• Author / Uploaded
• vmil

Citation preview

SUGAR SERIES Vol. Vol. Vol. Vol.

1. 2. 3. 4.

Standard Fabrication Practices for Cane Sugar Mills (Delden) Manufacture and Refining of Raw Cane Sugar (Baikow) By-Products of the Cane Sugar Industry (Paturau) Unit Operations in Cane Sugar Production (Payne)

sugar series, 4

unit operations in cane sugar production JOHN HOWARD PAYNE

ELSEVIER SCIENTIFIC PUBLISHING COMPANY Amsterdam — O x f o r d — N e w Y o r k

1982

ELSEVIER SCIENTIFIC PUBLISHING COMPANY Molenwerf 1 , P.O. Box 2 1 1 , 1000 A E Amsterdam, The Netherlands

Distributors for the United States and Canada: E L S E V I E R SCIENCE P U B L I S H I N G C O M P A N Y INC. 5 2 , Vanderbilt Avenue New York, N.Y. 10017

Library

of C o n g r e s s

Cataloging

in P u b l i c a t i o n

Data

Payne, John Howard, 1906Unit operations in cane sugar production. (Sugar series ; k) Bibliography: p. Includes index. 1. Sugar—Manufacture and refining. I. Title. II. Series. f TP37T.P38 1982 661* . 122 82-11373 ISBN 0-UUU-U210U-1

ISBN 0-444-42104-1 (Vol. 4) ISBN 0-444-41897-0 (Series) © Elsevier Scientific Publishing Company, 1982 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Scientific Publishing Company, P.O. Box 3 3 0 , Amsterdam, The Netherlands Printed in The Netherlands

ν

PREFACE This compilation is intended as a guide for the efficient performance of the several unit operations obtaining in a factory processing sugar cane. It is not a textbook, engineering handbook, or equipment operating manual. The purpose is only to present in simple form the basic principles involved and some of the means of reaching optimum results. The reader is assumed to be grounded in the fundamental chemistry, physics, engineering, biology, economics and common sense involved in sugar factory routine. He is also assumed to have had a reasonable length of experience in the industry. These practices are designed for an Hawaiian type of raw material. That is, two-year cane, rake harvested under both wet and dry conditions, requiring the use of wet cleaning plants. The product is a sugar averaging 99 pol, designed for delivery to a refinery. Although some of the Hawaiian conditions are not common, most of the practices are applicable to all cane sugar areas. The author salutes with respect his predecessors who developed the vast reservoir of knowledge which forms the basis of present technology. He thanks his contemporaries who have inspired as well as contributed. He greets the upcoming with the admonition that if they do not improve on this, we both have failed.

vi

ACKNOWLEDGMENTS

Acknowledgment goes to Amfac Inc. for initially suggesting the writing and close cooperation in the period of its development, to the Hawaiian Sugar Planters' Association for permission to use some unpublished material and to John W . Herkes who assisted in the preparation of the chapter in Equipment Maintenance. The comments of Hideo Idehara and J. R. Albert-Thenet are appreciated. Drawings are the work of James Kawamura. Specific recognition is gratefully extended to Misue Okino Sakamoto who assisted in editing and proofreading the manuscript and did much of the detailed assembly work.

vii

1

When Omer smote 'is bloomin

lyre,

,

He'd eard men sing by land an' sea; 1

An -what he thought 'e might

require,

Έ went an' took - the same as me. Rudyard

Kipling

Chapter 1

FACTORY CONTROL The monitoring, measuring, sampling and analytical procedures necessary for operational control and material accounting are outlined in the Schedule for Measuring, Sampling and Analysis at the end of this chapter. These are considered to be the minimum that will provide adequate information for all concerned in the factory operation. It is assumed that automatic devices are used wherever practicable. The laboratory load is reduced to a level where only one analyst is used per shift, exclusive of the analytical work necessary for field distribution and agricultural control. No provision is included for special analytical work of an experimental nature. The procedures are those of The Official Methods of the Hawaiian 1 Sugar Technologists. PROCEDURES Field Cane Received Measuring Weigh transport units at tolerance of ± 0.25%.

the

factory on a beam scale with a

Calibrate the scale annually, using known weights. Check tares each shift. No accumulation of dirt on the platform and dirt and water in the scale pit is permitted. Report the weight of cane received as the difference between the weights of the transport unit before and after discharging the cane. This weight is reported with the field area designation and the transport unit number. Sampling Core sample the cane on the transport. Equipment. A horizontal core sampling machine is used. The coring tube of 25 cm in diameter and revolves at 100 rpm. The tube enters 2 m into the cane through an opening provided in the transport approximately midway vertically in the load. Upon retraction, the sample in the tube is discharged into a conveyor leading to a 25-cm Rietz Prebreaker which chops the sample and discharges it into a revolving disc subsampler. This drops a portion into a subsample tray and discharges the remainder onto a disposal conveyor.

2

Procedure. The coring tube must enter the full distance into the load and the entire contents of the tube must go through the Prebreaker. The subsampler should approximately 250 g (0.5 lb).

be

adjusted

to

give

a

subsample

of

The entire subsample is placed in an airtight plastic container. Four cores from the sample area are composited in the container and constitute one sample for analysis.

same

Frequency. The number of cores taken is based upon the requirement to obtain a number of samples for analysis equal to the square root of the number of deliveries per unit area per week. Four cores are composited to make one sample for analysis in all cases where the number of deliveries exceeds four. The number of cores and samples needed is derived from the following table: No. of loads

Cores

Samples for analysis

1 2 3 4 5-16 17-30 31-42 43-56 57-72 73-90 91-110 111-132 133-156 157-182 183-210

1 2 3 4 5-16 20 24 28 32 36 40 44 48 52 56

1 2 3 4 4 5 6 7 8 9 10 11 12 13 13

900

120

30

Analysis Disintegrator Method for Fiber. Extract with the following variations:

Pol and Refractometer Solids %

1. The frozen sample is placed in a tared 12-gauge polyvinyl chloride (pvc) bag and weighed. The sample is broken up by beating with a mallet on the total contents of the bag (unless over 1000 g) are added directly to the disintegrator. The sample is not allowed to thaw prior to disintegration. With this precaution, mercuric chloride and sodium bicarbonate are not necessary. 2.

Pol is run on each sample.

3. Fiber is determined only on the basis of the square root of the number of samples, making sure that at least one fiber analysis is made on each area each day.

3 Prepared Cane Measuring Weigh prepared cane on a belt weighing device, with a precision of ± 2.0%, just ahead of the extracting unit. In diffuser installations, this is immediately following the Fiberizer. Sampling In the absence of a chute-gate time-cycle automatic sampling device, grab sampling is used. Samples are taken over a 5-minute period and held in a closed container until analyzed. No preservative is used, and the sample must be analyzed within 2 hours. Analysis Disintegrator Method for Fiber, Pol and Refractometer Solids, except that no sodium bicarbonate and mercuric chloride are used as preservatives. Stalk Cane Sampling 1. Cut and pull two stools of cane at benchmark locations in the area (not exceeding 12 ha). 2. Remove trash and weigh. Sample should include all the live stalks in the stools. A t this point, data on pest damage and cane condition may also be recorded. 3. Reduce the cane in the Prebreaker at the core sampling station, or in an ensilage cutter. 4.

Subsample the reduced material by quartering to approximately

500 g. 5. Place the subsample in an airtight container and preserve in a freezer until analyzed. Analysis Disintegrator Method for Fiber, Pol and Refractometer Solids % Extract. No preservatives are used. Bagasse Measuring Weigh on a belt weighing device, with a precision of ± 2.0%. Sampling Sample in the feed chute to the conveyor or to the pneumatic feeder. The sampling device consists of a time-cycle actuated, hydraulic slide gate which allows bagasse to drop into a container. The recommended timing is every 30 minutes, giving eight openings in a 4-hour period. The total amount collected should approximate 25 liters. The sample is taken to the laboratory and quartered to provide two 100 g samples.

4

Analysis Moisture Oven Method or Moisture Teller Method. Fiber Disintegrator Method. No preservatives are used. Pol Disintegrator Method. First Expressed Juice Sampling Collect separate samples from the crusher and the discharge roll of the first mill by hand, or by a continuous sampler. Analysis Crusher sample for Refractometer Solids and Pol. Purity of First Expressed Juice is calculated from these results. First mill sample for Refractometer Solids. This figure is reported for First Expressed Juice since it is less affected by extraneous water carried from the cleaning plant. Determine pH of the first mill sample. Absolute Juice Calculated from First Expressed Juice Pol, Refractometer Solids, Purity and Absolute Juice Factor. Analyzed as Juice in Prepared Cane Pol, Refractometer Solids and Purity. Last Expressed Juice Sampling Sample by hand from the last two rolls of the tandem, or from the dewatering device of the diffuser. Analysis Pol and Refractometer Solids. Press Return Measuring * Measure volume by means of a proportional weir at the point at which the press return enters the diffuser. Sampling Sample by hand from the press return weir box on top of the diffuser.

5

Analysis Pol, Refractometer Solids and Insoluble Solids. Mixed Juice Measuring Weigh on automatic beam scales with a precision of ± 0.1%. In order to maintain this precision, the rate of flow of juice into the scale tank must be uniform to insure a constant overshoot, i.e., the quantity of juice flowing into the tank after the scale is balanced and before the valve shuts off. Inlet and discharge valves are interlocked so that both will not be open at the same time. The scales are equipped with a recorder registering tank dumps. Tare is taken at least once a shift and the scales are calibrated annually and after any major maintenance. Sampling Sample automatically and proportional to the juice flow. In cases where hot diffuser juice is weighed, the sample must be protected from evaporation. Analysis Pol, Refractometer Solids, Insoluble Solids and pH. Filtrate Sampling Collect a sample from the filtrate return by hand. Analysis Pol, Refractometer Solids and pH. Clarified Juice Sampling Collect a sample from the evaporator supply tank by hand. Analysis Pol, Refractometer Solids and pH. Filter Cake Measuring and Sampling Cut measured areas of filter cake from two spots on the filter drum, just before the discharge blade. Weigh immediately. From the total filter area and the number of revolutions, calculate the total weight of cake discharged. Use the weighed samples for analysis. Analysis Fiber, Moisture and Pol.

6

Syrup Sampling Sample syrup continuously by collecting a drip sample from discharge of the syrup pump.

the

Analysis Pol, Refractometer Solids and pH. Massecuite Measuring Calibrate each pan for volume at various striking levels. Record the volume for each strike and estimate the weight from the calculated density. Sampling Collect a sample by hand from the flow of massecuite from the pan. Sample at intervals during the steady flow period of discharge. Analysis Pol and Refractometer Solids. Calculate Crystal content. Molasses, A , Β and C Sampling A - and B-molasses: Collect sample by hand from the centrifugal discharge trough as flow enters the pump feed tank. The sample should be taken at about the mid-point of centrifuging. C-molasses: Two samples are obtained, before and after the crystallizer. 1. Filter by pressure a portion of the C-massecuite obtained when the strike is dropped into the crystallizer.

sample

2. Filter by pressure a sample of the massecuite entering the heater after the crystallizer. Analysis Pol and Refractometer Solids. Molasses, Final Measuring Measure molasses in the storage at the factory with a Pneumercator with a precision of ± 0.5%. Calibration of Pneumercator: The average net area of tank is determined from circumferential measurements and the weight is calculated for a standard pressure of 2 2 0.71 g / c m (10 lb/in. ). The vertical height of the corresponding reading

7

on the manometer scale is then determined using a special calibrating jig. It is compared with the correct height of a column of mercury at this pressure to determine percent error. The manometer is then adjusted if necessary to bring the actual height into agreement with the correct height. Official weights are obtained from the terminal where molasses deliveries are weighed on beam scales with a precision of ± 0.1%. Sampling At the factory, sampling is by means of a continuous drip sample from the discharge of the pump handling the molasses to storage. At the terminal, sampling is conducted on each delivery. Analysis Samples collected Refractometer Solids.

at

the

factory

are

analyzed

for

Pol and

A weekly composite of the factory molasses is analyzed for Pol, Refractometer Solids, Sucrose, Reducing Substances and Ash, Carbonate. Purity Expected is calculated. The samples from the terminal are composited weekly and analyzed for Brix 1-1 dilution. Sugar - Remelt Sampling Sample continuously by collecting a drip sample from the discharge of the pump delivering the remelt to storage on the pan floor. Analysis Pol and Refractometer Solids. Sugar - Commercial Measuring Sugar is weighed at the terminal on beam scales with a precision of ± 0.1%. Sampling Factory: Each strike is sampled from the Composite for analysis on an 8-hour basis.

belt after

the

centrifugals.

Terminal: Each delivery is sampled automatically by a bucket sampler from the belt conveyor to the storage area. Analysis Factory and Terminal samples are analyzed for Pol and Moisture.

8 Condensate Sampling and Analysis Individual sources of condensate, evaporator cells, vacuum pans and juice heaters are monitored continuously by electrical conductivity. Response to high conductivity is to divert flow from boiler feedwater supply to house hot water supply - either automatically or manually. In the meantime, a sample should be collected and checks made with alpha naphthol to determine if high reading is caused by sugar carryover. A high reading may be caused by too low pH of clarified juice in the case of evaporator condensate. Feedwater Sampling and Analysis Feedwater is monitored continuously by electrical High conductivity water is discarded.

conductivity.

Boiler Water Sampling A sample is collected mannually every 8 hours from a sample line fitted with a condenser assembly. Analysis Alkalinity, pH, Phosphate Sulphite and Total Dissolved Solids. For high pressure desirable.

boilers,

Silica and Dissolved Oxygen are also

Instead of using Hawaiian Sugar Technologists Methods, standard procedures of companies specializing in boiler water control may be used. Condenser Water and Drains Sampling Condenser water from evaporator and pans, collection areas, are sampled manually once an hour.

and

drains

from

Analysis Colorimetric method for sugar. A suggested Schedule for Measuring, Sampling and Analysis is given in the following pages. Examples of Daily, Weekly and Recovery and Loss reports are also shown. REFERENCES 1

Payne, John H . , (ed.) Sugar Cane Factory Analytical Control, The Official Methods of the Hawaiian Sugar Technologists, Elsevier, Amsterdam, 1968 Revised Edition.

9

SCHEDULE FOR MEASURING, SAMPLING AND ANALYSIS Stream Cane Field Prepared Stalk Bagasse Final Juice First Expr. Absolute Last Expr. Mixed Filtrate Clarified Filter Cake

Measurement/Method Wt/Beam scale Wt/Belt scale

Core sampler Intermittent Grab

Wt/Belt scale

Intermittent

Wt/Beam scale

Prep, cane or calc. Grab Cont.(grab for insol.sol.) Grab Grab

Wt/calculated

Grab

Grab

Syrup Massecuite A, B, C

Sampling Method

Continuous Pan volume

Grab

Molasses A , B, C Final (Factory) Pneumercator Final (Terminal) Wt/Terminal scale

Grab Continuous Grab

Re melt

Grab

Sugar Factory Terminal Water Condensate Boiler Feed Boiler Water Drains

Wt/Terminal scale

Grab Continuous Monitor Monitor Grab Grab

10

Schedule for Measuring, Sampling and Analysis (continued) Analysis

Analytical Method

1/24 hr

Composite 4 cores Composite 1/4 hr Daily/area

Ref.Sol., Pol, Fiber Ref. Sol., Pol, Fiber Ref. Sol., Pol, Fiber

HST Pol Ratio HST Pol Ratio HST Pol Ratio

Bagasse Final

1/1 hr

Composite

Ref. Sol, Pol, Moist.

HST

Juice First Expr.

1/8 hr

1/8 hr

HST

Prep, cane or calc. 1/8 hr 1/4 hr

1/4 hr

Ref. Sol., Pol, pH Prep, cane or calc. Ref. Sol., Pol Ref. Sol., Pol, pH Insol. Sol. Ref. Sol., Pol, pH Ref. Sol., Pol, pH

HST

Stream Cane Field Prepared Stalk

Absolute Last Expr. Mixed

Sampling Frequency

Γ

\J Trucks/area/ day 1/1 hr

1/4 hr

Analysis Frequency

1/8 hr 1/4 hr

HST HST HST

Filtrate

1/24 hr

Composite 1/24 hr 1/24 hr

Clarified

1/8 hr

1/8 hr

Filter Cake

1/8 hr 1/8 hr

1/8 hr Composite 1/24 hr

Pol Fiber, Moist.

HST HST

Syrup

1/8 hr

Composite 1/8 hr

Ref. Sol., Pol, pH

HST

Massecuite A , B, C

Each strike

Each strike

Ref. Sol., Pol

HST

Remelt

1/8 hr

1/8 hr

Ref. Sol., Pol

HST

Each strike

Each strike

Ref. Sol., Pol

HST

1/8 hr

Composite 1/8 hr Composite weekly

Ref. Sol., Pol

HST

Ref. Sol., Pol, Sue, Red. Sugar, Ash, Brix 1-1

HST

Molasses A , B, C Final, Factory Terminal

Each truck

HST

HST

11

Schedule for Measuring, Sampling and Analysis (continued)

Stream Sugar Factory Terminal

Sampling Frequency

Analysis Frequency

Each strike

Composite 1/8 hr Composite weekly

Pol, Moist.

HST

Pol, Moist.

HST

Monitor Monitor 1/8 hr

Conductivity Conductivity pH, Alk., Phos. Suif., TDS Sugar

Cond. Cond. Boiler Control

Each truck

Water Condensate Boiler feed Boiler Water 1/8 hr Drains

1/24 hr

1/24 hr

Analysis

Analytical Method

HST

12 DAILY

FACTORY

REPORT

AMFAC FORM 1045B 2/71 Data & Tim* Started

[Data & Time E n d a d ( S I G N E D

PERFORMANCE PREPARED CANE NET lAVAIL.

DELAYS

(HOURS)

PRODUCTION

TIME

EXTRACT.

BOILING

POWER

EXTRACTING

AVAILABLE

PLANT

HOUSE

[GENERATION

TIME

TIME

SOLIDS REF. FIELD CANE

REC

IPREPARED CANE [STALK FILTER

CANE CAKE

1ST E X P . JUICE [LAST E X P . J U I C E ICLARIFIED IMIXED

JUICE

JUICE

[FINAL MOLASSES

DATE SAMPLED

FIELD NUMBER

Fig.

FIELD CANE RECEIVED

1-1. Daily factory report.

PREPARED CANE

13 WEEKLY FACTORY AM F AC FORM 1613 11/70

REPORT

FACTORY

FIELD

CANE

REPORT NO.

WEEK ENDING

SOLIDS REF.

PURITY

SIGNED

RECEIVED

TONS

PREPARED

TONS POL

POL %

FIBER %

CANE

TONS

TONS POL

SOLIDS REF.

POL %

FIBER %

This Week To Date EXTRACTING Extraction

CANE DILUTION %

Abs. Juka Prep.Cane|

Fiber

This Week

TONS PER HOUR E DAYS HOURS ieldCane Cetondar) Fiber p>rep.Cane

HOURS LOST

%

AREA HARV.

V A R I E T I ES %

%

%

I

To Date BAGASSE

DELAYS

TONS

POL %

TONS POL

Moisture %

FIBER %

NO CANE

Acres

EXTRACTING TIME

(HOURS) EXTN. Cleaner PLANT.

SOILING POWER GEN. HOUSE

HOURS Effciency AVAIL. I %

MISC.

This Week To Date MIXED

CLARIFICATION AGENTS

JUICE

GROSS TONS

NET TONS

TONS SOLIDS REF.

SOLIDS REF.

TONS POL

POL %

PURITY

Insoluble Solids %

LIME TONS

MgO TONS

Soda Ash TONS

This Week To Date JUICES FIRST EXPRESSED iofdsRef

POL %

LAST EXPRESSED

IPURITY Saids Ref. POL %

CLARIFIED

PURITY Saids Ref POL %

FILTERED PURITY Sofds Ref PURITY

ABSOLUTE (PREPARED CANE) Sofds Ref. P O L *

PURITY FACTOR

Γ I

This Week

I

To Dete SYRUP

pH

SOLIDS REF.

POL %

First

PURITY

FILTER Mixed J. Liming

Evap. Supply

SYRUP

CAKE

TONS

TONS POL

POL%

Moisture FIBER % X PREP. % SOLIDS CANE

This Week

I

To Date STRIKE S

FUEL A

Β

MASSECUITE eu. ft./ t. sua.

Sofds Ref PURITY

LOW GRADE

MASSECUITE MOL. PURITY Sofds Ref PURITY

J

?m

This Week

MOL. PURITY

MASSECUI ΓΕ eu. HJ

Sofds Ref. PURITY

MOL. PURITY

Remelt

Barrels

Purity

v#n Π

t

To Date F I N A L M O L A S S ES TONS

TONS SOLIDS REF.

TONS SUCROSE

SOLIDS REF.

TONS POL

POL %

Sucrose %

Per 100 Solids Ref. Sucrose Pol Pur RJSJAah Purity tfeleeded) R.S. ASH

HSJ Cond.

This Week To Date SUGAR

MANUFACTURED

TONS COM'L

TONS 96 DA

FINAL MOLASSES T O TERMIN AL

This Week Previously To Date

Fig.

1-2. Weekly factory report.

Moisture DET*N % FACTOR

TONS

BRIX 1 - 1

TONS 85 BRIX

Purity Above Expected

14 RECOVERY AND LOSSES REPORT STOCK

O N

Solids Ref %

H A N D

Purity

Pol %

C u Ft

W t Per C u Ft

X X X

X X X

Tons Solids R e f

Tons

Tons Pol

1 Juice 2 Syrup 3 Molasses 4 Remelt 5 Massecuite 6 7 8 Totals a n d A v e r a g e s RECOVERY

O N

STOCK

9 Pol

m

i(»- ) Tons

Tons Solids R e f

Solids R e f %

Pol %

Purity

Solids R e f %

Pol %

Tons

Tons Solids R e f

Tons Pol

10 Sugar, C o m m e r c i a l 1 1 Molasses FINAL

MOLASSES

12 W e i g h e d For C r o p / P e r i o d 13 A v a i l a b l e in Procoss This D a t e 14 Subtotal ι 15 In Process Beginning of C r o p / P e r i o d 16 Total D u e For These

Weeks Tons

BALANCE ENTERED I N T O M A N U F A C T U R E : 17 Cane, N e t 18 M i x e d Juice 19 A v a i l a b l e in Process Beginning ο

Crop/Period

X X X X X X

X X X

X X X

X X X Tons Pol In C a n e Out In

X X X

Pol %

X X X

2 0 C o m m e r c i a l Sugar P r o d u c e d 21 A v a i l a b l e in Process This D a t e 22 Loss in Manufacture; 23 Totals LOSSES

X X X X X X X X X

X X X

X X X X X X X X X

X X X X X X X X X

Tons

Pol %

Tons Pol in Mixed Juice Out In X X X

X X X X X X X X X

X X X X X X X X X

X X X X X X X X X Total Tons Pol

7.

24 Bagasse

%

Boiling House Tons Pol

X X X

X X X

25 26 27 28 29

Filter C a k e Molasses Undetermined Sum or T o t a l

X X X X X X

X X X X X X

RECOVERY 3 0 Total a n d Boiling H o u s e

m

31 A v a i l a b l e , s - j - m

*(i- ) i(s-m)

_

( I

) )

X X X

32 A c t u a l % of C a l c u l a t e d

X X X Period

GENERAL

To Date

33 Tons Sugar Produced a n d A v a i l a b l e 34 Tons 9 6 D A Sugar Produced a n d A v a i l a b l e

AMRAC FORM 1907 3/76

R & L Report N o .

.Factory .

For

the_

.Weeks

ending.

_I9 _Chemist

Fig.

1-3. Recovery and losses report.

15

Chapter 2

CANE CLEANING The essential steps in effective cane cleaning of pushrake harvested whole stalk cane are: 1. 2. 3· 4.

Thinning Rock, gravel and sand removal Washing Fibrous trash removal

A schematic diagram of a typical cleaning plant is shown in Figure 2-1. THINNING A thin mat of cane, 2 to 3 stalks thick, is necessary to obtain good cleaning. Thinning action is usually effected with a carding drum over which the cane passes. The drum, fitted with tines 30 to 40 cm (12 to 16 in.) in length, is set somewhat above and beyond the carrier and turns in the direction of cane flow at about 300 m (1000 ft) per minute peripheral speed. Carding is reasonably good if the cane is not presented in large bundles, some of which tend to be thrown over the top of the drum without carding. Partial preliminary load distribution is therefore necessary. An alternative to the carding drum is a tumbling conveyor. This is of pipe-slat construction hanging in a catenary arc of 55°. The angle is such that cane in some depth starts tumbling back near the top of the conveyor and will only pass over the top when the blanket is thin enough to give a low center of gravity. This type of action becomes less effective with unburned cane where tops and leaves bind the cane together. ROCK, GRAVEL AND SAND REMOVAL Rock, gravel and sand are the most deleterious materials in rakeharvested cane. Only by an hydraulic bath can an acceptable separation be made. Two types of equipment are suitable. In a velocity bath, a uniform rapid flow of water carries the cane in a flat arc through the bath to a drag-slat conveyor. Rock, gravel and sand (also tramp iron), being of higher density, drop in a steeper arc through the water and fall onto a conveyor in front and at right angles to the cane conveyor and here are conveyed to a truck for disposal. The importance of a thin blanket of cane is apparent, as a bundle of cane will raft the heavier materials across with the cane. In a sink-float bath a slurry of mud and water is maintained at a density that the cane will float and the dense materials sink. Mud density

Fig.

2-1. Cane cleaner.

16

17 3

3

should average about 1200 k g / m (75 l b / f t ) . Cane fed into the bath is floated across by some flow of slurry assisted by a revolving dunking drum. The same kind of cane and rock conveyors as outlined previously are used, and, as in the case of the velocity baths, a thin cane blanket is necessary to prevent rafting of rocks on the cane. WASHING Washing takes place initially on a drag-slat conveyor coming out of the hydraulic bath. The cascade principle is used with a large volume of water added near the top of a solid deck conveyor cascading downward in a turbulent flooding flow sweeping the soil out through a slotted deck above the water line of the bath. The cascade conveyor commonly has an angle of 40° and is run at a speed averaging 50 m/min (150 ft/min). FIBROUS TRASH REMOVAL The quantity of fibrous trash - tops, leaves and ground trash - is reduced by means of detrashing rolls. There are two types, Olsen and pipe or collared rolls. Olsen rolls are the most effective. These are constructed in banks consisting of several pairs of rolls made up of rectangular plates welded to a shaft. The banks are set at an angle 45° to 50° and the cane slides down over them by gravity. On revolving countercurrent to the cane flow at speeds in a range of 120 to 200 rpm, the corners of the plates enmesh, drawing leafy material down through and discharging onto a belt below, while the cane slides to a conveyor. Water jets help to push the trash between the rolls. Banks of Olsen rolls are often separated by single pipe rolls. Collared rolls may be used both before and after the hydraulic gap. Collared rolls are pipe rolls with welded circular discs which enmesh. They are set in banks as with Olsen rolls, but at a lower angle, usually 30°. As the cane slides over the rolls, water jets force loose material through the rolls while the cane travels over them. The rolls turn with the flow of cane, thus propelling it. Speed is of the order of 60 to 180 rpm. Effective trash removal requires that the cane pass over the detrashing rolls at single stalk depth. Thus, the cane blanket must again be thinned after cascader washing. This is done by means of combing drums which not only spread out the cane but also distribute it to the separate banks of detrashing rolls. Combing drums, 4 to 10 in number, are set on an upward slope of 20° to 30°. They are of relatively large tip diameter up to 1.8 m (6 ft) and are fitted with enmeshing serrated rings of steel plate. Rotating in the direction of the cane flow at progressively increasing tip speeds, they propel the cane with a bouncing action. Spacing is such that short cane will fall between the drums whereas long cane goes over the top of the last drum. By arranging the chutes below, distribution to the sets of

18

detrashing rolls is made. Jets of water are used in additional washing of the cane. OPERATION Cane cleaning is the most difficult of all factory operations to control and maintain. The one major problem is that of keeping a continuous, thin blanket flow of cane. Success with that gives good cleaning. Without it, there is poor cleaning and lost time caused by jams and equipment failure. The initial dumping of large loads of cane on the feeder table is the cause of most of the problems in distribution. The only reasonably satisfactory solution to this is one in which the cane is raked from the feeder table to the first cane conveyor. This is best followed by double carding, one after the first carrier and one ahead of the hydraulic bath. Without a raking system, care must be taken in feeding from the table to the conveyor to give partial distribution. From the carding drum on, the key is continuous steady flow. The necessity of keeping the clean cane conveyor full too often leads to a stop-go operation in which the cane may be flowing only half the time. This means that the cane blanket averages double the intended and the cleaning efficiency suffers accordingly. Controls are set so that any failure in the cleaner automatically stops all equipment behind in order to prevent jams. However, the conveyor after the hydraulic bath should always be kept running if possible, long enough to clear the cane from the bath. Conveyor speeds through the cleaner should be high enough to provide a minimum depth of cane. Water distribution should be concentrated at the points of maximum effectiveness. These are the cascader, combing drums and detrashing rolls. The quantity at each point must be sufficient to give a complete flooding action, otherwise effectiveness is lost. It must also be remembered that high pressure sprays are only effective at the point of impact. CANE SALVAGER In any hydraulic bath there are always some cane stalks, usually short lengths, that reach the bottom and are discharged on the rock and sand conveyor. These are recovered in a cane salvager which is essentially a small-scale velocity bath. Such a device should recover most of the cane which might otherwise be lost. EFFICIENCY OF CLEANING The quantity of extraneous matter delivered to the factory with the cane varies widely depending on harvesting conditions. With dry weather harvesting, a good burn will eliminate most of the dry leafy material leaving only tops and some ground trash. Soil and rock pickup will then be

19

a minimum varying with the terrain. In wet weather harvesting, with little or no burn, most of the fibrous trash will remain and large quantities of soil and rock will be present. Often during poor harvesting conditions the extraneous material can exceed the quantity of cane. Most cleaners are designed to handle cane with an average amount of extraneous matter and will not be effective under extreme conditions. In considering cleaner efficiency, it is only possible to look at the order of magnitude involved and to separate irrigated regions (relatively dry) from unirrigated regions (relatively wet). Even so it is only possible to get a reasonable handle on fibrous trash. 1

Early data by the Experiment Station of the Hawaiian Sugar Planters 1 Association ^ showed that the cane plant after two years of growth would have produced the following: Average makeup of total dry weight produced in two years: % Tops Leaves Roots Stalks

5 24 2 69

Composition at time of harvesting: Moisture % Tops Leaves Roots Stalks

76 34 — 70

Fiber % 20 64 — 13

During the two years of growth, leaves fall off and become ground trash, some of which decomposes. So at the time of harvest the leafy matter is considerably reduced. With a good burn most of the dry leaves burn leaving little fibrous trash except for the tops. The following figures are indicative of various conditions. Fibrous Trash Delivered to Factory % of Gross Cane Burned Tops Leaves Total (av.)

Irrigated Unburned

3.5 0.5

5.0 11.0 12

Unirrigated Burned Unburned 5.0 8.0

7.0 18.0 20

To these figures must be added dead cane, which is considered to be trash, and stools (root structure). The quantity of soil and rocks can range from a low of 1.5% under dry conditions to over 20% for wet conditions.

20

The cleaners in general perform well in removing soil and rock, the efficiency being in the range of 95%. Removal of fibrous trash, however, is rarely over 50% and usually can be considered to average 40%. Essentially no dead cane is removed. Stools, for the most part, are taken out in the rock bath. Removal of soil and rock is the most important function of the cleaner, since those materials place high wear loads on the equipment. Fibrous trash contributes only to increasing the quantity of fiber to be processed. With a high return value placed on fiber for fuel it can be economically sound to return all fibrous trash to the extraction plant. In this case the fiber should be given a secondary treatment, essentially the same as in the rock bath of the cane cleaner, to remove soil and rock fragments. WATER REUSE Clean water consumption can be cut substantially by multiple use. Extremely muddy water cleans equally with clean water except for a small residual. Minimum clean water use could therefore be only on the detrashing rolls, and perhaps a final rinse. That water is then collected and used on the combing drums. Then that from the combing drums on the cascader and finally in the hydraulic bath. Additional steps may also be used. Clean water requirements are considered to be about 4000 liters/ metric ton (1000 gal./short ton) clean cane per hour. This can be reduced with more efficient application and more stages of reuse. WASTE DISPOSAL Rock, gravel and sand are sent to land fill disposal. Fibrous trash may be returned to the fields but is useful to salvage for fuel. Two methods are used. One is to truck it to an open area, spread it out and allow air drying. The dry material is returned to the bagasse storage area, reduced in a shredder and fed to the furnace with bagasse. Another method is to treat the trash in a rock bath to remove rock fragments and sand and return it to the mill for dewatering. Waste water is allowed to settle in ponds or Hydroseparator-type equipment, with the effluent returned to the fields and the settlings pumped to landfill areas. LOSSES Losses in cane cleaning are in two categories - mechanical, loss of cane with rocks and with fibrous trash, and sugar loss by washing out of juice.

21

The mechanical loss can be kept at a level of 1% or below by use of a cane salvager and by proper operation and maintenance of detrashing equipment. The loss of juice (pol) is dependent upon the damage inflicted on the cane in harvesting, loading, hauling and cleaning. A single stalk of cane, sharply cut, can be washed thoroughly with negligible pol loss. A damaged or roughly cut stalk loses pol proportionally to the extent of the damage. Accurate quantitative data on pol losses are extremely difficult to establish because of the wide variation in damage and the extraordinary difficulty in the representative sampling of cane. Furthermore, measurements made on a weight basis are of limited value since the weight of the cane varies in each step of treatment - increasing with accumulation of soil, decreasing with removal of soil, increasing with accumulation of water and decreasing with loss of water and juice. Comprehensive investigations of the magnitude of pol loss in cleaning were made by the Sugar Technology Department of the Hawaiian Sugar 1 Planters Association on the basis of pol per unit fiber measurements. Although there was wide variability in individual cases the average statistically significant loss was 3.3%. Adding mechanical losses to this brings the value to around 4%. In economic appraisals a round figure of 5% would be a reasonable magnitude. Of course this is only the loss in cleaning. In the field operations during harvesting, the losses are of about the same magnitude bringing the total attributable to rake harvesting to the 10% range. REFERENCES 1 2

Borden, Ralph, Hawaii Planters Record, 46 (1942) 191-238. Stewart, Guy R., Assoc. Hawaiian Sugar Tech., Reports (1929) 221-230.

23

Chapter 3

MILLING INTRODUCTION Milling is basically an exercise in materials separation. In the simplest concept, cane consists of a solid - fiber, and a liquid phase - juice, which must be separated before sugar can be produced. This is done in a milling tandem in which juice is expelled from the fiber by successive applications of pressure as the cane passes between pairs of rolls. The efficiency of juice separation is determined by: Number of squeezes Effective pressure Degree of cell rupture Drainage Physical properties of fiber The capacity of a tandem is determined by the ability of the rolls to accept the cane presented and transport it by friction between the rolls. This is called feedability. The capacity is governed by the quantity of fiber since the juice alone puts no load on the rolls. In practice, a milling plant is designed for a nominal capacity at a nominal juice recovery. That is, the tandem should accept the quantity of cane desired in a unit time and expel a targeted percent of the juice. The actual results obtained depend upon how the tandem is set, how it is operated and how it is maintained. By pressure alone, it is impossible to expel much over 90% of the juice from the fiber because at some point the solid and liquid phases essentially coalesce into a mass which extrudes forward. Therefore, in order to recover more of the juice, it is necessary to add water. The water mixes with the juice and a certain percentage of the diulted juice is expelled in the next pair of rolls. By repeating this process, it is possible to recover substantially all of the juice. A conventional milling plant comprises equipment to prepare the cane for milling, which may be rotary knives, shredders, or combinations of knives and shredders; and a series of four to six 3-roll mills. In some installations a 2-roll mill, or crusher, may precede the mills. A crusher is considered both a preparatory machine and a mill. Each mill is normally equipped with a feeding device consisting of one or two rolls which compress the fiber before presentation to the mill. In some cases, these feeding devices become essentially a part of a mill resulting in 4-roll and 5-roll mills.

24

CANE PREPARATION Efficient juice separation in a milling tandem requires good preparation of the cane, whether obtained in preparatory machines, such as shredders, or in the mills alone. Good preparation means release of a high percentage of juice from the cellular structure of the cane without reducing the fiber size to such an extent that it will not feed well when presented to the mill. An approximate measure of the release of juice is given by a factor called Displaceability Index which gives the amount of juice that can be replaced by water in a fixed period of time. Preparation is best performed at the beginning of a milling tandem, and the best equipment is a heavy-duty shredder which essentially strips the fiber from the cane, releasing the juice, but still retaining a relatively long fiber structure. Such preparation also gives the best feedability. A well-designed shredder with heavy hammers (20 kg) running at a tip velocity of 6000 meters (19000 ft) per minute, should give a Displaceability Index of close to 88%. The horsepower necessary is about 13.2 per metric ton (12 per short ton) cane per hour. Running at lower speeds reduces the preparation and the horsepower. At 4500 meters (15000 ft) per minute, the Displaceability Index would fall to around 85% and the horsepower to 9 per metric ton (8 per short ton) cane per hour. In general, increasing the Displaceability Index by 4% would give a 1% increase in extraction in the same milling tandem. This is approximately the equivalent of adding one mill to a tandem. Lighter shredders are more effective if installed after a crusher, which provides protection from rocks. Also, by crushing the fiber and removing some of the juice, the power requirement is reduced. Such a shredder operating at a tip velocity of 4500 meters (15000 ft) per minute should give a Displaceability Index of 75% at a power requirement of 3.3 horsepower per metric ton (3 per short ton) cane per hour. TWO-ROLL CRUSHERS Two-roll crushers, formerly common in milling tandems, are rarely part of new installations. They are nevertheless simple and useful adjuncts. They serve the three functions of cane preparation, juice extraction and improved feeding for the next mill. Crushers are commonly the first unit of the milling train, although heavy shredders are often installed ahead of crushers. A well-operated crusher should give 50% juice extraction. By a combination of cane preparation, blanket compression and dewatering of the fiber, the first mill following (without maceration) should give an additional 27% juice extraction, giving a total of 77% by dry crushing. Crushers with coarse 7.5 cm (3 in.) circumferential grooving have good feedability and good drainage. Hard-faced, Krajewski-type corrugations perform a good job of rock crushing, which is an important function in protecting shredder hammers and subsequent mills.

25

MILLS The three rolls of a conventional mill are arranged in a triangle so that the fiber is squeezed twice between the top roll and the feed roll and the top roll and the discharge roll. The rolls have cast iron, grooved shells mounted on steel shafts. Fiber passing between the top and feed roll is conducted over a turner plate to the discharge roll. The rolls are pinion driven from the top roll which is driven at a speed of 3 to 6 rpm by a turbine through a gear reduction system. The feed and discharge rolls are fixed, while the top roll is free to move up and down by means of an hydraulic pressure system. Cane is moved between mills by means of intermediate conveyors. They are generally rake or drag-slat type which carry the fiber to a fixed chute leading to the next mill. FEEDING DEVICES The capacity of a mill to handle cane is governed largely by the ability of the rolls to accept the feed. Improved feedability can be obtained by the use of devices which compress the fiber blanket presented to the rolls, thus increasing frictional grab. These include: Overfeed roll Underfeed roll Pressure feeder Two-roll feeder Overfeed Roll A roll on top of the blanket just ahead of the mill compresses the fiber and thus improves feeding. These rolls are usually of smaller diameter 75% that of the mill rolls and are fabricated from sheet steel fitted with longitudinal angle iron strips. Sometimes cast iron rolls with grooving are also used. The rolls are driven by sprocket and chain at peripheral speeds 10% or more above that of the mill rolls. Underfeed Roll The underfeed roll is generally preferred to the overfeed roll. It operates beneath the blanket, compressing it against the top roll, which improves feedability. Construction is similar to that of an overfeed roll and drive is usually by sprocket and chain. The diameter is generally smaller - 75% that of the mill rolls and the peripheral speed is the same or somewhat higher than the mill rolls. In some regions the underfeed roll has been increased to the same size as the mill rolls, enmeshes with the feed roll and is driven by a crown wheel off of the top roll. The result is a 4-roll mill which has good feedability but has the disadvantage of poor drainage. At high mill speeds, expressed juice which cannot drain over the feed roll flows over the top roll and must be carried away from the discharged bagasse by means of a scraper trough. With good design these mills perform well at high grinding rates and offer a relatively easy method of increasing the capacity of an existing tandem.

26

Pressure Feeder A pressure feeder has two feed rolls the same size as the mill rolls set at an angle like a crusher. The rolls discharge into a pressure chute leading to the mill. The bagasse is thus kept compressed between the feeder and the mill. The feeder rolls are geared together with pinions and are driven by the mill turbine at a speed averaging 30% more than the mill rolls. The pressure feeder originated in Australia in order to improve the feedability of bagasse from hot maceration baths. Since a feeder extracts up to 40% of the juice from the fiber, it is not a feeding aid in the true sense. In reality the device converts a 3-roll mill into a 5-roll mill. Juice expressed by the top roll of the feeder is collected in a tray with scraper arrangement, while juice from the bottom roll flows both foward and backward into a collection trough. Because of juice extraction, the 3-roll mill is presented with a drier feed permitting the use of lower work ratios in the mill. The capacity of the 5-roll mill is up to 25% greater than that of a 3-roll mill of the same size, but power requirements are correspondingly increased. Two-Roll Feeder The 2-roll feeder is similar to a pressure feeder without a pressure chute. This is a true feeding aid as it does not extract juice. It is chain driven, and the bagasse discharges into a closed chute leading to the mill. The device gives good compression and has the advantage of good drainage. Because of lower power requirement and less cost, the 2-roll feeder is generally preferred over pressure feeders in application to existing installations. GROOVING Mill rolls are grooved to improve feedability and provide drainage. Grooves are of three types: Circumferential Juice (Messchaert) Chevron Circumferential Grooves Cutting grooves around the roll gives a corrugated surface of increased area which has better gripping action because of the compression of the fiber against the walls of the V-shaped grooves. Since the bagasse at the bottom of the groove is not so highly compressed, a drainage channel for juice is formed. The surface area changes with the angle of the groove, the sharper the angle the greater the surface. Larger pitch increases the speed differential between the tip and the channel. This gives a grinding action which will open up cells. Although sharper grooves give more surface, the effective surface is not much affected by very sharp grooves (30-35°)

27

since little pressure reaches the bottom of the channel. This is useful, however, as better drainage results. Sharper grooves are more susceptible to damage by rocks and tramp iron so angles less than 35° are not used. Normally, 45° is most practicable where damage potential is severe. Other conditions being the same, juice extraction decreases with increase in pitch of grooving. This is the composite effect of several factors, including reduced effective pressure, more groove wear caused by higher differential speed and more slippage. As preparation increases down the milling tandem, finer grooving is used. Good standard practice on the last mill is 1.25 cm (1/2 in.) pitch. With a good shredder, such grooving can be used throughout. With poor preparation, 2.5 cm (1 in.) pitch is effective. In the absence of a shredder, 5 cm (2 in.) pitch is common in the beginning mills. Juice (Messchaert) Grooves Juice grooves are narrow channels cut deeper than the circumferential grooves to give better drainage. They are normally not over 0.6 cm (1/4 in.) wide to prevent much fiber being compressed into them, and are of a depth sufficient to carry away the juice. This averages about 2.5 cm (1 in.). The pitch is also a function of the volume of juice and is normally 7.5 cm (3 in.). The juice grooves must be kept clean by scrapers. Wear causes problems by enlarging the grooves and letting excess bagasse enter with the juice. Narrow and shallow grooves are desirable for this reason. Juice grooves are most effective on the feed roll where juice flow is greater. They may also be used on the discharge roll but are not as efficient because they reduce the effective pressure area of the roll and increase the incidence of metal lost by fracturing of the circumferential groove adjacent to the juice groove. Chevron Grooves Chevron-shaped grooves are cut lengthwise through the circumferential grooving as an aid to feeding. They are nominally cut to one half the depth of the grooving and at a pitch of 25 cm (10 in.). Chevron grooves are effective only on the be detrimental on the discharge roll. Since effective pressure surface of the rolls and mixed juice, more efficient operation of chevron grooves by arcing maintenance of the

feed and top rolls and would such grooving decreases the causes excessive bagasse in a mill is obtained without circumferential grooving.

OPERATION Effective operation of a milling station involves proper setting of the individual mills; close control of the operational variables, chief of which are fiber feed rate, speed, hydraulic loading and imbibition; and finally, good maintenance.

28

Mill Settings The setting of a mill requires three measurements - the distance between the top roll and the feed roll, the distance between the top roll and the discharge roll and the distance between the top roll and the turner plate. The weight of fiber passing through the mill per unit time is the basis for calculating the proper setting. The following concepts are used in arriving at the initial setting. In the course of operation, changes from the initial are often necessary because of the many unknown variables involved. Work Opening: The work opening is the average distance between the rolls as measured on the common axial plane. It is calculated from the mean diameter of the two rolls, which is the diameter half way between the tip and bottom of the grooves. A fixed amount is added to the measured opening to allow for the lift of the top roll when floating in operation under the calculated fiber load. Mill Ratio: The mill ratio is the ratio of the feed and discharge settings. The mill ratio adopted is based upon the calculated discharge work opening which is determined by the fiber rate, size of the rolls, speed of the rolls and fiber content of the discharged bagasse. Once the discharge work opening is chosen, the feed work opening is chosen depending upon trash content of the cane, preparation, drainage and imbibition rate. The mill ratio should be kept at the minimum practicable. Nominal values for mill ratios are 2.5 for the first mill and 2.0 for mills thereafter. With 5-roll mills, ratios as low as 1.5 are possible. Escribed Volume: The escribed volume is the product of the roll length, the work opening and the roll peripheral speed. Fiber Index: Fiber index is the weight of fiber per unit escribed volume of bagasse from the discharge roll of the mill. It is a measure of blanket thickness and compression. Although it would appear that a high value of fiber index is desirable, there is an optimum value beyond which little change occurs for a given mill. The value increases from the first mill to 3 the last. Nominal maximum values of fiber index range from 560 k g / m 3 3 3 1 (35 l b / f t ) for a first mill to 880 k g / m (55 l b / f t ) for a fifth mill. Usual operating values are lower. Fractional Fiber Content: The fractional fiber content expresses the fiber on a weight basis, that is, the weight of fiber per unit weight of bagasse discharged. An indication of nominal values is shown by data from Mauritius. 2

29

Fractional fiber content of discharged bagasse % Crusher (2-roll) Mill 1 2 3 4 5

0.25 0.37 0.41 0.46 0.49 0.52 3

3

3

Using values of 1530 k g / m (95.5 l b / f t ) for fiber and 1028 k g / m (64.2 3 l b / f t ) for juice, these figures are considerably lower than the maximum values shown under Fiber Index. Reabsorption Factor; The reabsorption factor is a measure of the extrusion that occurs in a mill. Extrusion begins when the speed of the mills becomes high enough that the maximum pressure on the blanket of fiber begins to move ahead of the line of minimum distance between the rolls. Material is then propelled forward at a speed exceeding that of the rolls. A better term to describe the phenomenon is forward slip. When this occurs, the volume of the bagasse exceeds the calculated escribed volume. The increase in volume is caused largely by an increase in the amount of juice, since it moves forward preferentially to the fiber. Juice extraction decreases, therefore, as the amount of forward slip increases. The reabsorption factor is: No-void bagasse volume Escribed volume The factor increases with roll speed, increases with fiber rate and increases with a smaller discharge work opening. It decreases with the fineness of the fiber. In setting a mill this factor must be kept as low as possible. Because of the varying fiber load, the actual value of the factor can only be approximated. The optimum range for the factor, however, is 1.3 to 1.4. Procedures: Procedures adopted for setting mills, although based primarily on fiber throughput, must take into consideration many other factors - most important of which are: Mill strength Hydraulic loading Roll grooving - size and condition Cane quality - trash and soil content Cane preparation - Displaceability Index Drainage Imbibition - quantity of water and temperature The first step, however, is to calculate a tentative work opening for

30

the discharge roll. This is based on the fiber rate using a nominal Fiber Index as discussed earlier. A mill speed is selected which will keep the reabsorption factor at a reasonable level (below 1.5). The work opening for the feed roll is then chosen keeping the mill ratio as low as possible. For mills after the first, a nominal 2.0 mill ratio is standard. With a 5-roll mill, this can be lowered to 1.5. Better cane preparation will permit a lower mill ratio, as will more efficient feeding devices. Since feedability is impaired by hot maceration, higher mill ratios are necessary. In general, also, higher hydraulic loading, poor drainage and fine grooving require higher mill ratios. Turner Plate Settings: Turner plate settings are the ratio between the feed work opening and the opening at the toe of the plate. The opening at the toe of the plate is equal to the set opening plus one-half depth of the top roll groove plus hydraulic lift. There is only limited experimental evidence on which to base settings, so general experience is used. If the plate is set too high, the top roll will be lifted by fiber passing over the plate. This will increase power, increase wear and lower extraction. Also, the mill capacity will be decreased. If the plate is too low, there will be too high an angle of contact with the discharge roll which can cause choking. Experience indicates that an average fiber rate of 160 kg/m^ (10 lb/ft^) on preliminary mills and 240 kg/m^ (15 lb/ft^) on the last mill gives good performance. The ratio of toe of the plate to the feed opening is normally 1.8 for finely divided bagasse. With coarse bagasse, this ratio would decrease to 1.2. Maceration In the conventional compound maceration system, water is distributed laterally to the bagasse on the intermediate carrier feeding the last mill. Last expressed juice is then returned to the carrier to the next to last mill and so on up to the second mill. Juice from the second mill mixed with the dry crushing juice from the crusher and first mill constitute mixed juice. Although bagasse fiber will absorb about 650% its weight of liquid, it is not advantageous to use enough water to bring the liquid content to this point on the last mill. A t about 250% water on fiber, the effect on extracation levels off, so that steam and evaporator capacity requirements dictate little economic advantage of using more water. The water should be applied at as high a temperature possible without causing a feedability problem. The fiber becomes plastic at higher temperatures so is more easily compressed giving higher juice expression. Higher temperature also makes the fiber slicker so it does not feed as well. Mixing of liquid with the bagasse takes place mainly at the entrance to the mill where the increasing pressure forces the liquid backward through the fiber. Less lateral flow occurs, so the important consideration in applying the maceration is to insure uniformity across the width of the bagasse blanket.

31

Hydraulic Loading Juice extraction increases with increase in pressure applied to the top roll. The pressure that can be applied, however, is limited by the mechanical strength of the mill. Also, feedability decreases at higher pressures and power necessary increases substantially. The optimum pressure, therefore, is that which permits the top roll to float at the necessary feedability. The actual loading used on a given mill can only be determined by experience. General standards call for a nominal pressure of 14 mt/m (50 t / f t ) on a 2-roll crusher to 21 mt/m (75 t / f t ) on the last mill. Speed Other conditions, such as preparation and pressure, being the same, juice extraction decreases with increased mill peripheral speed. The reason for this is that extrusion (reabsorption, forward slip), meaning that juice is expelled with the fiber, is a function of mill speed. It is desirable, therefore, to operate at as low a speed possible, commensurate with the necessary fiber rate. The finer the preparation, the slower the optimum speed. This is because of better feedability. Also, there is the factor of cell rupture in the mill which, because of closer mill settings, increases with speed. It is standard practice to run all mills in a tandem at close to the same peripheral speed with a slight increase toward the last mill to improve feedability. General Effects The general effects of operating conditions on milling efficiency may be summarized as follows: Pressure Increase Increases juice expelled Increases power required Decreases feedability Roll Speed Increase Decreases juice expelled Increases feedability Cell Rupture Increase Increases juice expelled Increases feedability Imbibition Increase Increases juice recovery Decreases feedability Imbibition Temperature Increase Increases juice recovery Decreases feedability

32

Mill and Turner Plate Settings Decrease Requires mill speed increase to maintain throughput Increases cell rupture Increases power required CONTROL The primary control figures on milling are pol, moisture and fiber in final bagasse. These, of course, give only the total result and it is important to know the performance of individual mills. This requires bagasse analyses from each mill. But constant variation in the fiber load, because of trash, renders such figures of little value unless taken on a statistically significant basis. Therefore, they are commonly not justified from the cost standpoint. As a result, mill settings are estimated and then adjusted on the basis of experience. Juice Density Curves Useful guides to the behavior of each mill are juice density curves. These are made by an analysis of the refractometer solids in samples of juice from the feed roll, discharge roll and total from each mill. Discharge roll figures are plotted giving a curve which for a properly functioning tandem would appear as in Figure 3-1. Typically, the crusher value is lower than the first mill because of the extraneous water on the cane from the cleaner. The figures then fall in a smooth curve to the last mill. Since the refractometer solids content of the juice from the discharge roll of the last mill is about equal to that of the juice remaining in the bagasse, the lower the figure, the better the extraction. The actual value will depend upon the cane preparation and maceration efficiency as well as the performance of the individual mill. The individual values on the curve indicate the total performance up to that point. A deviation from a smooth curve is evidence of malperformance in that area. Usually the deviation is toward making the curve more horizontal and means that the preceding mill is to blame, either by total performance of the mill or poor extraction at the feed roll permitting lower density juice from the discharge roll. Rarely it might also be caused by very high extraction of the feed roll of the succeeding mill. A break, making the curve more vertical, is sometimes encountered. This could be caused by malfunctioning of the mill preceding the one where the break occurs either in the total extraction or poor extraction at the feed roll allowing low density juice from the discharge roll. It could also be the result of low extraction from the feed roll of the mill succeeding the break. Juice density curves are only significant on a routine basis to show deviations from previous curves. An individual curve, standing alone, is of little value, mainly because of the idiosyncrasies of individual mills. Once a curve with a break shows up, the test should be repeated to make sure that a real condition exists.

33

Fig. 3-1. Mill juice curve. The relative extraction of the feed roll and discharge roll can also be calculated from the juice density figures, as an example: Refractometer solids Feedroll Discharge roll Total juice

2.0 4.0 2.0

Calculating by the rectangle method: 2.0

1.5 2.5

4.0

0.5 2.0

34

1.5/2.0 χ 100 = 75% of the juice extracted by the feed roll and 25% by the discharge roll. Such figures would be indicative of good mill performance. REFERENCES 1 2

Van Hengel, A . and Douwes Dekker, K., Some Notes on the Setting and Operation of Mills, Proc. So. African Sugar Tech. Assn., 32nd Congress, 1958, pp. 57-67. Van Hengel, Α . , Some Additional Notes on Mill Settings, So. African Sugar J., 1958, pp. 855-861.

35

Chapter 4

DIFFUSION* INTRODUCTION In the separation of juice from the fiber in sugar cane by the process called diffusion, the juice is displaced from disintegrated cane by the countercurrent flow of water rather than being expelled by pressure as in milling. The unit operations involved are three: Cane Preparation Juice Displacement (Diffusion) Bagasse Dewatering CANE PREPARATION Cane preparation for effective juice displacement requires size reduction to give a compact permeable bed and rupture of close to 94% of the juice storage cells. This should be brought about with a minimum of grinding of the fiber and retaining a fiber bundle length of 10 to 15 cm (4-6 in.). The cell tissue should be stripped from the fiber bundles producing a mixture of shreds and pith tissue. Such preparation will yield a compact yet permeable bed for countercurrent extraction. The equipment most suited for this work is swing hammer shredders. All experimental investigations have shown that good preparation is best achieved in two steps. Shredders developed specifically for this purpose are the Silver Buster and the Silver Fiberizer. The Buster has 20 kg hammers rotating at about HOOrpm, working over an open grid screen grate with anvils between the hammers. Feed control to the machine is provided by a set of variable speed feed rolls. As distinct from conventional shredders, where the hammers are working over a solid ridged plate, cane is extruded through the screen so that no unshredded stalk can pass. Size reduction to a length of 5 to 10 cm (2-4 in.) is obtained with cell rupture up to 87%. Secondary preparation takes place in the Fiberizer which is more like a conventional shredder in that the hammers are working over a ridged 90° arc segment. Here, fiber length is reduced only to a limited extent and cell rupture is increased to 94%.

This discussion applies specifically to the Silver Ring Diffuser. However, the general principles involved are the same for any type of diffusion equipment.

36

JUICE DISPLACEMENT Displacement of juice by water in the prepared cane is attained by a countercurrent liquid flow in an advancing front similar to the phenomenon of saturated flow in porous media. The advancing liquid operates like pistons through the ruptured cell modules replacing the juice. Any macro breaks in the bed lead to piping, channelling and by-pass of liquid from the displacement process. Any mixing in the system reduces the effective countercurrent flow. Any pressing or squeezing will expel juice concurrently as well as counter, causing mixing. The ideal system, therefore, is an undisturbed bed of fiber and juice and plug type liquid flow. In the Silver Ring Diffuser, this mechanism is obtained with a rotating annular bed of prepared cane through which the liquid is pumped in 18 stages (Fig. 4-1). Cane is made to form a continuous bed about 1.5 m (5 ft) in depth, which is maintained for some 335 angular degrees. At this point, the bed is discharged vertically by means of multiple screws. Water is added from a distributor above the bed and several degrees ahead of the discharge screws. The water flows through the moving bed and into a compartment below and behind the distributor. From there, the liquid is pumped forward to the third distributor, passes through the cane bed and collects in the next juice compartment. Juice from bagasse dewatering enters the distributor just forward of the water distributor and flows into the second collecting tank. Flow continues forward to the point at which the cane enters. Here the juice is cycled ahead as first pass juice. On passing through the newly deposited cane, it picks up some suspended matter which may be removed by circulating through the bed. This second pass is called recycle, and pumping is in the direction of cane travel. The effluent recycle juice is analogous to mixed juice from a miling tandem and is sent to the boiling house. With large amounts of soil entering with the cane, this recycling can cause plugging by depositing soil on top of the bed, so should not be used. Efficiency of juice displacement is reduced by some channelling in the bed and by mixing at the overlapping boundaries of the stages in the bed. This is compensated for, to a large degree, by the use of more than the theoretical number of stages, and by application of more water. A typical extraction pattern obtains however. This is shown in Figure 4-2 illustrating the gradients occurring in routine diffuser operation. A rapid drop in pol takes place initially as juice is displaced. Then the decrease becomes more gradual and, finally, becomes asymptotic near the discharge point, being "supported" by the juice returned from bagasse dewatering. Factors which govern the forms of the gradient are: Cane Quality Cell Rupture Bed Depth Speed of Rotation Bed Permeability Quantity of Water Quantity and Quality of Juice Return From Dewatering

Fig.

4-1. Diffuser flow diagram.

37

Fig.

4-2. Diffuser gradients.

38

39

Operational conditions can be adjusted to give the optimum results consistent with the objectives. Pol extraction increases with increase in cell rupture and quantity of water. There is an optimum bed depth and speed of rotation, depending upon the cane rate and bed permeability. It is basic that the capacity of the diffuser is governed by the rate of the gravity flow of liquid through the bed. Cane quality, type of preparation, bed depth and dewatering juice quality are the variables affecting this flow rate. BAGASSE DEWATERING Following juice displacement, the bagasse must be dewatered. This step also is the final stage in the extraction process as some pol in unbroken cells and in the residual liquid is recovered and returned to the diffuser. Dewatering requires some sort of expelling mechanism, and most of the common types of machines are adaptable to such use. First thought usually starts with conventional 3-roll cane mills which are in general use for this purpose. In operating a mill after a diffuser, consideration must be given to the difference in properties between mill bagasse and diffuser bagasse. Firstly, it has a higher liquid content, averaging 650% on fiber. This compares with about half this for bagasse entering the last mill of a milling train at 200% imbibition on fiber. In fact, diffuser bagasse has about the same relation between liquid and fiber as whole cane. Special attention must, therefore, be given to drainage. Secondly, the bagasse is at a higher temperature than normal for mills. It is, therefore, more plastic and slicker, giving a lower coefficient of friction with the rolls. This results in poor feedability. Greater attention must thus be given to feeding devices and maintenance of mill grooving. Thirdly, diffuser bagasse is in a relatively fine state of subdivision. Meeting these conditions effectively is a 3-roll mill fitted with a heavy-duty pressure feeder. The settings should be more like that of a first mill rather than that of a last mill because of the high liquid-fiber ratio. The characteristics of diffuser bagasse are particularly favorable to dewatering in expeller type machines such as the screw press. In these, high liquid content and high temperature favor feeding and reduced friction in contrast to a roller mill. On the unfavorable side are the relatively high power requirements and high maintenance costs. PRESS JUICE TREATMENT Juice from the dewatering equipment is returned to the diffuser in order to recover some of the pol. Insoluble material in this juice may plug the bed and cause diffuser flooding, so treatment is necessary. Investigational work has shown that plugging is caused mainly by particles 40 microns or less in size. These fine particles form a layer of low

40

permeability by filling the interstices near the top of the bed. Juice percolation through the bed slows and juice accumulates on the top of the bed, a condition called flooding. Juice below the impervious layer drains away, leaving a porous bed which results in poor juice contact. A temporary flow can be induced by mechanically breaking up the layer, but a short time later, the layer will form again farther down in the bed (Fig. 4-3). It has been established that the particles that will cause plugging settle rapidly in the low density juice, so that simple gravity settling produces a relatively non-plugging juice. A conventional clarifier will normally give a good juice without chemical treatment with a retention time of about half that of lime defecation.

Fig. 4-3.

MINUTES Press juice percolation through prepared cane bed.

41

The underflow from a press juice clarifier is difficult to handle on a bottom feed rotary vacuum filter. The cake tends to form irregularly and slough off on the feed side. Top feed filters perform satisfactorily. Horizontal filters perform well and specially designed bottom feed rotary machines with increased hydraulic capacity are also functional if bagacillo is added. OPERATION At rated capacity, the Silver diffuser is designed for a target bagasse pol of 1.0 at a draft of 100 (juice % cane). A cell rupture in prepared cane of close to 94% is necessary. Based upon an average percolation rate of 2 2 250 liters/m (6.1 g a l . / f t ) at an optimum bed depth of 1.5 m (4.9 ft), the diffuser speed would be 1.5 revolutions per hour. Under these conditions, the fiber retention time in the diffuser is 37 minutes. The juice, however, 1 would have an average retention time of only half this (Fig. 4-4). Permeability, Bed Depth and Diffuser Speed When operated under design conditions (see flow diagram, Fig. 4-1), juice from a distributor percolates through the moving bed and drains into the next aft collection tank below. Deviation from standard conditions in permeability, bed depth and diffuser speed would have the following effects: 1.

If the bed permeability is lower, some juice is carried beyond the next tank. This mixes the countercurrent flow, moving richer juice toward the discharge end and reducing extraction. Corrective action may be: a. b.

reduce bed depth reduce diffuser speed

Both result in lower throughput. 2.

If the bed permeability is greater, some juice will recirculate and countercurrent flow will be disturbed in the forward direction. This, of course, is a good situation and can be corrected by increasing bed depth or diffuser speed, both of which result in increased throughput.

3.

If the bed depth is greater, the effect is similar to No. 1, low permeability.

4.

If the bed depth is lower, the effect is similar to No. 2.

5.

If the diffuser speed is higher, the effect is similar to No. 1.

6.

If the diffuser speed is lower, the effect is similar to No. 2.

In order to compensate for variation in these conditions, the diffuser has enough extra stages in its design to allow considerable intermixing of flows. Changes of the order of 15%, therefore, have little effect on performance. Changes of greater magnitude will show up in reduced efficiency.

42

REF SOL. FIBER INSOL SOLIDS WATER

REF SOL 2 . 5 4 FIBER 46.3 INSOL. SOL. 4 . 0 MOIST 47.2 POL 1.24 PURITY 48.4

D I F F U S E R FLOW BALANCE PREPARED CANE T/hr. 100 DIFFUSER JUICE % CANE 1 0 0 EXTRACTION % 97

PRESS R E T U R N CAKE 2 . 0 T/hr

Fig.

4-4. Diffuser flow balance.

43

Cell Rupture Cell rupture (Displaceability Index) below 94% gives higher bagasse pol because of the slow rate of diffusion of sugar out of unopened cells. Rich juice will, therefore, persist until the dewatering process, giving higher pol in the returned juice, thus raising the pol level at the discharge end of the diffuser. Good cell rupture is the single most important factor in reaching high extraction. This is obtained by careful maintenance of the shredder hammers and grids. As the surfaces of these wear and clearance increases, the horsepower required to drive the machines increases. Thus, horsepower is a good indicator of the type of preparation being obtained. Periodic analysis for Displaceability Index is, of course, necessary. Draft and Dilution Draft is extracted juice plus dilution water as a percentage on cane, so the 100 draft figure corresponds to a dilution of about 17% on absolute juice for average cane. With a good countercurrent flow pattern and high cell rupture, bagasse pol of 1.0 can be obtained with dilution rates below 17%. Under less favorable operating conditions, it is well to increase the dilution and hence increase draft to the extent that the evaporator station will handle the load. The inventory of liquid in the diffuser must be maintained at such a level that the bed is saturated. That means that the liquid level should be the same as the bed level. If the liquid level is too high, above-surface flow will result in unwanted mixing. If the level falls below the bed surface, plug-type flow will no longer be possible and air filled channels will result in poor liquid to fiber contact and lowered efficiency of extraction. The bed should never be allowed to drain completely because air pockets which form will remain, even if the bed is flooded again. In order to maintain a saturated bed, the quantity of liquid must be of 1 the order of 1200% on fiber (Fig. 4-5). This amount of liquid is maintained by sending to the diffuser all of the return from the dewatering unit plus fresh water to make up the total draft. Return press juice from the dewatering unit at 100 draft will be about 75%, requiring about 25% fresh water. The flow of these streams must be steady and as uniform as possible. When the diffuser is stopped, juice flow is automatically diverted so that it is pumped forward in complete countercurrent flow only at the discharge end but recirculates on itself with only a small forward flow at the feed end. This is to maintain a saturated bed. Shortly after a diffuser stop, press juice will stop and, with automatic control, fresh water will increase some four-fold to supply the running diffuser needs. The water should be kept on for a short time, but on stops of more than a few minutes, the water flow should be reduced to one-half rate after about 30 minutes. Press Return All juice from dewatering the bagasse is returned to the diffuser, just as last expressed juice is returned to the tandem in milling. The return must be relatively free of suspended solids and must have a minimum

Fig.

4-5. Juice flow as a percent of fiber.

44

45

temperature of 85° C. Quality with respect to bed-plugging material can be tested periodically simply by observing the flow of a sample through a 100-mesh screen. The juice should pass through without any holdup on the screen. The press juice distributor should be inspected periodically to ensure that there is no buildup of mud on the bottom. Although treatment, other than settling, is normally not necessary, it is advisable to add sufficient lime to bring the pH above 6 in order to reduce corrosion of the equipment. Settlings from the press juice clarifier need only rough filtration in order to return the juice to the press juice heating tank. Since the pol is low at this point, losses in cake are very small. Temperature Diffuser operation should be carried out under conditions in which microorganisms will not grow. A safe temperature is 70° C. This is held by keeping the press return and fresh water temperature at above 80° C and heating the recycled juice at the front of the diffuser to the same level. Growth of microorganisms in the diffuser will not only destroy sugar, but gum formation can seriously impair the permeability of the bed and cause flooding. It is well not to maintain too high a temperature as higher sugar losses through inversion will occur. A t 70° C, the average inversion loss through the diffuser, with unlimed juice (pH 5.2-5.5), is 0.14%. Also, lower temperatures give less color development and less solubilization of fiber components. This latter factor is responsible for the increase in nonsugar dissolved solids in diffusion with a consequent decrease in juice purity. Bagasse Dewatering The dewatering unit must deliver a constant flow of return juice to the diffuser and reduce the liquid content of the diffuser discharge from about 85% to less than 50%. The principal consideration is the maintenance of good feedability. In the case of a mill, the grooving must be kept in condition by frequent arcing. DIFFUSER FLOODING Failure of juice to percolate rapidly enough through the bed, resulting in accumulation of liquid on top of the bed and, particularly, in front of the discharge screws, is the principal problem encountered in operation. As indicated previously, it is usually caused by plugging with finely divided particles of soil. The first indication of flooding is often seen in the irregular discharge of bagasse from the screws. Liquid accumulating at the base of the screw by poor draining of the bagasse prevents pickup by the screws. Bagasse then, builds up until it reaches a high enough level that drainage has made

46

it dry enough to be picked up. Large quantities are then picked up until the bagasse level falls and pick up ceases. In cases of serious flooding, the bagasse may become so high that enough is picked up to overload the drive motors on the screws. The most common cause of flooding is return of press juice with substantial quantities of fine suspended particles. The remedy is proper control of overflow from the press return clarifier. Usually, the heated press juice settles rapidly and presents no problem. Juice from sour and deteriorated cane, however, may settle poorly. Moderate chemical treatment such as lime or poly electrolytes has not proven to be generally effective. Of course, extensive treatment such as high liming, followed by phosphatation, which give a clear juice, will be effective but expensive. The options available, therefore, are to slow down the diffuser or to discard some of the press return replacing it with clear water. Another source of flooding is at the feed end, from recycling the juice as previously indicated. This situation can be helped by stopping the recycle and sending first pass juice to the boiling house. The juice will not be as clear, but should not cause any difficulty in clarification. A third cause may be just too much soil coming in with the cane, so the whole bed is of low permeability. The obvious solution to this is better cane cleaning. If this is not possible, the only thing to do is slow down the cane rate. An uncommon cause of flooding is growth of microorganisms in the diffuser. This can only occur when temperatures fall below 70° C, permitting multiplication. Microorganism growth can give gum formation which can almost gelatinize the cane bed and block it completely. Devices which rake up the diffuser bed periodically have been used to aid in percolation. Such action does enable juice to go through the bed but poor extraction is the consequence. CONTROL Basic control parameters for a diffuser are: 1. 2. 3. 4.

Cane feed rate Displaceability index Diffuser speed - bed depth Dilution water rate

No sensing device has been developed for the governing element which determines how these parameters, with the exception of displaceability index, will be set, namely the bed permeability. This condition is judged solely by observation and operating set points plugged in accordingly. Cane feed is weighed on a belt scale, the feed to which is controlled by the chest pressure on the turbine driving the Buster. Feed to the Buster is controlled by the Buster feed roll. Cane in the chute to the Buster is

47

sensed by a level arm riding on the cane which actuates a switch to the carrier drive motor. Speed of the diffuser is set by the cane feed rate so as to maintain a minimum bed depth of 1 m and a maximum of 1.5 m. Displaceability index can only be determined by periodic sampling and laboratory analysis. However, a good idea of the degree of preparation can be had by visual observation. Also when the hammers become worn the horsepower of the drive increases. This means poor preparation and thus a low displaceability measurement. Rate of dilution water under standard conditions is that necessary for 100 draft. This means that the total of juice returned from dewatering bagasse (press juice) plus dilution water is equal to the cane rate. Since the return juice normally is around 75% of the cane, the water added is about 25% of the cane. It is often the case that the evaporator capacity is limiting. Under this condition, the quantity of water is that which the evaporator will handle, then the draft may be less than 100. Likewise, with ample evaporator capacity, additional water can be used to lower the bagasse pol. Water temperature should be 80° C. REFERENCES 1

Payne, J. H., Cane Diffusion - The Displacement Process in Principle and Practice, Proc. Intern. Soc. Sugar Cane Tech., 1968, pp. 103-121.

49

Chapter 5

BAGASSE MOISTURE Although the moisture content of final bagasse is governed primarily by the performance of the last mill, the composition of the material entering this mill has a significant effect. This is determined by the quality of the entering cane and the extraction in the preceding mills. Assume cane entering the tandem at 15% fiber and 15% dissolved solids, the entering moisture is 70%. Juice extracted by dry crushing in a 2-roll crusher and a 3-roll mill should be 77%. Then, the residual juice in first mill bagasse would be: (100 - 77) χ 0.85 χ 100

_

(100 - 77) χ 0.85+ 15

" ™ ™

Λ

and the moisture would be: 70 χ 0.23 χ 100 15 + (15 χ 0.23) + (70 χ 0.23)

,

i 4 Rf O 4

b

,

b

*

leaving fiber of

43.4%

moisture plus fiber

90.0%

and dissolved solids of 10.0% In the following mills of the tandem, the application of the countercurrent maceration system results in the replacement of most of the juice with water. If approximately the same relationship of solid to liquid (fiber to juice) is maintained in subsequent mills as in the first, then bagasse leaving the last mill with a juice of 3.0 dissolved solids would have: dissolved solids = 56.5 χ 0.03 = 1.7% and moisture = 56.6 - 1.7 = 54.9% Thus, the moisture content would increase as the dissolved solids decrease. In order to reduce the moisture, it is necessary to increase the fiber content progressively down the tandem. The mill settings must, therefore, be reduced, increasing the fiber, the last mill setting determining the moisture content of the final bagasse. With no mill slippage in a properly set mill, the water content of bagasse leaving a mill is independent of the water content entering. This is true because the load on the mill is determined by the fiber and not fiber plus liquid. However, with slippage, the water content of the expelled bagasse increases since the fiber entering the mill is less and space is available for more water to pass. Maintenance of feedability is, therefore, of prime importance on the

50 last mill. This requires constant attention to building up of roll grooving and effective operation of the feeding aid. Keeping in mind that the higher the dissolved solids in the residual juices (poorer extraction), the lower is the bagasse moisture at the same fiber content; the fiber content of bagasse is a more useful measure of last mill performance than moisture. Effect on Extraction Other conditions being the same, a reduction in the moisture in bagasse results in higher extraction. The actual increase is dependent upon the fiber, moisture, pol, and refractometer solids involved. By using average figures for these a general relationship as shown in Figure 5-1 is obtained. From this it will be seen that a decrease in bagasse moisture of 1% results in an increase in extraction of approximately 0.1%.

53

52

51

50

49

48

47

46

45

44

43

42

MOISTURE % BAGASSE

Fig. 5-1. Effect of bagasse moisture on extraction.

41

40

51

The calculations assume little change in the pol of the juice extracted and are illustrated in the following example: Assume pol in cane Bagasse fiber pol refractometer solids moisture

13.5 48.0 1.44 2.32 49.7

Then extraction =

97.0

Now increase fiber by one unit Bagasse fiber pol refractometer solids moisture

49.0 1.41 2.28 48.7

Then extraction =

97.1

If the bagasse fiber is coarse, the increase in pol extraction may not be as great because the pol of the juice within the fiber may be higher than that of the external juice. The applied imbibition water will not have reached equilibrium with the internal juice. This effect may be counter-balanced to some extent if the lower moisture level has been obtained by higher pressure on the mill or a greater preparation effect in the mill, both of which would tend to give closer to equilibrium conditions. Effect on Fuel Value The higher fuel value of bagasse at lower moisture content can be reasonably calculated but the practical value obtained varies considerably with the boiler and how it is operated. For rough estimations, however, a change of 1% in moisture changes the fuel value also by 1%. An illustrative example follows: At 50% Moisture Analysis

Per 100 Tons Bagasse

Fiber Moisture Soluble solids

%

Tons

48 50 2

48 50 2

Total

100

Approximate Net Calorific Value or

1853.7 1853.7 3336 χ 7760 χ

kg-cal/kg (3336 Btu/lb) (7760 kJ/kg) χ 1000 χ 100 = 185,370,000 kg-cal/100 metric tons 2000 χ 100 = 667,200,000 Btu/100 short tons 1000 χ 100 = 776,000,000 kJ/100 metric tons

52

At 49% Moisture Analysis

Per 100 Tons 50% Moisture Bagasse

Fiber Moisture Soluble solids

%

Tons

49 49 2

48 48 __2

Total

98

The weight of bagasse is now 98 tons. Approximate Net Calorific Value 1903.1 or 1903.1 3425 χ 7967 χ

kg-cal/kg (3425 Btu/lb) (7967 kJ/kg) χ 1000 χ 98 = 186,507,000 kg-cal/100 metric tons 2000 χ 98 = 671,300,000 Btu/100 short tons 1000 χ 98 = 780,766,000 kJ/100 metric tons

The increase in Net Calorific Value is therefore 186,507,000 - 185,370,000 = 1,137,000 kg-cal/100 metric tons 671,300,000 - 667,200,000 = 4,100,000 Btu/100 short tons 780,766,000 - 776,000,000 = 4,766,000 kJ/100 metric tons or

1,137,000 χ 100 = 0.61% 185,370,000

The actual value obtained in a boiler would be higher, approaching 1%. In calculating the fuel value of bagasse, it is customary to include the energy value of the sugar as well as the fiber. It should be kept in mind that this is valid only if the bagasse goes directly from the mill to the boiler. In storage, the sugar content is rapidly destroyed so that after several hours it can be considered to be zero.

53

Chapter 6

THE IMPACT OF EXTRANEOUS MATTER ON MILLING AND DIFFUSION Extraneous matter is the non-cane material in cane delivered to the factory. This is often referred to as trash, although the meaning of this term varies in cane sugar producing areas. In its original use, it referred to leaf material accompanying the cane stalk as harvested. This concept still obtains in many areas. A t present, the broadest definition is that officially used by the Hawaiian Sugar Technologists, which includes all extraneous material delivered with the cane stalk, from root to growing point, and covers not only cane leaves, but also the growing tip and roots, weeds, soil, rocks, casual matter such as dump refuse, structural material, machinery parts, livestock and even superficial water - an important item in washed cane. Purely quantitative figures on total extraneous matter do not provide adequate information for predicting the effect on operations. It is necessary to know the physical characteristics of the foreign materials. Five main categories, with subclassifications, cover most cases and give useful characterization. 1.

Fibrous material Dry leaves Tops Ground trash (partially decayed material) Roots Dead cane Weeds

2.

Soil Clay Loam Sand

3.

Rocks Gravel Stones

4.

Metal (tramp iron)

5.

Water

A consideration of the general effect of these on milling and diffusion follows.

54

MILLING In engineering calculations, the capacity of a milling tandem is customarily considered to be a function of the fiber throughput per unit time. Actually, the quality of the fiber also has a bearing but is difficult to quantify. Therefore, the calculated quantity of cane crushed in a mill is judged by the fiber content of the cane plus the fiber content of the extraneous material entering the tandem with the cane. In a given area, the fiber content of sound mature cane stalk varies within narrow limits (Hawaiian varieties 12-14%). The amount of fiber entering with the cane as extraneous matter is subject to wide variation, however, dependent upon the effect of burning, field and harvesting conditions, as noted in Chapter 2. The amount of fibrous trash entering with the cane ranges from 4% for well-burned cane on an irrigated plantation to 25% for unburned cane on an unirrigated plantation. The fiber content of the trash also varies from a low level of 10% for ground trash harvested under wet conditions, through 25% for green unburned tops to 80% for dry leaves. Using the estimated figure of 40% fibrous trash removal in the cane cleaner, the range of fibrous trash entering the mill is from 2.5 to 15% based on the entering moisture content. After the cleaner, however, the material is completely saturated with water regardless of the entering moisture. Fiber levels then are: Green tops Leaves Ground trash

25% 35% 10%

A rule-of-thumb value for the fiber content of fibrous trash entering the mill from the cleaner would be twice the fiber content of the cane, or 24 to 28%. Thus, in milling calculations, one ton of fibrous trash is approximately equal to two tons of stalk cane. If the fibrous trash is clean, the major effect of it would be to increase the fiber milled. Addition of 10% trash to clean cane would necessitate a 20% increase in capacity of the mill to handle the same amount of cane. Should the mill have sufficient capacity to handle the higher fiber rate, the setting can be adjusted so that the pol and moisture in bagasse can remain unchanged. Only the quantity of bagasse will increase, which means a greater loss in pol and lower extraction. Imbibition water would have to be increased proportionate to the fiber. The approximate magnitude of the loss can be illustrated example. Assume Cane

100 tons 12.5% pol 12.5% fiber

in an

55

Bagasse

2 5 tons 2.5% pol 50.0% fiber

.". Pol lost in bagasse = 25 χ 0.025 = 0.625 tons and Extraction = 12.5 - 0 625 x 100 1

ù

=

9 5

0 %

. 0

Now add 10% fibrous trash of 25% fiber Cane

110 tons 12.5x^100

Bagasse

=

1

.

I

3

6

%

12.5 + 2 . 5 x 100

=

1

3

p

>

ol

6 % 4

f i b re

30 tons 2.5% pol 50.0% fiber

. ' . Pol lost in bagasse = 30 χ 0.025 = 0.75 tons and Extraction = 12.5 - 0.75 x 100 12.5

94%

Therefore the combined effect of adding 10% fibrous trash would be to decrease the mill capacity by 20% and reduce extraction by 1%. A compensating factor is that the quantity of bagasse would be increased by 20%. The value of this as fuel for generating power would more than compensate for the loss in extraction. There are other effects associated with trash which may be of greater and even governing importance. Even with normal cleaning, fibrous trash, particularly ground trash, will carry into the mill soil, sand and rocks. These materials cause heavy wear on the equipment - knives, hammers, mill rolls, pipes and pumps. The result is a decrease in cane preparation and a lowering of the juice expression by the mills. Polishing of rolls will also decrease feedability and, thus, lower the capacity. The net effect can thus be a decrease in capacity of the mill substantially greater than that caused by the fiber alone. Extraction will also be lower than that caused by the fiber alone. Another factor is that soil carried into the mill decreases the permeability of the bagasse blanket holding the juice and lowers juice drainage, giving decreased extraction.

56

Trash is the cause of excessive water entering the mill from the cleaning plant since clean cane stalks can carry no more than about 2% superficial water. Green leaves, likewise, carry little water due to their waxy surface. Dry leaves and ground trash, however, can carry a large amount of extraneous water. This superficial water adds to the dilution of the juice without aiding extraction. In day-to-day operation of a mill, the effects discussed above will only be observed but cannot be measured. Experience alone will enable the operating staff to make the necessary judgment in appraising the daily results. DIFFUSION The discussion of the effects of extraneous matter on milling applies also for diffusion. The different technology presents a somewhat different picture, however. The capacity of a diffuser is determined primarily by the permeability of the bed of prepared cane, that is, the rate at which liquid can flow by gravity through the bed. The physical character of the fiber and the effect of the extraneous matter on this, therefore, are governing. The capacity then is the volume of juice that can pass through a quantity of cane per unit time. The size of the diffusion vessel, finally, is determined by the volume the bed occupies at an optimum bed depth. In diffuser design, a target extraction, say 98%, is set; then a practicable cell rupture value (Displaceability Index), say 94%, selected and the number of stages necessary calculated. A practicable bed depth has been found by experience to be about 1.5 m (5 f t ) . Bed permeability then determines the length of the bed at a given speed. To compensate for variability in permeability, more stages than calculated are normally designed. Some flexibility in operation is achieved by variation of bed depth and speed. Capacity of the diffuser then becomes a function of the third dimension, the width. Preparation of the cane was noted as the most important factor in diffusion efficiency. Maximum cell rupture, to free the juice, is necessary at the same time giving a fiber which is highly permeable. This was shown to be best obtained by a shredding machine which produces relatively long shreds resulting in minimum compaction when formed into a bed by gravity. Well-prepared cane will have a permeability averaging 250 liters/m^ (6.1 gal./ft2). - Fibrous trash, leaves and tops, generally make shredding of cane more difficult and greatly increases power required. Extraneous matter can have a controlling effect on diffuser performance mainly by the influence on bed permeability. Fibrous trash will increase the volume of the bed and so the capacity of the diffuser will be less in proportion to the volume occupied by the trash fiber. The effect is variable, depending upon the physical character of the trash. In general, shredded trash fiber occupies more volume per unit weight than shredded cane so that the capacity influence is greater than indicated by the weight of the trash.

57

The effect of clean fibrous trash on permeability is not great. Shredded dry leaves and green leaves can improve permeability. This is particularly true of green leaves, being waxy, improving liquid flow. Ground trash, being spongy, can have an opposite effect. However, unless the material is extremely fine (less than 300 mesh), fibrous trash does not have a major influence. Soil particles are the most serious suppressors of permeability. Clay type soils can almost completely seal a fiber bed and essentially stop diffusion if present in sufficient quantity. This leads to flooding - the most serious diffuser difficulty of all. With material finer than 300 mesh, the reduction in the permeability is proportional to the amount of the fine material up to a point - after which the amount is immaterial. This is because when an almost impermeable layer is formed, little liquid will pass through and the remainder will flow away on the surface of the bed. Other materials which can effect bed permeability are precipitated substances such as clarifier settlings. These, if returned to the diffuser, must be distributed through the bed and not placed on top. Sand, grit and rocks, of course, are without effect. In conclusion, as to the role of extraneous matter on diffusion, the fibrous trash has the effect of reducing capacity in proportion to its volume. Any other extraneous matter, soil in particular, of a fine particle size that decreases the bed permeability, can have a governing effect on the capacity of a diffuser. The impact on the preparatory and dewatering equipment is much the same as on milling. Because bagasse pol is lower in diffusion the effect of trash on pol loss is substantially less than in milling.

59

Chapter 7

A COMPARISON OF DIFFUSION AND MILLING WITH RESPECT TO RECOVERY AND LOSSES A question often raised, and obviously not subject to definitive answer, is how do diffusion and milling compare from the standpoint of recovery and losses. As with all equipment, it depends upon the capability of the installations and how they are used. There are, however, some basic principles to consider in looking at the technologies involved, and hence arriving at reasonable relative conclusions. Since the conventional extraction figure incorporates cane quality as well as extraction efficiency, it is more useful to look at bagasse pol as a measure of processing results. Basic diffuser design sets a target of 1.0% pol in bagasse. With good quality cane, this would correspond to extraction at the 98% level. Although in common practice mills are not designed to a bagasse pol target, the range for a good operation is 1.5 to 2.0 pol. Thus design consideration are based on losses in bagasse by diffusion 33% to 50% less than those for a mill. The extraction level for a mill is therefore 96% compared to 98% for diffusion. Diffusion takes place in closed equipment growth range for microorganisms. Loss by essentially nil. This contrasts sharply with perature system optimum for microorganism 1% have been reported.

at a temperature above the their activity is, therefore, milling - an open, low temactivity where losses above

Chemical inversion of sucrose, a product of temperature and pH, is negligible in a diffuser if a near neutral pH is maintained, since the average juice residence time is less than 15 minutes. With unlimed juice (pH 5.2), the calculated loss by inversion is 0.14%. With a mill, it would be less than 0.05%. It is true that the purity of diffusion juice is lower than the purity of absolute juice, as determined by direct disintegrator analysis of cane. However, it has been established that the purity change is not caused by a loss of sucrose but by an increase in dissolved solids (Brix). This is attributed to a solubilization of some high molecular weight components of the cane in the diffuser. No losses are involved therefore. It may be concluded from the above considerations that recovery by diffusion should be of the order of 2% greater than that by milling, in properly designed equipment operated at rated capacity. A good practical comparison of milling and diffusion is given by the results in Hawaii obtained by Hawaiian Commercial and Sugar Company in its two factories, Puunene and Paia. The factories are adjacent to one another and operate under as similar conditions as it is possible to find.

60

The Puunene factory had two milling tandems, each with four 198 cm (78 in.) mills preceded by a shredder and a 2-roll crusher. The Paia factory had a Silver 840 diffuser (2134 cm, 840 in. diameter) preceded by a Buster and Fiberizer, and followed by a 214 cm (84 in.) 5-roll pressure feeder dewatering mill. The results for the 1975 operating season were as follows: Extraction Bagasse pol Bagasse moisture Mixed juice purity Boiling house recovery Total recovery

Puunene

Paia

95.1 2.6 45.5 85.6 89.2 84.8

97.3 1.4 45.7 86.4 89.3 86.7

61

Chapter 8

CLARIFICATION The main objective at the clarification station is to increase the pH of the juice to a level where losses of sucrose by inversion are kept to a minimum in the subsequent sugar recovery process. Important, but secondary objectives, are the removal of insoluble material and some undesirable dissolved substances. It is fortuitous that adjustment of pH to the optimum level with the cheapest alkalizing agent, lime, gives satisfactory removal of much of the undesirable material in the juice and provides a suitable base for sugar recovery. Magnesium oxide behaves similarly to lime and is often used to reduce evaporator scaling. Target for pH adjustment of the juice is that which will give a syrup pH of 6.5. This is about the optimum for carrying out the crystallization steps which follow, by giving easy boiling massecuites, minimum development of undesirable compounds and color, small decomposition of reducing sugars and little loss of sucrose by inversion. A t higher pH levels, there is greater development of viscosity and color and substantial losses of reducing sugars, particularly fructose. A t lower pH levels sucrose inversion increases rapidly. Processing a syrup at 6.5 pH will usually give a final molasses pH of around 5.8, so the crystallization is conducted in the range of 6.5 to 5.8. The mixed juice will have to be raised to a pH of 7.5 in order to obtain a syrup of 6.5 because of the drop which occurs in the juice heaters, clarifiers and evaporators. This increase in acidity is caused by the relatively slow reaction with lime and particularly magnesium oxide in the cold, by formation of organic acids and by loss of ammonia from the decomposition of amino acids. The exact pH of juice liming required varies with the composition of the juice, so routine adjustment of set point is essential. Usually, with good quality cane, good clarification also occurs under this type of control. That is, there is good flocculation of suspended matter, rapid settling and overflow of limpid juice. With poor quality or deteriorated cane it is often impossible to get a clear juice and rapid settling. Milky-appearing juice is a sign of sour cane. It is caused by dextrans which, by protective colloid action, prevent good flocculation. In such cases, higher liming is sometimes helpful even though the effects on sugar crystallization are less favorable, so a compromise is necessary. In most cases, however, there is little that can be done to improve the situation. POLYELECTROLYTES Polyelectrolytes are the one class of additives, besides the alkalizing agent, that is so generally effective that its use in clarification is routine. In very small quantities, one to two parts per million of juice,

62

polyelectrolytes improve flocculation, give more rapid smaller volume of settlings. They have no effect on pH.

settling

and

Standard polyelectrolytes are partially hydrolyzed Polyacrylamides. The pioneer agents, used under the trade name of Separans, remain as standard. Preparation of the solution and point of application are important. High purity water, such as condensate, should be used to prepare the solution. Water with high dissolved solids or suspended solids will give a solution with inferior floe forming characteristics. The best point of application is after the juice is heated and a good place is in the flash tank of the clarifier. A metering pump is necessary to control the addition. The useful quantity of polyelectrolyte varies with the quality of the juice and the type of soil present. Usually, one to two parts per million of juice is sufficient. U. S. government regulations limit the quantity to not more than 5 parts per million. MAGNESIUM OXIDE Magnesium oxide (Magox) is an effective alkalizing agent for replacing lime becuse it leads to less scaling in the evaporators. Evaporator scale is composed mostly of calcium sulfate and silica. Since magnesium sulfate is soluble, less calcium sulfate is deposited on the tubes than when lime is used. Magnesium oxide has 40% more basicity per unit weight than quick lime, but even so is more expensive to use. When scaling is a problem, it can be used to replace enough of the lime to control the scaling. The amount must be determined by trial. Often, a 50-50 ratio is adequate. As the solubility of magnesium oxide is low, its rate of reaction with juice components is slow, making addition control by pH unsatisfactory unless a retention time of some 15 minutes is allowed. Although there is lag time in the heaters, better control is obtained from heated juice entering the clarifier. The quality of the magnesium oxide is extremely important as, depending upon conditions of calcination, the product can be highly absorptive (low temperature) or practically nonabsorptive (high temperature). For clarification use, the specifications are loss on ignition 2.5% and iodine number 20. PHOSPHATATION Clarification of juices deficient in natural phosphate is sometimes helped by increasing the phosphate content. In general, juices containing less than 0.03% phosphate are considered deficient. Adding phosphate to this level gives more calcium phosphate floe and often better clarity. Care must be exercised, however, because increased volume of settlings results and, usually, slower rate of settling.

63

Juices with excess natural phosphate (around 0.09%) are slow settling and give large volumes of settlings. Clarification of these is sometimes helped by reducing the liming pH. The best source of phosphate is phosphoric acid, a liquid which is expensive and presents a problem in handling. Normal practice is to use the cheapest and most available form which is fertilizer grade ammonium phosphate. SULFIT ATION Sulfur dioxide is widely used, both in cane and beet sugar processing, to reduce sugar color. It also has a secondary effect in improving the boiling characteristics of massecuites. The action of sulfur dioxide on color is complex, involving converting colored compounds to colorless, preventing color formation by oxidation and inhibiting color development from reducing sugars and amino acids by the browning reaction. Addition of the sulfur dioxide gas may be made before or after liming, with little apparent difference in effectiveness. The simplest application is to control the quantity entering the mixed juice relative to juice flow, then carry out the customary liming by pH control. In order to avoid the corrosiveness of sulfured mixed juice, partial preliminary liming may be used. To simply improve the quality of raw sugar, it is not necessary to have substantial absorption or retention equipment. Sulfur dioxide from a standard sulfur burner can be introduced proportional to flow into prelimed mixed juice by means of a venturi unit. A certain percentage of juice recycling will give uniformity. Liquid sulfur dioxide may be introduced directly into the mixed juice line, controlled by means of a flow meter. The quantity of sulfur dioxide used in improving the color of raw sugar is in the range of 100 to 500 ppm. This compares with around 1000 ppm used in the manufacture of a direct consumption plantation white sugar. JUICE HEATERS The objectives of juice heating are elimination of microorganisms by sterilization, completion of chemical reactions with the alkalizing agent, flocculation and removal of gases. Effective gas elimination is obtained by flashing as the juice enters the clarifier. The juice temperature must, therefore, be raised above the boiling point at atmospheric pressure which, at sea level, means a minimum of 103°C (217°F). If flashing does not occur, gas bubbles adhering to the floes decrease the rate of settling. Good steam economy is obtained by heating with vapors from the evaporator. Usually this is done in two steps with primary heating by vapors bled from the second or third effect and secondary heating with vapors from the first effect. The efficiency of heating is primarily a function of the heat transfer between the vapor on the outside of the

64

tubes and the liquid within. The optimum juice velocity is 2 m (7 f t ) per second which gives good heat transfer without too high pressure drop. The governing factor, however, is scaling of the tubes so periodic cleaning is required. Heater scale is relatively soft and can often be removed by cracking with steam followed by flushing with water. If this is not effective, flushing with caustic soda solution will usually remove the scale. Good heat transfer also requires removal of noncondensible gases and condensate discharge from the heaters. ROTARY VACUUM FILTERS Settlings from the clarifier, which contain 5 to 10% insoluble solids, are handled on a rotary vacuum filter to remove most of the insoluble material and wash the juice from it. The juice from the settlings, along with the washings, are returned to the incoming juice. A diagram of a vacuum filter station is shown in Figure 8-1. The filter is a rotating drum, the lower part of which is immersed in a tank of settlings. The drum is made of independent filter sections covered with a screen, usually of stainless steel with 0.6 mm (0.023 in.) perforations. The filter sections are connected by piping to a manifold at the end of the drum. The manifold opens to a rotary port valve which controls the connection to a vacuum system. See Operating Action Diagram, Figure 8-2. The filter sections immersed in the tank containing clarifier settlings mixed with bagasse fines are opened to a low vacuum of 18 cm (7 in.) of mercury called pickup. In this position, liquid flows through the screen and a cake is formed on the screen. After the formation of the cake, the bagasse fibers begin to retain the insoluble particles and the liquid passing through becomes relatively clear, as distinct from the first influx which is unfiltered. As the section continues to rotate through the settlings, the cake becomes thicker until it finally emerges from the settlings. Here, the valving connects with a high vacuum system of up to 50 cm (20 in.) of mercury and water is applied to the surface by sprays. The water passes through the cake, washing out the juice. A t the top of the drum, water is added by means of drip pipes which keeps the cake wet and the vacuum sealed until the vacuum is released and the cake scraped from the drum. The juice and washings are pumped to mixed juice and the cake is discharged to waste. Bagasse fines are used as a filter medium for retaining the suspended solids. These are obtained by screening bagasse and are mixed with the settlings before going to the filter tank. Bagasse Fines Fines for the filter are commonly obtained by means of a screen fixed in a bagasse chute. A suitable screen is flanged-lip type with 0.3 χ 0.5 χ 2.0 cm (1/8 χ 3/16 χ 3/4 in.) slot dimensions at head, tail and length.

Fig.

8-1 Vacuum filter station.

65

66

A. Β. C. D.

Fig.

PICK UP CLOSE WASH DISCHARGE

8-2. Vacuum filter operation action diagram.

67 2

2

Minimum area required is 500 c m (0,5 f t ) per ton prepared cane per hour capacity. The fines are pneumatically conveyed to a cyclone above the settlings mixer. A t the collection point, the stream of the fines is adjusted to fall past the intake in a manner so that sand drops out of the stream by air separation. Good quality screening should provide fines, 90% of which will pass through a 14-mesh screen. The material should contain fibers as well as pith in order to have good permeability. Mixing Fines with Settlings Fines are mixed with the settlings in a ribbon screw mixer. The fluidity of the mix must be maintained in order to prevent piling up of the bagasse on the surface. Since fluidity depends upon the consistency of the settlings, as well as the quantity of added fines, the ratio of fiber (fines) to insoluble solids that can be maintained is variable. A minimum ratio or 0.35 is considered standard and up to 0.5 is desirable. To maintain a 0.5 ratio, a quantity of bagasse fines equal to that of the insoluble solids must be added because the fines are 50% moisture. Thus, 100 tons of mixed juice at 1% insoluble solids would need 1 ton of fines. Pickup Vacuum At the filter pickup point, the vacuum applied should be just enough to hold the cake, but not enough to compress the cake that will reduce permeability; 18 cm (7 in.) vacuum is optimum. If a high vacuum is applied here, the cake will be compressed and the subsequent wash water will not penetrate in sufficient quantity to displace the juice. Wash Vacuum On the wash cycle, the vacuum must be high enough to provide juice displacement flow through the cake. A minimum vacuum of 38 cm (15 in.) is standard and 50 cm (20 in.) is the upper limit so that excessive flashing will not occur. Washing The cake must be uniformly covered with water during the wash cycle. It should not be allowed to dry out. Otherwise, the cake will crack, vacuum will be lost and cake penetration cease. The cake should always have a shiny, wet surface. Initial washing is with sprays set to give enough water to keep the cake wet without running down and eroding the surface. For these sprays to operate properly, clean water must be supplied at constant pressure. A pressure control valve is used on this line and a screen is essential. The water should be at a minimum temperature of 70° C. With a cake of good physical characteristics the quantity of wash water necessary can be as low as 0.5% on mixed juice and should not exceed 1.0%. Speed Standard filter speed is 10 revolutions per hour. The speed may be increased under some conditions and, in particular, with high ratios of

68

fiber to insoluble solids. In general, it is desirable to operate at the slowest speed compatible with a given composition of the settlings. Cake Pol Standard target for cake pol is 1.0%. conditions, the pol should be lower than this.

Under good

clean

cane

Cake Quantity A rule-of-thumb indicator of the quantity of cake is: Filter cake = 4 χ insoluble solids in mixed juice + 0.01 gross mixed juice

CLARIFIE RS A clarifier alkalizing step sterile product, give a syrup of of: 1. 2. 3. 4. 5. 6.

should provide the means of withdrawing juice from the in a condition suitable for sugar recovery. This means a relatively free from insoluble matter and at a pH level to 6.5. The equipment must, therefore, supply the functions

Gas removal Settling Scum removal Clear juice drawof f Settlings thickening Settlings removal

Present practice makes use of equipment in which the treated juice enters continuously with simultaneous withdrawal of clear juice, settlings and scums. The best design is that giving the minimum flow velocity at the entry and drawoff points, minimizing interfering currents. Units with multiple feed and drawoff points are less amenable to control. Flashing Clarifiers are equipped with a flash tank where the juice first enters and boils at atmospheric pressure, eliminating gases. The effectiveness is more a function of the area rather than the volume. Commonly, flash tanks have too small an area. The La Musse * formula is useful: Area for flash in m2 = 4.186 W ( T l - Tg) L Where W = Weight of juice in metric tons/hr T\ = Temperature entering, ° C T2 = Temperature at flash, ° C L = Latent heat of steam at T 2 , kcal/kg For a juice rate of 221 tons/hr and temperatures of 103 and 100°C, 2 respectively, the formula calls for 5.15 m area.

69

Settling Space The capacity of the clarifier is determined by the retention time necessary to allow the settling of a filterable mud. The most efficient design in this regard is the single mud compartment, multiple launder drawoff, of which the SRI (Australia) is a version. A retention time of 30 minutes or less is possible in these types. This compares with the multitray and multimud compartment units with retention time in the range of 2 to 2-1/2 hours. A disadvantage of short retention time clarifiers is that the juice flow must be relatively constant, for there is little surge capacity. The area involved in settling is a major factor in efficiency. The greater the area per unit volume, the more rapid the settling and the smaller the volume of settlings. Multicompartment clarifiers suffer more from uneven flow distribution and disturbing currents, even though total area may be adequate. Operation of Multicompartment Clarifiers Control of a multicompartment clarifier remains a manual operation. Drawoff of clarified juice is controlled by adjusting the drawoff valves to maintain the desired clarity, reducing the flow from those giving cloudy juice. This can be a time-consuming job since reducing the flow from one increases the flow from others in unequal amounts. Settlings drawoff is usually set to maintain a minimum consistency of 5% insoluble solids so that good operation of the filters is possible. With clarifiers having one mud compartment, this is relatively easy. For those with two mud compartments, it is difficult and for those with four, it is more difficult. Usually, more mud enters the first and last mud compartments than those in-between, so adjustments must be made accordingly. However, changing the amount of drawoff from one changes all the others and all the clarified juice drawoffs as well. As the quality of the incoming juice changes, so must the operation of the clarifier. The common solution to this problem is to provide excess clarifier capacity. This is not only capital expensive, but also increases losses because of longer retention time. The clarifier should be kept full at all times and not be used as a surge tank by allowing the level to drop. Holdover Juice Normal sucrose losses in clarification, exclusive of filtration, are of the order of 0.2%. The figure includes sucrose inversion, destruction and handling losses. Losses where juice is held in the clarifier for longer periods, such as weekend shutdowns, is subject to continued losses principally by inversion. These depend upon the temperature and the pH of the juice. In order to keep losses to a minimum, the temperature must be kept above 70°C to prevent the growth of microorganisms. The pH tends to

70

drop with storage, so addition of soda ash is useful to prevent a fall below 6.0. Normally, juice should not be held over 24 hours because of the difficulty of maintaining the temperature. The growth of organisms cannot be tolerated, as not only loss of sucrose occurs but subsequent sugar boiling operations are affected. Clarified Juice Screening Clarified juice usually contains a small amount of fine bagasse. This should be removed by flowing through a hillside screen of 80 mesh. REFERENCES 1

LaMusse, J. P., South African Sugar J. 61, (1977), 103.

71

Chapter 9

EVAPORATION The evaporation station performs the first step in the process of recovering sugar from juice - the evaporative removal of water. Standard practice is to concentrate the clarified juice to about 65 refractometer solids which requires removal of approximately 75% of the water. Steam economy dictates the application of the multiple effect principle. An appropriate installation makes use of a quadruple or quintuple effect arrangement with sufficient capacity to evaporate the water and is so designed to provide vapor for juice heating and vacuum pan operation. The station also furnishes boiler feed water from condensate. MULTIPLE EFFECTS In multiple effect evaporation the vapor from a vessel of boiling juice is used as the source of heat for a subsequent vessel. This can be done by reducing the pressure in the second vessel so that the boiling point is lower. In a series arrangement, or multiple effect, the Rillieux Principle states that one unit of steam will evaporate as many units of water as there are vessels, or effects. Thus, in a generally used four-unit series, or quadruple effect, one unit of steam would evaporate four units of water. The quantity is not absolutely correct but is close enough for routine estimations. Factors which cause deviations are: 1. Heat necessary to raise the temperature of the juice to the boiling point in the first effect. 2.

Heat losses by radiation and removal of incondensible gases.

3.

Increase in latent heat of the vapor as the temperature decreases.

4.

Decrease in specific heat of the juice as it is concentrated.

5.

Flash evaporation of the juice as it enters a lower pressure effect.

6.

Condensate flash.

The first three of these tend to decrease the total evaporation while the last three tend to increase the evaporation. The overall effect found in practice is that the total evaporation is somewhat less than the principle states. Calculations based upon estimations of the values of the above six factors show that the evaporation rates of the first three effects of a quintuple are less than shown by the Rillieux Principle whereas those of the last two, and in particular the last effect, are larger. However, because of scaling this is not found in practice. In order to have effective heat transfer, there must be an adequate temperature drop across each heating surface in order to transfer heat

72

through the tube to the juice, so the number of effects practicable is limited. The temperature in the calandria of the first effect is determined by the pressure of the exhaust steam available. Although it is advantageous to use higher pressure steam, general practice in cane sugar 2 2 operations is to utilize exhaust steam at not over 1.05 k g / c m (15 l b / f t ) (103 kPa) to enable better power production from the turbines. If the steam is saturated, the temperature would be 121° C. Actually, because of loss in the mains, the exhaust steam would enter the calandria of the first effect at around 116°C. Usually, the steam is superheated, and an average of 30° C superheat is not objectionable. Since it is not advisable to operate the last effect of the evaporator station at less than 55° C because of the poor heat transfer with high viscosity syrups encountered at lower temperatures, the total temperature drop across the station with saturated steam would thus be 61°C (116°C 55° C). With a quintuple effect, this corresponds to an average drop of 12°C across each body, a value that is about the minimum for satisfactory heat transfer, so six effects would not be desirable. The heat transfer would be better in a quadruple effect with an average temperature difference of 15° C, which is about optimum. The total effect of temperature increase in the juice caused by hydrostatic head also increases with number of effects making the average temperature difference smaller. The sum of all these factors leads to the generalization that converting a quadruple effect to a quintuple effect by addition of a body of the same heating surface does not raise the capacity of the evaporator unless the initial steam pressure is increased. A fifth body added at the front of the evaporator, and having a larger heating surface, would increase the capacity if the vapor produced by the added area is used for juice heating or pan boiling. The increase in capacity would be about equal to the quantity so used. In addition to temperature difference, heat transfer rate is proportional to the area of the tubes (heating surface), to the heat transfer coefficient and thickness of the tubes. The latter two have little import because with use scaling becomes the governing factor. The capacity of the evaporators can be increased by installing a juice preheater as an auxiliary ahead of the first effect. The heater, operating on exhaust steam, should raise the juice temperature close to the flash point. VAPOR BLEEDING As vacuum pans are single effect evaporating bodies, better steam efficiency can be obtained by heating them with vapor from one of the evaporator effects. The steam saving achieved increases with the downstream position of the effect from which vapor is bled, as represented in the formula: Steam saving =

73

Where m is the position of the effect and η is the number of effects. Thus, bleeding from the number one effect of"a quadruple would result in a saving of one fourth of the .weight of vapor withdrawn. It is generally desirable to maintain a positive pressure on the calandria of vacuum pans, so normal practice is to heat pans with vapor from the first effect. The same is true of secondary juice heaters where juice must be heated above the atmospheric flash point. Primary juice heating may be supplied by vapor from the second effect or, in some cases, the third effect. Vapor bleeding also reduces the quantity of condenser water required. CAPACITY The capacity of an evaporator station to remove water is established by the evaporation rate per unit area of heating surface, the number of effects and the location and amount of vapor bleeding. With no vapor bleeding, the capacity is determined by the performance of the least productive effect. The system is self-balancing. If a downstream effect cannot use all of the vapor developed by the preceding effect, the pressure in the preceding effect builds up and evaporation slows down until equilibrium is established. With effects of uniform heating surface, the productivity of the last effect is often the lowest on account of heavier scaling and high syrup viscosity. With vapor bleeding, the effect bled must have increased heating surface to provide for the vapor withdrawn, as well as the amount required for the downstream effects. Furthermore, all the bled vapor must be withdrawn, otherwise, as explained above, the system will drop to the evaporation level of the least productive effect. This means that the bled vapor, if not used in the pans or heaters, must be vented if the evaporation rate is to be maintained. This point must be carefully considered in controlling the station. If too much vapor is used by the pans and heaters, the pressure in the vapor will drop and there will be insufficient vapor for them and for the evaporators. Pressure can be maintained for the pans and heaters (vapor loop) by throttling vapor to the downstream effects. This, however, will reduce the total evaporation and in order to maintain evaporation, makeup steam must be added to the loop. OPERATION In operating the station, the exhaust steam supply to the first cell is controlled to give the total evaporation required to maintain the product syrup at the set range of 65 to 70 refractometer solids. Uniform juice feed is essential to good evaporator performance, especially with vapor bleeding. Regulation can be maintained by level sensing in the evaporator supply tank. Standard practice is to reduce the steam supply when juice level drops in the evaporator supply tank. With a demand remaining for first cell vapor, the vapor to the downstream cells is throttled. If the supply tank level falls below a set minimum, water can be automatically

74

opened to maintain evaporator operation. If the first cell vapor pressure drops too low, makeup with exhaust steam will be necessary in the vapor loop. The following factors evaporator station.

are

important

in

maintaining

an

efficient

Automatic Control Efficiency is enhanced by automatic essential elements controlled are:

control instrumentation.

The

Absolute pressure Syrup solids Liquid level Feed supply Absolute pressure is controlled by regulating the quantity of water to the condenser thereby maintaining a standard temperature of syrup in the last cell of about 55° C. The absolute pressure set point will depend upon the refractometer solids of the syrup. In the range of 65 to 70, the absolute pressure will be around 10 cm of mercury (4.0 in.). Syrup solids is controlled by regulation of the syrup discharge valve from the last cell. A minimum of 65 refractometer solids is standard. Concentrating above 70 should be avoided so that there is no danger of crystallization occurring. Liquid level in the tubes is an important factor in the heat transfer rate. If the level is too low, all the tube heating surface is not used and the tubes may dry out at the top. If the level is too high, the lower part of the tubes will be flooded with juice moving at too low a velocity to give maximum evaporation. The optimum level is that at which the liquid is just being carried by vapor bubbles to the top of the tubes with only a small flow over the tube sheet. It varies with the size of the tubes, temperature, heat transfer rate, scaling and viscosity of the juice. Under good conditions, the best level is around 25% of the tube length. Rarely is it over 40%. Since no flashing of feed occurs in the first cell, the situation differs from the others. If the entering juice temperature is much below the temperature of evaporation, the juice must be some distance up the tube before it starts to boil, resulting in higher levels. In the last effect, although the heat transfer rate is lower, the vapor expansion is much greater because of the lower vacuum giving a higher explosive force up the tubes. Feed supply should be kept uniform using the supply tank control. Dead band control is such that above a set level the supply is signalled to cut back. Below a set level, the steam reduced, and at a minimum level, a water valve is opened to evaporator going.

for surge source of supply is keep the

Condenser and Vacuum System With

satisfactory

condenser

design

and

adequate

vacuum

pump

75 capacity, the points of concern in operation are the quantity of water, water temperature and air leaks. Good condenser design will give the rated capacity with a difference of 3°C between the condenser discharge water and the entering vapor. The quantity of water required depends on its temperature - the higher the temperature, the more water necessary. Air leakage is usually the principal cause of evaporator malfunction. Individual vessels and the piping system must be checked periodically for leaks. Another common difficulty is air coming in with the juice feed. This would not be detected in testing the equipment for leaks. Condensate Removal Inadequate condensate removal can cause partial flooding of the tubes on the vapor side of the calandria with reduction of the effective heating surface. Since the first effect is always under pressure, condensate is readily removed by means of a steam trap. Condensates from the other cells, which are under vacuum, are removed by pumping. Condensates from individual cells are kept separate so that in case of contamination only the water from the leaking unit is discarded. Also condensate from the second cell contains volatile organic matter from the juice in the first cell. The principal substance present is ethanol but other alcohols, esters and acids are present and these are undesirable as feedwater for high pressure boilers. This condensate is better used for the house hot water supply than in the boilers. Noncondensible Gas A considerable quantity of noncondensible gas (air and carbon dioxide) may enter the calandria with the heating vapor. Air enters also through leaks in the vacuum vessels and carbon dioxide is generated in the juice. If not removed, these will collect and interfere with the condensation of steam on the tube surface. Calandrias under pressure can be vented to atmosphere. Those under vacuum must be vented to the vacuum system. Although it is easy to vent to the body of the same effect, the gas is then transferred to the next calandria and so on until removed in the condenser. Better practice is to vent the effects individually to the vacuum system. This requires careful valving to prevent appreciable vapor losses. A good guide is to monitor the temperature of the vent line. The closer it approaches the calandria temperature, the more vapor is lost. Calandria venting should be from both the top and bottom of the calandria and at points opposite the vapor entrance. Scaling Cane juice becomes saturated with respect to calcium sulfate and silica before the concentration of dissolved solids reaches the desired syrup level of 65 refractometer solids. The precipitation of these two compounds, together with small amounts of other substances, causes a build up of hard scale, principally in the last cell. Heat transfer is then greatly impaired. The quantity of scale deposited depends on the total concentration of

76

precipitable components in the juice, but the major constituent is calcium sulfate. Its concentration can be reduced by using magnesium oxide instead of lime in clarification, since magnesium sulfate is soluble. A l though scaling cannot be prevented completely, control to workable levels is possible by substituting magnesium oxide for part or all of the lime. Scaling is greatest near the vapor entrance and at the bottom of the tubes. Scale builds up with use and evaporation rate diminishes so that cleaning becomes necessary. The most effective cleaning agent is caustic soda which can be used directly as a 50% commercial solution. The efficiency of the cleaning solution decreases rapidly with dilution so care must be exercised to keep dilution to a minimum. Standard practice is to spray the caustic soda over the tube sheet with sufficient steam in the calandria to heat it. Treatment time averages two hours. The caustic solution may be reused but becomes less effective because of dilution and concentration of dissolved material. After caustic soda treatment, a powdery film will remain on the tubes. This is easily removed by washing with water followed by dilute sulfamic acid (0.25%) rinse and water wash. Entrainment Entrainment, or the carrying over of liquid with the vapor from one effect to the calandria of the next or to the condenser of the final effect, not only results in the loss of sugar but also causes contamination of condensate for boiler feedwater and pollution of the condenser water discharge from the factory. Juice is expelled from the top of the tubes with a velocity sufficient to atomize the liquid and project droplets to a considerable distance vertically. The velocity increases from cell to cell reaching a maximum in the final body where calculated velocities for 5 cm (2 in.) diameter tubes may reach 18 m (60 ft) per second. This means that some droplets have a velocity sufficient to project them 18 m. This is a hundred times that in the first effect. The problem is, therefore, most serious in the last effect and an efficient separator is essential. Separation, for the most part, works by effecting the impingement of the droplets on a surface from which the liquid can be returned to the cell. The main problem encountered with separators, if their design is adequate, is keeping them clean, as scaling takes place rapidly. Saddle or ring-packed separators, approximately 30 cm (12 in.) in depth extending across the top of the cell, are most effective provided they are also cleaned when the evaporator is cleaned. This involves spraying the caustic soda solution through the separator. Similarly, packed separators in the vapor pipe need also to be cleaned with the evaporator. They are less effective because of the high vapor velocity in the small diameter pipe. They are better installed in the enlarged sections. Stainless steel mesh separators extending across the body are effective if kept clean. Zig-zag plate units, also placed across the body, are effective but more expensive. With sufficient height above the calandria, simple flow reversal baffle-type separators in the dome serve well. Condensates should be monitored by conductivity routinely for entrainment. Condenser water should be tested for sugars on a scheduled basis.

77

Malfunctions Evaporator problems arise from many causes. The principal ones are: Low steam pressure Air leaks in system Condenser water supply Vacuum pump Noncondensible gas removal Condensate removal Scaling Vapor bleeding Difficulties with the steam supply, vacuum and water system are easily detected. Problems in individual cells with gas and condensate removal and scaling are best detected by observation of the temperature drop across the cells. Temperature and pressure measurements in each should be recorded on a regular basis. A malfunction can be traced by changes in these measurements. For example, if the temperature gradient across a cell increases, while the total drop across the set remains the same, then the drop across the other cells will decrease. This means a problem in the abnormal cell that requires investigation. It could be caused by failure of condensate removal or noncondensible gas removal. Of course, it could be caused by scaling, but this would occur gradually and would only occur to a great extent in the last cell. A sudden decrease in evaporation over the whole set might be caused by limited withdrawal of vapor to heaters and pans. Unless vapor is vented, pressure will build up in the vapor, reducing evaporation rate. This can be detected by pressure readings. EVAPORATOR CALCULATIONS In the design of a new evaporator station, many assumptions are made in order to arrive at the area of heating surface required in the multiple cells. Often, the designer is given only the one figure - tons cane to be handled per hour. He must, therefore, estimate the following: Weight of juice to be evaporated per hour. Refractometer solids content of juice. Temperature of juice. Exhaust steam pressure. Exhaust steam temperatures. Quantity of vapor to be bled. Heat transfer rate at the heating surface of each cell. Boiling point elevation of juice in each cell. Refractometer solids of syrup. Condenser water quantity and temperature. Heat losses. For an operator on a day-to-day basis, such estimations are of little value. He is faced with conditions in which the quantity and composition of the juice varies continuously, as does the exhaust steam pressure, the requirements for bled vapor, heat transfer rate, and the quantity and temperature of the condenser water. He can, however, make use of

78

approximations to arrive at problem areas in the evaporator set. Knowing the exhaust steam and juice flow to the evaporator by measuring temperatures, refractometer solids and pressure in the cells, calculations may be made on the basis of assumed heat transfer values. The Dessin-Coutanceau tion is: E

= v

1

method is useful for this purpose. The equa-

(100 - B ) ( t v - tj)(tj - t c ) + ( t v - tj) T670ÜÖ

Where

Ev Β tv tj tc

= = = = =

2

rate of evaporation in lb/ft /hr average brix of juice in the cell temperature of the vapor in ° F average temperature of juice in ° F temperature at vacuum in last cell in ° F

The constant of 16,000 applies for evaporators with tubes 1-3/4 χ 72 2 in., 5 lb/in. exhaust pressure and 26 in. Hg vacuum in the last effect. The values from the last cell are set into the Dessin-Coutanceau equation using measured values of solids, temperature of syrup, vapor to condenser and calculated temperature of vapor to the cell. Calculations for the other cells follow in order. The following example is from an actual case of an evaporator showing limited capacity. Basis To evaporate 157,400 kg (347,000 lb) water per hour from 196,860 kg (434,000 lb) juice at 13.0 solids. Quadruple effect with vapor bleeding from first cell. Cell number 1 2

Heating surface, m 2 ft

1,394 15,000 2

Evaporation rate, kg/m /hr 2 lb/ft /hr Vapor bled, kg/hr lb/hr

39.0 8.0

2

3

1,324 14,250

1,231 13,250

25.4 5.2

4

27.3 5.6

975 10,500 34.2 7.0

22,680 50,000

Water evaporated, (without bleeding) kg/hr 33,566 lb/hr 74,000

33,566 74,000

33,566 74,000

33,566 74,000

Total water evaporated, kg/hr lb/hr

33,566 74,000

33,566 74,000

33,566 74,000

56,246 124,000

Solids out

18.3

24.1

35.3

65.0

Solids average

15.7

21.2

29.7

50.1

0.7 1.2

1.1 2.0

2.0 3.6

4.3 7.8

Boiling point elevation, ° C °F

79

Vacuum °F

mm Hg

Pressure

Cell 4

°C

in. Hg

Vapor t ( t c ) Juice t4 ( t j 4 ) tc - tj4 At (drop)

54.0 58.3 4.3 20.0

129.0 136.8 7.8 36.0

64.8

25.5

Vapor t4 ( t V 4 ) 58.3 + 20 136.8 + 36.0

78.3

172.8

42.9

16.9

Juice t 3 (tj3> 78.3 + 2.0 172.8 + 3.6

80.3

176.4

tc - t j 3 80.3 - 54.0 176.4 - 129.0 At

26.3 8.3

47.4 15.0

Vapor tß ( t V 3 ) 80.3 + 8.3 176.4 + 15.0

88.6

191.4

26.2

10.3

Juice t2 (tj2> 88.6 + 1.1 191.4 + 2.0

89.7

193.4

t c - tj2 89.7 - 54.0 193.4 - 129.0 At

35.7 5.8

64.4 10.5

Vapor t2 ( t V 2 ) 89.7 + 5.8 193.4 + 10.5

95.5

203.9

11.4

4.5

Juice t\ ( t j i ) 95.5 + 0.7 203.9 + 1.2

96.2

205.1

tc - tjl 96.2 - 54.0 205.1 - 129.0 At

42.2 6.9

76.1 12.5

Vapor t i ( t v i ) 96.2 + 6.9 205.1 + 12.5

103.1

217.6

kg/cm

2

lb/in.

Cell 3

Cell 2

Cell 1

0.12

1.7

2

80

These calculations indicate that the evaporator should operate if the 2 2 exhaust pressure were 0.12 k g / c m (1.7 lb/in. ) and the exhaust steam were supplied at the rate of 56,246 kg/hr (124,000 lb/hr). However, a very high rate of evaporation was used for the fourth cell because of its small 2 2 heating surface - 34.2 kg/m /hr (7.0 lb/ft /hr). This is higher than normally encountered by about 50%. The conclusion would be that the evaporator cannot handle this quantity of juice. The limiting cell is number 4, and it should be replaced with one having more heating surface. A traditional heat balance can be developed at this point. However, it must be kept in mind that conventional heat balances are based on so many assumptions that they are of only limited value to the operator. The proper approach is to measure the conditions in the evaporator. Because of the numerous conditions which cause an evaporator to perform below calculated capacity, it is advisable to design with ample provision to take care of all possible situations ranging from low steam pressure to heavy scaling. REFERENCES 1

Staub, S. and Paturau, M. Principles of Sugar Technology. P. Honig (ed.), Elsevier, Amsterdam, 1963, Vol. Ill, p. 58.

81

Chapter 10

COMMERCIAL RAW SUGAR CRYSTALLIZATION Raw sugar of commercial standards (Hawaii 1980) requires crystals of reasonably uniform size 0.8 to 1.0 mm in length and 98.8 to 99.3 pol. Of other quality factors, color, ash and filterability, only color remains of major importance to the refiner. Marketable sugar meeting these specifications can be produced efficiently by crystallization from syrup in two steps. These are necessary because of the physical limits imposed by the centrifugal separation of the sugar crystals from the mother liquor or molasses. In the first step, A-massecuites utilizing syrup in the purity range 83 to 88, when more than 60% of the sucrose is crystallized, the massecuite becomes close to a solid mass which cannot be centrifuged. Crystallization is therefore kept below this point, and the molasses is returned for the second step or B-massecuites. In this crystallization, the limit for the percentage of sucrose in crystal form is lower because of the higher viscosity of the mother liquor, so that the maximum limit is not over 50%. The molasses from the B-massecuite is also returned for further crystallization, but the product cannot be used as marketable sugar. This C-massecuite sugar, or low grade, is too small in crystal size and too low in pol and must, therefore, be dissolved and used in the feedstock for the B-massecuite or, in a practice which is disappearing, as a magma of seed crystal for the B-massecuite. VACUUM PAN CRYSTALLIZATION In generating sugar crystals of relatively uniform commercial size, use is made of the fortuitous property of sucrose in forming extraordinarily stable supersaturated solutions. This reflects the reluctance of concentrated sucrose solutions to develop crystal nuclei, a property making it possible to establish a fixed degree of supersaturation in a footing of sugar liquor. Then a predetermined number of nuclei, as seed crystals, can be added, and by maintaining the supersaturation relatively constant with simultaneous feeding of sugar liquor and by evaporating, crystals of the average desired size can be grown. Pan operations are normally conducted at a standard constant vacuum of 125 mm Hg (5 in. absolute pressure). A-Massecuite A-massecuites are boiled on an initial footing of evaporator syrup. Syrup purities will vary from time to time, the usual range being from 80 to 90 (refractometer pol purity). In order to have relatively uniform conditions for seeding, it is generally advisable to establish a standard purity of the graining charge - one that can be maintained most of the

82

time. A reasonable value in most locations is 85. In case the syrup is of higher purity, Α-molasses is added to the footing syrup. Of course, if the syrup purity is below 85, nothing can be done but adjust the standard graining procedure. The footing is then concentrated to the degree of supersaturation optimum for establishing the grain. The degree of supersaturation can only be sensed by indirect means. Those commonly used are: Refractometer solids Boiling point elevation Electrical conductivity Consistency Refractometer Solids, as measured by a pan refractometer, is a simple method of estimating the supersaturation. Data are available in the literature showing the effect of temperature and purity on the 1 refractometer readings at various supersaturations (Figs. 10-1 and 10-2).

62

64

66 68 70 72 TEMPERATURE °C

74

76

78

Fig. 10-1. Purity-refractometer solids relative to temperature at saturation.

80

83

TEMPERATURE ° C

Fig. 10-2.

Purity-refractometer solids relative to temperature at 1.2 supersaturation.

Since the refractive index is highly temperature temperature correction or compensation is essential.

sensitive,

good

Boiling Point Elevation (ΒΡΕ) is difficult to measure and for that reason is the least used method. Proper instrumentation requires a sensing of the vapor temperature at the surface of the boiling massecuite compared with the temperature of boiling water at the same absolute pressure. Literature data relating ΒΡΕ to temperature and purity are available, however, and make this a useful technique (Fig. 10-3). Electrical Conductivity is simple and is reasonably self-compensating with regard to temperature. There is a problem in locating the electrodes and keeping them free from scale. Its sensitivity to ash content makes it difficult to use in areas where there are wide variations in the inorganic salt content of the incoming juice.

84

Fig. 10-3.

Nomograph of boiling point, refractometer solids and purity relationships at 1.3 supersaturation.

Consistency, up to the time of seeding, is a measurement of viscosity. Like conductivity, it is somewhat temperature self-compensating. Since the sensing rotor of a consistency probe is affected by the quantity of vapor bubbles if placed above the tube sheet and by the movement produced by a mechanical circulator if placed below the tube sheet, a suitable location is difficult to find. A serious disadvantage of the method is that viscosity is increased greatly by small quantities of dextrans and gums, and a false indication of supersaturation is given when deteriorated cane is being processed. In choosing pan instrumentation for seeding, it would be desirable to make a selection that could be used also after the grain is established. The possible choices are electrical conductivity and consistency, since refractometer solids and boiling point elevation both measure only the molasses and do not make necessary allowance for the effect of the crystals which is governing near the end of the strike. Both electrical

85

conductivity and consistency are sensitive to the crystal content, but they suffer the drawbacks described. The conclusion is that it is better to use one type of instrumentation for seeding and another type for the remainder of the strike. An effective system is the use of refractometer solids for seeding and consistency for the remainder of the boiling. Regardless of the type of instrumentation adopted, empirical relationships between the apparent supersaturation and the instrument readout must be established under the conditions obtaining. Considerable trial and error experience is thus essential. In addition to sensing and control instruments for supersaturation, it is important to have automatic absolute pressure control in order to maintain stable temperature conditions. Massecuite level sensing is useful for programming the changes inherent as crystallization continues. These include purity decrease of the mother liquor, which means an increase in viscosity, increase in crystal content and effect of the massecuite pressure head. Automatic steam pressure control is advisable also. The supersaturation at seeding point for an 85 purity footing should not exceed 1.15 to be in a zone safe from spontaneous nucleation. This corresponds to about 80 refractometer solids. A t this reading, the seeding slurry is added. This is prepared by grinding good quality sugar in isopropyl alcohol in a ball mill to a uniform size (see Chap. 12). The amount of crystals necessary for the completed strike can be calculated as shown in the example at the end of this chapter. Experience has shown, however, that 2 to 3 times the calculated amount must be used to give the right number of crystals. The reasons for this are multiple. The average size of the seed particles made by the method shown is 4.5 microns. Particles so small tend to conglomerate on standing. Another factor seriously reducing the effective number of particles is the flash that occurs when the alcohol slurry is introduced well above its boiling point. In the flash, some of the particles cake together; others are lost by sticking to the tubes and to the sides of the pan. Finally, some are dissolved when they reach superheated zones near the tube walls. The correct amount of seed must therefore be determined experimentally. It will vary with the type of pan and the manner in which the pan is operated. Once the correct quantity is determined, however, the procedure can be standardized. After seeding, the steam to the calandria should be shut off while the grain xbecomes established. This takes about 5 minutes, which allows the introduced crystals to reach a size where they have enough surface to absorb the sucrose introduced when the syrup feed is started. After the waiting period, steam flow is resumed but only at a reduced rate. A t this point water may be introduced for a short time to insure against reaching too high a supersaturation. When the crystals have grown to a size that they appear close together, automatic syrup feed can be started. This is a critical time and feeding must not be forced at a rate faster than the crystals can accept the sucrose, or new crystals may develop. Feeding is continued until the pan is full (1.5 m above the tube sheet). Massecuite is then cut to the finishing pans or a massecuite holding vessel.

86

The final sugar crystals will never be of uniform size, not only because the original ground seed particles lack uniformity, but also because apparently uniform crystals grow at different rates. This is caused by non-uniform conditions in the pan and also by the fact that crystals with molecular dislocations grow at a faster rate than those without. Fast growing crystals tend to be more uniform in size than slow growing ones for this reason. Also, regardless of how carefully conditions are controlled, there will always be some new nuclei formed in the boiling at all stages and stray crystals will appear. The best that can be done is to keep the quantity of these to a minimum. Good circulation is the major factor in preventing a large number of stray crystals. Normally, the volume of the massecuite at striking time must be a minimum of 8 times the volume at seeding to give crystals of average standard size. When the crystal size is suitable, feed is stopped and evaporation is permitted to continue until the correct solids content for centrifuging is reached. The refractometer solids at this point is approximately 92.5. The massecuite should have a crystal content above 50%. With a corresponding pan drop of close to 20 points, the A-molasses from the centrifugals will be in the range of 65 purity. Since remelt (dissolved C-sugar) is a mixture of relatively pure sucrose crystals and final molasses, it should not be used in the Α-strikes but should be sent to the B-strikes. In this manner about 50% of the total commercial sugar is not subject to crystallization in the presence of final molasses. Quality, in particular color, should therefore be better. B-Massecuite The purity at which B-massecuites are boiled depends on the purity of the Α-molasses and remelt. With an Α-molasses purity near 65 and remelt in the 80 to 85 range, the B-massecuite purity can be at the 75 level. A suitable footing is usually remelt. Pan procedures are similar to those for Α-strikes. Viscosity is greater and rate of crystallization is slower so more time is required. Because of the lower purity, crystal content will be less and a target of 40% is reasonable. Refractometer solids at the end of the strike should be at the 93 level. Crystal Content and Pan Drops It is important to obtain the highest yield practicable at each stage of the crystallization process. The yield is the percent of the sucrose in the massecuite which is in crystal form, as shown by the formula: massecuite purity - molasses purity 100 - molasses purity

m

a

s

s

e

c

u ei

f r a c to m e t e r solids

tr e

Massecuite purity minus molasses purity is usually called pan drop. This figure is the one commonly observed in comparing pan work, but the crystal content is a more useful value because of the sensitivity of yield to purity levels. For example, compare the crystal content at the same pan drop (20) for massecuites of 80 and 85 purity.

87

Crystal content 80 Purity = 85 Purity =

?n — — χ 92.5 = 100 - 60 — — χ 92.5 = 100-65

46.3 52.9

So the higher the purity the less the pan drop that is necessary to obtain the same yield. As a rough guide, it is not necessary to multiply by the refractometer solids so relative yields may be noted from the ratio of the pan drop to 100 minus molasses purity. Vacuum Pan Design For efficient operation, the vacuum pan must be designed for good circulation, essential to both rapid and uniform crystallization. In calandria pans, natural circulation by expulsion of the massecuite upward through the tubes by the action of expanding bubbles of vapor, and drainage down through the center well give effective distribution in a well-designed vessel. Circulation can be improved, nevertheless, by means of a mechanical circulator at the bottom of the center well. The critical points in design are short tubes, low head of massecuite above the tube sheet at time of striking, minimum volume between the bottom and the tube sheet, belt diameter the same as the calandria and proper heating surface in the tubes. Optimum length for standard 10 cm (4 in.) diameter tubes is 75 cm (30 in.). In longer tubes, the percolation effect is diminished. Shorter tubes do not provide enough heating surface. The tubes should be arranged around a center well of approximately 40% the diameter of the pan. Circulation diminishes as the level of massecuite above the tube increases, so the level at the time of striking should not be over 1.5 m (5 ft). This volume capacity should not be obtained by flaring the pan to a greater diameter above the tube sheet (so-called low head design) because the inventory of massecuite beyond the tube sheet suffers from poor circulation. The belt, therefore, should be the same diameter as the calandria. The pan bottom should be streamlined so that the massecuite flow is directed from the center well outward with a minimum volume of massecuite below the tubes. The necessary heating surface of the tubes depends to some extent on the pressure of the heating vapor used. In pans boiled on vapor from the 2 2 first evaporator effect, the pressure is a normal 0.5 k g / c m (7 lb/in. ). A good design would, in this case, provide a ratio of heating surface to 2 3 2 3 volume of 5 m / m (1.5 f t / f t ) . The location for a mechanical circulator is on the center line of the bottom tube sheet. The design should not be that of a ship's propeller to give forward thrust but to force the massecuite laterally under the tubes.

88

The speed depends upon the diameter but must be left low, 40 rpm is 3 optimum for a 600 hi (2000 f t ) pan. Procedures Instrumentation - Semiautomatic control - Graining by refractometer Boiling by consistency or conductivity 1. Automatic absolute condenser. Recorder. 2.

pressure

control

by

flow

of

water

to

Pan refractometer. Recorder.

3. Automatic feed control, sensing by means of consistency or electrical conductivity to maintain maximum supersaturation. Recorder. 4.

Automatic level control. Recorder.

5.

Pan vapor thermometer. Indicator.

6.

Pan microscope.

7.

Pressure gauge calandria. Indicator.

Operation - typical example 1. Draw in footing of just sufficient volume to cover the tube sheet when evaporated to the seeding point. Footing should be syrup if its purity is 85 or below. If syrup purity is above 85, add enough A-molasses to lower the footing purity to 85. 2. sure.

Concentrate to 80 refractometer solids, at 125 mm absolute pres-

3.

Shut steam valve. Keep mechanical circulator running.

4.

Add 30 ml slurry per 283 hi (1000 f t ) final massecuite.

3

5. When grain is established (5 min) open steam valve half normal and start syrup feed. 6. After 10 minutes, open steam valve to normal and maintain supersaturation of 1.2 until massecuite is 1.5 m above the tube sheet. 7.

Cut seed to striking pans or seed storage.

8. Continue boiling in striking pans until crystal reaches average size of 0.8 mm feeding Α-strikes with syrup and B-strikes with Α-molasses and remelt. If crystals have not reached 0.8 mm when pan volume reaches 1.5 m above the tube sheet, another cut must be made. 9. Drop strikes at 92.5 and 93.0 refractometer solids, respectively, for the A - and B-strikes. Calculation 3

Seed slurry required for 283 hi (1000 f t ) massecuite. Massecuite: 3

283 hi (1000 f t ) 93 refractometer solids

89

Crystal content 55% refractometer solids Crystal size 0.8 mm Weight of massecuite at 1.4 kg/1 = 39,620 kg Weight of crystal in massecuite = 20,265 kg 39,620 χ 0.93 χ 0.55 Seed required: 20,265 χ (M045)3 \ 0.8 ' 20,265 χ 0.0000001778 = 0.0036 kg or 3.6 g Slurry has 1000 g in 2 liters alcohol Total volume = 2630 ml 1000 ·+ 2000 = 630 + 2000 1.587 lmlhas

3

S - S =0

38

6

' = 9.5 ml 0.38 Estimated practical quantity is 3 times this or 28.5 ml (round number 30). Quantity of Massecuites The quantity of massecuites boiled is a minimum in a straight-forward boiling system with the only return being C-sugar to the B-massecuites. Any inboiling, that is, return of low purity molasses to higher purity massecuites, increases the total. The quantity also becomes greater as the syrup purity decreases. The following calculations illustrate this effect in examples of 85 and 80 purity syrups. For simplicity, the sugar has been calculated at 100 pol and no losses in processing are included. Straight A , B, C boiling is used with C-sugar returned to B-massecuites. Basic assumptions include:

85 Purity syrup 80 Purity syrup

A-Massecuite crystal yield

C-Massecuite purity

50 45

58 56

85 Purity syrup Tons Syrup Refractometer solids Pol Nonsugars

100.00 65.00 55.25 9.75

Final molasses purity 35 36

90

A-Massecuite Crystal yield 50% assumed Sugar crystal, tons = 27.63 0.5 χ 55.25 Refractometer solids to B-massecuite, tons = 37.37 65.00 - 27.63 Α-molasses purity = 70 85 - purity u 100 - purity '° C-Massecuite Assume:

58 purity massecuite 35 purity final molasses Crystal yield = 35% ό

100 - 35 "' * Nonsugars, tons =9.75 Pol, tons = 13.46 Pol - π c o 5 8 P o l + 9.75 " ° Refractometer solids, tons = 23.21 9.75 + 13.46 Sugar crystal, tons = 4.71 0.35 χ 13.46 Sugar (solids) at 80 purity returned to B-massecuite, tons = 6.81 4.71 0.692 100 \ 35

/

80

45 / \ 20

x

1

00 =

6

20 χ

1

00 =

3

ët

9

·

2 %

0

%

c r

Y

m

o

s t al

l

a

s

solids

ss e

s o l i sd

D D

Molasses (solids) returned to B-massecuite, tons = 2.10 6.81 - 4.71 Pol returned to B-massecuite, tons = 5.45 4.71 + (0.35 χ 2.10) Final molasses Nonsugars, tons = 9.75 Pol, tons = 5.25 _P£l = 0 35 U e DJ Pol + 9.75 Refractometer solids, tons = 15.00 9.75 + 5.25

91

B-Massecuite Tons Refractometer solids from A-massecuite Refractometer solids from C-massecuite Total Pol from A-massecuite Pol from C-massecuite

37.37 6.81 44.18 27.63 5.45

Total

33.08

Purity = 74.9 33.08 χ 100 44.18 40.2% Crystal yield 74.9 - 58 0.402 100 - 5 8 Quantity of massecuite Tons refractometer solids A-massecuite B-massecuite C-massecuite

65.00 44.18 23.21 Total

132.39

80 Purity syrup Tons Syrup Refractometer solids Pol Nonsugars

100.00 65.00 52.00 13.00

A-Massecuite Crystal yield 45% assumed Sugar crystal, tons = 23.40 0.45 χ 52.00 Refractometer solids to B-massecuite, tons = 41.60 65.00 - 23.40 Pol to B-massecuite, tons = 28.60 52.00 - 23.40 Α-molasses purity = 63.6 80 - purity 0.45 100 - purity C-Massecuite Assume

56 purity massecuite 36 purity final molasses

Crystal yield = 31.3% 56 - 36 = 0.313 100 - 36

92

Nonsugars, tons = 13.00 Pol, tons = 16.55 Pol ηcß 5 6 Pol+13.00 - ° Refractometer solids, tons = 29.55 13.00 + 16.55 Sugar crystal, tons = 5.18 0.313 χ 16.55 Sugar (solids) at 80 purity returned to B-massecuite, tons = 7.53 5.18 0.688 100 Ν

36

44 Ü

44 80

x

1

00 =

6

8

·

8 %

c r

y

s t al

s o l l sd

/

~

| ^ χ 100 = 31.2% molasses solids

Molasses (solids) returned to B-massecuite, tons = 2.35 7.53 - 5.18 Pol returned to B-massecuite, tons = 6.03 5.18 + (0.36 χ 2.35) Final molasses Nonsugars, tons = 13.00 Pol, tons = 7.31 =

3 6

Pol+13.00 °" Refractometer solids, tons = 20.31 13.00 + 7.31 B-Massecuite Tons Refractometer solids from A-massecuite Refractometer solids from C-massecuite Total Pol from A-massecuite Pol from C-massecuite

Crystal yield i l d = 33.0% 70.5 - 56 = 0.330 100 - 5 6

49.13 28.60 6.03

Total Purity = 70.5 34.63 χ 100 49.13

41.60 7.53

34.63

93

Quantity of Massecuite Tons refractometer solids A-massecuite B-massecuite Omassecuite

65.00 49.13 29.55 Total

143.68

These calculations show that with 85 purity syrup the quantity of solids handled is about double that incoming with the syrup. A t 80 purity it is of the order of 10% additional. Most of the increase is on the C-massecuite where the solids handled is greater by 27% and the final molasses is greater by 35%. REFERENCES 1

Honig, P. (ed.), Principles of Sugar Technology, Vol. 2, Elsevier, Amsterdam, 1959, pp. 358-359 (from Thieme).

95

Chapter 11

LOW GRADE SUGAR CRYSTALLIZATION Traditionally, the final step in sugar recovery is allowed to take place by cooling in crystallizers rather than during evaporation in vacuum pans. The reason for this is that the rate of crystallization becomes progressively slower as the molasses purity falls. A more cost effective procedure, therefore, is to boil low grade massecuite in a vacuum pan for a limited period of time, then discharge it into atmospheric crystallizers where sugar can be allowed to crystallize at length without evaporation. CRYSTALLIZATION BY COOLING The technology of crystallization by cooling differs markedly from that of crystallization by evaporation, and although crystallizers appear to require little attention, success in recovery depends to a great degree on the details of their operation. Two major factors are important in the technique. 1. A t a temperature range of 50 to 60° C and below, the rate of diffusion of sucrose molecules to the surface of a crystal exceeds the rate of deposition on the crystal; whereas at higher temperatures, the rate of diffusion is less than the rate of deposition (Dedek). 2. In this same temperature range (50-60° C), saturated molasses (and 1 supersaturated molasses also) has a minimum viscosity (Fig. 11-1 J. Application of these following conclusions:

facts

to crystallizer operation leads to

the

1. Mixing in the crystallizer is necessary only to prevent the crystals from settling and to aid in heat transfer. 2. Maximum crystal surface sucrose deposition. 3.

is essential

to a favorable rate of

The optimum holding temperature is in the range of 50 to 60° C.

These points must be superimposed upon the one basic fact that water is the principal component that keeps sucrose in solution. A t a given temperature, the saturated molasses purity varies directly with the water content. In order to obtain a lower purity molasses water must be removed, increasing the saturation temperature. But as water is removed molasses viscosity rises sharply, giving proportionately higher massecuite consistency. As the crystal content is relatively constant, molasses viscosity becomes the limiting factor. Since consistency must be kept at a level that permits massecuite pumping and separation of crystals from the molasses by centrifuging, molasses viscosity must be carefully controlled.

96

Fig. 11-1.

Viscosity-temperature relationships for saturated molasses at varying water content.

With the optimum viscosity-saturation relationship occurring in the 50 to 60° C temperature range, the water content of the massecuite should be targeted to give the maximum consistency of massecuite practicable. Such a practice will give the lowest attainable molasses purity for a given material composition in a fixed period of time. The control procedure for the crystallizers, therefore, is determined by the composition and physical character of the massecuite discharged from the pans. Crystallizer Design Design criteria for a crystallizer are directed toward heat transfer from and to the viscous mixture of sugar crystal and molasses. The most

97

effective design is that of rotating pipe coils through which water is circulated. These give enough movement to prevent crystals from settling and provide reasonable opportunity for heat transfer. They are more effective for heating than for cooling. When the coils are hotter than the massecuite, the molasses viscosity is less and the massecuite slides off the surface permitting access of cooler material. When the coils are cooler than the massecuite, the cooled material at the surface of the coil becomes more viscous and clings to the coil insulating it from the body of the massecuite. For this reason plate-type crystallizers are not effective for cooling heavy massecuite. The coils must be far enough apart that massecuite is not carried as a body between them. The coils must move through the massecuite giving time for exposing new material to them. The coils must be heavy and well supported. Since little agitation is required, a speed of 12 rph is sufficient. Massecuite level should be above that of the coils to prevent incorporation of air which causes an increase in the viscosity of the massecuite. Operation Procedure Instrumentation. 1. 2.

Thermometers. Recorder. Drive motor ammeter. Alarm.

Batch crystallizers. 1.

Cool the massecuite to the range 50-55° C. In an efficient water-cooled crystallizer, this will take some 15 hours. With cooling water at ambient temperature, it is practically impossible to cool a properly boiled massecuite rapidly enough to cause spontaneous new grain.

2.

Hold the massecuite at the chosen temperature by circulation of water at a slightly higher temperature for a minimum of 15 more hours. After this period, the supersaturation and molasses purity have become so low that the crystallization rate is exceedingly slow. It is, therefore, of little practical value to hold longer, although it can be done if capacity is available.

3.

Start reheating in the crystallizer by circulation of about 60°C water. This can be done for a variable period of time depending upon the need for the crystallizer for subsequent strikes and the availability of massecuite heaters. The object of reheating is to reduce the supersaturation to close to zero before centrifuging. With the use of massecuite heaters, reheating in the crystallizer can be kept to a minimum sufficient to improve the flow of massecuite from the crystallizer. Otherwise, several hours reheating will be necessary. The temperature to which massecuite is reheated, either in the crystallizer or heater, will depend upon the

98

saturation temperature of the molasses. Since it is not practical to determine the saturation temperature on each strike of massecuite, a general temperature is chosen, based upon periodic measurements on typical samples and experience. Continuous crystallizers The same principles hold for the operation of continuous crystallizers with necessary variations in procedure required by the individual installation. For continuous operation, the massecuite must be discharged from the pan into a strike receiver which acts as a reservoir to feed the bank of crystallizers. The preferred number of units in a bank is 7, permitting 2 or 3 for cooling, 2 or 3 for holding and 2 for reheating. In order to make coil repair possible without emptying two units, flow is from the top at one end, out the top of the other. Surface flow is prevented by a horizontal baffle plate located one third of the way from the entry point and extending not more than 50% of the depth of the massecuite. Troughs between the units should be wide and shallow to give a channel rather than a tube-like flow. A possible water flow diagram is shown in Figure 11-2. Cooling water, at the chosen temperature, flows through the second crystallizer and then through the first. The reason for counter current flow is to avoid, to some extent, too rapid initial cooling causing solidification around the coils and bypassing of massecuite. Heating water enters the last body and flows through the next to last, then out to a tank where it mixes with cooling water and is temperature controlled for holding water which goes through the third, fourth and fifth bodies in series. Temperature control is required only at two points - the holding water and the heating water. Quantity of water must be controlled relative to the flow of massecuite. In the operation of continuous crystallizers it is important to keep in mind that the retention time is a function of the rate of flow, which is set by the quantity of massecuite that must be handled per unit of time. The quantity of low grade massecuite is governed largely by the purity of the syrup. As an example of the wide variation, the quantity produced from 80 purity syrup would be about double that at 88 purity per unit of pol in the syrup. Therefore, unless the rate is controlled, the retention time with massecuite from 80 purity syrup would be only half that from 88. For optimum operation, therefore, the flow through the continuous system must be monitored and controlled. VACUUM PAN CRYSTALLIZATION Low grade boiling procedures are set by the crystallizer operation discussed above, the objective being to provide a massecuite of a consistency and purity which will give maximum recovery in the crystallizers.

99

Ο m LÜ

U

Ο LÜ

3

Lü

Hl

< 5 ο χ

CO

(Τ Ο

ΙΟ

"Mr (Τ LÜ *< ο

Ο Ο

If)

C M

rÜJ

ϋ Ο LÜ

ω