ENGINEERING RESEARCH INSTITUTE UNIVERSITY OF MICHIGAN ANN ARBOR PROGRESS REPORT V PROCESSING OF SUGAR BEETS -By L. E. BROWNELL S. A. ZIEMINSKI TEH LEE Project 2047 US DEPARTMENT OF AGRICULTURE CONTRACT NO. A-ls-334164 May, 1954

v&- 9 g V N-,Po- (- g v.. 57-, ~

TABLE OF CONTENTS Page LIST OF TABLES LIST OF FIGURES v ABSTRACT I. REVIEW OF THE PREVIOUS WORK 1 II. GENERAL OUTLINE OF THE PROPOSED METHOD OF RECOVERY OF JUICE FROM SUGAR BEETS 2 A. The Steam-Pulp Separator 4 B. The Juice Separator 4 C. The Exhausted Pulp 4 III. DESCRIPTION OF THE EQUIPMENT USED 6 A. The Feeding Device 6 B.o Ejector 9 C. The Disintegrator Pipe 9 D. The Impingement Baffle 11 IV. THE EXPERIMENTAL TECHNIQUE 11 A. Preparation of Cossettes 11 B. Preparation of 'a Uniform Sample 13 C. Disintegration Test 14 V. PRELIMINARY RECOVERY TESTS 15 VI INFLUENCE OF THE TYPE OF IMPINGEMENT BAFFLE ON THE DEGREE OF DISINTEGRATION AND-ON SUGAR RECOVERY 17 VII. T E INFLUENC F C ICAL PRET ON THE INFLUENCE OF CICAL PRETREATMENT ON THE DEGREE OF DISINTEGRATION AND SUGAR RECOVERY 21 VIII. INFLUENCE OF STEAM PRTREATMENT OF COSSETTES ON THE DEGREE OF DISINTEGRATION AND SUGAR RECOVERY 23

TABLE OF CONTENTS (cont ) Page IX. QUALITY OF THE JUICE PREPARED FROM PULP AS COMPARED WITH THAT OF THIE DIFFUSION.JICE 25 X. RECOVERY TEST AT HIGHER CENTRIFUGAL FORCE 29 XI. CONCLUSIONS AND PLANS FOR FUTURE 31

LIST OF TABLES Numboer Title Page I Mixing Test 13 II Preliminary Recovery Tests at Low Capacity 16 III Influence of Type of Impingement Baffle on Sugar Recovery 19 IV Influence of Chemical Pretreatment 'on Sugar Recovery 22 V Influence of Steam Treatment on Sugar Recovery 24 VI Juice Prepared by the New Method vs Diffusion Juice 26 VII Total Nitrogen in Raw and Thin Juices 29 VIII Recovery Test at Higher Centrifugal Force. 30 iv

LIST OF FIGURES Number Title Page 1. General Flowsheet of the Proposed Method 3 2. The Cylindrical Baffle 5 3. The Disintegrator and Feeder '6 4. The Feeder 5. The Disintegrator.without Feeder 10 6. The Ejector Chamber with the Nozzle and Inlet Pipe Withdrawn 10 7. The Slicer 12 8. The Mixer 12

ABSTRACT This report presents the results obtained using continuous equipment for.disintegrating sugar beets. The principle of the proposed method of recovery of juice from sugar beets is briefly reviewed. The operation of the disintegrator is described together with. details of the experimental technique. A series of experiments have been performed to determine the degree of disintegration and.sugar recovery as influenced by the design of the impingement baffle and by the chemical and heat pretreatment of cosstettes. Three comprehensive experiments have been carried out in which the juice from the pulp, as well as the diffusion juice, were submitted to the standard purification process. The characteristics of raw and thin juices prepared by these two methods were then compared. In neither of these runs was the purity of the thin juice prepared by the new process lower than that of the thin juice obtained by the usual method. The actual processing capacity and steam consumption have been brought close to the corresponding design values given by Progress Report IV. A sugar recovery of 99 percent has been reached at a draft of approximately 130 percent by weight. At the end of the report suggestions are made concerning the further study and development of the proposed method for the recovery of juice from beet pulp. vi

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN PROGRESS REPORT NO.V PROCESSING OF SUGAR BEE.TS I. REVIEW OF THE PREVIOUS WORK The main point of the proposed method for the recovery of juice from sugar beets is the preparation of a beet pulp sufficiently fine to en-. able a good recovery of juice by a direct method of separation such as centrifuging., filtration, etc. followed by washing. Therefore, the previous studies, as well as the work presented in this report, deal mainly with the method ofr disintegration of sugar beets. The feasibility and economy -of the preparation -of beet pulp on a commercial scale depends mainly on the following factors: -1. The quality of the juice should not be decreased by the process. 2. The equipment should be low in cost and easy to maintain. 3. The time of 'retention should be short so as to reduce the size of the equipment and prevent or decrease the action.of'microorganisms. 4. The process should be continuous to facilitate the supervision of the equipment and to decrease the costs of labor. 5- The cell structure of the beet should be ruptured as completely as possible without, however, producing a great amount:of fines. This last point is of importance for easy centrifuging or filtration. 6. The pulp should be sufficiently disintegrated to permit good recovery of the sugar from the beets.

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN 6. The pulp produced should not undergo appreciable dilution during the process sof disintegration. Such dilution would result in a higher draft and a decrease in the ecohnomy of the process. 7. The steam or mechanical energy consumption should be within reasonable limits and correspond to the conditions existing in an average sugar factory. In previous experiments the cossettes were preheated to a tempera"ture of 10OC and then brought in contact/ (for about 2,seconds) with saturated steam (50 psig) and the pressure suddenly released The cossettes were accelerated by the expanding ste and wer then allowed to strike against an impingement baffle. It was believed that the stress caused by the sudden -evaporation of a part of the cell content togther with the stress developed on impact was responsible for disintegration. However, in later experiments (described in Progress Report III) it- was found that evn with the use of.ompressed airy instead of steam -a reasonably good disintegration could be obtained. To minimize dilution and the possibility of a decrease in the quality of the juice, it was decided to utilize rupture by impact in preference to rupture by inflation and to design a continuos disintegrator in which the particles would be accelerated to high velocity and ruptured essentially by a sudden change of momentum. The constructional details of the disintegrator and accessory equipment were described in Progress Report IV. The description of the disintegrator and its operation will be reviewed in the next pages of this report. II. GENERAL OUTLINE OF THE PROPOSED METHOD OF RECOVERY OF JUICE FROM SUAR BES The flow sheet given in Fig. 1 presents the general idea of the proposed method of production of raw juice. Cossettes or beet particles of other shape are introduced into the hopper of the disintegrator from which they a cae carried away by the impeller of the air locke A part of the expanded steam leaving the separator is recycled into the lower part of the air lock to fluidize th~ cossettes. The fluidized mixtur-e of cossettes and exhaust steam enters the ejector chamber. In the ejector chamber the high pressure steam issuing from the nozzle transfers some of its:momentm to the mixture of vapor and coasettes which enters the chamber as a result of the vacuum created by the high-velocity stream. The cossettes -are -accelerated to a high velocity and strike an impact surface arranged in the separator where they disintegrate. The resulting pulp is then removed from the separator either by keeping it at a pressure Slightly higher the atmospheic or by a suitable revolving scraper. The expanded steam is partially recycled to the air lock and the rest is used for juice heaters, etc. The juice is

Slicer Hopper To Heaters Exhausted Air Lock Concentrated Pulp Exhaust Recycled To Drier Exhaust Raw Juice Baffle and H P Steam Steam Separator Nozzle Pulp Fig. I Centrifuge, Filter or other Separating Device General Flow Sheet of the Proposed Method

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN removed from the pulp in a centrifuge, filter, or-other separating equipment with the aid of a wash and the exhausted concentrated pulp sent to a drier.,In the experiments presented in this report the amount of cake (dehydrated pulp) varied from 30 to 21 percent beet, depending on the centrifugal force used. This pulp needs little or 'no pressing before it is sent to the drier. The cost of the Juice separator therefore would be partially offset by the cost saved through the elimination of pulp presses. As has already been mentionedu the study of this new process was concerned 'mainly with the production -of pulp, with the qualities of the obtained juices, and with the sugar recovery from pulp in a centrifuge..The following points of the general flow sheet have not yet been studied. A. The SteamiPulp Separator A separator which would also act as -an impingement baffle was designed and constructed (see Progress Report IV, page 16 and Fig. 2 of this report ),.owever, the short.duration Of each experiment (30 to 60 seconds) did not allow the adjustment of the separator during a run. As the construction of the separator depends to a great extent on the type of baffle used, it was decided to postpone the study of this problem until more data are available on the optimum design of the baffle. B. The Juice Separator In all the experiments a centrifuge was used for separating the juice from the pulp. The results obtained were very encouraging. The juice separated easily and good recoveries were obtained at reasonable drafts. It is, however, possible that a rotary filter may prove more advantageous. C. The Exhausted Pulp The; drying properties of the concentrated..exhausted pulp have not yet been studied, It has been noticed, however, that the pulp cake when left.in the open air dries fast:and even after partial dehydration the particles stick firmly together giving a layer of cake with considerable mechanical strength. Therefore, -it may be advantageous to pass the- pulp through a pair of rollers to produce flakes before introductiont the drier.

Fig. 2. The Cylindrical Baffle Fig. 4. The Feeder

e IL Induced Air d ~~m F~~~~~~O h IISteam~ IIp Ps g Fig. 3 Disintegrator and Feeder

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN -III. DESCRIPTION OF TTHE EQUIPMENT USED The design of the disintegrato and feeder is shown in Figs. 3, 4, 5, and 6 of this report and Figs. 7 to 29 of Progress Report IV. A. The Feeding Device The purpose of the feeder (Figs. 35 and 4) is to obtain a constant and uniform flow rate of the particles so that the ejector can be uniformly and continuously charged without appreciable pulsation in the flow. This point is of special importance because of the small cross sections of the conduits: in the experimental plant and the consequent danger of choking. Increasing the dimensions of the disintegrator pipe and diffuser throat was impracticable as that would require a processing capacity too great for laboratory conditions. This problem, although of lesser importance in a largescale installation, is of great importance in a -smallscale plant. Feeder A (Fig. 3) consists of three parts made of bronze. The upper cylindrical part, 12 inches in diameter, can easily accomodate about25 lb of cossettes. The bottom of the upper part consists of a system of two cones. The outside cone,.b. is made of stainlesssteel screen to allow the passage of condensate in the case of steam pretreatment. The closing cone, a, is fixed on the stirrer rod, c, which makes a reciprocating motion up and down. The length of the stroke, the number of strokes per minute, as well as the position of the cone, a- on the stirrer rod can be adjusted. The reciprocating motion of the cone loosens the meass of cossettes and allows the partcles to flow through the central cylindrical section of the feeder down to the conical bottom From there the coassettes are transferred by the rotating blades to the lower part of -the rotary feeder, B, where they meet the trem of the induced fluid introduced through the front plate of the casing (-ee Fig. 4). As a result of the vacuum existing in the ejector chamber, C, the cossettes- together with the induced fluid (steam or air), are carried away into the ejector chamber. In the experiments presented in this report air was used as induced fluid instead of steam. There were several reasons for introducing this change. First, it was noticed that especially at higher capacities there was a tendency on the part of cossettes to choke the entrance to the rotary feeder. This was caused by the small size of the equipment as well as by the surface characteristics of cossettes. To. prevent this harmful effect air was introduced tangentially at the loest part of the cone, e, to kee the | cossettes in a swirling motion, and thusallow a easy charging of the rotary feeder. If steam were used in the rotary feeder and the pressure fell belaw stmospheric, the air used fr stirring would bled t past the clearances of the rotary feeder and becme mixed with the steam. The..~~~~

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF- MICHIGAN determination of -the' amount -of the inducefd wfluidould then become difficult and unreliable. Also, if the pressure in the feeder increased ab-ove atmospheric a leakage of steam would occur in the direction upards to the cassettes resulting in some stea r condensation on the cossettes. As there was no time for improving the air-locking property-of the rotary feeder the following procedure was adopted: a known _amount of air* from a critical flow nozzle was introduced into pie,i, (Fig 1) from which a-pt was used for stirring of c-ossettes at the bottom of the cone a the rest was passed directly to the lower 'part of the feeder. Since the upper part -of the feeder 'was tightly closed, the part of the air which was used for stirring eventually entered the ejector chamber passing thrgh the cl-earanes of the rotary feeder. As the pressure of the air in the rotary feeder.was always below critical with rspect to the pressure upstream from the nozzle the regulation of flow was simple-. accurate and indpendent of changes of pressure in the feeder. Thus, the use of air not only.simplified the experimental procedure, but also permitted aecurate metring of the induced fluid. Obviously, in commercial operation steam should be used not air to minimize dilution of steam with air. As the feeding device is -an -important part of the disintegrating unit 'a critical survey- of its operation is given. The chief requirement for good operation of a feeder is the production of a uniform flow rate of par.-ticles so as to decrease the pulsations.. As a result of the reciprocating motion of the stirrer (150 to 250 strokes per minute) the charging of the compartments of the rotary feeder was not uniform. To improve the uniformity of flow the speed of th impeller was inicreased above the value necessary for a givn capacity and an orifice was inserted in the outlet of the rotary feeder (see Progress Report IV, Drawing 12). The purpose of this orifice was to pass at a given flow of air a required -amount of cossettes.. When. as a result of the fluctuation of the flow of cossettes some of the compartments of the impeller were overcharge a part of the cossettes would then be recirculated to compensate f or any subsequent undercharging of compartments. But ',even assuming a perfect operation of this arrangement there will always be a certai onpoSatifon.o discharge from the rotary feeder. At the usual velocity.(160 rpm) of the impeller consisting of -six compartments -the frequency, of- the discharge pulsation 'will be 16 pulses per second. Even though this frequncy is small it cannot be neglected considering the high velocity of particles, especially at the outlet of the disintegrator pipe (see Fig. 3,F). Therefore.; it is most probable that at a given average capacity the ejector and the baffle may be temporarily overcharged It is believed that further imrovement in the uniformity of flow will considerably increase the steam economy of this system. 2.41 'llb per minute in all expriments presented in this report,. I~~~~~

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN _B. Ejector The feeder was connected to the ejector (Figs. 3, 5, and 6 of this report and Drawing 20.of Progress Report Iv) with a 1-inch standard brass pipe 90 inches long. In all experiments presented in this report a dry saturated steam of t10 psig was used as a motive fluid. The flow rate of the steam was 3.53 lb per minute in each of the experiments. The steam nozzle (Fig. 3, h) was of convergent-divetrgent type with a 0.200-inch throat diameter and 0.3-17-inch exit diameter. The total angle of:convergence of the entrance to the diffuser throat (Fig. 3, t) was 25~ and the diameter of the diffuser throat was 0.625 inch. An importat factor for good operation of 'th ejector -is the way in which the cossettes:and air are introduced into the chamber. Logically the best way would hae been to introduce the cossettes axially with the direction of flow of the motive steam to avoid or decrease the change of direction of the flow of the cossette-air mixture. There are several ways of solving this problem, however, each of them presents serious constructional difficulties when applied to small-scale equipment. One possibility is to use an annular nozzle instead of the one described and to introduce the cossettes throug the enterline of the nozzle. For an annular nozzle of 2-inch diameter and having the same cross-sectional arzea of the throat as that used in these experiments,, the width of the annulus would be only 0.005 inch. This idea was disgarded because of the difficulty of construction and ease of damage and wear to an annulus of such a close tolerance. The use of a ring of small separate nozzles presented similar constructional difficwulties. The actual solution is presented in Figs. 3 and 6. The cassette inlet pipe was introduced deep into the ejector chamber and was slightly flattened and bent in the direction of -the flow of steam. For other constructional details see Progress Report IV. C. The Disintegrator Pipe The purpose of the disintegrator pipe is to provide a necessary length of path for acceleration of the particles. Two different pipes were used in the experiments presented in this report: (1) a 3/4-inch standard brass piPe 37-5/16 inches long which was connected with the ejector chamber by a short diffuser, D, (Fig. 3); and (2) a 0. 5-inch standard brass pipe 16-5/8 inches long which was connected directly with the chamber by means of an adapter. In the latter case no diffuser was used, with the smallest cross section of the ejector being equal to the cross section of the pipe. 9~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Fig. 5. The Disintegrator without Feeder Fig. 6. The Ejector Chamber with the Nozzle and Inlet Pipe Withdrawn 10

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN D. The Impingement Baffle The main purpose of the baffle is to provide an efficient impact surface for the cossetteso As the baffle is the part of the equipment where the particles undergo disintegration its design and construction affects the efficiency of the whole system. On the basis of the observations made a good baffle should fulfill the following requirements: 1. It should permit every particle to strike its bare surface. 2 Its shape and construction should decrease or prevent the deflection of the incoming particles caused by those rebounding from the baffle. 3. The pulp should be removed from the surface of the baffle as soon as it is formed to decrease the harmful cushion effect..4. It is also believed that the impact surface should exhibit a cutting or abrasive action on the particles to assist in their disintegration. In general a baffle of low efficiencywould require a higher' velocity of particles and a greater steam consumption to give an adequate disintegration as compared to that obtained with a more efficient baffle. Several types of impact surfaces have been investigated and the results are presented and discussed in the later part of this report. IVo THE EXPERITENTAL TECMIQIE A. Preparation of Cossettes As a result of the small diameter at the diffuser throat (0.625 inch) the length of the particles used in the experiments varied between 1/4 to 1/2 inch. In order to prepare cossettes-of- approximately equal length all the beets used for anexperiment were first cut to the same length (6 inches) and then a number of incisions were made in the direction perpendicular to` the long axis of the beet. The distance between the incisions was: equal to the required length Of cossettes. A small portion Of the beet was left uncut so that after removal of the beet from the cutter the slices were held together by the uncut portion.; After the incisions were made the beet was transferred to the slicer (see Fig. 7 and Figs. 1 to 5 of Progress Report IV) where it 11

~i~:ii-ii~iiiii:iiiii~ii:,:::iiiii:::::: —:::iiiii:gii::~ii ~:ii:ii: iiiiiiiiiiiii~11ii~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~............!iliii:"iiiii::.-ii iiii ---:i.:::::::::~~::::::::::::; ii ~::'::iiiiiiiiii iJir~i.iiiiiiiiiiiiii RD Fig. 7.The Slicer F'ig. 8.The Mixer

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN was cut with a beet slicer knife with vertical splitter, 46 divisions in 165 mm length. The setting of the knife was the same in all experiments. The V-shaped cossettes had a total length of 13.3 mtr per 100 gm and the leigth of particles varied from 6 to 12 mm. It was difficult to obtain a better uniformity of the length of the particles due to the irregular shape of the beet. The particles prepared from each beet were then passed through a screen to remove slabs and chunks. B. Preparation of a Uniform Sample As a portion of the prepared cossettes was used for the actual experiment and the rest for preparation of reference juices it was imperative to have a uniform sample of cossetteso For this purpose the cossettes were placed ina tumbling type of mixer (see Fig. 8 of this report and Fig.. 6 of Progress Report IV) and mixed for 6 minutes. This typeof mixer was found to be very reliable. It has a good mixing efficiency and does not damage the comparatively soft particles. It was made entirely of brass and adjusted to 10 rpm. Table I gives the results of a test in which 20 lb of cassettes (length 6 to 12 mm) were mixed for 6 minutes. The sugar:content of the cossettes was then determined in six small lots taken at random from different areas when the whole mass was spread on a flat surface. Table I shows that the difference in sugar content was within the experimenrtal error of deter. mination. TABLE I MIXING TEST Sample Pol.1 18.27 2 18.30 3 18.23 4 18..30 5 18.20 6 18o23 13XI

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN C. Disintegration Test In the upper compartment of the feeder 6000 and 7000 gm of cossettes were placed and the cover tightly closed,.Dry saturated steam was introduced into the nozzle of the ejector and the presure adjusted to 110 psig. The required amount of 'inducd air was then introduced into the upper and lower part of the rotary feeder and its speed adjusted to 160 rpm. By starting the reciprocating stirrer of the feder the ceossettes were introduced into the ejector The pulp was colleted in,a brass container placed under the baffle. To decrease the variation in feed rate the experiment was stopped before a11 cosstes charged in the feeder were passed through the ejector. After the experiment the cosettes remain ing in t er compartment above the closing coe were removed from the fder and weighed to determine the actual processing capacity The duration of the experiment varied from 30 to 60 seconds depending on the capacity usedo The temperature of the pulp was detind and a well-mixed sample of about 2 lb was collecte-d in -a closed j cooled in ice, and used for the recovery test. In experiments where flat -or screen baffles were used the outlet of the- disintegrator pipe including the baffle were overed with an aluminum hood to prevent splashing of the pulp and to direct the pulp-stream mixtue into the receiver, It has already been mentioned that in -experiments resented in this rert n steam separator was used. As a result of not using a steam separator, some of the exhausted ste condeon the pulp and on the walls of the receiving container causing a dilution of the pulp. This hdilution varied from 6:5 to 10 percent per beet and was calculated for each experiment on the basis of -the sugar-content difference in cossettes and in pulp. Since by the use of a steam separator this dilution could "have been substantially decreased, a correction was introduced. In.addition to the draft obtained with the dilute pulp a corrected draft is reported for all experiments. The corrected draft was calculated by subtracting the dilution in percent per beet from the draft experimentally determined. In case a steam separator wer used the actual draft would be between those two values and would approach the value of the "'corrected draft" with more efficient steam separation The pulp obtained had a consistency of a semilquid paste from which a part of the juice separate -easily by gravity Therefore, it is imperative to mix the pulp well before a sample is taken for centrifuging or sugar determination. For the separation of juice from, the pulp the. following procedure was used: The well-mixed pulp was weighed into four stairnes:ssteel cyl inders, each containing a filtering medium consisting f ne layer f cloth and. two bronze screens resting on a suprt (see Progress eport II, Fig. 6). The cylinders were then inserted into centrifuge tubes and the juice spun off at;a relative centrifugal force of 666 gravities. After 5.minutes the 14

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN cake was washed with small lots af hot water and each wash was folled by 3 minutes of spinning. When the required amount of water had been added~ the cynlinders wer. e evacuated. throuh a smll e the e asteamed with we steamof atmospherie presure for 1. to 2 minutes and centrifuged again. This steaming proceudre was used in all eeriments presented in this report except in the preliminary recovery tests (see Table II). The draft), the amount of wash-'water use-d, and the cake were determined by direct weighing and the corresponding values expressed in percent per beet. The sugar content in the cake and the apparent purity (Refr.) of the juice were also determined In. preliminary experiments the purity of the juice was compared with that of a hot digestim juice prepared by digesting a sample of cossettes (taken from the mixer) with water for 50 minutes at 75~800Cq The details of this method were given in Progress Report II. In case of complete tests in which qualities of thin juices were compared,. the diffusion juice was used as a standard of comparison The methods of preparation of the diffusion juice, its subsequent purification and the analytical methods usedwere described in the Progress Report III. A short review is given below: The diffusion juice was prepared in a small*scale diffusin bttery consisting of 10 cells, each with a capacity of 200 gm of cossettes. The time of diffusion was 60 minutes and the maximum temperature of the juice about 80C. The diffusion juice and the raw juice prepared by the new method from the same lot of cossettes were then quickly heated to 900C and submitted to a hot progressive preliming, followed by main liming at the same: temperature The first carbonation wascarried out by Silin's method with a known amount of CO2. The first carbonation juice obtained (Ph = 11) wa then adjuste to 75C and its rate of filtration determined by the Sengler, T6dt and Bttger method. The filtered juice was then saturated with C02 to a reaction neutral to phenolphthalein, boiled for 10 minutes, and filtered, and the- resulting thin juice was analysed The reproducibility of the above method of purification has been studied in previous work:(Progress Report III) and found to be good. Vo. PRELIMINARY RECOVERY TESTS In the preliminary recovery tests presented in Table II, a 0.5 inch.standard brass pipe was used as the disintegrator pipe (Fig. 35,F) This pipe was connected directly to the ejector ch'amber C, and had a total length of 16.5/8 inches and an inside diameter of 0.625 inch. The imping' met baffle used in this group of experiments was of a double screen type. The characteristics of this baffle are described in the following section of thLs reporto About 90 percent of the cossettes striking the baffle were 15

TABLE II PRELIMINARY RECOVERY TESTS AT LOW CAPACITY 'low Rate of Steam, 3553 lb/min Flow Rate of Induced Air, 2.41 lb/mm 1 2 5: 4 5 67 8 9 10 11 12 15 14 15 1617 % Sugar Juice Temp. Dilu- Wash- Cor- Sugar Juice un raf xet-I Cake * lost in Average un of tion Water, Draft, Arected Caacp Chamer( in in /Pulp, gm/100 lo by Draft, % of Sugar Cake, eght % by ine f x CPolake, ity Ixe setesPulp 0 gm Beet gm Beiet Cake Bt slettes Bei~t;Betj s17.52 16.10 75 8.8 55 127.5 118.5 33.7 1.55 0,45 14.97 15.52 90.5 6.9 I PFroce~ss (vac.) Rference Juice 8,19 7.56 89.,8 * Dieot Pigestl~i 7o216,022 6-3 f52 9 5 130 D5 121..9.30, 5 7,;T lr48 b37 7.2 -~1.62 i6.22 63 8.6 44,3 12~5oL2~ 1,24 0,1 15.0. 1 1 2 -4 L5 72 vc: Juice (Rot~ ~ ~~~~~lw W 7.36 90.3:7 IDigest ion)

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN for~ced through. The rebounding part (about 10 percent) of less finer pulp was collected separately and was not used in the tests. In Run No, I the wash-water, which amounted to 53 Percent per beet, was aplied in sev equal lots and after washing the cake was not steamed. The increase of temperature to 73~C resulted mainly from the condensation of steam on the pulp in the receiver. Columns 8 and 9 in Table II give the actual nd corrected drafts respectively. It is expected that with efficient steam separation a value closer to the corrected draft could have been obtained. The amount of sugar lost in cake (col. 12) was 0.45 percent per beet. The comparison of the apparent purities (col. 15) of the raw juice and of the hot digestion juice shows that the former has a slightly better purity. Run II of the same table was carried out in the same way as Run I except that in Experiment 2 (see colP 7) a smaller amount of wash-water was used (applied in six lots) and the cake steamed. In this experiment after washing with water the cylinders were removed from the centrtfuge. The lower part of the centrifuge cylinder was evacuated through a hole located below the cake and the uper part was connected to a flask containing 200 ml of water at 920C. As a result of the decrease in pressure a certain amount of water flashed from the flask, condensed on the cakey, and passed to the lower evacuated part of the cylinder. After steaming the cylinder was centrifuged again for 5 minutets. For Experiment 2 the amount of washwater given in column 7 includes the amount of steam condensed on the cake. It can b-e seen from columns 9 and 12 that by using steam as a last wash (Exp. 2) a lower draft and a smaller loss Of sugar in the cake was obtained with the same lot of beets. It can also be seen from column 15 (Exps. 1 and 2) that the steaming did not affect the purity of the juice which in both cases was practically the same. The pracessing capacity used in Runs I and II (see col. 16) was adjusted to about 7 lb per minute, which was less than one-third of the design capacity (25 lb per minute) of the disintegrator. VI. INFLUENCE OF THE TYPE OF IMPINGEMENT BAFFLE ON THE DEGREE OF DISINTEGRATION AND ON SUGAR REC0VERY In this series of experiments a 3/4-inch standard brass pipe was used as the disintegrator -pipe (Fig. 3F). This pipe, 37-5/16 inches long and 0.822 inch inside diameter was connected with the ejector chamber through a diffuser (Fig. 3,D) having'a throat diameter 0.625 inch. The following four different types of impingement baffles were investigated in this group Of experiment s (1) The Double-ScE:reen Baffle consisted of two 9-mesh brass screens (0.045-inch-wire thickness)placed one on the other so that the openings of 17

ENGINEERING RESEARCH INSTITUTE * UNIVERSITY OF MICHIGAN the screens partially overlapped The double creen was then arranged at an angle of about 45- to the direction of flow. (2) The1 Double-Screen Baffle Covered with Fine Screen was the same as described under (1) except that a fine screen, 285 openings per squae inch (0'.Ol0.inch wire thickness), was put on the side of the double screen facing the outlet of the disintegrator pipe (3) The Flat Baffle consisted of a steel plate machined with fine grooves in a diamond pattern. The plate. was fix:ed at an ange 45 to4 the' direction of flow. (4)- The Cylindrical Baffle (see Fig. 2 and Drawing 25 of Progress Report IV) consisted of a bronze cylinder with vertical grooves along half of its inside surface. The cossettes were introduced tangentially in the diretion perpendicul ar to the gros dalled to travel along the grooved portion. A helical partition arranged inside the cylinder directed the pulp to the:conical outlet of the baffle. In the two runs made in this groupf experiments (see Table III) the pulp was centrifuged at 666 gravities and the same washing pr'cedure was used in each test. The wash-water was applied in seven portions. After the last wash the centrifuge- cylinders wer-evacuated through a side hole bt 1ow the level of the ae and wet saturated steam at atmospheric pressure introduced throuh a thin glass tube above the surface of the cake. The duration of steaming was minute-, After steaming, the tubes were' centril fuged for 5 minutes. Column 7 gives the total amount of sh-water,ing the condensate formed during steaming. The- latter, determined from the difference of weight of the cylinder before and after steaming, amounted to 18 rcent per beet. This steaming procdur was used in the- est of the tests presented in this report except that in a fw tets the time of st-eam ing was diferent. Only a part of the total amount of condensate formed during steaming was formed diretly on the cake, a prt was fd on the inside wall of the cylinder and the rest was introduced in the form Of small droplets by the stream of wet steam. It should be expected that in a continuous operation the efficiency of teaming would be better This point will be discussed in a l;ater part of the report Both the runs presented in Table III were carried out with a different sample of beets. In order to eliminate thhe influence of the different beet material and to allow a better comparison of the types of baffles used, the experiment with the double screen baffle was repeated in each run. It can be seen from Run I of Table III th tha the d'oublescrelen baf* fle is decidely better than the cylindrical type 18

TABLE III INFLUENCE OF TYPE OF IPING4EiENT BAFFLE ON SUGAR RECOVERY liTow Rate of Steam, 5.53 lb/min Flow Rate of Induced Air, 2.11 lb/mm 1.2 53 4 5 6 7 8 9 10 11 12 15 11 % Tem: Sugar Av e.rage Tem.p D Washa racuum m eilut ion Draft, rected Cake, lost in Ar ingemnt % Sugar of ~110 Water jb hat or Sugar Cr CapaqBaffleP Sugar in Sua f gim1lO Water by Draft, Of in Cake, t- Ejco ~~un Sugar in Pu~~~~~~~~ gm~/100 in itBaff le -in Pulp gm S Cossettes Pulp c gi Beet gm Beet Cake lof lb/min Chamber, Weight Beet In, Hg Double Screen 17.47 15.69 75 11.5 55.5 156.7 1251.4 30.1 0.69 0.21 10,6 Cylindrical 17447 15.88 75 10.0 55.2 126.4 116.4 1 6.8 2.34 o.86 10.6 1 Double Screen 17.70 15.92 67 11.2 55 5 1 9 1211.7 50.5 0.90 0.27 9.8 Doudble: J Screen and 17.70 16..20 70 9.5 55.7 151.7 122.4 31.2 1.10 0.34 9.5 Fine Screen Flat Baffle 17.70 16.55 115 8.3 55.,8 127.6 119-.5 511.11 561.1.-21.9.26

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN Using the same lot of cossettes and the same processing capacity the amount of cake was 36.8 percent per beet for the cylindrical baffle and only 30.1 percent for the doublet-screen baffle. The amount of sugar lost in the cake was four times greater when the cylindrical baffle was used (see -col. 10 and 12). The smaller draft in case of the cylindrial baffe was caused mainly by the greater amount of cake. Although the res-ults oint against the use of the cylindrical type, it is possible that by decreasing the area of the grooved surface the efficiency of the baffle could be increased.o It is believed that this would facilitate the passage of the pulp and keep the grooved portion cleaner, and therefore more efficient. The advantage of this type of baffle is that it (if successfully developed) -ould be used at the same time as a steam separator. A comparison of the flat with the double-scr:een baffle (see Run II) shows that in the case of the flat baffle the amount of cake (see lo. 10) is higher.and the loss of sugar in the cake almost five times greater than when operating with the Xdouble. screen baffle. The results with a double-screen baffle coverred with a fine screen do t show much differencet when ea with the results ob tained with the doublescreen baffleo This can be explained by the fact that just at the bginning Of the e iment a hol was puntured in the fine screen by the stream of cossettes so that essentially the double** screen arrangement was restoredo. Due to the lack of time no furthr baffle experiments were made and the double-screen arrangement which gave the be'st results was used in the rest of the experiments. It is possible that the better -efficieny of the screen baffle may be due to its self-cleaning,action which decreases the cushion effect. The importance of the baffle design cannot be overemphasized. The baffle is the part of equipment where the actual disintegration takes place The active surface of the baffle is comparatively being approximately equal to the surface area of the projection of the disintegrator outlet on the baffle. It may be assumed that in case f a scren bafle the majority of the disintegration occurs on this- projected area. In the experiments with 3/4*-inch pipe this small surface aa of less than 1 squ inch h to take care of a pressing capacityf about 10 b per minute A further study of the design of -baffles (stationary and rotary types) should be considered as a most important point in the future development of this process. 20

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN VII, THE -I.FLU'CE OF CHEMICAL PRETREATNENT ON THE DEGREE OF DISINTEGRAT0ON AND SUGAR RECOVERY The purpose of this group of experiments was:t investigate to what extent the changes in the hardness and the brittleness of; the particles affect their disintegration~ In Run I reported in Table IV, 7000 gn wer taken from a. well-mnixed sample of cossettes and mixed with 700 ml of 3 percent milk of lime (0.3 percent CaO per beet) for 10 minutes in the cossette mixer. After treatment thte cssettes rwere transferred to the feeder an disinteggratei!d in the usual The treated cossettes were hardnd greenish in color and there was no spare liquid left in the mixer- To have a standard of comparison a second disintegration test was made from the same' lot of cassettes,$ but without pretreatment Run 2 was carried out in the same way as Run 1 except that 700 ml of 2 percent milk of lime was used for 700.0 gm of cossettes (0.2 percent per beet). In Run 3 the cossettes were treated with a solution of aluminum sulfate. The procedure was essentially the same as, that used for lime pretratment. For 10 minutes 7000 gm of cass:ettes were mixed with 700 ml of 1 perent solution of aluminum sufate (0.1 pereent of aluminum sulfate per beet) There was no spare solution left after mixing. The- cassettes retamined their natural color and were slightly harder than untreated cossettes. A 3/4-inch disintegrator pipe (37-5/16 inches long) and a double-. screen baffle was used throughout this group of experiments.: The pulp was centrifuged at a relative centrifugal force of 666 gravitiesO The washwater was applied in seven lots and after washing the cake was steamed for 1 minute. -The percent of wash-waterper beet given in column 7 of Table IV includes the condensate formed. during steaming. Its amount as determined by the difference of weights before and after steming was slightly less than 2 percent per beet. In case of chemical pretreatment the-dilution given in column 6 is the total diutlion cased by pretreatment and condensation of steam on the pulp in the receiving container of the disintegrator. The draft was corrected only by the dilution resulting from the latter cause o - 21

TABLE IV IFLUENCE OF CHEMICAL PRETREATMEIT ON SUGAR RECOVERY Flow Bate of Steam, 5.53 lb/min Flow Rate of Induced Air, 2.41 lb/mm 1 2 5 4 5 6 7 8 9 10 11 12 15 14 Clor- Sug~ar % Temp' iuto Wash. lost in Aea Tep.Wsh.CoaSua Average Ejector inSugar Sugar of Water, Draft, rected Cake, lost inac Bun Me dSugar of i/l100 % / by Draft,. %.of Suga Cake ut~n Iflethod gm/ in Pulp, in onf ity Vacuum, Cossettes gm Beet Weight by Beet % Ofg Pulp ~C gm Beet Cake lb/mI Weight Beet Liming 0.3 % CaO 17.52 142.6 62 21.4 61.7 148.5 157.4 34.6 1.15 o.4o 7.2* I per Beet No Liming 17.52 15.69 67 10.4 56.7 15531.6 125.2 51,5 o8 0.25 7.5 6 Liming 0.2 % CaO 17.54 14.81 76 17,1 58.o 145.8 136.3 31.3 1.o6 0.55 8.2* 7 II f per Beet INo Li~ming 17.4 120 75 7.0 53.8 11.6 124.6 29.1 o.64 0.18 10.5 5 Aluminum Sulfate 18.85 15,80 74 19.5 6o.5 146.5 154.9 55.5 1.24 0.41 7.5* 6 0.1 % per -Beet Without Pretreat- 18,85 17.66 75 6.7 53.5 127,9 121.2 32.3 1i.47 0.47 12-.8 s inent *of wet pretreated cossettes

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN In Runs 1 and 2 the disintegration of limed cos6settes was decidely less satisfactory than that of untreated cossettes. The amount of cake per beet (cal. 10) was higher and in spite of a greater amount of wash-water per beet the sugar loss in cake was almost twice as great as in the case of untreated coassettes. In Run 3 where ccassettes were treated with 0.1 percent per beet of aluminum sulfate the difference in the amount of cake per beet and sugar loss in cake is too small to warrant any deductions. In addition, the amount of sugar lost in cake in the case of untreated cossettes was exceptionally high, 0o.47 percen-t per beet which was much higher than in Runs 1 and 2, which were 0.25 and O.18 percent per beet respectively. In summary of these tests it may be said that the chemical treat-, ment in the above described set of experimental conditions was not advantageous and gave poorer disintegration and smaller recovery. It is, however, -possible that with other types of impingement baffles the increased brittleness of cossettes ay' be of advantageo There are some other reasons against the use ofchemical pretreatment apart from the additional complication of the process there will always be an unavoidable dilution of the juice if the chemical is added in the form of a solution and it will be more difficult to fluidize the wet and sticky part ic les VIII INFLUENCE OF STEAM PRETREATMEiET OF COSSETES ON. T1E DGREE OF DISINTEGRATION AND SUGAR ECOVERY In the experiment presented in Table V the cossettes were first steamed for 3 minutes at 1 psig. For this the feeder containing a known amount of cossettes was evacuated to 18 inches Hg vacuum and saturated steam was introduced into the upper compartment of the feeder ntil the pressure increased to 1 psig. The condensate formed, together with the excess steam, was removed through the side tube, m, (see Fig..3) mounted above the dise of the gate valve~ W. After 3 minutes steaming the feeder was quickly evacuated to 18 inches Hg vacuum. The cooled cossettes were then disintegrated in the usual way. In this experiment a.O-inch standard brass pipe (16*5/8 inches long, 0.625 inches ID).was used. It was connected directly to the ejector chambero The baffle was of the doublescreen tiype previously described. The pulp was centrifuged at a relative centrifugal farce of 666 gravities. The value for wash-liquid given in column 6 (62.5 percent per beet) includes the total amount of the sugar-containing condensate (24k96 percent per beet), 25

TABLE V INFLUENCE OF STEAM TREATMENT ON SUGAR RECOVERY Flow Rate of Steam, 3.53 lb/min Flow Rate of Induced Air, 2.41 lb/min ~1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 % Sg jiSugar % j ] % Con- Wesh- Cake, % Sugar Da Average in in % Pulp densate Liquid Suga"Dft, Condensate Juice Capac- Chamber ossettes Pulp per Beet per Beet gm/100 et in b Cake, Weigth | " B E ecto gml Beet, Cake % Beet Bx B x Pol lb nan New Process (Coss ettes 18.34 17.40 93.22 24.96 62.5 40.6 5.22 2.12 115.20 9.42 8.49 90.1 15.78 14.33 90.8 9.06 1 Steamed) Reference Juice (iot 18.34 - - - - _ _ - - - - 8.31 7.56 90.9 - (Hot1. Digestion)

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN which divided into two lots was used as a first wash. The remaining part of the wash-liquid was water which was applied in six lots. No steaming of cake was used in this experiment. The disintegration was decidely less satisfactory as compared with other experiments made in similar. conditions with unsteamed cossettes. The amount of cake was: exceptionally high (40o6 percentper beet) a so was the sugar loss in the cake, The comparatively low draft of 115.2 (col. 10) can be explained by the greater amount of cake and by evaporation of some of the wat;er during evacuation-f.the feeder after steaming. This exceptionally poor disintegration of steam treated cossettes was also found in some of the preliminary experiments. This ~poor disintegration of steam treated cossettes needs a more comprehensive investigation before a satisfactory explanation can be given. The steaming procedure does not seem to affect the apparent purity of the juice which shows the same value as that of the hot digestion juice (col... 16). IX. QUALITY OF TE JUICE PREPARED FROM PULP AS COMPARED WITTHTAT OF THE DFFI.SION JUICE In order to obtain a better comparison of the qualities of juices prepared by these two different methods three complete -experimental runs have been carried out in which both the raw juices were submitted to the usual purification process. The experimental technique has already been described in section IV of this report. Table VI in which the results of these tests are presented is div. ided in two parts The upper part gives data concerning production of the pulpand snugar recovery. The lower part presents the characteristics of juices prepared in each run from the same lot 'of beets. Before the results, of the experiments are discussed, an explanatian concerning the washing procedure and the setting of the equipment is given. In Runs 1 and 2 a 0,,5-inch standard brass pipe (16-5/8 inches long) was used as the disintegrator pipe This pipe was coected directly with the ejector chamber. In Run 3 a 3/4-inch standard pipe was used (length 37-5/16 inches) which was connected with the ejector chamber by means of a diffusfer with a throat diameter.of 0G625 inches. A double-screen baffle was used in Runs 1:an 2 and a single-screen baffle (9 mesh, 0.045 wire thickness) in Run 3. Both baffles were placed at an angle of about 450 to the direction of flow and in all runs the pulp

TABLE VI (Part 1) JUICE PREPARED BY THE NEW METHOD VS DIFFUSION JUICE q Rate of Steam 3 53 lb/m Flowm Bate Of Iuced Air, 2.41 lb/m 2 5 4 5 6 7 8 9 10 11 12 15 14 16 17 18 Sugax Temp. Cor- Sugar Average Pressure Pressure sure ID Sugl Tem Temp.C Dilu- WashBt rectedWCaket Sugarlost in C_ _ _IH n in Hg. Proc *i of tion WaterDraft Draft % Of in Cake, ity, Diffuser, Ejector Rotary Pipe, esettes COf Tm Dl Wr d, r in C i in in of es os- iCo sos- Pulp, gm/1 gm gm/100 % by % by Beet Cake % Beet lb/min psg Chamb-er, Feeder, Ins settes O C C s Beet Igm Be~et Weight Wegh -n H New 17.,74 i6,i 17 71 9 8 48.9 126.9 117.01 1.7 0.95 0.30.8 Q;3vo )25 2-3 New 17.15 15,95 17 68 7.5 54,i 1 128.1 26.3 o051 0,153 7'( 1 to 2 0.625 (vac) New 17.69 16.54 18 6;5 6.,95 54.6~ 132 D3 125.4.29.2 0.92 0.27 23 2,-3 0 to 2,2

TABLE VI (,Part- 2) 2 5 4 5 6 7 8 9 10 11 1213 14 15 6 17 18 Raw Jui'uice 1st Carbondtion Juice 2nd Carbon-tion Juice~ Pectin Alkali- Lim Sp. Co-nd. A (Silin) nit-y, ate of Filtration.-LgT salts, at 18'C 1 GaO Bx Pal Q gr/100 % cao Bx Pi Q at 560 mg/loo and5Bx Phenol-BMethy (refr,) ml 0Ml 100 ml 150 ml (refr.) Ml' Bx mhos/cm plthal New 15.28 13i.4 89,9 0.24 o0o6 56 351011 7'321' 14.83 15.93 95.9 28 0.019 0.024 Di f fusion 12.69 11.54 90.9 0.50 0,o6 50" 2,125 5t 45 11. 83 11.12 95.9 54 - 0.019 0.025 New 13.96 12.52 89.-7 0.28 o.o66 49" 2'31" 6'08" 15.00 12.22 94.0 0.002 22 0.00155 0.015 0.026 Dif11*fusion 15.24 11.95 90.2 o0. o.o6 411" 2104" 5'588" 12.21 11.41 93.4 0o036 41 0.00146 0.014 0.019 New 14.75 12.94 90.2.064 14.45 15.5994.0 o0oo8 26 0.00152 0.021 0.025 Diffusion 15,66 12-.52 90.2 0.079 - 12.46 11.65 93.3 0.044 21 0.00142 0.017 0.017

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN was centrifuged at 666 gravities. The washing procedure was the same except that in Run I the wash-water was divided into 6 lots and the time of steaming was 2 minutes, while in Runs 2 and 3 the wash-water was applied in seven lots and the time of steaming was cut to 1 minute. The processing capacity in Runs 1 and 2 (see col, 14) was small, about onewthird of the design processing capacity (25 lb per minute) and the amount of coarse pulp which did not pass through the screen was approximately 10 percent. In Run 3 where -the capacity was increased almost to the design value the amount of -coarse pulp increased to almost 30 percent. This coarse pulp was not used in the recovery tests. The smallest draft (cols. 9 and lO) as obtained in Run 1 where a smaller amount of water was used for washing the cake ("e col. 8), but the sugar loss in cake (col. 13) reached the highest valu of 0.3 percent per beet. The best recovery was obtained in Run 2 where the capacity was small (col. 14) and the amount of wash-water was increased (col. 8). The amount of cake and the loss of sugar in cake in Run 3 was comparatively small (29.2 and 0.27 percent of beet respectively) in spite of the fact that the capacity was inreased about three times and a single-screen baffle was used. A comparison of purities of raw juices (col.. 5, Part 2 of the table) shows that the apparent purity of the diffusion juice is slightly higher, except in Run 3 where the purities have the same value. The situation becomes reversed when the purities of thin juices are compared (col. 13) Except in Rn I where the purities are the samel the purity of the thin juice prepared from the pulp juice was higher. This fact is in good agreement with observations made in a similar series of experiments carried out last year and presented in Table I of Progress Report III. As for the rate of filtration of the first carbonation jwuce, the juice prepared from diffusion juice gives a better rate of filtration although the difference is not great. Previous experiments (Progress Report III) indicate that the use of cold preliming of pulp juice may considerably improve the rate of filtration. As can be seen from column 14 the color of the thin juice prepared from pulp juice is much lighter. In Run 2 the thin juice prepared from pulp was almost colorlesso 'The difference in lime salts content (col. 15) of thin juices was small, except in Run 2 where the thin juice obtained from diffusion juice contained a much greater amount of lime salts. In Run. the total amount of nitrogen was determined by Kjeldahl method in raw and thin juices. The results are given in Table VII. 28

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN TABLE VII TOTAL NITROGEN IN RAW AND THIN JUICES Process Total Nitrogen in mg/lOO Bx Raw Julice Thin Juice New 559 437 Diffusion 505 -441 It can be seen from the table that the comparatively small difference in total nitrogen in raw juices disappears after purification. Both the thin juices: contain practically the same amount of total nitrogen.. Xo RECOVERY TEST AT HIGHER CEETRIFUGAL- FORCE In the experiment presented in Table VIII the setting of the equip ment was the same as in Run 3,. Table VI, that is a 3/4-inch standard pipe and a single-screen (9 mesh) baffle were used. The pulp was centrifuged at a relative centrifugal force of -1460 gravities, more than twice the value used in previous exeriments. The amount of washwater used for washing the cake (see col. 6) was 45 percent per beet and divided into eight lots. The condensate formed during steaming of the cake was about 2 percent per beet and is included in the amount of wash-water 'used The use of the higher centri-: fugal force resulted in a smaller amount of cake which dropped from the usual value of about 30 percent per beet-to 21.4 percent with a corrected draft o 125,7 perent, Because of the smaller 'am ount of cake it was possible with a smaller amount of wash.:water to decrease the sugar loss in cake to 0,14 percent per beet. The purity of the pulp juice was slightly higher than that of the diffusion jui, although the differen as rather sal. 29

TABLE VI II RECOVERY TEST AT HIGHER CENTRIFUGAL FORCE.ow Rate of Steam.53,, lb/min Flow Rate of Induced Air 2)41, lb/mm 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 % Sugar 1Cr- Sugar Average Raw Juice Pressure Vacu Pressure Temp~ Dilu- Wash. rected in in in oc- of Draft, Cake, % Capacin Pulp tion Water, Draft, lof Sugar Cake, ity, ifuser, Chanberl Rotary Cos- 0C /100 gm/lOO g by in BeOf lb1m BX Pol B Ieeder, Pulp Weight Be3;eet settes Beet gm Beet Weight Cake Beet pg In. Hg. w 17.95 16.62 70 7.8 45.0 133.5 125.7 21.4 o.68 0.14 18.9 14.70 15.41 91.2 5 1 to 0 2 to 5 sion 17.93 - - - 15.70 12.44 90.8 - -

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN XIo CONCLUSIONS AND PLANS 'FOR FUL-E In the experiments discussed in this report the greatest proc:essing capacity used was 23 lb of cossettes per minute-,.which at a flow rate of steam equal to 35.53 lb per minute gives a steam -requirement of 15 lb per 100 lb of beets. Further improvement on the economy of the system was not studied. The steam consumption of the disintegrator depends on a number of factors. One of the most important variables influencing the degree of disintegration and steam consumption:is the baffle design. A more efficient baffle will require' a lower velocity of particles and will result; in a smaller steam coons sumption for the same degree of disintegration. The baffle should also be designed to simplify the construction and installation of an efficienrt steam separator to avoid an unnecessary dilution of pulp. The great differences in efficiencies of the four type of baffle investigated show that there is great opportunity for improvement in baffle design. Another important factor for the economy- of the 'disintegrator is the uniformity. of flow. As- the flow rate of steam through the system is constant any decrease in the flow of -cossettes will result in a higher steam consumption per weight of beet. A considerable saving on steam can be expected by further improvement in the uniformity of flow of cossettes. The efficiency of the thermocompressor should be improved to increase the ratio of the induced steam to motive- steam. The manner in which the cossettes are introduced into the stream of steam, as well as the length of the disintegrator pipe, should be studied further to determine the optimrum conditions. The use of higher steam pressure should also be investigated. The filtering and washing properties of pulp prepared in the disintegrator were very good._ With a comparatively small amount of water (about 50 percent per beet) a sugar recovery of more than- 99 percent was obtained at drafts of about 130 percent, The amount of -cake was about 30 percent when centrifuged at a relative centrifugal force of 666 gravities and 21.4 percent at 1460 gravities. In all experiments presented in this reort aa centrifuge was used for the s-eparation of the juice from the pulp. AlthoUgh this method of separation proved successful, it is possible that the use of rotary filters or rotary filters with pressing rolls would be advantageous. Also-, the juice separator (of whatever type it might be) will concentrate (dehydrate) the cake and the saving on pulp presses will to a certain extent oSfset the cost aof the juice separator. 31

ENGINEERING RESEARCH INSTITUTE ~ UNIVERSITY OF MICHIGAN The problem of selection of equipment for juice separation was beyond the scope of this project and has not been studied It is expected that a continuous washing of the cake with a uniform spray of ater may give en higher recovery or a smaller draft than obtained by washing with separate lots as was the case in the experiments reported. However with the type of equipment used a uniform and continuous washing would ha been difficulty if not impossible~ The steaming of the cake although advantageous in the experiments:reported should not be considered as a nec-esasity for good recovery. The quality of juices prepared by the new method compared favorably.with that of juices obtained by the usual diffusion process., Chemical pretreatment of 'raw cossettes with lime' and with aluminum sulfate did not im-. ~prove. the disintegration, but on the contrary appeared to be harmful -in the experimental conditions studied.~ The experiments reported using continuous equipment have demonstrated the feasibility of the new process with regard-to quality.y of the juice, draft and sugar recovery. 32

UNIVERSITY OF MICHIGAN 3 90111111111111111111111111111111111111111111 1111110 3 9015 02539 7210