United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
EPA-600/2-79-208
December 1979
Research and Development
oEPA
Low Wastewater
Potato Starch/
Protein Production
Process
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into.nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution-sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-208
December 1979
LOW WASTEWATER POTATO STARCH/PROTEIN
PRODUCTION PROCESS
by
John R. Rosenau and Lester F. Whitney
Department of Food Engineering
University of Massachusetts
Amherst, Massachusetts 01003
Grant No. R-803712
Project Officer
Harold W. Thompson
Food and Wood Products Branch
Industrial Environmental Research Laboratory
Cincinnati, Ohio 45268
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory, U. S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the
views and policies of the U. S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
ii
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on our
health often require that new and increasingly more efficient pollution
control methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
This report presents pilot plant results on a modified potato starch
production process which uses significantly less water than commercial
processes. The proposed process also recovers two byproducts streams for
use in animal feeds. An economic analysis shows the process to be very
promising. Further information on the subject can be obtained by contacting
the Food and Wood Products Branch, Industrial Environmental Research Labora-
tory-Cincinnati .
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
ill
-------
ABSTRACT
While potato starch has been an item of commerce for many years, tradi-
tional processing methods have incurred large volumes of high BOD effluents.
The research summarized by this report has lead to a modified process which
upgrades the soluble components formerly discarded in the effluent to animal
feed materials in an economical manner.
The process developed starts by grinding and sieving as in the tradi-
tional process with the exception that recycled juice rather than fresh water
is used to flush the starch granules from the pulp in the sieving operation.
The pulp is pressed and dried as in traditional processes. The starch is
separated from the juice and refined by an "elutriation" type liquid cyclone
and a basket centrifuge. Water - at the rate of one kg per four kg of input
potatoes - is introduced at the basket centrifuge and flows in a counter-
current manner through the cyclone system. Excess juice is heated to preci-
pitate the heat coagulable protein. The protein is centrifuged from the
juice and spray dried; the deproteinated juice is concentrated to a molasses-
like feed material by reverse osmosis and multiple effect evaporation.
The report concludes with an economic summary based on a new plant. The
process is shown to be economically promising with a return of investment of
40% with an assumed starch value of $0.22/kg ($.10/lb), an assumed cull
potato cost of $1.65/100 kg ($0.75/cwt), and an input rate of 455 metric tons
(1 x 10^ pounds) per day.
This report was submitted in fulfillment of Grant No. R-803712 by the
University of Massachusetts under the sponsorship of the U. S. Environmental
Protection Agency. This report covers a period from July 7, 1975 to July 6,
1978, and work was completed as of September 1, 1979.
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables vii
Acknowledgments viii
1. Introduction 1
2. Conclusions 3
3. Starch Production 4
4. Juice Processing 12
5. Economics 20
References 28
Appendix - Photographs of Pilot Scale Implementation 30
-------
FIGURES
Number
1 Initial starch/protein meal process flowchart
2 Starch and pulp processing flowchart 7
3 Juice processing flowchart 14
Al Washing with perforated basket and power washer .... 30
A2 Grinding with Fitzmill with SO addition 31
A3 Feeding of slurry with peristaltic pump 32
A4 Vibratory screen used to separate pulp from starch
and juice 33
A5 Screw press used to dewater pulp prior to drying .... 34
A6 Flow scheme at underflow of elutriation cyclone .... 35
A7 Steam infusion heating of juice 36
A8 Air actuated back-pressure valve 37
A9 Spray dryer used to dry protein precipitate centrifuged
from the heated juice 38
A10 Vacuum evaporator used to concentrate deproteinated
juice 39
vi
-------
TABLES
Number Page
1 Solids Balances and Starch Whiteness Values for Last
Five Pilot Scale Runs 8
2 Expected Mass Balance for Industrial Application 11
3 Spray Dried Protein Meal Composition and
Digestibility 17
4 Glycoalkaloid Concentration and Mass Balance 18
5 Product S02 Levels 19
6 Estimated Fixed Capital Investment for Processing 455
Metric Tons of Cull Potatoes per 20-Eour Day 21
7 Annual Costs, Income, and R.O.I 23
8 Estimated Fixed Capital Investment Under Livestock
Option 25
9 Annual Costs, Income and R.O.I, for Livestock Option ... 26
10 R.O.I. Sensitivity to Fluctuations in Income, Major
Costs and Plant Size (Non-Livestock Option) 27
vii
-------
ACKNOWLEDGMENTS
In addition to the principal funding agency, i.e., the Industrial
Pollution Control Division of the EPA, the authors would like to thank Agway,
Inc., the Maine Potato Commission, and the Agricultural Experiment Station
of the University of Massachusetts for their support of this project.
In addition, special thanks are due to Mr. Harold w. Thompson, Project
Officer, EPA, for his interest, patience, and special administrative coopera-
tion; and Mrs. Georgia P. French, Chemist, Agricultural Experiment Station,
University of Massachusetts, for her skill and willingness to conduct protein
digestibility and lysine availability analyses.
viii
-------
SECTION 1
INTRODUCTION
Cull potatoes (i.e., those misshapen, too large or small, or with hollow
heart or other defects) have been processed into starch and pulp for many
years (1-6). Typically such processing involves grinding (with sulfite addi-
tion to prevent browning) and wet sieving with added water to separate starch
from pulp. The pulp is dewatered by centrifuging or pressing (sometimes with
added lime which improves its dewatering properties) and then dried in a
rotary or flash drier for cattle feed. The starch is refined by centrifugal
or multistage liquid cyclone methods. Vacuum filtration or automatic basket
centrifugation is used for final starch dewatering prior to drying, grinding
and bagging.
Non-starch constituents washed from the starch granules produce large
BOD loadings in the effluent stream. For example, a traditional plant pro-
cessing 450 metric tons of cull potatoes per day would release daily about
4500 m3 (1.2 x 106 gallons) of effluent containing about 18000 kg (40,000
pounds) of dissolved solids. All but one of the starch processers in the New
England area using cull potatoes (as opposed to starch slurries obtained from
chip or french fry operations by liquid cyclone methods) as input material
have closed principally due to the prohibitively expensive costs of treating
these wastes.
Closing of these plants has lead to additional problems. Cull potatoes
have been dumped into landfills and gravel pits as well as back onto growers'
fields (which leads to potential disease propagation). Culls have been used
directly for animal feed but their availability is variable and animal pro-
duction facilities are often located too distant for profitable shipping and
handling. Culls have been ground and dried for animal feed but, as fuel costs
have risen, this outlet has become less and less attractive. Moreover, the
value of such animal feed is low compared to that obtained by processing the
culls into starch and by-products.
It is difficult to place estimates on cull potato production - especially
since it varies drastically from year to year. Various experts have suggested
that 10% of the crop is a realistic figure. This corresponds to an annual
U. S. production of well over a million metric tons.
Several groups of researchers have recently investigated recovery pro-
cesses for potato protein (7-14). Heat induced coagulation has been the most
intensely researched method. Dutch researchers have implemented such methods
to production scale but they have not publicized particulars of their methods.
Researchers at the Eastern Regional Research Laboratory (ERRL) of the USDA
-------
have extensively investigated heat coagulation methods for upgrading process-
ing wastes from traditional starch plants (8, 12). Meister and Thompson
investigated Fed , lime and H PO , and lime and FeCls as well as heat coagu-
lation methods for treatment of potato chip processing wastes (13, 14).
Haight et al. (9) in work supported by this project, investigated the effects
of pH, coagulation temperature, time and concentration on the dewatering pro-
perties of coagulated juice protein.
In general, the methods have shown that about a third of the crude pro-
tein in the juice can be heat coagulated into a 70% protein (d.b.) concen-
trate. While acidification has been shown to enhance protein precipitation,
this benefit is reduced at high temperature and solids concentration levels.
Section 3 of this report discusses modifications to the traditional
starch production process leading to juice production with a minimum of
dilution. As discussed therein, water requirements have been reduced to
about 25 kg water per 100 kg of potatoes processed as opposed to former
levels as high as 1000 kg of water.
Processing of the juice to yield protein meal and a molasses like con-
centrate and economics of the processes are discussed in Sections 4 and 5,
respectively.
-------
SECTION 2
CONCLUSIONS
1. The elutriation liquid cyclone system developed and piloted
in this project can produce starch of acceptable quality at
a fresh water input of 25 kg per 100 kg of input potatoes.
2. High temperature heat coagulation can produce a 70% protein
meal of high lysine content from excess juice produced by
the liquid cyclone system.
3. Glycoalkaloids are not concentrated to hazardous levels in
the protein meal.
4. Deproteinated juice can be concentrated by reverse osmosis
and multiple effect evaporation into a molasses-like
ruminant feedstuff.
5. The economics of the developed systems are sufficiently
promising to warrant industrial scale interest.
-------
SECTION 3
STARCH PRODUCTION
The flow chart shown in Figure 1 summarizes the process initially inves-
tigated to minimize water usage in starch production. As shown, washed pota-
toes were ground with sulfite (1000 ppm SO ) in a hammer mill fitted with a
screen with 1.6 mm (1/16 inch) holes. The slurry was dewatered in a decanting
centrifuge, dried, ground, sieved through a 74 ym (200 mesh) screen to remove
peel fragments, and air classified to remove fines. As an option, the cake
leaving the centrifuge was slurried with tap water and recentrifuged.
The process as outlined can be thought of as a combination of 2 existing
concepts. First, decanting centrifuges have been installed in a number of
existing full scale operations. As pointed out by Strolle et al. (8), about
60% of the undiluted juice can thus be spun off; by adding an equal weight of
water to the resulting cake and recentrifuging, about 93% of the soluble
solids can be removed from the starch/pulp mixture. This mixture is then pro-
cessed by traditional methods to yield pulp (animal feed) and refined starch.
Strolle et al. (8) also point out that the centrifuged liquids, when combined,
yield a liquid containing 3.4-4.4% solids.
The second concept utilized has been outlined by Shaw and Shuey (15).
Therein, potatoes are dried, ground, sieved (to remove peel and other cellu-
losic materials) and air classified to remove protein containing fines from
the starch. Problems with the method, however, include that the resulting
starch is not sufficiently clean to compete with other starches currently
being marketed and nearly all of the water initially in the potato is removed
in a rather energy intensive manner; i.e., by drying rather than by pressing
or centrifuging followed by reverse osmosis and/or multiple effect evaporation
of the juice.
By combining the two concepts discussed above, it was anticipated that a
cleaner starch would be obtained, energy efficient juice processing methods
could be used, and dilute process effluents would be completely eliminated.
The process was implemented with a Pfauldler ZIL decanting centrifuge and
an Alpine American 100 MZR air classifier. The cleanest starch made with the
process (using 40 kg of wash water per 100 kg of potatoes) tested 1.2% protein
(d.b.) and had an Agtron reflectance of 75%. (In comparison, commercial
potato starch produced by conventional methods is about 95% reflectance and
has only traces of protein; the starch produced by the Shaw and Shuey process
(without the benefit of counter-current washing as also included in their
work) tested 73-91% reflectance and 4.0% protein.
-------
01
WHOLE POTATOES
WASHER
fSULFITE,
GRINDER
PRESS OR CENTRIFUGE
PULP
DRJER
SCREEN
I
AIR CLASSIFIER
STARCH
JUICE
-fc-PEEl
-»-FEED
EVAPORATOR
SPRAY DRIER
PROTEIN MEAL
Figure 1. Initial starch/protein meal process flowchart.
-------
Inspection of the air classified starch with a low power microscope
revealed particles of pulp that were small enough to pass through the sieve
with the starch but were not removed by air classification - even when pre-
ceded by an extra grinding step with an Alpine American cross flow mill
adjusted to its most intensive setting. Evaluation of the starch by indus-
trial personnel (Colby Starch Cooperative, Calibou, ME) confirmed that it
would be difficult to market.
A critical review of the above results - especially in view of the
ability of the traditional process to produce starch of extreme purity - com-
bined with excellent results obtained by tangential elutriation liquid cyclone
methods in the mining industry (16) led to a revised process that drastically
improved starch quality, lowered wash water requirements still further and
reduced capital equipment costs.
As shown in Figure 2, the revised process begins by hammer milling of the
washed cull potatoes with direct addition of sulfite (see below and Section 4)
to reduce browning. The slurry is then sieved, in either one or two stages,
as in the traditional process. Instead of introducing water at this point to
help flush the starch from the pulp, recycled juice is used. The starch/juice
slurry passes through the 105 ym (140 mesh) screen(s) while the pulp passes as
overflow to a dewatering press where it is dewatered to about 30% solids. It
is then dried for cattle feed. It may also be fed directly without drying if
a feed lot is nearby.
The starch/juice slurry is pumped to an elutriation type liquid cyclone
wherein the starch is spun to the underflow while the overflow is largely
recycled to the screening system. A defearning device such as the Cornell
Machine Company's (Springfield, NJ) "Versator" can be used to control foam-
ing - a problem of major importance if frozen or especially off-grade tubers
are to be processed. Silicone based defearning agents such as Dow Coming's
FG-10 are also effective but add a significant cost. The underflow from the
liquid cyclone is pumped to a second liquid cyclone for thickening prior to
passing to an automated basket centrifuge for final washing (with sulfited
water) and dewatering. Dilute juice and wash water from the centrifuge are
recycled in a counter-current fashion to the thickening cyclone arid overflow
therefrom is used as the elutriation flow into the primary cyclone (the
design of which has been patented by the Bird Machine Company (So. Walpole,
MA). Centrifuged starch (60-65% solids) would be dried in a flash dryer in
commercial practice to obtain maximum whiteness. Acceptable whiteness (see ,
Table 1) has been obtained however even with simple tray drying at 105C
(220F).
The revised process was tuned and confirmed with pilot scale runs of
about 500 kg (1,100 pounds) . As shown in more detail by the Appendix photo-
graphs, cull potatoes obtained locally from a fresh market packer were washed
by placing about 10 kg (22 pounds) of potatoes at a time in a plastic garbage
can modified by cutting holes in the bottom. This was turned by hand to
tumble the potatoes while spraying them with water from a 40 atm (600 psi)
power washer.
It should be stated parenthetically that washing (and/or the practice of
moving potatoes by fluming) does not appear to give rise to troublesome waste
-------
WASHED POTATOES
MECHANICAL
DEFOAMER
CAKE
PUMP
CONTINUOUS
TRAY DRIER
PULP
*
PUMP
CYCLONE
BASKET CENTRIFUGE |
J!
CAKE
DRIER
STARCH
Figure 2. Starch and pulp processing flowchart
7
-------
TABLE 1. SOLIDS BALANCES AND STARCH WHITENESS VALUES
FOR LAST FIVE PILOT SCALE RUNS
oo
Run
F
G
H
I5
J6
Feed
(kg)
171.2
151.5
(159. 5)4
160.9
173.4
Pulp
(kg)
32. 53
41. 63
25.9
24.0
27.0
Protein
Cake1
(kg)
7.7
7.3
9.1
6.8
7.0
Starch
(kg)
73. 03
59.33
82.9
87.0
78. 67
Yield2
(%)
713
673
89
92
737
Cone.
Juice
(kg)
43.7
32.6
35.9
36.9
37.6
Losses
(kg)
14.3
10.7
5.75
6.2
23. 27
so2
Level
(ppm)
1000
250
0
250
250
Reflectance
(Hunter L)
91.8
92.5
84.0
89.4
92. 27
1. Measured as dry weight of cake leaving the centrifuge, i.e., before
spray drying.
2. Determined by estimating starch in feed as percent solids in feed less 6%.
3. Due to improperly adjusted elutriation flow rate, starch carried over into
the juice. This was removed and added to the pulp stream.
4. Missing sample; solids estimated by assuming the percent solids to be
**the same as run G.
5. Potatoes in poor condition.
6. Potatoes in very poor condition.
7. Starch cycloned an additional time to determine effect on starch whiteness.
-------
streams in existing commercial practice. The material removed is largely
mineral in nature and settled quickly in simple dug basins, settlers, or
liquid cyclones. Odor problems are not common; apparently there is suffi-
cient aeration of the recycled water (coupled with its usual low temperature)
to prevent them.
The washed potatoes were ground in a model M Fitzmill (Fitzpatrick
Company, Elmhurst, IL) fitted with a screen with 1.6 mm (1/16 inch) holes.
Sulfite was mixed with the potatoes by pumping a solution of sodium metabi-
sulfite directly into the mill with a variable speed peristaltic pump.
The ground potato slurry was then pumped with another (much larger) peri-
staltic pump at a rate of 180 kg (400 pounds) per hour to a mixing tee where
it was mixed with four volumes of recycled juice. The mixture then flowed to
a 61 cm (24 inch) diameter vibrating screen (Sweco, Inc., Los Angeles, CA)
which served as a single stage wet screening system. Juice and starch
passed through the 105 Urn (140 mesh) screen while the pulp was retained. It
was observed (see Table 1) that the fraction of starch granules separated
from the pulp by the screening operation was about the same as the juice
fraction removed; i.e., about 90% for the 1 to 4 ratio of feed to recycle
juice used. While this would imply that a two stage counter-current screen-
ing system would be expected to extract about 99% of the starch granules, a
more realistic expectation would be somewhat lower than this.
The starch/juice slurry was then pumped into a 10 cm (4 inch) diameter
elutriation type cyclone (Bird Machine Company, So. Walpole, MA). This size
was the smallest elutriation cyclone available. As it requires an input
stream of 190 liters per minute (50 gpm) a recirculation system (including a
heat exchanger to prevent heat buildup) was added to match the cyclone to the
pilot scale operation.
Underflow starch was pumped, along with wash water from the following
basket centrifuge, into a 10 mm diameter secondary thickening cyclone (Dorr-
Oliver, Inc., Stamford, CN) with an air actuated diaphragm pump (Warren Rupp
Company, Mansfield, OH). An air chamber was used to dampen the pressure
fluctuations of the pump. Overflow from the secondary cyclone was used as
the elutriation flow of the primary cyclone. Underflow starch from the
secondary cyclone was metered therefrom with a peristaltic pump to overcome
flow instabilities incurred by simple throttling. The thick starch slurry
was dewatered and washed in a basket centrifuge (International Equipment
Company, Needham Heights, MA) and dried in a tray drier at 105C (220F).
Pulp from the vibrating screen was dewatered to about 30% moisture in a
Beloit-Jones (Dalton, MA) 7.6 cm (3 inch) diameter screw press. Juice pressed
from the pulp was returned to the top of the vibrating screen. The dewatered
cake was dried in a tray drier at 105C (220F).
Some comment should be made toward various difficulties encountered in
the tuning and confirming pilot runs. One of these involved attempts to co-
dry concentrated de-proteinated juice (see Section 4) with the pulp. This
was not successful for several reasons. The concentrated juice proved to be
difficult to mix with the pulp (either as wet press cake or as a dried,
-------
ground flour). The mixture dried very slowly and stuck badly to the trays.
The final material was quite hygroscopic.
Another difficulty encountered was periodic blinding of the vibrating
screen. This seemed to happen especially if too much juice was bled from the
cyclone/sieve system allowing air to enter the cyclone causing a foamy con-
dition in the recycle juice. Two useful preventative measures were to care-
fully watch the liquid level in the balance tank under the vibrating sieve
and to add a nylon scrub pad to the vibratory screen which revolved with the
overflow pulp.
It was anticipated that adding the ground potato slurry directly to the
vibrating screen and spraying recycled juice over the pulp rotating on the
screen would give the best removal of starch. This was not the case; better
results were obtained by simple mixing of feed and recycle juice followed by
screening of the mixture. Apparently, the pulp entraps the starch if there
is too little juice present and prevents its migration to the screen surface.
This is further confirmed by the observation that starch was not present in
the juice separated from the pulp by the screw press.
Various attempts were made to eliminate the thickening cyclone and
operate only with the elutriation cyclone. Such operation seemed unstable -
either starch would back up in the cyclone to the point of appearing in the
overflow juice or, upon increasing the underflow rate, too much of the juice
was removed with the starch causing problems of low whiteness due to the
presence of too much colloidal material not removed by the centrifugal wash-
ing step. When the secondary cyclone was used as depicted in Figure 2,
problems with stability were greatly reduced. Even in this mode, however,
flow settings had to be very closely monitored to prevent starch back-up or
poor whiteness. Table 1 presents results of the last fine pilot scale runs
and shows that even after previous experience with five tuning runs, starch
overflowed the primary cyclone with the juice due to operational errors dur-
ing portions of runs F and G. Some of these errors might be expected; during
start-up of each run, flow rates have to be constantly adjusted. Moreover,
it was not early appreciated how much the potatoes tended to segregate in the
holding cans used between grinding and processing. In the last runs, the
entering potato-slurry was mixed to minimize variations in the feed material.
In commercial practice, density controls are available which simply adjust
the underflow rate to control the starch concentration of the same.
In conclusion of this section, a few explanatory comments should be made
concerning trends shown in Table 1 which includes solids balances as well as
starch reflectance values of the last five pilot runs: (1) whiteness has
been measured as the L value obtained with a Hunter color difference meter.
In our comparisons with the reflectance values obtained with £he commonly
used Agtron readings, we found the Hunter to read about a percentage point
lower; (2) it can be noted by comparing the first three runs that the S02
content of the feed and wash water (both were sulfited to the same level)
can be reduced to 250 ppm without reducing whiteness; (3) runs I and J were
made with potatoes in poor and very poor condition; (4) to determine the
effect of an additional cyclone step on starch whiteness, the starch of run J
was passed through the cyclone system a second time. In this run, the mass
10
-------
balance closes poorly and the starch yield dropped substantially. While it
is tempting to blame both results on the extra cyclone step, carry-over of
starch into the juice was not observed. Moreover, the solids contents of the
pulp, protein meal, and concentrated juice streams were not increased. The
high level of losses thus remains undertermined and cannot be simply
explained by the extra cyclone step.
In the juice processing system, 35% of the crude juice protein is
precipitated to yield a 70% protein (d.b.) meal and the residual juice is
concentrated to produce a molasses-like material (see next section). If
potatoes of 19% solids are used, commercial implementation of the process
should achieve the yields shown in Table 2 assuming a starch extraction
efficiency of 96%.
TABLE 2. EXPECTED MASS BALANCE FOR
INDUSTRIAL APPLICATION
„„„„„,.,,. MASS SOLIDS
MATERIAL ... .„ , .
(kg) (%, w.b.)
Input feed 100. 19.
Output starch 15.22 82.
Output pulp 3.09 90.
Output protein meal 1.04 90.
Output cone, juice 4.31 65.
11
-------
SECTION 4
JUICE PROCESSING
Various options are available for concentrating and upgrading the 3.9% solids
juice produced by the starch processing system described in the previous sec-
tion into animal feeds. One obvious system would include ultrafiltration to
separate and concentrate a protein stream in a similar manner as is now widely
used for upgrading cheese whey. Potato protein, however, like soybean protein,
contains trypsin inhibitors which should be inactivated with heat. Also, the
juice often contains a large microorganism load which can easily be reduced
with heat. Finally, with little more heat than required for trypsin inhibitor
inactivation and pasteurization, the heat coagulable protein - roughly one
third of the total crude protein of the juice - can be precipitated. This
protein can be centrifuged from the juice as a 20% solids material ready for
spray drying. It thus seems pointless to include a superfluous ultrafiltra-
tion step.
Another system for juice processing consists of concentration without
fractionation by multiple effect evaporation (with or without a reverse osmosis
preconcentration step) and drying with a spray or drum dryer. The ERRL has
investigated this concept at some length. The dried juice is very hygroscopic
unless lime is added before the drying step. Sufficient heating again must be
done to inactivate the trypsin inhibitor. Chick feeding trials performed by
R. Gerry at the University of Maine have shown that inclusion of this material
at levels above 6% of the diet slows growth rates.
E. O. Strolle (ERRL), in attempting to pinpoint the cause of this poor
performance, has measured losses of available lysine of as high as 80% in
batch evaporation at 60C. In earlier work, however, Strolle performed feeding
•trials with coagulated protein obtained by steam injection heating and
obtained protein efficiency ratios equal to casein. This suggests that, if
rapidly heated, the coagulable protein will precipitate fast enough into a
form wherein Maillard browning is inhibited thus yielding protein of high
lysine availability. Braverman's text on food chemistry (17) supports the
idea that reaction rates are slow for insoluble materials as compared to
rates for the same materials in solution. In addition, Maillard browning is
reduced under dilute, acidic conditions. Thus, if the protein is separated
from the juice before concentration at or below its natural pH of 6.0,
lysine availability should be enhanced.
The concepts discussed above suggest that rapid heat coagulation be in-
cluded in systems upgrading juice to animal feed. As pointed out in the in-
troduction, several groups of researchers have investigated this area. The
general results of these investigations show that protein is best separated by
12
-------
acidification to pH 5.5 and below, heating to 100C or higher, and using as
high a solids concentration as possible. Haight et al. (9) found that these
same conditions enhanced the dewatering properties of the curd. They found,
however, that more solids were precipitated with the curd if the pH was not
reduced by acid addition. Strolle et al. (8) and Meister and Thompson's data
(13) point out that the benefits of acid addition on protein precipitation
are reduced by high temperature coagulation. Acid addition also make in-
creased demands on equipment corrosion resistance - especially if a concen-
tration step is included.
A disadvantage of high temperature coagulation systems is that they are
energy intensive unless some manner of heat recovery is devised. This has
been included in the flowchart of Figure 3. As shown, juice is heated to 60C
(140F) by regenerative heating in a plate heat exchanger. While many indus-
trial plate heat exchanger systems have 90% or greater heat recovery,
attempts to obtain this with a rented plate heat exchanger lead to excessive
fouling. By keeping the heated juice temperature at or below 60C (14CF) ,
this fouling would be eliminated. Following regenerative heating, steam is
mixed with the juice to heat it further to 121C (250F) to coagulate protein
and the mixture is piped to a second heat exchanger where it is cooled to 66C
(151F). The heat recovered at this stage is used to heat the pulp drier.
The 66C juice then is passed to a nozzle centrifuge to separate the protein
precipitate. The hot protein sludge centrifuged from the juice is spray
dried to give a 70% protein (d.b.) powder. The deproteinated overflow is
returned to the first plate heat exchanger, cooled therein to 19C (66F), and
pumped into the reverse osmosis system for concentration. If the reverse
osmosis system is operated at 27 atm (400 psi), it will concentrate the juice
to 10% solids. (Pilot operation at 41 atm (600 psi) concentrated the juice
to 15% solids but led to apparent membrane compaction - the lower pressure
should ensure a one-year membrane life.) The reverse osmosis system produces
73.8 kg of water for every 100 kg of cull potatoes processed. This can be
used as wash water at the centrifugal starch washing step (25 kg per 100 kg
of potatoes) and as boiler make-up water (10.6 kg per 100 kg of potatoes).
Excess water (38.2 kg per 100 kg of potatoes processed) should be suitable
for direct discharge as it will have been produced by passage through a
highly retentive membrane. The juice is then further concentrated to 65%
solids in an evaporator, thus producing a high crude protein (43% d.b.)
molasses suitable for ruminant feeding. An additional 23.7 kg of water per
100 kg of cull potatoes processed is produced in this step. This water
should also be suitable for direct discharge having been produced by an eva-
poration/condensation process. As an option, the 10% solids concentrate
from the reverse osmosis step could be fed to cattle directly if the plant
were located near a feedlot.
Various aspects of the pilot scale implementation of the flowchart of
Figure 3 should be mentioned. In early pilot runs, a rental plate heat
exchanger (American Heat Reclaiming Corporation, New York, NY) was used as
a conventional regenerative heater using 121C (250F) juice to preheat the
incoming cold juice prior to steam injection. The system was started using
tap water and after steady state was achieved, switched to juice. While the
unit operated successfully, regenerative performance declined in the course
of the run and the pressure drop through the unit increased markedly. At the
13
-------
JUICE (3.9% SOLIDS)
IO°C
|PLATE HEAT EXCHANGER fc
I9°C
60°C
INFUSER
I2I°C
[PLATE HEAT EXCHANGER ^ZZTHEAT EXCHANGE FLUID
66°C
NOZZLE CENTRIFUGE
SLUDGE, 20% SOLIDS
SPRAY DRIER
T
PROTEIN POWDER
(70% CRUDE PROTEIN, D.B.)
(TO 8 FROM PULP DRIER);
REVERSE
OSMOSIS
EVAPORATOR
CONG. JUICE
(65% SOLIDS, 43%
CRUDE PROTEIN, D.B.)
Figure 3. Juice processing flowchart
-------
finish of the run, the plate exchanger was opened and inspected. Extreme pro-
tein fouling was noted. The turbulence generating plate configurations were
not sufficient to prevent fouling. The fouling could not be removed by clean-
in-place procedures but had to be laboriously cleaned manually. All of the
fouling took place on the heating side of the plates. This implies that the
hot juice could be used in a plate heat exchanger to heat a heat exchanger fluid
which in turn could be used to heat the pulp drier. The heat required by the
pulp drier is such that the hot juice would be reduced to about 66C (151F).
At this point, it could be used as shown in Figure 3 to heat the incoming cold
juice to 60C (140F). Under these conditions, fouling due to protein precipi-
tation would not occur.
In light of the above, the later pilot scale tests were conducted without
the plate heat exchanger using a steam infuser (Crepaco, Inc., Chicago, IL) to
heat the juice directly to 121C (250F). An air actuated valve (controlled by
a float in the infuser) was used to provide the necessary back pressure. The
juice was allowed to flash to atmospheric pressure upon passing through the
valve. While the infusion system worked well, it had to be closely adjusted
and monitored to prevent sticking of the float control valve. In retrospect,
it would seem that a simple, relatively inexpensive, steam injector would work
as well.
Following heating, the juice was allowed to cool overnight. Most of the
protein precipitate was removed with a basket centrifuge fitted with a nylon
liner. Overflow juice was recentrifuged using a cream separator modified by
removing the disks. The protein cake was ground in a blender - a process which
completely changed its physical form from a dry cake to an easily pourable
liquid without adding additional water except for a small amount used for
start-up purposes. The protein cake slurry was dried in a small spray drier
(Bowen Engineering, Somerville, NJ) using a rotating disk atomizer.
In the early pilot runs, the deproteinated juice was concentrated with a
3.0 m2 (33 ft2) reverse osmosis module fitted with a membrane rated at 97%
NaCl retention (Osmonics, Inc., Minneapolis, MN). The module was operated at
41 atm (600 psi) - the highest pressure suggested by the manufacturer. In the
first few runs, the flux rate varied with concentration in agreement with the
manufacturer's literature for glucose solutions. Concentration was continued
with the retentate returned and mixed with the feed until the solids in the
juice reached 15%. The juice temperature was maintained by cooling the jacketed
tank with tap water. After the first few runs, however, flux rates declined
and could not be reestablished even with extensive enzymatic cleaning as sug-
gested by the manufacturer. It is believed that simple membrane compaction is
the cause of the decline in flux rates. Supporting this belief is the observa-
tion that the membrane is used commercially at 27 atm (400 psi) rather than
41 atm (6OO psi) . When used at this reduced pressure, the manufacturer guaran-
tees a one-year life. At this pressure, concentration to 10% solids should be
straightforward.
Following the reverse osmosis step, the retentate was further concentra-
ted to 65% solids in a single effect evaporator (Arthur Harris, Chicago, IL).
(In the last pilot runs, due to the decline in flux rate of the reverse osmosis
15
-------
membrane, the unconcentrated juice was taken directly to the evaporator with-
out preconcentrations.) Evaporation was straightforward with only a moderate
tendency towards foaming.
An attempt was made to mix the molasses-like 65% solids concentrate with
the pulp followed by co-drying. This was not very successful; the resulting
mixture was slow drying, glassy, and hygroscopic. It seems more logical to
handle dry pulp and concentrate as separate solid and liquid feed ingredients.
The concentrate has remained stable with no sign of fermentation after several
months of room temperature storage.
To further examine the concept that high heat coagulation should preserve
lysine availability, samples of spray dried protein meal were analyzed for
lipid content, crude protein content, protein digestibility and available
lysine. The results are shown in Table 3. The digestibility values obtained
are comparable to those usually obtained with commercial soybean meal. The
levels of available lysine obtained are about the same as the total lysine of
commercial soybean meal. .The concept of rapid heating would thus seem to be
valid. The value of such potato protein meal should be at least equal to that
of soybean meal on a price per pound of protein basis.
In addition to these aspects of protein quality, possible concentration
of toxic materials into one of the product streams must be considered. The
principal toxic materials associated with potatoes are the glycoalkaloids
solanine and chaconine. ("Solanine" is also used to denote the whole family
of approximately twenty different compounds.) Dutch researchers found up to
half of the glycoalkaloids associated with the heat coagulated protein. The
results obtained in this work, however, do not confirm this (see Table 4).
Rather, the glycoalkaloids seem to act as if they were soluble (although they
are not) and can be found in the various output streams in rough proportion to
their water content at the points of separation. In particular, the glycoal-
kaloid content of the protein meal is about the same, on a dry weight basis,
as the parent potatoes. The deproteinated juice is enriched in glycoalkaloids
but not so extensively as to cause problems. For example, if the pulp and mo-
lasses were blended, the mixture would contain about twice the glycoalkaloids
(on a dry weight basis) as the parent potatoes. This, according to personal
conversations with T. J. Fitzpatrick and S. F. Osman (ERRL) who have worked
extensively with glycoalkaloids, should not cause problems in ruminant feeding
especially when blended with other feedstuffs.
In addition to naturally occurring toxic substances, care must be taken
to ensure that the added SO in its various forms is not concentrated to
hazardous levels. The reverse osmosis membranes used for deproteinated juice
concentration are highly retentive toward SO" and HSO~. Moreover, no acid
(which would tend to drive off SO in the evaporation step) ispadded in the
process. Thus, it is reasonable to assume that the added SO appears in the
output streams in proportion to their water content at the various points of
separation. The results of this mass balance are shown in Table 5 wherein SO
contents of the product streams are given on a dry weight basis for an in-
process level of 250 ppm SO (w.b.) in both the ground potato slurry and the
starch wash water stream.
16
-------
TABLE 3. SPRAY DRIED PROTEIN MEAL COMPOSITION
AND DIGESTIBILITY
Run
F
G
H
I
K
Lipid
% d.b.
1.9
2.1
2.9
5.0
4.1
Protein1
% d.b.
67.8
79.7
75.1
69.9
69.2
Digestibility2
%
84
90
85
83
84
Available lysine
g/lOOg crude protein
6.31
6.14
6.07
6.20
6.20
1. 6.25 x Kjeldahl nitrogen.
2. As measured by Pepsin digestion, AOAC,
llth ed.. Section 7.040.
3. As measured by microbial assay, AOAC,
llth ed., Section 39.128.
17
-------
00
TABLE 4. GLYCOALKALOID CONCENTRATION
AND MASS BALANCE
Run
G
H
I
J
Feed
ppm,d.b.
711
_2
331
310
Pulp
ppm,d.b. %
411 16
489 -2
353 16
425 21
Protein Meal
pprn,d.b. %
230 1.6
449 -2
457 5.9
209 2.7
Molasses
ppm,d.b.
2781
1493
887
1102
t1
84
_2
61
77
Losses
-2
2
17
-1
1. Percent of total glycoalkaloids of feed.
2. Missing feed sample.
-------
In evaluating the hazard associated with the listed SO levels, it is
necessary to include how the products might be used and what levels of SO in
the final diet are likely to cause problems. It is fortunate that rather
large amounts of SO can be injested without toxic effects. For example,
rats have been given 750 ppm SO drinking water for three generations without
effect (18). In other experiments, diets including 615 ppm SO had no effect
on rats (18).
All of the products shown in Table 5 would be used by blending with other
ingredients to the point where there should be no hazard due to high SO
levels. For examples, in dairy rations, molasses is commonly mixed with
grain at levels up to 10%. The grain mixture forms about half of the total
ration. Thus, even if the concentrated juice were used to replace molasses
at its highest use level, it still would not be likely to cause toxic effects.
TABLE 5. PRODUCT SO LEVELS
_ , SO_ Concentration
Product 2
(ppm,d.b.)
Starch 147
Pulp 695
Protein Meal 1070
Juice Concentrate 9450
*
For SO levels in the ground potato slurry
and the starch wash water of 250 ppm., w.b.
19
-------
SECTION 5
ECONOMICS
Implementation of the methods described in this report would most likely
be performed by established potato processing companies. For such companies,
the process would probably not entail construction of a new facility but be
installed as an addition to existing processing equipment and buildings. The
economics of such an addition are somewhat different from those of an indepen-
dent operation. Nevertheless, it is useful to perform an economic analysis as
if a new independent facility were to be constructed. Departures to fit
existing operations can easily be made.
The analysis that follows is based on a cull potato input of 455000 kg
(500 short tons) per day (20 hours) for a 200-day season. Such a plant would
process roughly a tenth of the U. S. cull potato production and would appear
to be the largest practicable - especially in light of recent advances in
potato flour and granule processes which upgrade small potatoes to food use.
On the other hand, the 69 metric tons of starch produced by the described
plant is small when compared to corn starch plants in this country or potato
starch plants in Europe. Another perspective of the chosen plant size is
given by noting the few remaining, traditional style, starch plants in the
U. S. each produces about 50 tons of starch per day. These plants are loca-
ted in processing centers such as northern Maine, California, and Idaho and
handle culls from several (competing) plants producing various frozen potato
products. Thus, the plant discussed herein would be a replacement for such
existing plants.
Estimated capital costs based on references (9-23), personal estimates,
and telephone calls to selected suppliers are shown in Table 6. An M&S cost
index of 520 was used (19)- Table 7 shows the resulting annual costs and
returns as well as the resulting return on investment (R.O.I.) of 40%.
As an option to the above, if the processing center were located near a
ruminant feedlot of the equivalent of at least 2500 dairy cows (or vice
versa), the pulp and concentrated juice from the reverse osmosis system could
be used directly without drying or vacuum concentration. The jwater in the
concentrate would substitute for the normal daily intake of 60 liters per
animal. The concentrate would contain all the crude protein required by the
animals. (The concentrate protein would not be well balanced in amino acids
but ruminants do not require balanced protein.) The pulp drier and vacuum
evaporator as well as some ancillary equipment would be eliminated by this
option which would require trucking of 42 metric tons of wet pulp and 130
metric tons of 10% solids concentrated juice from the plant to the feedlot
daily.
20
-------
TABLE 6. ESTIMATED FIXED CAPITAL INVESTMENT FOR
PROCESSING 455 METRIC TONS OF CULL
POTATOES PER 20-HOUR DAY
ITEM
COST, $
COST, $
I. Process Equipment:
Flume pump
Washer pump
VJasher
Destoner
Mill
Metering feeder for Na S 0
Sieve system
Mechanical defoamer
Press
Pulp drier
Pulp mill
Pulp air conv. system
Pulp bins (2)
Cyclone pumps (5)
Elut. cyclones (8)
Thick cyclone pumps (5)
Thick cyclones (4)
Pump, centrif. to cyclones
Basket centrif.
Starch drier
Starch mill
Starch conv. system
Starch screen
Plate heat exchanger
Steam injector i
Boiler (12000 Ib/hr)
Fuel tank
Nozzle centrif.
Spray drier
Rev. osmosis modules
Rev. osmosis feed pump
Rev. osmosis cir. pumps (3)
Evaporator
Concentrate pumps (2)
Concentrate tank
Balance tanks (4)
Protein meal bin
Truck scale
Total Process Equipment
5,000
3,000
20,000
20,000
50,000
4,000
60,000
24,700
50,000
60,000
10,000
5,300
10,000
15,000
5,600
15,000
800
1,000
98,800
16,800
10,000
5,500
6,000
20,000
2,000
25,400
16,900
100,000
23,700
87,000
15,800
12,000
118,600
8,000
16,900
25,200
5,000
9,500
983,000
(continued)
21
-------
TABLE 6. (continued)
COST, $ COST, $
II. Installation (30% of I) 295,000
III. Piping, Wiring, and Control 511,000
Systems (40% of I and II)
IV. Engineering and Contractors' 358,000
Fees (20% of I, II, and III)
V. Contingencies (15% of I, II, 268,000
and III)
VI. Trucks:
Concentrate truck 70,000
Protein truck 40,000
Pickup truck 7,000
Trucks Total 117,000
VII. Land 40,000
VIII. Buildings:
Office 35,000
Incoming storage 63,800
Process area 168,000
Warehouse 60,000
Buildings Total 327,000
IX. TOTAL FIXED COSTS 2,900,000
22
-------
TABLE 7. ANNUAL COSTS, INCOME, AND R.O.I.
ITEM COST OR INCOME, $
I. Costs:
Rev. osmosis module replacement 42,000
Direct labor (10 months, 3 shifts of 5) 185,000
Indirect labor (12 months, 1 shift of 7) 140,000
Potatoes (at $0.75/cwt.) 1,500,000
Oil ($13.40 per barrel, 75% effic.) 153,000
Electricity ($0.025 per kwhr) 28,000
Maintenance (land and buildings at 2%) 7,000
Maintenance (equipment at 6%) 152,000
Taxes and insurance (3% of fixed costs) 87,000
Depreciation on buildings (4%) 13,000
Depreciation on equipment (10%) 253,000
Interest on working capital 20,000
General plant overhead (50% of 172,000
maintenance and direct labor)
Total Costs 2,750,000
II. Income:
Starch ($0.10 per pound) 3,040,000
Pulp ($100. per ton) 309,000
Protein meal ($215. per ton) 224,000
Cone, juice ($78. per ton) 335,000
3,910,000
III. R.O.I.: 40%
23
-------
The cost figures for this option are shown in Tables 8 and 9, and yield a
43% R.O.I, under the assumption that the feedlot would pay for the feed value
of the pulp but not for that in the concentrated juice.
It can be seen from Tables 7 and 9 that the major portion of income is
from starch. As shown in Table 10, departure from the assumed $0.22/k
-------
TABLE 8. ESTIMATED FIXED CAPITAL INVESTMENT
UNDER LIVESTOCK OPTION
ITEM COST, $
I. Equipment: Same as Table 1 less pulp drier, 772,000
hammer mill, conv. sys., and bins; less
evaporator; use of smaller sizes of boiler
and fuel tank; but with larger cone, juice
tank.
II. Installation (30% of I.) 232,000
III. Piping, wiring, and control systems 401,000
(40% of I and II)
IV. Engineering and contractors' fees 281,000
(20% of I, II and III)
V. Contingencies (15% of I, II, and III) 211,000
VI. Trucks: Same as Table 1 117,000
VII. Land 40,000
VIII. Buildings (somewhat smaller processing 297,000
area than used in Table 1)
X. TOTAL FIXED COSTS 2,350,000
25
-------
TABLE 9. ANNUAL COSTS, INCOME AND R.O.I.
FOR LIVESTOCK OPTION
ITEM COST OR INCOME, $
I. Costs:
Rev. osmosis module replacement 42,000
Direct labor (10 months, 3 shifts of 5) 185,000
Indirect labor (12 months, 1 shift of 7) 140,000
Potatoes (at $0.75 per cwt.) 1,500,000
Oil ($13.40 per barrel, 75% effic.) 66,000
Electricity ($0.025 per kwhr) 26,000
Maintenance (land and buildings at 2%) 7,000
Maintenance (equipment at 6%) 121,000
Taxes and insurance (3% of fixed costs) 71,000
Depreciation on buildings (4%) 12,000
Depreciation on equipment (10%) 201,000
Interest on working capital 20,000
General plant overhead (50% of maintenance 156,000
and direct labor
Total Costs 2,550,000
II. Total Income: Same as Table 3 but less 3,570,000
income from cone, juice
III. R.O.I.: 43.%
26
-------
TABLE 10. R.O.I. SENSITIVITY TO FLUCTUATIONS
IN INCOME, MAJOR COSTS, AND PLANT
SIZE. (NON-LIVESTOCK OPTION)
CHANGE1 IN R.O.I. FOR
ITEM A 10% CHANGE IN
LISTED ITEM (%)
Starch 10.5
Cone. juice 1.2
Pulp 1.1
Protein meal 0.8
Potatoes -5.3
Process equipment -1.9
Labor -1.4
Oil -0.5
Electricity -0.1
Plant size2 3.7
The change listed is the actual (as opposed
to relative) change in the R.O.I. For
example, a change in R.O.I, from 40% to 41%
is shown as 1% rather than 2.5%.
2Based on equipment costs assumed proportional
to capacity raised to the 0.6 power; utilities,
proportional to output; and labor, constant.
27
-------
REFERENCES
1. Hemfort, H., Huster, H., and Heimeier, 1975. Low water consumption in
preparing potato starch. Reviewed by Peterson, N.B. Edible Starches
and Starch-Derived Syrups. Noyes Data Corp., Park Ridge, NJ.
2. Hicks, C. P., 1970. Starch refining 2-quality, yields and equipment
Process Biochemistry 5(7):30.
3. Howerton, W. W., and Treadway, R. H., 1948. Manufacture of white potato
starch. Industrial'and Engr. Chem. 40(8):1402.
4. Radley, J. A., 1976. "Starch Production Technology." Applied Science
Publishers, Ltd., London.
5. Treadway, R. H., 1959. Potato Starch. Potato Processing, ed. by Talburt,
W. F., and Smith, O., AVI Publishing Company, Westport, CN.
6. Treadway, R. H., 1967. Manufacture of Potato Starch, ed. by Whisler,
R. L., and Paschall, E. F., in "Starch: Chemistry and Technology
Vol. II", Academic Press, NY.
7. Knorr, D., 1977. Protein recovery from waste effluents of potato
processing plants. J. Food Technol. 12:563.
8. Strolle, E. O., Cording, J., Jr., and Aceto, N. C., 1973. Recovery
potato proteins coagulated by steam injection heating. J. Agric. Food
Chem. 21(6):974.
9. Haight, J. R., Rosenau, J. R., and Whitney, L. F., 1977. Process optimi-
zation for dewatering coagulated potato juice protein. Paper #77-6505
presented at the 1977 Winter Meeting of the American Society of Agri-
cultural Engineers, Chicago, IL.f Dec. 13-16.
10. Anon., 1968. Protein recovery from potato starch. Process Biochemistry,
May, 1968. 51 pp.
11. Oosten, B. J., 1976. Protein from potato starch mill effluent. In:
"Food from Waste", ed. by Birch, G. G., Parker, K. J., and Worgan, J. T.,
Applied Science Publishersf Ltd., London.
12. Stabile, R. L., Turket, V. A., and Aceto, N. C., 1971. Economic analysis
of alternative methods for processing potato starch plant effluents.
Proceedings of the Second National Symposium on Food Processing Wastes,
Denver, CO.
28
-------
13. Meister, E., and Thompson, N. R., 1976. Physical-chemical methods for
the recovery of protein from waste effluent of potato chip processing.
J. Agric. Food Chem. 24(5):919.
14. Meister, E., and Thompson, N. R., 1976. Protein quality of precipitate
from waste effluent of potato chip processing measured by biological
mechods. J. Agric. Food Chem. 24(5):924.
15. Shaw, R. and Shuey, W. C., 1972. Production of potato starch with low
waste. Amer. Potato J. 49(1):12.
16. Dahlstrom, D. A., 1954. Fundamentals and applications of the liquid
cyclone. "Mineral Engineering Techniques". American Institute of
Chemical Engineers. 50(15).
17. Braverman, J. B. S., 1963. "Introduction to the Biochemistry of Foods".
Elsevier Publishing Company.
18. Anon., 1975. Sulfites as food additives. Food Tech. 29(10):117.
19. Anon., 1978. Economic indicators. Chem. Engineering. 85(2):7.
20. Enochian, R. V., Edwards, R. H., Kuzmicky, D. D., and Kbhler, G. O.,
1977. Leaf protein concentrate (Pro-Xan) from alfalfa: an updated
economic evaluation. ASAE paper #77-6538. Presented to the Winter
Meeting of the American Society of Agricultural Engineers, Chicago,
IL, Dec. 13-16.
21. Vosloh, C. J., Jr., Edwards, R. E., Enochian, R. L., Kuzmicky, D. D.,
and Kohler, G. 0., 1976. Leaf protein concentrate from alfalfa: an
economic evaluation. National Economic Analysis Division, Economic
Research Service. Agricultural Economic Report No. 346.
22. Peters, M. S., and Timmerhaus, K. D., 1968. "Plant Design and Economics
for Chemical Engineers, 2nd ed." McGraw-Hill Book Company, NY.
23. Sohns, V. E., 1971. Cost analyses for new products and processes
development in USDA laboratories. J. Am. Oil Chemists' Soc.
48(9):362A.
29
-------
APPENDIX
PHOTOGRAPHS OF PILOT-SCALE IMPLEMENTATION
Figure Al. Washing with perforated basket
and power washer.
30
-------
Figure A2. Grinding with Fitzmill with
SO addition.
31
-------
Figure A3. Feeding of slurry with peristaltic pump.
Also shown is the overflow juice tank, recycle
piping, and the tubular heat exchanger used
to prevent temperature rise due to pumping.
32
-------
Figure A4. Vibratory screen used to separate
pulp from starch and juice.
33
-------
Figure A5. Screw press used to dewater
pulp prior to drying.
34
-------
Figure A6. Flov; scheme at the underflow of the elutriaticn cyclone.
The underflow, together with wash water from the basket centrifuge,
is pumped through the surge vessel into the small secondary cyclone.
Overflow from the secondary cyclone provides the elutriation flow of
the primary cyclone.
35
-------
Figure A7. Steam infusion heating of juice. The juice is metered
with the peristaltic pump through the diaphragm pump into the in-
fuser shown at the right. The level of hot juice in the infuser is
maintained by a float which controls the actuating air of the back-
pressure valve shown in Figure A8. The hot juice leaves the infuser
through the pipe at the lower right.
36
-------
Figure A8. Air actuated back-pressure valve. The hot juice
(121C) flashes to atmospheric pressure as it leaves
the valve.
37
-------
Figure A9. Spray dryer used to dry the protein
precipitate centrifuged from the
heated juice.
38
-------
Figure A10. Vacuum evaporator used to concentrate the
deproteinated juice to a 65% solids molasses-
like material.
39
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-79-208
2.
3. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
LOV7 WASTEWATER POTATO STARCH/PROTEIN PRODUCTION
PROCESS
5. REPORT DATE
Decemb er 1 9 7
------- |