, sealant pressure will be less than the hydraulic injection pressure monitored at the injection gun. Remember that sealants are compressible materials and that pressure drops drastically as the distance from the point of injection increases. In most cases, the leaking orifice is very small, usually little more than a hairline crack. Sometimes, however, large leaking orifices occur. And when the process is very hot, the sealant may initially

thin before it has time to cure. When this happens there is a greater possibility of sealant getting into the process stream. There are a number of indications of mainlining and ways to prevent or minimize the possibility of extruding sealant into the process stream. These include, but are not limited to, the following twelve considerations. 6.1. Monitoring the Injection Pressure Either a hand-operated sealant injection gun or a pneumatic injection gun can be used. Both have a gage by which pressure is monitored to control the injection pressure. Pressure can also be felt on the injection gun handle as the sealant is being released. An experienced technician can tell when the sealant is traveling and when it is bridging. Tables 1 through 4 list the maximum allowable injection pressures for standard flanges at service temperature below 650°F (343°C). 6.2. Monitoring the Curing Rate Pressure on the injection gun handle indicates to the technician whether the sealant is curing rapidly or slowly. If necessary, the injection rate can be adjusted

Page 574 TABLE 1. Maximum Allowable Injection Pressures for Standard Flanges (below 650°F): 150-lb Rated Flange Wire wrap or tongue Crunch, packing, or tubing clamp clamp Nominal System pressure System pressure pipe @ 150 lb @ 275 lb @ 150 lb @ 275 lb size 6669 6625 3830 3805 5317 5264 3338 3165 1 4365 4303 2720 2682 3533 3429 2217 2152 2 3917 3801 2609 2532 2904 2791 1807 1737 3 3104 2924 1812 1706 4 4444 4249 2762 2641 6 5625 5306 3599 3395 8 3941 3578 2361 2144 10 5762 5406 3899 3658 12 3740 3411 2517 2295 16 4136 3823 2917 2696 20 6958 6442 4417 4090 24 6534 5994 4486 4115 To get a true maximum pressure, add the static gun pressure to the results from the table.

to correspond with the curing rate of the sealant to ensure that the sealant is in a cured state when it reaches the leaking orifice. The injection pressure and curing rate are thus monitored together to prevent sealant from entering the process stream. 6.3. Abrupt Drop in Injection Pressure An abrupt drop in injection pressure indicates possible extrusion

into the process stream. This would be the case only if no external extrusion is observed and the sealant cavity is nearly full. If extrusion is suspected, injection can be stopped to allow additional time for the injected sealant to cure. Injection can be resumed after confirming that adequate curing has taken place. 6.4. Injection Considerations 6.4.1. Distance from Leak Initial injections into a sealant-retaining device should begin as far away form the leak location as possible. This provides the sealant more time to cure before

Page 575 TABLE 2 Maximum Allowable Injection Pressures for Standard Flanges (below 650°F): 300-lb Rated Flange Wire wrap or tongue Crunch, packing, or tubing clamp clamp System pressure System pressure Nominal pipe @ 300 lb @ 450 lb @ 300 lb @ 450 lb size 5173 5132 3183 3158 5345 5305 3338 3313 1 4920 4867 3095 3062 4843 4771 3058 3012 2 7229 7101 4277 4202 7374 7260 4750 4676 3 6532 6382 4216 4119 4 5018 4837 2772 2671 6 4633 4418 3022 2882 8 4529 4289 3100 2935 10 6649 6363 4452 4261 12 6392 6088 4055 3862 16 5935 5666 4087 3901 20 5708 5365 3633 3415 24 5907 5565 3852 3629 To get a true maximum pressure, add the static gun pressure to the results from the table.

reaching the leaking orifice (Fig. 12). To eliminate the possibility of extrusion into the process steam, only cured material should reach the leaking orifice. 6.4.2. Hand-Operated Injection Gun The final phase of injection is always conducted using the handoperated injection gun. In this fashion the technician precisely

controls the injection pressure and the volume of sealant injected to finish the procedure. When the final injection phase is finished, to confirm that the sealant is cured the sealant should be drilled out of the injection valves. 6.5. Use of Injection Valves The injection procedure begins by opening all injection valves and leaving them open until sealant extrudes through them. The open injection valves allow the leaking process fluid to vent and be pushed out of the retaining device ahead of the injected sealant. This condition minimizes the pressure in the sealant retainer cavity. Consequently, the cavity can be filled with sealant at a minimal injection pressure. In the vast majority of cases, actual injection pressure required to fill

Page 576 TABLE 3 Maximum Allowable Injection Pressures for Standard Flanges (below 650°F): 600-lb Rated Flange Wire wrap or tongue Crunch, packing, or tubing clamp clamp Nominal System pressure System pressure pipe @ 600 lb @ 900 lb @ 600 lb @ 900 lb size 5090 5008 3132 3082 5265 5184 3288 3238 1 4815 4710 3029 2963 4698 4552 2966 2874 2 6373 6717 4126 3974 7146 6918 4603 4456 3 6231 5930 4022 3827 4 4883 4619 2984 2822 6 5827 5526 3754 3560 8 5528 5164 3573 3338 10 6113 5778 4186 3957 12 6699 6288 4824 4528 16 6326 5916 4633 4333 20 7622 7119 5565 5197 24 8313 7741 5971 5560 To get a true maximum pressure, add the static gun pressure to the results from the table.

the retaining device will be less than line pressure. This prevents sealant extrusion into the process. The sealant will then be allowed to cure partially before additional injection pressure is applied. The added pressure is needed to compact and compress the cured sealant. Additional compaction is necessary to negate the effects of shrinkage and porosity that might otherwise result during the curing process.

6.6. The Stopgap Plug or Covering To prevent extruding sealant into the process stream, the leaking orifice should be plugged or covered before installing a sealantretaining device (Fig. 13). The injected sealant then compresses the plug or covering more tightly against the leaking component, and extrusion through the leaking orifice is either prevented or severely restricted. This technique can be used effectively on piping and fittings, but has limited applicability on leaking flange gaskets. This procedure will be applicable, however, if the gap between flanges is large enough to install a stopgap covering of some type.

Page 577 TABLE 4 Maximum Allowable Injection Pressures for Standard Flanges (below 650°F): 900-lb Rated Flange Wire wrap or tongue Crunch, packing, or tubing clamp clamp Nominal System pressure System pressure pipe @900 lb @ 1350 lb @ 900 lb @ 1350 lb size 7028 6956 4524 4478 6362 6263 3896 3836 1 6827 6720 4154 4088 6601 6434 3998 3897 2 6362 6193 4217 4106 6209 6047 4337 4224 3 5454 5165 3599 3408 4 5978 5688 4223 4018 6 5273 4942 3764 3528 8 5233 4892 3777 3531 10 5094 4732 3823 3551 12 5331 4906 4048 3725 16 7135 6574 5222 4811 20 7060 7339 5755 5306 24 8647 8068 6128 5717 To get a true maximum pressure, add the static gun pressure to the results from the table.

6.7. Selection and Use of the Proper Sealant Careful, accurate selection and use of the proper sealant is critical to preventing extrusion into the process. The sealant should cure in the required amount of time in the process environment established (chemical, pressure, and temperature). It should also be able to travel, as necessary, through the sealant channel(s).

6.8. Sealant Bridging or Sealant Ability The bridging or sealing ability of the sealant is somewhat dependant on its flow characteristics. Sealant can be evaluated by the amount of pressure required to force it to flow from the injection gun and through an open injection valve (static gun pressure). Field mixing of a sealant may be conducted to modify and enhance its bridging or travel characteristics. 6.9. Controlling Sealant Volume The sealant volume can be estimated from the known volume of the cavity in the sealant retainer plus the compression ratio of the sealant to be injected.

Page 578

Figure 12 Initial injections should begin as far away from the leak location as possible.

Limiting the volume of injected sealant to only that which is necessary to fill the cavity will assist in limiting the amount of extrusion into the process stream. The calculation used to determine the required amount of sealant will depend on the particulars of the component that is leaking. Take, for example, an 8-in. flange that is rated at 300 pounds and has gap between the flange faces. It has an outside diameter of 15 in. and a raised-face diameter of An engineer can use these figures, plus the information on the stud size, stud hole diameter, flange widths, and sealant compression ratios, to compute accurately the volume of sealant required to seal the leak. Sealant injection can then be controlled and limited to the proper pressure and amount. If there is still some indication of leakage after the estimated sealant volume has been injected, the injection process should be reevaluated

before proceeding further. 6.10. Maintaining Continuous Injection Once the sealant injection process has begun it should be continued without interruption until completed. Stopping the injection process has the potential of leaving voids and pockets of the process chemical entrapped in the sealant. A

Page 579

Figure 13 Various methods of plugging a leaking orifice.

sealant-retaining device that has been partially injected with uncompressed sealant and then left overnight requires a much higher injection pressure to restart the procedure for sealing the leak than was originally required. As a result of the higher injection pressure, the possibility of extrusion into the process stream is greater. Another consideration is that sealant that has not been properly compacted and that is left saturated in the process fluid could react and result in total sealant breakdown. In addition, process fluid trapped in the sealant has the potential of honeycombing, flashing off, or creating other problems. If it flashes off, the resulting vapor pressure could be very hazardous. Again, once started, the injection process should not be stopped until completed. 6.11. The Perimeter Seal Enclosure

To keep sealant completely away from the leak source, the design of the sealant-retaining device may incorporate perimeter-sealing grooves. When these grooves are injected, the sealing is accomplished at the perimeter of the retainer. This eliminates the need to inject sealant into the central void cavity covering the leak orifice (Fig. 14). The retaining device now becomes a pressure vessel, and the pressure boundary is relocated from the piping component to the retaining device.

Page 580

Figure 14 Sealant is sometimes injected into a left-in-place retaining device rather than into the joint itself.

6.12 Limiting Reinjections The leak-sealing process stops leaks, but it does not correct any of the design or other basic problems that caused the leak in the first place. Typical problems include the integrity of the piping and its components and piping system movement and bending brought about by temperature fluctuations in the process or perhaps by components that were bent and forced into alignment during construction. Initially these leak-causing problems are easily resolved, but they will almost always occur again. To continually assess these problems, it is necessary to reevaluate any sealed joint that has required numerous reinjections. An initial reinjection of a sealed component may be necessary, because injection process problems were not obvious at first, or perhaps the sealant experienced some shrinkage during the curing process. If subsequent reinjections are needed, they should be examined for

causes outside the sealant process. A rule for flange systems practiced by the Central Electricity Generating Board of Great Britain is to examine the joint if more than four reinjections are required. Team policy recommends reevaluation after three. Plant rules or the nature of the process may warrant an earlier evaluation. If the evaluation determines that the integrity of the joint is acceptable and that the fasteners or flange studs are in good shape, it is possible to continue reinjections. In any event, an evaluation should be conducted if the need for

Page 581

reinjections continues, because the reinjections could be forcing sealant into the process stream or the component could be failing. Other factors, such as temperature, pressure transients, or cyclic system operations, may cause the leak to reoccur. In some instances, reinjection and resealing may be required as an ordinary maintenance practice. Such a practice would apply to repacking a motor-operated valve. 6.13. Comments When the preceding items are taken into consideration and included in the procedure for injection of sealants, the probability that a sealant will extrude into the process stream is extremely low, and in most cases, any extrusion is prevented. Care must be taken to ensure that unrealistic restrictions based on operation pressure are not placed on injection pressure. Remember, minimum injection pressure is static gun pressure plus one psig. It is possible to restrict injection pressure to such a low value that flow and proper sealant compaction will not occur. 7. Flange Clamp Problem Areas When dealing with flanges, problems can occur that require a cooperative effort between the technician and the design engineer. Figure 15 shows the problem and solution when the flange OD's are tapered. Valve end flanges in particular

Figure 15 Potential flange clamp problem: Flanges are not square, as shown on the left. Solution: Use a tongue and packing clamp such as that shown in Fig. 5.

Page 582

are frequently left as cast and are not machined. This often causes the flanges to be out of round or tapered on the OD surfaces. Flange mismatch, as shown in Fig. 16, is a commonly encountered problem due to the bolt hole clearances. When mismatch is in excess of .030 in. a tongue clamp will solve the problem. If insufficient gap exists for a tongue, then a clamp must be designed with offset diameters. This means that the two clamp bores are machined to different centers and are not concentric. Out of roundness of flange OD's is another problem that requires the use of a tongue clamp to solve the problem. If the out of roundness is not excessive, then a packing or tubing clamp will seal the leak. The out-of-roundness problem is illustrated in Fig. 17. Often, valve end flanges or pump flanges may be larger than the companion flange that is bolted to it. In this case, special clamps with different bores are frequently designed. This is shown in Fig. 18. In this case a strongback must be included in the clamp design. When the flange gap varies as shown in Fig. 19, the solution is to use a crunch teeth clamp, a tubing clamp, a packing clamp, or a tongue clamp with a specially machined tongue. Obstructions are generally a problem that can affect the design of the clamp. As seen in Fig. 20, the I-beam close to the flange necessitates the use of spread ears in the design of the flange clamp. 8. Flange Bolt Stresses Since the flange bolts are also located in the enclosed sealant cavity, it would seem that sealant pressure in that cavity would

produce an additional load on

Figure 16 Potential flange clamp problem: Flanges are mismatched, as shown on the left. Solution: Use a tongue clamp such as that shown in Fig. 2.

Page 583

Figure 17 Potential flange clamp problem: Flange OD is out of round, as shown on the left. Solution: Use a tongue clamp.

Figure 18 Potential flange clamp problem: Mating flanges have different diameters, as shown on the left. Solution: Use crunch teeth at different diameters, supported by a strongback, as shown on the right.

Page 584

Figure 19 Potential flange clamp problem: Flanges are misaligned, creating a variation in the gap between them (left). Solution: Use a packing clamp, as in Fig. 3, or a tubing clamp (Fig. 4).

Figure 20 Potential flange clamp problem: An obstruction prevents the use of a full clamp. In this sketch, for example, the flange is located too close to an I-beam. Solution: Use spread ears in the design of the clamp, as shown here, or use a three-piece clamp.

Page 585

the bolts and result in increased stud bolt stress. It would also seem that the magnitude of the additional load would be the product of the sealant pressure and the area into which sealant is injected. These conditions could actually occur only if (1) the bolts in the flange are not tensioned or preloaded at the time the sealant pressure is applied, and (2) the sealant is a true hydraulic fluid. Neither is true in actual practice. Since the bolts are preloaded the actual effect of sealant pressure on the bolt stress is negligible until the sealant pressure becomes high enough to produce a force nearly equal to the gasket preload. From that point, additional sealant pressure will produce a proportional increase in stud stress. When the flange is initially assembled and the bolts tensioned, with no pressure inside the line, the bolts are at preload conditions. As the line is pressurized, the total bolt load remains constant. The pressure inside the line causes the gasket pressure to decrease. In effect, the gasket load decreases by the same amount as the load produced by the internal pressure. The same thing occurs when sealant pressure is applied between the flanges. More gasket load is transferred to the sealant again, leaving the bolt load relatively unchanged as long as the bolt

Figure 21 Joint diagram showing the effect of line plus sealant pressure on the tension in the bolts.

Page 586

load conditions are not exceeded. Once the bolt preload conditions are exceeded, additional factors need to be evaluated. The foregoing has been verified in tests conducted form 1980 to 1982 and is the basis for computations to determine the maximum sealant pressure that may be used for injection without significantly affecting stud stress. A good approximation for computing this sealant pressure is as follows: BL = total bolt load = PiAi + PgAg - PsAs Pi = internal pressure Pg = gasket pressure Ps = sealant pressure Ai = inside cross-sectional area of pipe Ag = gasket area As = sealant area To determine the maximum sealant pressure (for a flange) that will not increase stud stress, the preceding equation is solved for Ps. To solve this equation, it is assumed that all of the gasket preload will be transferred to the sealant so the PgAg term in the equation will be zero. The equation becomes: BL = PiAi - Ps + As For a 6-in. 600-lb rated flange with a raised-face gasket, the following data and calculations can be made. Twelve 1-in. 0diameter studs at 30,000 lb each are used as bolts. Bolt holes are

in diameter. Flange outside diameter = Raised face diameter = Internal line pressure = Pipe inside diameter for Schedule 80 Bolt circle diameter = Bolt load = BL - total bolt load = = 12 × 30,000 = =

14 in. 8.5 in. 600 psig 5.76 in. 30,000 lb each bolt preload* 360,000 lb 15,640 lb = 97.2 in.2

Maximum sealant pressure (Ps) = *Up to point A total bolt load equals bolt preload.

The actual stress in the studs must be estimated to determine the total bolt preload. As a rule of thumb, based on NB3232.1, ASME Section III [2], the

Page 587

actual bolt stress is assumed to be two times the allowable stress shown in Section III, Table 11.3 for safety-related items. Using this value, the maximum sealant pressure is then determined. The calculated value is that amount of sealant pressure that would not appreciably increase the stud stress. It should be pointed out that for the purpose of the calculation, it is assumed that the sealant is acting as a hydraulic fluid pressurizing the entire circular cavity. Tests have been performed that verify that this condition does not occur for high-viscosity sealing compounds. The sealant injection procedure for all flange leaks calls for an initial fill of the sealant cavity with injection valves venting to minimize cavity and injection pressures. Once the cavity is filled, the sealant is allowed some time to cure. Compaction of the cured sealant is then achieved by applying a significantly high-injection pressure at each injection port on the clamp. Once cured, the sealant in the cavity will not transmit hydraulic pressure, so the calculation that is made on stud stress assumes the most conservative case. When high-pressure processes are involved, there are numerous cases where the on-line repair of a leak is not advisable. To do so would introduce the possibility of adding significant load to the flange bolts. When this situation in encountered, it is advisable to use a total flange enclosure instead of a flange clamp. When the total flange enclosure is used, there is no need to evaluate stud stress because sealant pressure in the enclosure cavity actually reduces stud stress and increases gasket compression.

References 1. ASME Boiler and Pressure Vessel Code, Section VIII. New York: American Society of Mechanical Engineers, 1989. 2. ASME Boiler and Pressure Vessel Code, Section III. New York: American Society of Mechanical Engineers, 1989.

Page 589

INDEX A Accuracy, torque vs. preload, 524, 525 Aging of gasket, effect on its behavior, 149, 150, 247-262 AHOT test (see PRVC tests) American Society for Testing and Materials (see ASTM) Anaerobic gaskets (see Chemical gaskets, Formed in place gaskets) ARLA test (see PVRC tests) Asbestos-free gaskets (see Nonasbestos gaskets) Asbestos gaskets, 2, 28, 29, 30, 143, 258, 312, 313, 384, 409 ASME (see also ASME Code, Gasket constants, and Standards) Special Working Group on Bolted Flanged Joints, 145 ASME Boiler and Pressure Vessel Code (see ASME Code) ASME Code Appendix 2, 89, 424, 425, 427, 447-456 Appendix S, 463, 465 Appendix Y, 490 historical notes, 143, 424, 425, 508

in general, 88, 89, 293, 403, 489 nomenclature used in, 427-430 ASME flange design (see Flange design) Assembly efficiency (see Efficiency, assembly) Assembly of gasketed joint (see also Accuracy, Efficiency, Elastic [Assembly of gasketed joint] interactions, Installation of gasket, Preload, Stress on gasket, seating) general discussions of, 13-15, 83-85, 333-336, 339-341, 358, 359, 523-539 minimum bolt tension for, 406, 407 potential problems in, 339-341, 377, 378, 524-532) ASTM standards (see Standards) ASTM, 38, 89 ATRS test (see PVRC tests) Automotive gasketing (see also Exhaust gaskets, Head gaskets, Intake gaskets, Manifold gaskets) in general, 2-85 types of, 49-53, 58-64, 69, 72

B Beater addition gaskets, 93, 94, 99, 100 Behavior of gasket (see also Mechanical behavior of gaskets, Properties of gaskets, Temperature, elevated, Thermal effects)

at elevated temperature, 128-130, 262-267 in general, 1-33, 147, 390, 391 long term, 23-32, 262-267 Behavior of gasket, mechanical for automotive gasket, 47 for pressure vessel gaskets, 147, 198-204, 404-406 Blowout of gasket causes of, 139, 140 in general, 139, 289-291

Page 590

[Blowout of gasket] testing for, 289-291 Boiler and Pressure Vessel Code (see ASME Code) Bolting area, computing, 436-438 Bolts, specifying, for gasketed joint, 368-370 British Standards (BS) (see Standards)

C CAF gaskets (see Asbestos gaskets) Chemical environment, effect on gasket, 3, 330-332 Chemical gaskets advantage of, 125-127, 129, 130, 134 design considerations for, 128-132 in general, 123-126 limitations of, 127, 128, 130 product selection, 135 Classification systems, gasket, 39-41, 443, 444 Code (see ASME code) Compressed asbestos gaskets (see Asbestos gaskets) Computer (see also KLINGERexpert®)

used for gasket calculations, 394-406 Constants, gasket (see Gasket constants) Costs of leakage, 303 Creep relaxation in general, 6-9, 26-28, 157, 159 long term, 270-280 short term, 157, 178, 182-191, 413 CRUSH test (see PVRC tests) Cured in place gaskets (see Chemical gaskets) Cylinder head gaskets (see Head gaskets)

D Design of flanges (see also Flange design, ASME Code) examples, 451-460, 469-472, 496-508, 510-515, 519, 520 Design of flanges in full-face contact Blach-Bazergui-Baldur method, 513-515 design curves for, 518 improved method for, 515-520 in general, 507-520 Design of gasketed joints (see also Design of flanges) assembly considerations during, 406, 407

design goals for, 404-406 example using VDI rules, 415-420 introduction to, 403-422 VDI procedure for, 406-420 Design of noncircular flanges equivalent circular flange method, 489-492 examples of, 496-505 frame bending flange method, 492-496 in general, 488-505 DIN (see Standards) Double jacketed gaskets, 90, 409 (see also Gasket, types of)

E École Polytechnique, 147, 294, 313 Efficiency, assembly, 445, 446, 467, 468 EHOT test (see PVRC tests) Elastic interactions, 413, 528-532 Embedment, 413 Emission leakage, 88, 139, 141-143, 288, 289, 303 End thrust, hydrostatic (see Hydrostatic end thrust) Engine gaskets, 56-76

Evaluation criteria for gaskets, 38, 39, 150-153, 280-292 Exhaust gaskets, 77-80 Expanded PTFE (see PTFE) Eyeleted gaskets, 50, 119

F Factors, gasket (see Gasket constants) Failure, reasons for, 92 Fiber gaskets, 8-10, 144, 254, 285, 287 Finish (see Surface finish)

Page 591

Fire, effect on gaskets and joint, 219-227, 324 FIRS test (see PVRC tests) FITT test (see PVRC tests) Flange insulating, 379, 380 thermal profile of, 373-375 Flange design Blach-Bazergui-Baldur method for flanges in full face contact, 513-515 for chemical gaskets, 128-132 for flanges in full faced contact, 507-520 for noncircular flanges, 488-505 goals for, 404-406 historical notes, 424, 425, 507-509 Taylor-Forge method, 510-513 (see also Taylor Forge) Flange design, ASME Code, proposed by PVRC examples of, 451-460, 469-484 general discussions of, 293, 340, 423, 424 graphical procedure for, 469-472 tightness-based rules, 431-447

Flange design, ASME Code, traditional design rules for, 430, 431 general discussions of, 143, 419-421, 423, 424 Flange design, for flanges in full face contact (see Design of flanges in full face contact) Flange parallelism, 526-528 Flange rotation, 11-14, 32, 336-337 Flanges in full-face contact (see also Design of flanges) Flatness of flange (see Surface flatness) Flexible graphite gaskets creep relaxation of, 8 degradation of, 382 gasket constants for, 441 general discussions of, 8, 105-109, 145, 313-315 predicting life of, 286-288 [Flexible graphite gaskets] properties of, 109 service temperature limits for, 383, 384 stiffness of, 409 thermal decomposition of, 30, 254, 256, 314, 315 Fluid Sealing Association (see FSA) Formed in place gaskets

advantages of, 125-127 application procedure for, 132, 133 applications of, 124 design considerations for, 128, 129 FSA, 32 Fugitive emissions (see Emission leakage) Full-faced gasketed flange (see Flanges in full-face contact) Function of gaskets, 38, 39, 90-93, 124, 390, 391

G Gas law, ideal, 174, 175 Gas viscosity vs. temperature, 177 Gas, properties of, 176 Gasket classification systems, 39-41, 443, 444 Gasket constants Gb, a, and Gs, 89, 227-230, 232-246, 307, 392, 393, 423, 425, 426, 436, 439-442 m and y, 89, 227, 392, 395, 416, 417, 425 Gasket factors (see Gasket constants) Gasket materials (see also Gaskets) asbestos (see Asbestos gaskets) asbestos free (see Nonasbestos gaskets)

CAF (see Asbestos gaskets) chemical (see Chemical gaskets) DIN 306, 385 fiber, 8-10, 144, 254, 285, 287 KLINGERit, 30 metallic, 90, 97, 98, 122, 441, 442 mica laminates, 116, 117 PTFE (see PTFE gaskets) rubber-coated metal, 109, 110

Page 592

Gasket, selection of, in general, 303-387 (see also KLINGERexpert®) Gaskets (see also Gasket materials; Properties of gaskets) classification systems for, 39-41, 443, 444 flowchart for selection of, 381 industrial, 87-122 installation of, 83-85 processing of, 98-122 PVRC identification codes for, 443, 444 requirements for, 38, 39, 90, 91, 124, 390-392 selection of, in general, 303-387 stiffness of, 409-411 Gaskets at elevated temperatures (see Temperature, elevated and Thermal effects) Gaskets, automotive (see Automotive gasketing) Gaskets, behavior of (see Behavior of gasket and Properties of gaskets) Gaskets, chemical (see Chemical gaskets; Formed in place gaskets) Gaskets, eyeleted, 50, 119 Gaskets, properties of (see Properties of gaskets) Gaskets, qualification guides for, 280-292

Gaskets, segmented, 117, 118 Gaskets, semimetallic, 120, 121 Gaskets, spiral wound (see Spiral wound gaskets) Gaskets, types of, 56-85, 89-90, 93-98, 123, 372 German standards (DIN) (see Standards) Graphite, flexible (see Flexible graphite)

H HALR test (see PVRC tests) HATR test (see PVRC tests) Head gaskets combustion sealing, 62-68 in general, 57-73 liquid sealing, 68-73 History of flange design, 424, 425, 507-509 History of gaskets, 1, 2, 36-38, 59-61 HOBT test (see PVRC tests) HOMT test (see PVRC tests) HOTT test (see PVRC tests) Hydrostatic end thrust, computing, 398, 399

I Inspection of in-service joint general discussion of, 541-556 by measurement of clamping pressure, 52-56, 554-556 traditional techniques for, 542-544 using special fasteners, 545-552 using ultrasonics, 552-554 Installation of gaskets (see Assembly of gasketed joints and Stress on gasket, seating) Installation stress (see Stress on gasket, seating) Insulation of gasketed joint, 378, 379 Intake gaskets, 76, 77

J Joint diagram, 47 (see also Behavior of gasket, mechanical) Joint surfaces (see Surface finish, Surface flatness)

K KLINGERexpert® softwear calculations made by, 401, 402 in general, 389-402 input data for, 396-401

L LCMT test (see PVRC tests) Leak, measurement of, 173-175 Leak rate allowable, 310, 315-317, 400, 401 calculating, 402 correction factors (vs. ideal gas flow),

Page 593

[Leak rate] 174, 175 measurement of, 173-175 vs. tightness parameter, 148-150 vs. type of fluid, 175-177 Leak sealing disassembly of wire wrapped flange, 570 effect on bolt stresses, 580-585 flange clamp techniques, 560-575 general discussion of, 559-585 potential problems with, 580-585 sealant injection procedures, 570-580 standard applications, 559, 560 wire wrap methods, 567-571 Leakage (see also Leak sealing, Emission leakage, Leak measurement, Leak rate, Tests, gasket) forecasting amount of, 402 general discussion of, 138-143, 400, 401 PVRC studies, 137-301 selecting acceptable level of, 304-311, 435, 436 standard, 305, 435

stopping, in-service, 559-586 Leakage behavior, testing for (see Testing gaskets) Leakage, stopping, in-service flange clamp techniques for, 560-575 in general, 559-586 potential problems with, 580-586 sealant injection, 571-581 when flange cannot be clamped, 565-567 wire wrap method, 567-571 Load factor (VDI), 407-411 Load-bearing characteristics, 15-23 Loads on gasket, 15-23, 330-332

M Manifold gaskets, 76-80 Manufacturing gaskets, 98-122 Materials, gasket (see Gasket materials, PTFE, Asbestos, Flexible graphite) Mechanical behavior of gaskets, 47, 147, 197-203, 404-406 Metallic gaskets, 90, 97, 98, 122, 409, 410, 441, 442 Mica foil gaskets, 116, 117

N Nonasbestos gaskets, 30, 93, 318-320, 383, 384, 386, 409, 410 Nonmetallic gaskets classification of, 39-41 fabrication of, 41-52

O Oil pan gaskets, 80, 81

P Parallelism, flange (see Flange parallelism) Performance of gasket (see Behavior of gasket) Preload (see also Assembly, and Stress on gasket, seating) control of, 524, 525, 532-538 maximum acceptable, 414, 415 minimum acceptable, 406, 407, 460, 461, 463 in service, 415 vs. torque, 524, 525 Preload, relaxation of (see Creep, Relaxation of gasket; Relaxation of gasketed joint) Pressure on gasket (see Stress on gasket)

Pressure Vessel Research Committee (see PVRC) Processing gaskets, 98-122 Production processes, 98-122 Properties of a typical gas (see Gas, typical: properties of) Properties of gaskets (see also Spiral wound gaskets, Flexible graphite gaskets; PTFE gaskets) biaxially oriented PTFE gaskets, 114-116 compressed material gaskets, 104, 105

Page 594

[Properties of gaskets] expanded metal gaskets, 104 expanded PTFE gaskets, 113, 114 flexible graphite (see Flexible graphite gaskets) material properties, 43 mechanical, 147 mica laminate gaskets, 117 PTFE (see PTFE gaskets) rubber-coated gaskets, 110 spiral wound (see Spiral-wound gaskets) PTFE expanded, 113, 114, 441 PTFE gaskets biaxially oriented, 114-116 creep/relaxation of, 8, 11, 189, 190, 332, 333 expanded, 113, 114 gasket constants for, 441 general discussions of, 11, 93-95, 110-113, 144, 145, 344 hot performance of, 270-280, 325 properties of, 111 service temperature limits for, 384

stiffness of, 409 thermal decomposition of, 30, 254, 256 tightness of, 270-280 PVRC tests acronyms defined, 151, 296, 297 AHOT test, 209, 210, 254, 272, 286, 315 ARLA test, 151, 183, 191-196, 247, 272, 283, 315 ATRS test, 151, 183, 186-191, 247-249, 254, 261, 280, 283, 285, 315 CRUSH test, 204-206 EHOT test, 151, 210, 213, 281, 315 FIRS test, 151, 280, 282, 283, 315 FITT test, 151, 280, 282, 283, 315 fixtures and procedures, 154-172 general discussion of, 137-301 HALR test, 192, 280 [PVRC tests] HATR test, 188, 193, 254, 256, 257, 283, 315 HOBT test, 151, 217-219, 271-275, 315 HOMT test, 151, 181, 182, 184 HORT test, 151, 213-217, 291 HOTT test, 151, 207-213, 254, 280, 283, 286, 315

LCMT test, 151, 184, 186, 292 leak measurement systems, 154-172 leakage studies, in general, 137-301 nomenclature used for, 298 quality parameters, 284, 285 ROMT test, 151, 181, 182, 184, 280, 292 ROTT test, 3, 32, 151, 196-204, 228-234, 280-282, 315 screening, 182-191

Q Qualification parameters, 284, 285 Quality criteria for gaskets, 2-4, 248-254, 284, 285 Quality parameters Ae, 254-257 mechanical, 248-253 Qp, 248-253 tightness, 253-254

R Relaxation of gasket, 6-9, 26, 27, 70, 158, 178, 181, 190-192, 270280, 326-330 (see also Creep relaxation) Relaxation of gasketed joint, 405-408

Relaxation of PTFE gaskets, 270-280, 327 Requirements of gaskets, 38, 39, 90, 91, 124 Rocker and cam cover gaskets, 81-82 ROMT test (see PVRC tests) Rotation of flange (see Flange rotation) ROTT test (see PVRC tests) Roughness (see Surface finish) RTV silicone gasketing (see Chemical gaskets) Rubber-coated metal gaskets, 109, 110

Page 595

S SAE (see Standards) Screening tests, 182-191 Sealing theory, fundamentals of, 390-394 Selection of gaskets (see Gaskets, selection of) Service life of a gasket, 28-30, 260-267 Service temperature (see Temperature, elevated and Thermal effects) Society of Automotive Engineers (see SAE) Solid metal gaskets, 97, 98, 122 Spiral-wound gaskets elevated temperature behavior of, 311, 312, 316, 317, 385 failures of, 360-366 gasket constants for, 442 general discussions of, 120, 121, 154, 452, 453, 455 service temperature limits for, 385 stiffness of, 409, 411, 412 tightness vs. gasket stress of, 311, 312 Standard leakage (see Leakage, standard)

Standards API 605, 410, 411 ASME B 16.20, 87 ASME B 16.21, 87 ASME B 46.1, 356 ASME Code (see ASME Code) ASTM D2000, 38-40 ASTM F 36, 150 ASTM F 37, 150 ASTM F 484, 150 ASTM F 586, 150 ASTM F104, 38, 39, 41 ASTM F38B, 17, 18, 26, 138, 393 BS 1560, 410, 412 BS 7531, 9, 17, 18, 26 DIN 2505, 390, 392, 509 DIN 28090, 3, 18, 20-22, 26, 27, 392, 400 DIN 28091, 20 DIN 3535, 150, 400 DIN 3754, 150 [Standards]

DIN 52913, 17, 26, 150, 393 SAE J2000, 38-40 SAE J90B, 38 U.S. Navy, for flange parallelism, 527 Stiffness of flange, 411, 413 Stiffness of gasket, 409, 410 Stiffness ratio, bolt to joint, 410-412 Stress distribution testing, 52-56, 554-556 Stress on gasket cyclic, 330-332 installation (see Stress on gasket, seating) measuring, 52-56, 554-556 vs. tightness parameter, 313, 315 Stress on gasket, seating (see also Installation stress, Assembly) crash point, 22, 23 critical, 23 crushing, 204-206, 247 in general, 229, 236-240, 307-309, 397 maximum permissible, 393, 384, 397, 402 minimum required, 391-393, 406, 407, 438 Stress relaxation (see Creep, Relaxation of gasket)

Stress, bolt, 460, 461, 582-586 (see also Preload) Stress, flange, 426, 451, 459, 460, 474-484, 492-495, 499, 501, 510, 515, 527, 582-584 Stress, installation (see Assembly, Preload; Stress on gasket, seating) Stress, maximum permitted (see Stress on gasket, seating) Surface finish, 74-76, 355, 356 Surface flatness, 74-76 Surface pressure (see Stress on gasket) Surface waviness (see Surface flatness)

T Taylor-Bonney (see Taylor-Forge) Taylor-Forge method of flange design, 455, 498, 509-513 (see also