• Author / Uploaded
• Jorge Luis Hernández

Citation preview

SimSci®

DYNSIM® 5.3.2 Distillation Column Modeling Guidelines

December 2016

All terms mentioned in this documentation that are known to be trademarks or service marks have been appropriately capitalized. Schneider Electric Software, LLC. cannot attest to the accuracy of this information. Use of a term in this documentation should not be regarded as affecting the validity of any trademark or service mark ActiveFactory, ArchestrA, ARPM, AssayCentral, Autonomous, Avantis, Avantis.PRO, Connoisseur, CrudeManager, CrudeSuite, DATACON, DATAPREP, DT Analyst, DYNSIM, DYNSIM Power, Esscor, EYESIM, Factelligence, FactorySuite, FactorySuite A², Foxboro, Foxboro Evo, FSIM Plus, HEXTRAN, Hot Links, I/A Series, IMPACT, IMServ, InBatch, InControl, Industrial Portal, IndustrialSQL Server, InFusion, INPLANT, InSQL, IntelaTrac, InTouch, IPact, Invensys, Logic Validator, MBM, NETOPT, PIPEPHASE, PRO/II, PROVISION, QI Analyst, ROMeo, SCADAlarm, SIM4ME, SimCentral, SimSci, SimSci-Esscor, Skelta, Skelta BPM.NET, Skelta Sharepoint, Skelta Sharepoint Accelerator, Spiral Software, SuiteVoyager, TACITE, Tactical Batch, The Simulator, TRISIM Plus, VISUAL FLARE, VISUAL FLOW, VISUAL FLOW SUITE, VISUAL NETWORK, Visual Solutions, VOYAGER, WindowMaker, WindowViewer, Wonderware, Wonderware InTouch, and ZI ZHI are trademarks of Schneider Electric SE, its subsidiaries and affiliated companies. © 2016 by Schneider Electric Software, LLC. All rights reserved. No part of this documentation shall be reproduced, stored in a retrieval system, or transmitted by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of Schneider Electric Software, LLC. No liability is assumed with respect to the use of the information contained herein. Although precaution has been taken in the preparation of this documentation, Schneider Electric Software, LLC assumes no responsibility for errors or omissions. The information in this documentation is subject to change without notice and does not represent a commitment on the part of Schneider Electric Software, LLC. The software described in this documentation is furnished under a license agreement. This software may be used or copied only in accordance with the terms of such license agreement. ArchestrA, Avantis, DYNSIM, EYESIM, Foxboro, Foxboro Evo, I/A Series, InBatch, InduSoft, IntelaTrac, InTouch, PIPEPHASE, PRO/II, PROVISION, ROMeo, Schneider Electric, SIM4ME, SimCentral, SimSci, Skelta, SmartGlance, Spiral Software, VISUAL FLARE, WindowMaker, WindowViewer, and Wonderware are trademarks of Schneider Electric SE, its subsidiaries, and affiliated companies. An extensive listing of Schneider Electric Software, LLC trademarks can be found at: http://software.schneiderelectric.com/legal/trademarks/. All other brands may be trademarks of their respective owners. Schneider Electric Software, LLC 26561 Rancho Parkway South Lake Forest, CA 92630 U.S.A. (949) 727-3200 http://software.schneider-electric.com/ .

Table of Contents Distillation ...................................................................................................................... 5 Column ........................................................................................................................... 5 The Column/Tower Model in DYNSIM .......................................................................... 6 Column (Legacy) ....................................................................................................................................... 6 Tower ........................................................................................................................................................ 6

DYNSIM Tower Model Building..................................................................................... 7 Introduction ............................................................................................................................................... 7 Step 1 – Thermo Specification .............................................................................................................. 7 Step 2 – Reduce Number of Components ............................................................................................ 8 Step 3 – Main Tower ............................................................................................................................. 9 Step 4 – Overhead System ................................................................................................................. 10 Step 5 - Reboiler Configuration ........................................................................................................... 12

An Alternate Method to Increase the Holdup on Each Tray..................................... 13 Data from PRO/II..................................................................................................................................... 14

Typical Tray Parameter Values ................................................................................... 14 Data Entry for Packed Bed Column ........................................................................... 17 Data from PRO/II..................................................................................................................................... 17

Typical Packing Parameter Values............................................................................. 20 Checklist....................................................................................................................... 20 What to Tune........................................................................................................................................... 20 Confirm the Following ............................................................................................................................. 20 Check after Modelling ............................................................................................................................. 21

Chimney Trays ............................................................................................................. 21 Chimney Tray Modeling .......................................................................................................................... 22

Product Draw Sumps/Side Draws .............................................................................. 22 Increasing Number of Trays ....................................................................................... 22 Tower Relief Study Hints ............................................................................................ 23 General Suggestions for Tower Modeling in Relief Analysis .................................................................. 23

Phase Separation and Level Calculations ................................................................. 24

Good to Know .............................................................................................................. 24

Disclaimer This document is based upon proven project work performed using earlier versions of DYNSIM® application. New capabilities introduced in DYNSIM 5.0 and 5.1 are not yet covered in the Best Practices documents. These documents will be updated in future releases of DYNSIM application.

Distillation Distillation is the separation of key components by the difference in their relative volatility or boiling points. It is also called fractional distillation or fractionation. Distillation is favored over other separation techniques such as crystallization, membranes, or fixed bed systems when; •

The relative volatility is greater than 1.2.

•

Products are thermally stable.

•

Large rates are desired.

•

No extreme corrosion, precipitation, or sedimentation issues are present.

•

No explosion issues are present.

Column A Column is a pressure node that is used to model distillation columns and fractionators. A Column unit consists of a vapor holdup on the top, for Column pressure calculation, and multiple tray sub-models beneath. A tray is a sub-model that represents an equilibrium stage within a Column. At least one tray must be present in a Column, although a realistic Column will normally have a much higher number of trays. The trays are linked with the vapor from the next lower tray and liquid from the next higher tray. The trays are numbered in such a way that the top tray is numbered as one, the next tray is numbered two and so on as you go down the Column. Each tray includes a liquid holdup to model the liquid inventory of the tray. A Column uses a theoretical tray approach, with an adjustment for liquid holdup, based on the ratio of actual trays to modeled trays that adjust the liquid holdup on each tray. Trays may also represent packed stages. For a packed stage, each tray represents a single equilibrium stage. The height of the tray or stage should represent the height of a theoretical packed transfer unit. Reactions can also be connected. Typically, reactions occur with packed sections of a column, but it can be added to plate stages as well. Multiple feeds and products can be connected to both plate and packed trays. Products from trays are optional although the top tray should have a vapor product and the bottom tray should have a liquid product. The tray sub-model contains only one liquid holdup. Vapor from the tray will go directly to the next higher tray or to the Column top vapor holdup, if it is a top tray. The Column includes the cylindrical section of a distillation column only. All peripheral equipment such as condensers, reboilers, accumulators, and side-strippers must be modeled with separate equipment models. The Column base can be modeled as the bottom tray. Alternatively, a Drum or a Separator model can be used to model the Column’s base. If the base has a partition for a thermo siphon reboiler, use a vertical Separator with a weir orientation. Column accounts for heat transfer from fluid to the metal and metal to surroundings. Column permits heat transfer from external sources directly to the metal or fluid through heat streams that can be connected to any of the tray’s liquid holdup.

The Column/Tower Model in DYNSIM Column (Legacy) A Column simulates distillation columns and fractionators. A Column model consists of a vapor holdup on the top, for Column pressure calculation, and multiple trays beneath. The tray contains only one liquid holdup. Vapor from the tray will go directly to the next higher tray or to the Column top vapor holdup, if it is a top tray.

Tower A Tower simulates distillation towers and fractionators. A tower model consists of a total holdup of vapor and liquid on every tray, unlike the column model where there is vapor holdup on the top tray. Furthermore, every tray is a pressure node in the pressure flow solver. Tower is a very robust model for stable startup and shutdown operation and is also very accurate for relief load calculations. A tray represents an equilibrium stage within a Tower. A minimum of one tray must be present in a Tower although a realistic Tower will normally have a much higher number of trays. The trays are linked with the vapor from the next lower tray and liquid from the next higher tray. The trays are numbered in such a way that the top tray is numbered as one, the next tray is numbered two, and so on as you go down the Tower. Multiple feeds and products can be connected to any tray. Products from trays are optional although the top tray should have a vapor product and the bottom tray should have a liquid product. Each tray includes a holdup to model the vapor and a liquid inventory of the tray. The Tower includes the cylindrical section of a distillation column only and all peripheral equipment such as condensers, reboilers, accumulators, and side strippers must be modeled with separate equipment models. The Tower uses a theoretical tray approach with an adjustment for liquid holdup based on the ratio of actual trays to modeled trays, which adjusts the liquid holdup on each tray. The Tower accounts for heat transfer from fluids to metals and metals to its surroundings. It permits heat transfer from external sources directly to the metal or fluid through heat streams that can be connected to the liquid holdup of any of its trays. The DYNSIM® tower allows for vapor holdup on all trays and based on project experience is the preferred configuration for a tower.

DYNSIM Tower Model Building Introduction Building a DYNSIM Tower may be difficult for the new user. If a PRO/II® simulator is available, using the PRO/II to DYNSIM translator will certainly help. However, it is often useful to know how to build a Tower using only DYNSIM application. The first time users often try to build the entire model and then perform a startup. While DYNSIM application can perform Tower startups, starting up a model that has never been operated at normal conditions is extremely difficult. Such Towers may have numerous sizing and configuration errors which must be identified one by one during the course of the startup. It is not clear if the Tower requires more time to reach steady state or if the Tower is not properly configured so the user attempts to be patient hoping for steady state. Towers should be built at normal operating conditions and then shutdown to perform a startup. A steady state heat and material balance is required before starting. If not available, it is recommended that a PRO/II simulation model is built first.

Step 1 – Thermo Specification Confirm the thermodynamic system that you wish to use before you start. If this is a new process or thermo system that you have not used before, request for a thermodynamic consultation with SimSci-Esscor® Technical Support. Confirm the thermodynamic methods using PRO/II application first. One of the primary reasons Tower models do not operate as expected is inappropriate thermodynamic and flash specifications.

Step 2 – Reduce number of components Dynamic simulation may require a reduced component slate, as column flash calculations tend to be CPU intensive. The best way to reduce the component slate is by using a corresponding PRO/II model. Adding or removing components from a configured DYNSIM model can be time consuming. Typically, PRO/II application uses a large number of Petro components in refining applications to accurately model distillation curves such as TBP, D86, and D1160. The end points are very susceptible to number of components. However, dynamic simulation seldom has this requirement for rigorous distillation curve calculations. Often, just matching tray temperatures is good enough. In PRO/II application, lump ideal components as required and decrease the number of TBP cuts. Continue to do this until the Tower Summary in the PRO/II output report starts to show unacceptable change. Then, back off back to an acceptable solution. Use this revised component set in the DYNSIM simulation. Typically, DYNSIM models can use 20 to 30 components and still have acceptable accuracy.

Step 3 – Main Tower Build the Tower model using constant feed, reflux, and "Boil up" as follows. Reflux can be taken from steady state heat and material balance. Boil up may be taken from a Pseudo product in PRO/II application. This approach allows the main Tower to come to the proper steady state solution without worrying about interaction with the overhead or reboiler equipment.

In the example given above, stream sets have been used to force flow in the beginning, flow control valves are used more frequently instead of the stream sets.

Step 4 – Overhead System Complete the overhead system but do not connect reflux yet. Be sure that the system is lined out and the overhead accumulator contents match reflux conditions. The following figure describes how such a model might look:

Confirm that the overhead product flow in this configuration matches the steady state heat and material balance reflux plus overhead product. Then connect reflux as follows:

Since the overhead system was already commissioned, it was connected with minimal disturbance to the actual Tower. Consider how difficult this would be if the reflux was added to the main Tower without commissioning the overhead system first. A poorly configured overhead system would cause severe disturbances to the main Tower which would affect the main Tower overhead vapor stream conditions. The Tower overhead vapor would affect the condenser and so on. It would be difficult to determine the source of the problem. You will have to wait long periods of time hoping the Tower will come to steady state. You may replace the stream set with a pump, valve, and flow controller as well.

Actually, it will be better to commission the reflux pump, valve, and flow controller while disconnected from the main Tower. Save an initial condition.

Step 5 - Reboiler Configuration On the reboiler, use an approach that meets your dynamic simulation application requirements. A simple way to configure the reboiler is with a Heat Stream directly to the bottom tray as follows. Remove the

Boil up source and StreamSet models. Then, commission a Utility Exchanger to calculate the heat stream duty as follows:

Save an initial condition. If a more complicated thermosyphon reboiler is required, use the same approach

used to commission the overhead system. Keep the Boil up source and StreamSet models. Add a Separator model to model the Tower sump with partition. Allow the model to run while commissioning the thermosyphon reboiler until the flow, temperature, and composition to Snk2 is very close to that of the Boil up. Then, connect the thermosyphon reboiler as follows:

Note that a thermosyphon reboiler has steam and condensate on the shell side. More detailed modeling may require a Drum unit for the steam/condensate side of the reboiler. The steam can be introduced on the flow control with the condensate released with a steam trap. The steam trap can be modeled with a level controller with the set point very close to the bottom of the drum. Steam pressure floats up and down to increase temperature-driving force, as required. Alternatively, the steam can be on pressure control with the condensate on flow control. The condensate level will float up and down, as required, to uncover the required tubes. In any event, a Drum model is required to simulate the steam/condensate side of the thermosyphon reboiler.

An Alternate Method to Increase the Holdup on Each Tray Redistribute the stage height to keep the column height constant. The feed port and product port elevations (the port elevation within the stage and the port elevation with respect to reference) have to be conserved. Also, redistribute the sum of all weir height to the theoretical stages. These steps will conserve the liquid holdup and vapor holdup. In this case, the Holdup Factor remains 1. When the throughput of the column is large, there may occasionally be numerical instabilities. In this situation, increasing the column volume often helps to stabilize the calculation. Do not increase the column volume more than 2 times the actual volume.

Data from PRO/II •

No of stages, (subtract reboiler and condenser stages)

•

Tray efficiency

•

Data from Drawing

•

T-T length, sump spacing, Empty weight

•

Draw-off / pump around height

•

Level transmitter tapping

•

Elevation

•

Port height

•

Data from datasheet

•

Weir height

•

Weir length

•

No of passes

•

Efficiency 0.1 to 0.01 for Chimney trays

•

Hole area fraction

Typical Tray Parameter Values •

Tray spacing – 600 mm (24 inch)

•

Weir Height – 50 mm

•

Downcomer clearance – 25 mm

•

Maximum width of a tray panel – 400 mm

•

Void Fraction is used to calculate holdup for every stage. Minimum value is 0.9 and maximum value is 0.98

•

Surface area is actually m2/m3 (1/m). In reality, this decides the HETP. Important in DYNSIM application during reactions. Minimum value is 64 and maximum value is 750

•

Avoid more than 18 stages per bed or a bed of height more than 6m in case of packed column

Keep default values for rest.

The basic tab in column DEW (Data Entry Window) is as shown below.

Data Entry for Packed Bed Column •

Correct number of stages. (Since the theoretical number of stages is taken from PRO/II application, stage efficiency is 1).

Data from PRO/II •

No of stages, (subtract reboiler and condenser stages)

•

Data from Drawing

•

T-T length, sump spacing, and Empty weight

•

Draw-off / pump around height

•

Level transmitter tapping

•

Elevation

•

Port height

•

Data from datasheet

•

Packing type

•

Specific surface area

•

Void fraction

Stichlmair coefficients (if available directly) else use the table given in the help file. Stichlmair Coefficients (C1, C2, and C3), Specific Surface Area, and Void Fraction for commonly used packings are available in the table. Taken from Stichlmair, J., Bravo, J.L. and Fair, J.R. , "General model for prediction of pressure drop and capacity of counter current gas/liquid packed towers," Gas Separation & Purification, Vol. 3 (March, 1989), pp. 19-28

The table displays the DYNSIM help file: Packing Type/size

a

void

C1

C2

C3

B1 300

300

0.97

2

3

0.9

B1 200

200

0.98

2

4

1.0

B1 100

100

0.99

3

7

1.0

2A

394

0.92

3

2.4

0.31

3A

262

0.93

3

2.3

0.28

Mellapak 250Y(plastic)

250

0.85

1

1

0.32

Mellapak 250Y (metal)

250

0.96

5

3

0.45

BX-packing

450

0.86

15

2

0.35

10

472

0.665

48

8

2.0

10

327

0.657

10

8

1.8

15

314

0.676

48

10

2.3

15

264

0.698

48

8

2.0

30

137

0.775

48

8

2.0

35

126

0.773

48

8

2.15

25

192

0.742

10

3

1.2

25

219

0.74

1

4

1.0

35

139

0.773

33

7

1.4

35

165

0.76

1

6

1.1

Reflux Rings

50

120

0.78

75

15

1.6

Hiflow Rings

20

291

0.75

10

5

1.1

Berl Saddles

15

300

0.561

32

6

0.9

35

133

0.75

33

14

1.0

20

300

0.672

30

6

1.4

25

183

0.732

32

7

1.0

35

135

0.76

30

6

1.2

25

255

0.73

19

1

0.85

50

120

0.75

10

8

0.75

12

416

0.94

60

1

7.5

15

317

0.924

40

1

6

25

215

0.94

0.05

1

3

Structured packing: Montz

Gempack

Sulzer

Dumped ceramic packings: Raschig Rings

Pall Rings

Intalox Saddles

Torus Saddles

Dumped metal packings: Raschig Rings

Pall Rings

35

130

0.95

0.1

0.1

2.1

Bialecki Rings

25

225

0.94

50

7

2.5

Nutter Rings

50

96.5

0.978

1

1

2.65

Cascade Mini Rings

25

230

0.96

-2

-2

2

Supersaddles

25

165

0.978

1

1.6

2.1

Pall Rings

90

71

0.95

-5

-4

2.3

NSW-Rings

25

180

0.927

1

1

1.35

Leva

1

190

0.92

1

1

2.0

2

143

0.94

1

1

2.3

Dumped ceramic packings:

The basic tab in the column DEW (Data Entry Window) for Packed Bed is as shown in the figure below.

Typical Packing Parameter Values Void Fraction is used to calculate holdup for every stage. Set Minimum value as 0.9 maximum value as 0.98. Surface area is actually m2/m3 (1/m). In reality, this decides the HETP which is important in DYNSIM application during reactions. Minimum value is 64 and maximum value is 750. Avoid more than 18 stages per bed or a bed of height more than 6 m in case of packed columns. Set default values for rest.

Checklist What to Tune •

Tune column conductance factor for pressure drop. (Suggested values range is from 0.5 to 1.5)

•

The reboiler and condenser conductance, and heat transfer coefficients for temperature profile. (For steps 4 and 5)

•

Tune the controllers in the following order Flow, Pressure, Level, and Temperature

Confirm the Following •

Reflux enters as liquid.

•

Boil up enters as vapor.

•

Tower pressure, temperature, and vapor and liquid flow profile.

•

Save an initial condition.

Check after Modelling •

MW / P and T profiles should match the PRO/II profiles.

•

There is some flow for either liquid or vapor in tower viewer.

Chimney Trays Chimney trays may be used as draw trays, transition trays, and/or protection against leakages. Use a Chimney tray if the residence time is required for pumping or start-up. This type of tray frequently maintains a fairly high level of liquid on it. The basic tab in the column DEW (Data entry window) is as shown below:

The selection of Chimney Tray in DYNSIM application •

Sets the stage feed port location at weir height

•

Sets the liquid weeping (drain) flow to zero

•

Modifies the vapor flow calculation based on the downcomer area instead of the holes area.

•

Bypasses all vapor flow from the stage below, but you can use StageEff to control the contact between vapor and liquid holdups. The bypass vapor goes to the vapor holdup directly.

Chimney Tray Modeling •

Setting stage efficiency to zero seems to make the model bounce around a bit at normal conditions. Increasing the efficiency settles the model down.

•

For the Chimney Tray, set Conductance Factor (KJ) to 0.001 so that there is very little vapor feed.

Product Draw Sumps/Side Draws Product draw sumps are required in several types of columns. These may be for partial or total draw of the liquid from the down comer. They are configured in DYNSIM application by setting the values for sump height (DRAWHEIGHT) and area fraction (DRAWAREA). This option is available only through the OEV window as shown in the figure below.

Increasing Number of Trays Occasionally, it may be necessary to change the number of trays after you have finished building the column. If you choose to reduce the number of trays, then simply delete the trays from the column, the removed trays will disappear and initial conditions for the rest of the trays will retain the configuration as before, but when you increase the tray number, the trays you added are set to ambient temperature and in equal mole percentage which will result in skewed calculation in DYNSIM application. In this case, you should run the DYNSIM model for a few seconds, pause the model and load IC, then run the model and

pause and load IC again. Normally, after three times, the newly added trays will be in good calculation condition.

Tower Relief Study Hints Some key simplifications can be made when building a model specifically for Tower Relief analysis. Note that this model is used for one specific purpose: to calculate the relief load. It will not be used for startup. Such modeling must be performed in accordance with API RP 521. Here are some such assumptions: It is not necessary to model product cooling. Most controls do not need to be modeled exactly as actual plant/design since they should be placed in manual at the start of the relief scenario. Credit may not be taken for automatic control action. Reflux can be simplified with just a StreamSet, since reflux will either 1) Operate at constant flow -because the controls would be placed in Manual. 2) Flow is set to zero for reflux failure.

We plan to develop a more systematic procedure for Tower relief simulation in a future document. If you have more questions, please contact Cal Depew at [email protected]

General Suggestions for Tower Modeling in Relief Analysis •

Use the same number of theoretical stages as PRO/II application.

•

Set efficiency to one.

•

Set holdup factor to the ratio of actual trays to theoretical stages.

•

Set aeration factor to 1 so as to have clear liquid on every tray.

•

Clear liquid height is a good assumption for relief analysis. This is only available through OEV.

Phase Separation and Level Calculations Each stage of the Tower makes use of InternalPhases, ExternalPhases, and ExternalPhasesStage to determine the type of separation it performs. InternalPhases of the flash object can be VLE, Free Water, VLLE, or Decant. ExternalPhases is a global parameter for a Tower and ExternalPhasesStage is a vector that is used to control the external phase option on each stage. ExternalPhases and ExternalPhasesStage can be TWO or THREE. To separate out the second liquid phase on a stage, both ExternalPhase and ExternalPhasesStage [stageXX] need to be THREE. If one of ExternalPhases and ExternalPhasesStage [stageXX] is TWO, there will be only one mixed liquid phase on stageXX. The liquid level calculations are independent for each stage. The bottom of each stage corresponds to the reference zero level. If the internal liquid draw sump exists, the bottom of sump corresponds to the reference zero level. The liquid level is calculated based on this reference level. The maximum liquid level in any stage is limited to the stage spacing.

Good to Know •

Paths determine the weir length and that determines the height of the liquid over weir. Set “other” for Paths, and specify the weir length fraction to get the desired height of liquid in the tray. For packed columns, you can use two or four paths if it does not cause instability problems.

•

Aeration factor controls the amount of vapor in the liquid phase. The higher the value, the more there is vapor in the liquid and hence the higher the liquid volume. Normally leave it as default.

•

For valve tray column, the weep vapor flow will be about 10 % of steady state flow. For sieve tray column, normally, it is 40% of normal vapor flow.

•

In the edit window of a tray, select all, which will enable the DPLiqFactor that can be used to scale the liquid static head. Use 0 for a chimney tray. The default is 1.

•

The whole area fraction determines the vapor flow delta pressure across the tray. The available whole area for liquid is for liquid draining.

•

When simulating a chimney tray, do not assign efficiency = 0. If the efficiency is set to zero, the pressure drop across the tray will be very high. For example, for T115TOP tray no.6, do not set efficiency to zero. Efficiency value of 0.001 has proven to work well.

•

When two column sections need to be integrated, generally an extra volume setting for the two columns is required.

•

All the columns should have a trend for all the controllers associated with the column. For example, pressure, level, temperature etc.

•

To configure the liquid side draw and separate out the second liquid phase, a) Define liquid side draw sump by providing DrawArea and DrawHeight through the OEV.

b) Set InternalPhases (from DEW) to Free Water, VLLE, or Decant, and set both ExternalPhases (from DEW) and ExternalPhasesStage [stageXX] (from OEV) to THREE.

Invensys Systems, Inc. 26561 Rancho Parkway South Lake Forest, CA 92630 United States of America http://iom.invensys.com Global Customer Support Inside U.S.: 1-866-746-6477 Outside U.S.: 1-508-549-2424 or contact your local Invensys Representative. Email: [email protected] Website: http://support.ips.invensys.com