EPA-600/2-77-241
December 1977
Environmental Protection Technology Series
CLOSED PROCESS WATER LOOP IN
NSSC CORRUGATING MEDIUM MANUFACTURE
industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
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-600/2-77-241
December 1977
CLOSED PROCESS WATER LOOP
IN NSSC CORRUGATING MEDIUM MANUFACTURE
by
Gerald 0. Walraven
William R. Nelson
Peter E. DeRossi
Richard L. Wisneski
Green Bay Packaging Inc.
Green Bay, Wisconsin 54305
Grant No. S-800520
Project Officer
Ralph H. Scott
Food and Wood Products Branch
Industrial Environmental Research Laboratory-Cincinnati
Corvallis Field Station
Corvallis, Oregon 97330
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 Environ-
mental Research Laboratory-Cincinnati, U. S. Environmental
Protection Agency, and approved for publication. 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 endorse-
ment or recommendation for use.
11
-------
FOREWORD
When energy and material resources are extracted, processed,
converted, and used, the related pollutional impacts on our en-
vironment and even on our health often require the new and in-
creasingly more efficient pollution control methods be used. The
Industrial Research Laboratory - Cincinnati (IERL-CI) assists in
developing and demonstrating new and improved methodologies that
will meet these needs both efficiently and economically.
This report describes means and methods employed to develop
a nearly closed system for wastewaters produced from the manu-
facture of neutral sulfite semi-chemical corrugating medium, a
significant portion of the effort concerning in-plant recycling
and control. This program has been supplemented by the first
full-scale commercial reverse osmosis system operating in the
pulp and paper industry. The reverse osmosis facility is employ-
ed to process those accidental losses not controlled by the in-
plant recycling system. The combination of tight in-plant con-
trol and reverse osmosis processing has permitted the industry to
satisfy discharge permit requirements without the use of any ex-
ternal treatment for wastewaters. The information will be of
value to other segments of the industry, consultants and reverse
osmosis equipment suppliers. For further information please con-
tact the Food and Wood Products Branch of the Industrial Environ-
mental Research Laboratory, Cincinnati.
David G- Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
111
-------
ABSTRACT
From 1972 to 1976, the Green Bay Packaging corrugating mill
was redesigned to reuse enough water to nearly close the pro-
cess water system. Less water enters the system during the
cooking of chips and repulping of secondary fiber than leaves
the system with the sheet to the paper machine dryers. Many
small, dilute water streams, therefore, can be brought into the
system without upsetting the water balance.
When extraneous water inputs do upset the system balance,
the condition is correctable by removing water, using either
thermal evaporation or reverse osmosis, The reverse osmosis
plant design operating performance and economics are described.
Although many reverse osmosis operating problems have been
solved, flux rates are somewhat lower than had been predicted.
Other system additions and revisions for process water entrap-
ment, recycling, and surge protection are described.
A monitoring system is in use for early spill detection
and problem correction. Included in key areas is standby
equipment for use to correct failures quickly.
Levels of BOD loss have been reduced from the 9072 kg per
day range (20,000 Ib/day) of 1971 to less than 454 kg per day
(1000 Ib/day) as a monthly average for 1975. The daily maximum
of 1814 kg (4000 Ib) had not been exceeded in any mill-operating
day during 1975.
This report was submitted in fulfillment of Grant Number
S-800520 under (partial) sponsorship of the Office of Research
and Development, U.S. Environmental Protection Agency. This
report covers the period from July 19, 1972, through February 2,
1976. Progress will be reported annually for an additional
five-year period.
IV
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables vii
Abbreviations and Symbols viii
Acknowledgment ix
1. Introduction 1
2. Conclusions 3
3. Recommendations 5
4. Materials and Methods 6
Process description 6
Process water system closure 9
Mill system water recycling 26
Internal water monitoring and alarm system . . . .28
Reverse osmosis system 31
5. Operating Experience 42
In-plant environmental effects of mill water
system closure . 42
Process water system additives 43
Process water system ash buildup 44
Reverse osmosis plant operating philosophy . . . .46
Reverse osmosis module fiberglass membrane
support tube failure 48
Reverse osmosis membrane fouling 55
Reverse osmosis pump and pump seal problems . . .58
6. Results and Discussion 60
Resume of mill operating experience 60
Medium strength characteristics and
conversion behavior 66
Economic evaluation of reverse osmosis vs.
evaporation 68
References 71
Appendices
A. Analytical and test procedures and equations ... .72
B. Institute of Paper Chemistry summary on medium
quality 76
C. Project visitors 79
Glossary 82
v
-------
FIGURES
Number Page
1 Mill process 7
2 Pulp mill—combustion plant liquor flow 8
3 Historical mill discharge performance 10
4 Mill discharge stream—1971 11
5 Mill discharge stream—1975 12
6 Paper machine floor collection system—1971 14
7 Paper machine floor collection system—1975 18
8 Pulp mill floor collection system—1971 20
9 Pulp mill floor collection system—1975 21
10 Mill effluent sources 25
11 Process water reuse system 27
12 Paper machine effluent system—alarm points 30
13 Reverse osmosis plant interface 34
14 Reverse osmosis plant 35
15 Pump assembly rack 36
16 Module rack 38
17 Reverse osmosis flow diagram 39
18 Reverse osmosis plant control 41
19 Grit removal system 45
20 Flux performance comparison 49
21 Cleaned module flux data . . . 50
22 Photomicrographs—new and aged tubes 52
23 "Original" vs. "improved" module failure experience. . 54
24 Relationship of percent dissolved solids 61
25 Paper machine productivity 62
26 Total mill cost factors 64
VI
-------
TABLES
Number Page
1 Mill Discharge Streams—1971 13
2 Mill Discharge Streams—1975 13
3 Inventory of Potential Problems as of 1971 and
Disposition as of 1975—East Machine Room
Discharge 15
4 Inventory of Potential Problems as of 1971 and
Disposition as of 1975—West Machine Room
Discharge 16
5 Inventory of Potential Problems as of 1971 and
Disposition as of 1975—Dilution Streams 17
6 Inventory of Potential Problems as of 1971 and
Disposition as of 1975—Pulp Mill Discharge 22
7 Inventory of Potential Problems as of 1971 and
Disposition as of 1975—Recovery Plant Area 23
8 Inventory of Potential Problems as of 1971 and
Disposition as of 1975—Wood Room Discharge 23
9 Inventory of Potential Problems as of 1971 and
Disposition as of 1975—Tank Storage Area 24
10 Total Reverse Osmosis Plant Average Operating
Parameters 61
11 Strength Comparison—Open to Closed System 67
VII
-------
LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
BOD 5
Btu
E.D.T.A. Na4
gfd
gpm
HP
MGD
NSSC
PH
psi
#
SYMBOLS
HC1
Mg
Na2C03
Na2S04
five-day biochemical oxygen demand
British thermal unit
sodium salt of ethylenediaminetetra acetic acid
flux units—gallons per ft2-day
gallons per minute
horsepower, electrical
million gallons per day
Neutral Sulfite SemiChemical
hydrogen ion concentration, logarithmic form
measure of pressure—pounds per square inch
pounds, mass
hydrochloric acid
magnesium
sodium carbonate
sodium sulfate
viii
-------
ACKNOWLEDGMENTS
This project was performed under the direction of
Mr. Ralph H. Scott, Chief, Wood Products Staff, U. S. Environ-
mental Protection Agency, federal Project Officer. Project
Director for Green Bay Packaging was Mr. William R. Nelson;
Project Manager was Mr. Gerald 0. Walraven. Other Green Bay
Packaging personnel participating in the project were Messrs.
David C. Morris (now with Weyerhaeuser Co.), Scott Brown (now
with Simons-Eastern Co.), and Peter E. DeRossi, Chemical
Engineers; Messrs. Orv Rautmann and Dave Meverden, Chemists;
Mr. Richard Wisneski, Engineering Technician; Mr. Frank Piechota,
Process Technician; and Mrs. Kathy Pitts, Secretary.
The close cooperation and assistance of Universal Oil
Products, Fluid Systems Division (previously Fluid Sciences
Division), is gratefully acknowledged. Key personnel were
Messrs. Kenneth E. Anderson, Richard A. Walker, Richard B.
Hartupee, and Dr. Jay Sobel of Corporate Research.
The long-standing and excellent support of the staff of the
Institute of Paper Chemistry is also acknowledged.
The privilege to present project interim reports through
both the Technical Association of the Pulp and Paper Industry
and American Institute of Chemical Engineers, Environmental
Division, is gratefully acknowledged.
IX
-------
SECTION 1
INTRODUCTION
Green Bay Packaging Inc. operates a pulp and paperboard mill
adjacent to the Fox River in Green Bay, Wisconsin. The mill pro-
duces approximately 270 metric tons per day of corrugating medium
from a combination of virgin mixed hardwood, neutral sulfite
semichemical (NSSC) pulp, and recycled corrugated box plant waste
fiber.
The organic products of reaction (pulping liquors) from the
sodium NSSC pulping operations are separated at their highest
concentration from the virgin pulp fibers produced. The strong
pulping liquors are screened to minimize the loss of usable
fiber, inventoried, and eventually burned in a Dorr-Oliver fluid
bed combustion plant to destroy the organics and produce a mar-
ketable inorganic ash. The residual (approximately 25%) pulping
liquor organics remaining with the pulp fibers subsequently leach
into the process water, which conveys the fibers through the'con-
ventional fiber processing and papermaking process.
The corrugated box plant wastes are returned to the paper-
board mill process for reuse by repulping in excess paper machine
process water at elevated temperature. Residual soluble mate-
rials become redissolved in the process water as the waste fibers
are separated and resuspended. The resulting recycled fiber is
eventually combined with the NSSC pulp fiber and passed through
the normal processing stages, finally being made into corrugating
medium at the paper machine.
Earlier construction and operation of the fluid bed combus-
tion plant had satisfactorily solved the strong NSSC pulping
liquor waste problem. It has been necessary to address the
problem of residual pollutional losses as they appear in both
variable and constant excess process water losses containing low
concentrations of solubles from both the virgin NSSC and the
recycled pulping and papermaking operations.
The efforts of the company to further reduce the pollutional
effect of the mill discharge to the Fox River have been directed
primarily at identifying, in both quality and quantity parame-
ters, the many uses of the process water system and to identify
methods of collection and control, reuse, or treatment to prevent
the losses of such excess process waters to the river.
-------
Encouraging progress had been made in the reuse of excess
process waters at the Green Bay mill concurrent with a pilot
plant evaluation project on the reverse osmosis process for con-
centrating such excess process waters under a Demonstration Grant
from the Office of Research and Monitoring of the U. S. Environ-
mental Protection Agency. The information generated at that time
has been reported in Water Pollution Control Research Series
12040 FUB, published by the U. S. Environmental Protection Agency
in January, 1972. The results achieved during that time are the
basis for the project here reported.
In June, 1972, a Research, Development, and Demonstration
Grant offer was made to Green Bay Packaging Inc. in support of a
mill scale, multifaceted effort to demonstrate a "Closed Process
Water Loop in NSSC Corrugating Medium Manufacture." The project
objective was to accomplish the maximum closure of water use loop
in an integrated pulp and paperboard mill by recycling contami-
nated process waters for direct reuse; by providing a protective
collection and surge system for such excess surge volumes as
might occur during both normal operations and process upset con-
ditions; by providing a continuous system of loss monitoring
capability to support operator decisions; and by providing a
reverse osmosis plant of appropriate capacity to separate pollu-
tional constituents from such excess volumes as might occur,
prior to their discharge from the mill, while recycling permeate
so produced to the maximum practical extent.
A secondary objective of the project was to demonstrate the
results of paperboard making operations under such high water
reuse conditions and to report useful techniques of operating.
The project was visualized as a one-year operation under maximum
closure conditions to fully demonstrate the effects on production
efficiency, maintenance, product quality, and mill pollutipnal
control. A further objective of the project was to report the
operating performance of the reverse osmosis facility over one
year of continuous operation, together with capital and operating
costs of the reverse osmosis facility.
Although the project completion was anticipated in 32
months, actual completion has required 48 months due to time con-
suming problem analysis and correction in certain areas. In
general, the time extensions were required to allow for exper-
ience in proprietary reverse osmosis equipment corrections, which
in turn allowed for a more comprehensive evaluation of the over-
all system performance. The application of membrane technology
in this instance is a direct substitution for conventional ther-
mal evaporation procedures. In addition, the project objectives
present an opportunity to evaluate membrane separation perfor-
mance, as well as a cost competitive technology, in comparison
with water removal operations of another form.
-------
SECTION 2
CONCLUSIONS
Intensive recycling to minimize system losses has proven to
be the simplest and best course for our mill. For three years,
the program to entrap, contain, and reuse water has proven suc-
cessful, as demonstrated by continued low losses while maintain-
ing production of a quality corrugating medium. This approach
can apply generally, but even a simple change in equipment—batch
to continuous digesters, screw presses to washers—may change the
water balance to preclude complete recycling, through either cost
or physical impracticability.
All constituents of a closed process water system tend to
increase in concentration to a new equilibrium, depending on
system conditions. Calcium can reach a concentration where
objectionable scale will form in the process water or liquor
systems. Sodium chloride, at increased concentrations, can
accelerate corrosion to the point where replacement materials of
construction are required, or it may interfere with a liquor
burning operation. The modifications necessary for recycling
will necessarily be highly dependent on the existing plant and
process configuration. An option of excess contaminated process
water collection, containment, and reuse open to one plant may
not be a valid option for another.
A small commercial reverse osmosis plant has been able, on
demand, to maintain the reuse system volume within the limita-
tions of the available surge storage. Although most reverse
osmosis operating problems have been solved, flux rates are
somewhat lower than had been predicted. The actual operating
flux was 3.02 liters per square meter-hr (5.13 gal/ft^-day)
compared to the predicted flux of 3.94 liters per square meter-hr
(6.68 gal/ft2-day).
The life expectancy of the reverse osmosis module hardware
has been greatly improved through the efforts of our supplier.
The project objective to identify module and membrane life
expectancy has not yet been realized. The substitution of
reverse osmosis for conventional steam evaporation has been
successfully demonstrated.
The low energy need for water removal from the process water
system has been offset by a higher capital cost for equipment.
-------
The relative economics prevailing at a given plant site will
determine whether membrane separation can offer advantages over
thermal evaporation.
The large number of process modules necessary to build
reverse osmosis capacity presents serious problems to the appli-
cation of membrane technology to high volume, continuous duty
applications. The experience of this project suggests continued
efforts be directed towards improving reliability of reverse
osmosis support structures and their associated equipment.
The issue of trace contamination in tightly closed mill
water reuse systems deserves further attention. Relatively rapid
changes in unexplained membrane fouling and the forms of fouling
experienced suggest that transient contaminants can seriously
affect the overall performance (and costs) of a reverse osmosis
operation processing internal plant reuse waters.
The project experience demonstrates the ability to recycle
process waters in the Green Bay Packaging mill without serious
detrimental effects on-production or product quality. Changes in
raw materials, such as a high calcium water supply or inputs of
higher amounts of sodium chloride in cooking chemicals, will
create new equilibrium conditions in a closed system. This con-
clusion only applies to the solubles present in this water reuse
system. Application to another system must be predicated on
pilot experience on the reuse stream existing in that system.
Therefore, it should not be presumed that our experience is di-
rectly applicable to all unbleached board mills or even those
producing the same grade of NSSC corrugating medium because of
the vast differences in the raw materials used and the construc-
tion and age of many of these facilities.
-------
SECTION 3
RECOMMENDATIONS
To become more economically attractive, vendors of reverse
osmosis equipment should produce membranes both capable of
higher flux rates and functional over wider ranges of pH and
temperature. These advances in membrane technology would lead to
reduced capital costs for equipment and lower plant operating
costs, hopefully to a point where much broader application will
be realized. Also, even though much progress has been made in
membrane support structures, there is still a need to improve
reliability and increase module life by improving the chemical
resistance of the materials presently used. Module maintenance
and replacement costs, although not yet available from our
project data, will represent a large part of overall operating
costs.
The causes and mechanisms of membrane fouling are not well
understood. Additional efforts are needed to develop more under-
standing of fouling and methods of prevention and effective,
rapid cleaning.
Development of better, higher volume high pressure pumps is
needed for all applications, particularly for feed streams which
act chemically to deteriorate existing industrial materials of
pump construction.
When planning new installations or expanding old ones, more
consideration should be given to higher consistency pulp pro-
cessing methods. The costs and advantages of maintaining a
balanced water system with internal disposal of pollutants should
be weighed against existing conventional systems, typically
utilizing end-of-the-line treatment systems.
The corrosion problems associated with tightly closed water
reuse systems, inherently more severe because of increased tem-
peratures and electromotive activity, should be addressed by
basic equipment manufacturers and process design engineers.
Improved chemical resistance of all materials exposed to recycled
process water is a necessary objective to improve system reli-
ability and to reduce maintenance and operating costs.
-------
SECTION 4
MATERIALS AND METHODS
PROCESS DESCRIPTION
The mill produces 272 metric tons (300 short tons) per day
of corrugating medium from NSSC pulp with mixed hardwoods and
from recycled box plant clippings. Figure 1 shows the mill pro-
cess. Four rotary digesters produce about 181 metric tons (200
short tons) per day of pulp using a vapor phase cook. Paper
machine process water is added to the chips during the blow-down
period to aid in the recovery of spent liquor. Partially pulped
chips leave the screw presses at about 55% solids. Paper machine
process water is used for washing in the presses. This process
water is also used to dilute the pressed chips to 26% consistency
for primary refining. Secondary fiber, slurried in process
water, is added to the pulp prior to finish refining. Further
dilution to headbox consistencies with paper machine process
water precedes both a coarse cleaner and a two-stage final cen-
trifugal cleaning system before the paper machine.
Figure 2 shows a more detailed layout of the process streams
in the pulp mill, as well as those going to the fluidized bed
reactor liquor system. In the pulp mill, liquor is separated
from the cooked chips by two screw presses, combined with
digester wash and blow liquor, and screened to reclaim usable
fiber.
The screened liquor, containing about 18% dissolved solids
and 8.4 grams per liter (70 lb/1000 gal) of very fine suspended
solids, is stored in a 227.1-cubic meter (60,000-gal) insulated
tank before going to the evaporators. The smaller portion is fed
to the Aqua-Chem spray-film evaporators and concentrated to about
25% solids. The evaporated liquor can be stored in either of two
large insulated tanks. The smaller of the two is the normal pro-
cess storage and can hold a maximum of 557.7 cubic meters
(150,000 gal). The other tank is used for long-term storage for
reactor shutdowns and can store up to 946.2 additional cubic
meters (250,000 gal). The combined liquor storage will allow for
a reactor shutdown of as long as ten days. The normal turnaround
time for reactor shutdown, cleanout, and repairs is about five
days.
-------
GASES - WATER - CO,
SALT CAKE
32 TONS/DAY
STEAM
WOOD
250 TONS/DAY
CHEMICALS
SODIUM SULFITE
30 TONS/DAY
SODIUM CARBONATE
8 TONS/DAY
FLUIDIZED
BED REACTOR
CENTRIFUGAL
CLEANERS
CORRUGATING
MEDIUM
300 TONS/DAY
PRIMARY
REFINERS
FINISHING
REFINER
CORRUGATING
WASTE
FIBER
100 TONS/DAY
REPULPER
DILUTION
PROCESS
WATER
Figure 1. Mill process.
-------
oo
CLARIFIED
WHITE WATER
R-0 CONCENTRATE
PULP MILL WASH
DILUTION TANK
I
DIGESTERS
LIQUOR RETURN
TANK
CHEMICAL
MIXING TANK
SCREW
PRESSES
(HL) HIGH LEVEL ALARM
FLUIDIZED
BED
REACTOR
PULP MILL
WASTE LIQUOR
STORAGE TANK
EMERGENCY
WASTE LIQUOR
STORAGE TANK
EVAPORATED
WASTE
LIQUOR
STORAGE
TANK
Figure 2. Pulp mill—combustion plant liquor flow,
-------
Pulp mill liquor is combined with the 25% liquor, and the
mixture is fed to the Venturi scrubber/evaporator and concen-
trated to 43% by the reactor hot gas stream, A portion of the
43% liquor is fed to the reactor as fuel. The ratio of 18% and
25% liquor is adjusted for solids and water inventory control in
the reactor Venturi loop. Additional information on the Dorr-
Oliver Fluidized Bed Reactor system can be found in the
August 17, 1970, issue of Paper Trade Journal.(1)
PROCESS WATER SYSTEM CLOSURE
Much progress has been made since 1971 in reducing BOD dis-
charge from the mill by a series of steps leading to an essen-
tially closed process water reuse system (Figure 3). Levels have
been reduced from the 44,091 kilograms per day (20,000 Ib/day)
range in early 1971 to less than 2200 kilograms per day
(1000 Ib/day)—monthly average—for 1976. The total mill BOD
generated approximates 99,000 kilograms per day (45,000 Ib/day).
Note also the progress in reducing daily maxima so that 8800
kilograms per day (4000 Ib/day) has not been exceeded in any mill
operating day in the last year.
Figures 4 and 5 are averages of the discharge streams from
the operating segment of the mill for 1971 and 1975, respec-
tively. Keys for the losses for each stream are in Tables 1 and
2. A major step had already been accomplished by the spring of
1971, with process water successfully replacing fresh water on
most of the paper machine showers. Table 1 indicates the magni-
tude of the losses that still existed in 1971 after the shower
conversion. That year, the evolution of a closed process water
system started in earnest with the preparation of an inventory of
discharge streams according to priorities. The improvements ob-
tained from 1971 to 1975 are observed by comparing the differ-
ences in each outfall and the total mill discharge in Table 2
with those of Table 1.
Figure 6, the paper machine basement trench system diagram
in 1971, is an example of the type of inventory for losses which
proved extremely useful in conquering the problems that existed
at a given outfall. The steps taken to eliminate losses and re-
vise the paper machine system are indicated in Tables 3, 4, and 5.
When completely revised, the paper machine water collection
system appeared as is shown in Figure 7.
The highest priority was assigned to those contributions
whose elimination was necessary to reach 1973 permit limits.
These included pit and tank overflows and gutter or floor drain
flows that could be recycled by the addition of more pumps and
tankage. Uncontaminated cooling water was repiped to bypass the
recycle loop (see Table 5, Inventory of Dilution Streams). Addi-
tional screened process water washup hoses were added at the
paper machine so that city water is used only at the final step
-------
26,000
24,000
-------
MILL DISCHARGE
FLOW - 1.80 MGD
TOT. SO LI D S - 2847 7 */ DAY
SUS. SOLIDS- II22*/DAY
BOD 5 - 59 3 2*/ DAY
TO RIVER
FLUIDIZED BED
REACTOR
RIVER WATER IN
FLOW - 1.20 MGD
TOT. SOLIDS - 3450?VDAY
SUS. SOLIDS - I90*/DAY
BODs - 115V DAY
Q-SEE TABLE-1
BOILER FLYASH
POND
CITY WATER IN
FLOW
TOT. SOLIDS-
SUS. SOLIDS -
BOD5
0.7 MGD
75O*/DAY
O^DAY
0*/DAY
1
CORRUGATED
WASTE
REPULPER
MACHINE ROOM
1
Figure 4. Mill discharge stream—1971.
-------
MILL DISCHARGED
FLOW - Tio^MGD
TOT, SOLIDS- 7343*/DAY
SUS. SOLIDS- 420*/DAY
BOD 5 - I008#/DAY
FLUIDIZED BED
REACTOR
N)
RIVER WATER IN
FLOW - 1,20 MOD
TOT. SOLIDS - 2968%DAY
SUS. SOLIDS - 281*/DAY
BODs - 83*/DAY
Q- SEE TABLE-2
BOILER FLYASH
POND
ROOM
*AGE
PULP MILL
CITY WATER IN
FLOW - 0.5 MGD
TOT. SOLIDS- 700 */ DAY
SUS. SOLIDS- 0*/DAY
BOD 5 - 0*V DAY
CORRUGATED
WASTE
REPULPER
L- 1
1
1
1
1
1
1
1
1
1 MACHINE ROOM
1
1
1
Figure 5. Mill discharge stream--1975.
-------
TABLE 1. MILL DISCHARGE STREAMS—1971
(Key for Figure 4)
1.
' 2.
3.
4.
5.
6.
7.
Stream
East machine
room
West machine
room
Pulp mill
Tanks
Wood room
Fluidized bed
reactor
Total mill
discharge
Total
Flow solids
(MGD) (f/day)
1.070 11,887
0.029 2,100
0.287 2,100
0.047 9,929
0.040 393
0.327 2,031
1.800 28,477
Total
suspended
solids
(#/day)
377
137
184
329
45
50
1,122
BOD5
(i/day)
2,350
490
327
2,400
0
365
5,932
TABLE 2.
MILL DISCHARGE STREAMS — 1975
(Key for Figure 5)
1.
2.
3.
4.
5.
6.
7.
Stream
East machine
room
West machine
room
Pulp mill
Tanks
Wood room
Fluidized bed
reactor
Total mill
discharge
Total
Flow solids
(MGD) (#/day)
1.140 5,637
(Discontinued)
0.250 610
0.001 430
0.040 108
0.170 558
1.600 7,343
Total
suspended
solids
(#/day)
125
281
14
0
0
420
BODs
(tt/day)
832
19
107
0
50
1,008
13
-------
REPULPER
TO REUSE
KEY REFER TO TABLES 3,4,5
NUMBER - POTENTIAL LOSS
H> LETTER-DILUTION
PROCESS WATER
COLLECTION PIT
FINISHING
REFINER
WEST MACHINE
ROOM DISCHARGE
OQO
VACUUM PUMPS
MAIN EFFLUENT LINE
EAST MACHINE
ROOM DISCHARGE
Figure 6. Paper machine floor collection system—1971.
-------
TABLE 3. INVENTORY OF POTENTIAL PROBLEMS AS OF 1971
AND DISPOSITION AS OF 1975—EAST MACHINE ROOM DISCHARGE
(In Order of Priority)
(Code for Machine Room Discharges, See Figures 6 and 7)
1971 - Problem 1975 - Status
Number 1 Priority: Urgent—elimination necessary to meet 1973 limits
1) Seal pit and paper machine process Combined pits; installed larger pumps;
water collection pit overflow; shut- installed spare pump; installed high
down and startup surges or pump level alarm
failure
Amount: 3000 gal
Frequency: once/day
2) Losses from paper machine gutters All losses directed to process water
and floor drains from normal seepage pit by new piping
and washup
Amount: 2-5 gal/min
3) Periodic entrainment of process water Installed conductivity alarm and manual
in Centri-Cleaner vacuum pump seal vacuum breaker; modified vacuum pump
water , seal water system
Amount: 5 gal/min
Frequency: 30 min/wk
Number 2 Priority: Needed to ensure stable operation and better closure
None in this category
Number 3 Priority: May be postponed until after 1973—more information is required
4) Overflow of Centri-Cleaner reject Discharges to enclosed sewer and is
tank; shutdowns and failure of pump recycled with sump pump
Amount: 50 gal
Frequency: each shutdown,
2-3 times/wk
5) Couch and wet press vacuum pump • Unresolved
carryover
Amount: 3-5 gal process water/min
6) Trap dumping Magna Cleaner finishing Now dumps to closed sewer and is
refiner recycled by sump pump
Amount: 20 gal process water
Frequency: once each.8 hr
7) Trap dumping Magna Cleaner secondary Now dumps to closed sewer and is
Centri-Cleaner system recycled by sump pump
Amount: 20 gal process water
Frequency: once each hr
8) Overflow from overfilling paper Now is in closed sewer area; stock is
machine stock storage chests manually washed back to sump pump for
Amount: 600-1000 gal process water recycle
and fiber
Frequency: once/mo
Number 4 Priority: Insignificant or poorly defined
9) Paper machine area pumps and side Most leaks are located in area of
agitator seal leaks enclosed sewer and are recycled with
Amount: small but continuous sump pump
10) Seepage from area under paper Unresolved
machine wet presses
Amount: small but continuous,
some fiber loss
15
-------
TABLE 4. INVENTORY OF POTENTIAL PROBLEMS AS OF 1971
AND DISPOSITION AS OF 1975—WEST MACHINE ROOM DISCHARGE
(In Order of Priority)
(Code for Machine Room Discharges, See Figures 6 and 7)
1971 - Problem 1975 - Status
West sump pump is not used presently.
Number 1 Priority: Urgent—elimination necessary to meet 1973
limits
12) Kraft clippings repulper Installed bar screen for large
sump pump failure; me- junk; can now be diverted to
chanical or plugged with process water pit for complete
large junk pump failure
Amount: 5000 gal
Frequency: once/two mo
13) Losses of process water Eliminated at end of pilot
and fiber from pilot DSM work
screen for reverse osmosis
pilot work; coarse fraction
only
Amount: 5 gal/min
14) DSM screen supply tank Installed interlock system
overflow high level; installed larger
Amount: 2000-3000 gal pump to DSM screen; added one
Frequency: twice/mo screen section to DSM screen
Number 2 Priority: Needed to ensure stable operation and better
closure
None in this category
Number 3 Priority: May be postponed until after 1973—more
information is required
None in this category
Number 4 Priority: Insignificant or poorly defined
15) DSM supply tank area; Added new sewer trench to
pumps and agitator seal discharge to recycle system
leaks
Amount: small but
continuous
16
-------
TABLE 5. INVENTORY OF POTENTIAL PROBLEMS AS OF 1971
AND DISPOSITION AS OF 1975
Inventory of Dilution Streams Entering Process
East Machine Room
(Code for East Machine Room Dilution Streams)
(See Figures 6 and 7)
1971 - Problem
1975 - Status
A) Air conditioner cooling
water, three units
Amount: 25 gal/min
B) Oil coolers and air
compressor cooling water
Amount: 17 gal/min
C) Washup hoses, used on
paper breaks
Amount: 1000 gal/day
D) Centri-Cleaner vacuum pump
seal water
Amount: 9 gal/min
Piped out of closed system
Piped out of closed system
Two hoses connected to clari-
fied process water; city water
hose used only for final
washing
Amount: 300 gal/day
Piped out of closed system but
alarmed for process water en-
trainment; see Item 3—East
Machine Room Discharge
West Machine Room
None in this category
Pulp Mill
(Code for Pulp Mill Dilution Streams, See Figures 8 and 9)
A) Oil coolers
Amount: 10 gal/min
B) Digester blow line sample
condenser
Amount: 8 gal/min
C) Pulp mill testing station
sinks
Amount: 2 gal/min
Rerouted directly to main
effluent line
Rerouted directly to main
effluent line
Rerouted to sanitary sewer
17
-------
00
WEST MACHINE
ROOM DISCHARGE
(DISCONTINUED)
KEY -REFER TO TABLES 3,4,5
(H) NUMBER - POTENTIAL LOSS
ffft LETTER - DILUTION
PROCESS WATER
COLLECTION PIT
i
bOO
CHESTS
FINISHING
REFINER
\ w Ah
000
VACUUM
MAIN EFFLUENT LINE
PUMPS
EAST MACHINE
ROOM DISCHARGE
Figure 7. Paper machine floor collection system—1975.
-------
in washup. City water, in small amounts, is still used on the
cut squirts due to their tiny nozzles and at other critical
points, such as the breast roll shower and suction roll box
seals.
The same type of approach was used in the other operating
areas of the mill. Figures 8 and 9 and Tables 5 and 6 are the
result of a similar study made for the pulp mill. Tables 7, 8,
and 9 indicate the problems and changes made in the reactor area,
wood room, and tank storage area, respectively. No diagrams are
included, but most items are self-explanatory.
In a lower priority category were steps that were needed to
ensure stable operation and better closure. These included such
steps as collection of evaporator cleanout streams and steam
tracing and insulating of storage tank transfer lines formerly
dumped to avoid freezing.
In a third priority category were items which could be
deferred until after 1973 or which required more information.
Some of these are now being evaluated, for example, a cyclonic
mist separator on the couch vacuum pump. Others are unresolved
at this time. Some have been taken care of more recently,
including collection and recycling of both the acid scale removal
waste stream from the bimonthly cleaning of recovery plant piping
and the caustic from the evaporator demister pad wash. An unre-
solved problem is the acid condensate from the evaporator, adding
22.7 kilograms per day (50 Ib/day) of BOD to the effluent.
Another continuing problem is leaking of strong liquor from pump
and agitator packing in the storage tank area.
The list of problems and solutions may not be complete, and
will not apply directly to other mills, but should convey that
reaching complete closure is an evolutionary process consisting
of a great many separate steps, requiring diligence in seeking
out and identifying even the minor contributions. Figure 10
summarizes the mill effluent sources as they now exist.
Although the total flow was reduced by only 11% since 1971,
the total suspended solids and BOD loads were reduced by 63% and
83%, respectively. Most of the mill discharge is made up of
river water which receives no treatment other than coarse screen-
ing. It is used, without recirculation, as seal water on vacuum
pumps, for evaporator cooling, and for wood slab washing. The
remainder of the effluent is uncontaminated city water used for
cooling of air conditioners, air compressors, and oil coolers.
Present losses are characterized either as routine during
stable operation or as upset losses. The largest routine loss is
the mist of process water carried into the vacuum pump seal
water. The upset losses arise during nonroutine conditions, such
19
-------
to
o
SCREW
SUMP PRESSES
REFINERS
DDDQ
SUMP
HYDRASIEVE
DO Din
REFINERS
|
\5)
±
KEY -REFER TO TABLES 5 and 6
x
NUMBER - POTENTIAL LOSS
LETTER-DILUTION
RIVER
WATER
SCREEN
MAIN EFFLUENT LINE
Figure 8. Pulp mill floor collection system—1971.
-------
N)
SCREW
SUMP PRESSES
TO
SANITARY
SEWER
REFINERS /
KEY -REFER TO TABLES 5 and 61
NUMBER - POTENTIAL LOSS
LETTER-DILUTION
RIVER
WATER
SCREEN
MAIN EFFLUENT LINE
Figure 9. Pulp mill floor collection system—1975.
-------
TABLE 6. INVENTORY OF POTENTIAL PROBLEMS AS OF 1971
AND DISPOSITION AS OF 1975—PULP MILL DISCHARGE
(In Order of Priority)
(Code for Pulp Mill Discharge, See Figures 8 and 9)
1971 - Problem
1975 - Status
Number 1 Priority:
Urgent—elimination necessary to meet 1973
limits
1) Periodic overflow of press
and digester process water
wash tank
Amount: 100-200 gpm
Frequency: 5 min/wk
2) Weak liquor seepage in
screen area from splashes
and leaks; concentrated
liquor
Amount: 1/4 gpm,
continuous
3) Screw press tank overflow
from plugging with fiber
and chips
Amount: 25 gpm
Frequency: 5 min,
twice/day
4) Overflow from pulp mill
tank for fiber recovered
from waste liquor
Amount: 150 gal
Frequency: each mill
shutdown,
2-3 times/wk
Control of makeup to tank
improved; overflow to closed
pulp mill liquor sump
Directed to closed pulp mill
liquor sump
Directed to closed pulp mill
liquor sump
Directed to closed pulp mill
liquor sump
Number 2 Priority:
Needed to ensure stable operation and better
closure
5) Dump of Hydrasieve effluent
tank to clean out sand
Amount: 100 gal/wk
6) Stock pump pit seepage
Amount: small but con-
tinuous, stock and
process water
Directed to closed pulp mill
liquor sump; installed contin-
uous sand removal system
Directed to closed pulp mill
liquor sump
22
-------
TABLE 7. INVENTORY OF POTENTIAL PROBLEMS AS OF 1971
AND DISPOSITION AS OF 1975—RECOVERY PLANT AREA
(In Order of Priority)
•**
(Not Diagramed)
1971 - Problem 1975 - Status
Number 2 Priority: Needed to ensure stable operation and better closure
1) Dumping of Merco motor driven filter Installed small collection tank and
sludge pump to divert to unclarified process
Amount: 2 gal 40% liquor water surge tank
Frequency: once/hr
2) Flushing of gorator pump for repairs Flushes to recovery plant collection
Amount: 100 gal tank
Frequency: once/120 days
3) Evaporator cleaning with water and Pump directly to unclarified process
NaOH water surge tank or weak liquor tank
Amount: 300-500 gal
Frequency: once/wk
Number 3 Priority: May be postponed until after 1973—more information is required
4) Acid used to remove scale from Pumped to recovery plant collection
liquor feeding lines tank '
Amount: 400 gal
Frequency: once/2 mo
5) Caustic wash for evaporator demister Pumped to recovery plant collection
pads , tank
Amount: 250 gal
Frequency: once/wk
6) Acid condensate from evaporators Unresolved; evaporator usage is about
Amount: 20 gpm 1/3 of what it was in 1971—501
Frequency: 150-200# BOD/day BOD/day
TABLE 8. INVENTORY OF POTENTIAL PROBLEMS AS OF 1971
AND DISPOSITION AS OF 1975—WOOD ROOM DISCHARGE
(In Order of Priority)
(Not Diagramed)
1971 - Problem 1975 - Status
Number 3 Priority: May be postponed until after 1973—more information is required
1) Losses of fiber, bark, and small Sweco screen and sump added to remove
chips debris
23
-------
TABLE 9. INVENTORY OF POTENTIAL PROBLEMS AS OF 1971
AND DISPOSITION AS OF 1975—TANK STORAGE AREA
(In Order of Priority)
(Not Diagramed)
1971 - Problem
1975 - Status
Number 1 Priority:
1)
Urgent—elimination necessary to meet 1973 limits
Seal pit and paper machine process
water collection pit overflow; shut-
down and startup surges or pump
failure
Amount: 3000 gal
Frequency: once/day
Combined pits; installed larger pumps;
installed spare pump; installed high
level alarm
2)
All losses directed to process water
pit by new piping
3)
Losses from paper machine gutters
and floor drains from normal seepage
and washup
Amount: 2-5 gal/min
Periodic entrainment of process water
in Centri-Cleaner vacuum pump seal
water
Amount: 5 gal/min
Frequency: 30 min/wk
Number 2 Priority: Needed to ensure stable operation and better closure
None in this category
Number 3 Priority: May be postponed until after 1973—more information is required
Installed conductivity alarm and manual
vacuum breaker; modified vacuum pump
seal water system
4)
5)
6)
Overflow of Centri-Cleaner reject
tank; shutdowns and failure of pump
Amount: 50 gal
Frequency: each shutdown,
2-3 times/wk
Couch and wet press vacuum pump
carryover
Amount: 3-5 gal process water/min
Discharges to enclosed sewer and is
recycled with sump pump
Unresolved
Now dumps to closed sewer and is
recycled by sump pump
Trap dumping Magna Cleaner finishing
refiner
Amount: 20 gal process water
Frequency: once each.8 hr
7) Trap dumping Magna Cleaner secondary
Centri-Cleaner system
Amount: 20 gal process water
Frequency: once each hr
8) Overflow from overfilling paper
machine stock storage chests
Amount: 600-1000 gal process water
and fiber
Frequency: once/mo
Number 4 Priority: Insignificant or poorly defined
9) Paper machine area pumps and side Most leaks are located in area of
agitator seal leaks enclosed sewer and are recycled with
Amount: small but continuous sump pump
10) Seepage from area under paper Unresolved
machine wet presses
Amount: small but continuous,
some fiber loss
Now dumps to closed sewer and is
recycled by sump pump
Now is in closed sewer area; stock is
manually washed back to sump pump for
recycle
24
-------
RIVER
WATER
RIVER
WATER
SCREEN
to
WET PRESS
VACUUM PUMP
SEAL WATER
REVERSE
OSMOSIS
PERMEATE
CITY WATER
FOR COOLING
AIR CONDITIONERS
OIL COOLERS
AIR COMPRESSORS
COUCH ROLL
VACUUM PUMP
SEAL WATER
WOOD ROOM
SLAB WASH
FILTRATE
BOILER
FLYASH
POND
FLAT BOX
VACUUM PUMP
SEAL WATER
PAPER MILL DISCHARGE
fcl CONDUCTIVITY
EVAPORATOR
CONDENSER
WATER
ACID
CONDENSATE
COOLING
WATER
MAIN EFFLUENT LINE
Figure 10. Mill effluent sources.
-------
as shutdowns for machine clothing changes or failures in controls,
equipment, and/or human judgment.
MILL SYSTEM WATER RECYCLING
The fresh water input to the process system was reduced from
3.21 cubic meters per minute (850 gpm) a few years ago to 0.491
cubic meters per minute (130 gpm) today. The 0.491 cubic meters
per minute of water that does enter the process water system is
either carried with the raw chips or is from certain small four-
drinier showers and many pump shaft seals.
All the excess water produced at the paper machine is
directed to the main collecting point, which is the process water
collection pit (Figure 11). This pit also receives the water
from the paper machine floor washup drains and the floor gutters
along the fourdrinier and wet presses. The collected water from
the process water collection pit, containing about 2.16 grams of
suspended solids per liter (18 lb/1000 gal), is pumped to our
unclarified process water surge tank and is then supplied for
direct reuse to the areas insensitive to suspended solids. These
uses are waste paper repulping, machine broke repulping, all
stock dilution, and cooking liquor makeup.
The excess unclarified process water not needed for the
above operations is metered, based on unclarified process water
surge tank level, into a large stream being used for corrugated
waste repulping. This dilute stock is processed over a vacuum-
less thickener. Due to the gentle driving force removing water
on the thickener, the thick mat containing long kraft fibers acts
as an excellent filter medium for the retention of fiber, and the
filtrate contains only 0.66 grams of suspended solids per liter
(5.5 lb/1000 gal) with very little long fiber. The excess por-
tion of this clarified filtrate is then pumped at a constant
nozzle pressure of 2.81 kilograms per square centimeter (40 psi)
over a four-bank, 150-micron DSM slotted screen for removal of
any remaining long fiber down to 0.0096 grams per liter (0.08 lb/
1000 gal). The treated water will still contain from 0.24 to
0.60 grams per liter (2 to 5 lb/1000 gal) of extremely fine sus-
pended solids. The clarified water is stored in our clarified
process water tank and is reused on all paper machine showers
except cut squirts, the breast roll shower, and the seal water
showers on the suction roll boxes.
Process water showers were selected after extensive testing
of a variety of designs. The Bird Aqua-Purge shower was selected
for its simplicity and absence of parts subject to deterioration.
A key element in the successful operation of these process water
showers is the completely closed design of tanks and other ele-
ments following the DSM screen cleaning step. This assures the
absence of debris that might otherwise cause plugging.
26
-------
CLARIFIED
PROCESS
WATER TANK
to
FIBER TO PROCESS
FIBER
THICKENER
FIBER TO
PROCESS
DSM
FILTRATE
TANK
PAPER
MACHINE
SHOWERS
PROCESS WATER
COLLECTION PIT
VACUUM BOX
SEAL PIT
REVERSE
OSMOSIS
PLANT
T>
w
CONCENTRATE
TO PULP MILL WASH
DILUTION TANK
PERMEATE TO
EAST PAPER MILL DISCHARGE
(fift HIGH LEVEL ALARM
^F> PUMP FAILURE
k
.M*
M £
SUP
5CREEN
PLY
RECYCLED
FIBER
REPULPER
UNCLARIFIED
PROCESS WATER
SURGE TANK
HEAT
RECOVERY
SCRUBBER
REFINING
WATER SUPPLY
TANK
REFINERS
TO
BROKE
PULPER
DILUTION
TO PROCESS
UNCLARIFIED
PROCESS WATER
STORAGE TANK
Figure 11. Process water reuse system.
-------
A third very large tank—946.2 cubic meters (250,000 gal)—
serves as long-term storage. This tank, designated as the un-
clarified process water storage tank, is held at a constant level
by water flowing in a loop from the unclarified process water
surge tank and back to the same tank at equal rates.
Two precautions are necessary in a closed water system. The
first is to keep the process water hot; the second, to have no
dead storage areas where cooling can take place and sludge can
form. An area that is cool and/or dead will promote aerobic or
anaerobic bacteria growth.
If in the course of operating an excess of water occurs
either in the unclarified process water surge tank or clarified
process water tank, the flow in the loop from the unclarified
process water storage tank is lessened or stopped until the
excess is removed from the system and stored in the unclarified
process water storage tank. Then the normal equal flow in and
out of this tank will resume. Likewise, a deficiency in either
the clarified process water tank or unclarified process water
surge tank will cause an increase in the returning flow rate from
the unclarified process water storage tank until the need is
satisfied.
Although there has been a substantial increase in wire pit
suspended solids (from 0.2% to 0.3% as a result of closure),
there has not been a loss of machine speed. Higher process water
temperature—71°C (160°F)—and higher average sheet moisture
(computer control) have undoubtedly been compensating factors.
Furthermore, the increase in water extractables in the final
sheet—increased from 20 to 41 kilograms per metric ton (40 to
180 Ib/ton) due to closure—has been accompanied by a reduction
in fiber content at the same basis weight and, presumably, a
corresponding reduction in drainage resistance. No stiffness-
related strength loss is associated with this change in compo-
sition.
INTERNAL WATER MONITORING AND ALARM SYSTEM
An important aspect of running a tightly closed mill is
knowing when and where process problems occur, as system upsets
and resulting spills remain a potential source of excessive dis-
charge from a normally tight system. A good monitoring system
is necessary to know as quickly as possible the location and de-
gree of BOD and suspended solids losses.
The monitoring system put into use at Green Bay Packaging
contains generally two types of sensors: storage level alarms
and effluent concentration measurement by conductivity.
The monitoring and alarm system provides the following
information at a central location:
28
-------
(1) A continuous record of the total mill effluent
dissolved solids level.
(2) A continuous monitoring of potentially high pollutant
discharges at selected locations throughout the mill.
Mill personnel are informed by an annunciating alarm
panel of the upset conditions.
(3) A continuous record of each upset occurrence and its
duration; also, "on time" event and duration of
standby equipment, such as backup pumps.
(4) A continuous record of the dissolved solids concen-
tration measured by each of the conductivity sensors.
In addition, this system is flexible to revision and is a
source of dependable information which can be relied upon for
overall supervision of a closed mill water balance.
It is worth noting that much work was done to determine the
relationship between conductivity and dissolved solids concen-
tration of the mill process water. This was necessary as con-
ductivity measurement was to be the tool for the system monitor-
ing. Once conductivity was selected as the means of system
monitoring, the number and locations of the sensors were deter-
mined .
Four locations have been selected for the conductivity
sensors. The sensors are presented in sequential order as they
occur in the mill effluent system. The first sensor is on the
vacuum pump discharge from the centrifugal cleaners. Although
this discharge is normally fresh water, the vacuum seal can
break, allowing process water with 5% dissolved solids to be dis-
charged to the main clear water sewer in the mill. The next
conductivity sensor is in the trench receiving the paper machine
vacuum pump discharge. This flow is primarily river water; how-
ever, the main sewer trench also directs its flow past this
sensor so that upsets occurring in the mill basement can be de-
tected (see Figure 12). Another sensor is in the recovery plant
discharge leading to the main effluent line. Ordinarily, this
stream consists of cooling water from the evaporator condensers;
however, an occasional strong liquor spill in the recovery plant
can reach this stream (see Figure 10, page 25). Finally (on the
same diagram), a conductivity sensor is at the flume. Its pur-
pose is to monitor total mill effluent concentration. This
sensor provides information on the relative strength of the
effluent reaching the river.
In addition to relative strengths of streams within the mill
system, liquid levels are monitored on five tanks and two sumps.
High level alarms using conductance probes alert mill personnel
of impending problems prior to overflow. Enough advance notice
29
-------
REPULPER
SUMP
TO REUSE
SUMP
WEST MACHINE
ROOM DISCHARGE
(DISCONTINUED)
HIGH LEVEL ALARM
F> PUMP FAILURE
(T> CONDUCTIVITY
PROCESS WATER
COLLECTION PIT
MACHINE CHESTS
ooo
VACUUM PUMPS
MAIN EFFLUENT LINE
EAST MACHINE
ROOM DISCHARGE
Figure 12. Paper machine effluent system—alarm points,
-------
is given to allow for corrective action to be taken; thus, if the
spill can not be averted, its impact will certainly be dimin-
ished. The seven items monitored are:
(1) Chemical mixing tank.
(2) Clarified process water tank.
(3) Unclarified process water surge tank.
(4) Refining water supply tank.
(5) Pulp mill wash dilution tank.
(6) Process water collection pit.
(7) Repulper sump.
Two pumps are also included in this monitoring system. The
first is a pump supplying process water to the DSM screen for
clarification. Should this pump fail, an alarm will be triggered
by a pressure switch indicating low discharge pressure.
The second is the backup pump for the main sewer effluent
pump. This pump is monitored for "on-time" only, an indication
of primary pump failure or excessive flow to the flume.
Figure 11 (page 27) and Figure 2 (page 8 ) indicate the lo-
cation of alarms and sensors schematically.
As with any complex mechanical-electrical system, a period
of operational development must be endured. As an example, poly-
shrink tubing and time delays had to be installed on the tank
high level probes to keep them from reporting falsely high read-
ings caused by splashing.
Although the monitoring system has not, and will not, elimi-
nate upsets, it is safe to say that the adverse impact of many
potentially severe upsets has been either averted or greatly
minimized by this system.
REVERSE OSMOSIS SYSTEM
By the spring of 1973, the revision of sewers and surge con-
trols had stabilized the mill process water system to the degree
that the losses caused by variable water inputs taxing the volume
limit of the system became evident. Over a period of months of
analyzing loss data, it was determined that during these times
when excess volume was entering the system the normal rate of
buildup was from one to a maximum of 56.7 liters per minute
(15 gpm). It was subsequently decided that, with the surge
storage available for sudden shifts of inventory, a reverse
31
-------
osmosis plant capable of 75.7 liters per minute (20 gpm) of
product water would be suitable to keep the long-term system
volume under control. The plant feed flow of 340.6 liters per
minute (90 gpm) of clarified process water was then set by the
Fluid Sciences Division of Universal Oil Products Company, based
on striking an optimum between feed pumping power requirements
and limiting solids increase through the plant, because the
higher the solids the less driving force for water production.
Plant Feed Characteristics
The feed stream to the reverse osmosis plant is clarified
process water taken from the line supplying the paper machine
showers. Chemically, it is a solution of 4-6% dissolved solids
containing wood extractives and sodium lignosulfonates character-
istic of high yield NSSC hardwood pulping. The stream also con-
tains an average of 300 ppm of suspended solids, which are by
nature a mixture of semicolloidal solids and small fiber debris.
The feed is supplied to the reverse osmosis plant from the mill
process at a temperature of 60-65°C (140-150°F), a pressure of
7.03 kilograms per square centimeter (100 psi), and a pH ranging
from 5.6 to 6.0.
Reverse Osmosis Theory
Osmosis is defined as the tendency of a fluid (water) to
pass through a semipermeable membrane (cellulose acetate) into a
solution of higher concentration, so as to equalize concentra-
tions on both sides of the membrane. If a pressure is exerted
on the more concentrated solution side, the flow will be de-
creased. As the pressure on the concentrated solution is in-
creased, a point will be reached at which the osmotic flow has
stopped entirely. This pressure is called the osmotic pressure
of the solution and is a property of the solution. The osmotic
pressure is a function of not only the type of solution but also
the soluble solids concentration and temperature. For example,
our process water has an osmotic pressure at 35°C (95°F) of about
12.3 kilograms per square centimeter (175 psi) at 5% soluble
solids and 24.6 kilograms per square centimeter (350 psi) at 10%,
whereas a sodium chloride solution at the same temperature has an
osmotic pressure of 42.4 kilograms per square centimeter (603
psi) at 5% and 94.2 kilograms per square centimeter (1340 psi) at
10% soluble solids.
Increasing the pressure beyond the osmotic pressure at a
given solution condition of temperature and percent solute causes
a reversal in osmotic flow of the solvent and is the basis of
the reverse osmosis process.
The peculiarities of the osmotic passage of a solvent
through a membrane is the most important difference between the
filtration and reverse osmosis processes. Osmosis depends on the
32
-------
selective property of a membrane, that is, solvent can pass
through the membrane, while one or more of the soluble components
can not do so. Therefore, in the reverse osmosis process, re-
moval of the solvent increases the concentration; and, therefore,
the osmotic pressure of the concentrating solution increases with
a subsequent decrease of the flow of the solvent under a fixed
pressure.
In a plant where the goal is to remove water and not to gain
concentration, it can easily be seen why a maximum plant feed
flow is necessary to minimize this concentration effect. This
also is the basis for a lower limit on velocity passing the mem-
brane surface so that a concentration gradient is not created by
laminar flow. Velocity effects will be discussed further under
fouling problems.
Reverse Osmosis Plant Interface
The role of the reverse osmosis plant in the general mill
scheme has been previously described. Figure 13 illustrates the
interface of this operation to other mill processes. As the
temperature of the module membrane is limited to 40°C (105°F),
the plant feed is cooled to 38°C (100°F) in a plate-style heat
exchanger; boiler makeup water is used as the cooling medium.
There are two product streams from the reverse osmosis plant
permeate (product water) and a concentrate. Initially, the
permeate was to substitute for a portion of the city water used
for pump and agitator seals. In order to avoid corrosion, this
was not done as all small pipelines in this system are steel or
copper; and although low in BOD, the permeate has a pH of 4.5
due to a trace of acetic acid.
The concentrate, or the second stream, can be routed back,
when necessary, to the process water collection pit; but it
normally flows to the pulp mill wash dilution tank to be used in
the countercurrent liquor washing operation. This use helps
increase final liquor solids and promotes minimum evaporator
usage.
Reverse Osmosis Hardware
The equipment received from Universal Oil Products
(Figure 14) consisted of three separate packages: the pump
assembly (right side), control center (center), and six module
racks (left side). The pump rack consists of eight Goulds multi-
stage booster centrifugal diffuser-type pumps of which two are
25 Hp 47-stage feed pumps and six are 10 Hp 23-stage recycle
pumps. Also included in the pump package are the magnetic flow
meters, valves, and sensors (Figure 15).
The control center is primarily electronic. The right panel
consists of amp meters, pump controls, and electrical switches
33
-------
CLARIFIED PROCESS WATER
TO MACHINE SHOWERS
DISCHARGE
PERMEATE
BOILER MAKEUP
(CITY WATER)
1
HEAT
EXCHANGER
REVERSE
OSMOSIS
PLANT
PULP MILL WASH
DILUTION TANK
CONCENTRATE
HEATED
MAKEUP
BOILER
WATER
PROCESS WATER
PIT
Figure 13. Reverse osmosis plant interface.
-------
.
Figure 14. Reverse osmosis plant.
-------
Figure 15. Pump assembly rack
36
-------
for starting and stopping the reverse osmosis unit. The left
panel has the alarm annunciator panel; power, flow, and opera-
tional hour totalizers; adjustable timers; recording charts; and
necessary controllers.
The 288 modules are incorporated equally on the six racks
with 16 rows on each rack and three modules in each row. Each
module contains 18 porous tubes which are fiberglass, filament
wound, and resin bonded. The membrane, formed from cellulose
acetate, is on the inner surface of the fiberglass tubes and pro-
vides 1.55 square meters (16.7 ft2) of membrane surface per
18-tube module. These tubes, approximately 2.23 meters long
(88 in), are contained within a hexagonal shroud and connected
internally in series by means of two molded heads.
/
Feed is piped into either of the two ports at the high pres-
sure end; and after nine complete round trips (18 tube lengths),
it exits at the same end as it entered. The other port in the
feed end becomes the concentrate exit (near the wall, Figure 16).
Water molecules permeate outward from the 18 pressurized tubes
through the semipermeable membrane and the porous tube wall into
the shroud. Permeate is collected inside the shroud and exits
from the outlets at the opposite end under low pressure.
Reverse Osmosis Process Plant Flows
The clarified process water feed flow, under conditions of
full plant operation (Figure 17), is 340.6 liters per minute
(90 gpm). The two feed pumps pressurize this literage to 31.6
kilograms per square centimeter (450 psi) before it combines with
the recycled concentrate; the combined flows—946 liters per
minute (250 gpm)—are boosted by the six recycle pumps to 42.2
kilograms per square centimeter (600 psi). This module feed is
supplied to the six identical module rack assemblies arranged in
parallel. After passage through the 96 module rows, the indi-
vidual rack concentrate flows are recombined in the return
header; a portion, the concentrate discharge—255 liters per
minute (67.5 gpm)—is returned to the pulp mill, and the remain-
der, still under pressure, is recombined with fresh feed to main-
tain the high velocities necessary to minimize fouling.
The permeate is collected in small tanks above the six
module racks before combining as the total permeate flow. The
tanks are necessary to return permeate to the modules during the
pause or rest cycle, during which osmotic backflow removes accu-
mulated material (fiber debris) from the membrane surface. The
modules are sloped to ensure that the individual tubes are
surrounded by permeate (Figure 16).
37
-------
Figure 16. Module rack,
38
-------
CLARIFIED
PROCESS
WATER
FEED
vo
PERMEATE
RESERVOIR
CONCENTRATE
RETURN
PERMEATE
FEED PUMPS
DISCHARGE
CONCENTRATE
RECYCLED
CONCENTRATE
RECYCLE
PUMPS
MODULE
FEED
MODULE RACK
TYPICAL I OF 6
MODULE ROW
TYPICAL
I OF 16
PER RACK
Figure 17. Reverse osmosis flow diagram.
-------
Reverse Osmosis Plant Controls
The control and monitoring instruments for the plant are
relatively straightforward (Figure 18). The feed flow and total
module feed are measured, controlled, and totalized. The per-
meate is measured and totalized. Pressure on the modules is
maintained by controlling concentrate flow. Other necessary
variables are measured and recorded—electrical power consump-
tion, pH, temperature, conductivity of the permeate, etc. All
the indicating and recording devices were provided to ensure
ample data for reporting the performance of this demonstration
plant. Also, adjustable timers controlling the pause and
operating cycle times are included.
40
-------
CLARIFIED
PROCESS WATER
FEED
PERMEATE
PH SENSOR
TEMPERATURE SENSOR
SOLUBLE SOLIDS ANALYZER
MAGMETER
CONDUCTIVITY
SENSOR
TURBINE
METER
|_FLOW _CO_NTRO_L_
SIGNAL
FEED
PUMPS (2)
&4-^
> ^
MAGMETER
RECYCLE
PUMPS (6)
T
|_FLQW __ CONTROL^ J
SIGNAL
MODULES
(__ PRESSURE
CONTROL
CONCENTRATE
Figure 18. Reverse osmosis plant control.
-------
SECTION 5
OPERATING EXPERIENCE
IN-PLANT ENVIRONMENTAL EFFECTS OF MILL WATER SYSTEM CLOSURE
From 1968 to 1973, the total dissolved solids in the mill
process water rose from 0.7% to 5.2%. This was a consequence of
intensive process water reuse.
Several effects from the increased dissolved solids were
noticed. First, the buildup of dried liquor solids on paper ma-
chine surfaces increased. This was particularly true on the
underside of the machine above the wire pit. A first approach
to correcting this problem was to install under-machine, side
entry, atomizing steam showers to increase the humidity and thus
prevent the drying of soluble solids. Because of adverse mist
outside the machine, this eventually gave way to installing vinyl
side curtains on the machine and abandoning the atomizing steam
showers. The side curtains appear to avoid air circulation and
trap the necessary humidity under the fourdrinier.
By reusing process water, the system temperature also in-
creased. The rise was approximately from 60°C (140°F) to 71°C
(160°F). This temperature rise made necessary the reconstruction
of the machine room ventilating system. A canopy had to be
placed above the fourdrinier to keep condensed water vapor from
falling onto the paper machine wire. More air was needed to re-
duce the room air temperature as well as remove excess humidity.
An additional 30 Hp was required to bring in fresh air and prop-
erly distribute it, and another 30 Hp was required for a second
roof fan above the fourdrinier paper machine to expel warm, humid
air. In addition to the above changes, the machine tender aisle
ventilation system was reworked, with another 25-Hp fan capacity
added to this system.
These changes have not only improved employee comfort but
they have also reduced the in-house fallout of process water
solids. These modifications were made over a period of several
years at a cost of approximately $115,000.
42
-------
PROCESS WATER SYSTEM ADDITIVES
Slimicides
The potential for slime growth is significantly enhanced by
closing the process water loop. This is mainly due to the in-
creased nutrient concentration in the system. Bacteriological
slime, a gelatinous secretion of undetermined type bacteria,
under certain conditions can detach from vessels and pipe walls
to create runnability problems for the paper machine.
The general level of slime growth can be controlled by the
use of slimicides and by maintaining a sufficiently high system
temperature. In the early weeks of system closure, little in-
crease in slime growth was noted. However, the first time the
mill's recovery plant had a lengthy shutdown after closure, the
process water system temperature dropped due to lack of makeup
heat from the recovery plant. With the cooling of the now en-
riched process water, the bacterial growth mushroomed. The re-
sultant increase in loosened slime globules caused numerous
breaks of the wet web leaving the couch roll on the paper machine.
The immediate remedy was to return to fresh water on the paper
machine showers, followed by a thorough cleaning of the entire
process water system.
Slimicide was then used to suppress new growths of slime-
forming bacteria. The level of slimicide addition was then ad-
justed to the minimum amount while just maintaining control of
bacterial growth.
Currently, two types of slimicide are used, an organic
dibrominate and methylene-bis-thiocyanate. These materials are
alternated from drum to drum. The slimicide addition occurs
every four hours with 22.7 to 45.4 kilograms per day (50 to 100
Ib/day) being added to the DSM screen supply tank.
Samples of process water are taken at four locations: the
inlet to the headbox, the clarified process water storage tank,
the process water collection pit, and the reverse osmosis plant.
Normally, the desired level for total colony count is 100,000
or less at the headbox and 500,000 or less in the rest of the
system. Although these are desired levels, actual counts may
vary from 1000 to several million for short periods without any
visible indication of slime.
The daily process water samples are cultured in 100 x 15
millimeter Petri dishes using tryptone glucose extract agar as
the culture medium. The cultures are incubated at 37°C (100°F)
for 48 hours.
43
-------
Defoamers
The closing of the process water loop caused the temperature
to rise to a high of 71°C (160°F) at the paper machine headbox
and a low of 62.8°C (145°F) at the surge capacity storage tanks.
The defoamer used prior to the closed system was a combination of
oils (steric fats and fatty acids) and isopropyl alcohol. The
defoamers had to be heated to affect solution and were water
emulsifiable. The combination of higher temperatures and dis-
solved solids made this type of defoamer ineffective.
After an extensive investigation, a defoamer was found which
met all the necessary pH range and temperature requirements of a
closed system. Basically, the defoamer was a mineral oil carrier
with a silicone coated silica suspension. The defoamer was sur-
face active and immiscible with water. The application was on
and around the paper machine with a daily rate of usage at
approximately 13.6 to 59.1 kilograms per day (30 to 130 Ib/day),
depending on machine conditions. The system for addition of the
defoamer is completely automated and will respond according to
existing conditions.
A different defoamer was used in the secondary fiber re-
pulping system, specifically on the water dropleg on the Impco
thickener. Foam develops here due to the cascading of process
water into the standpipe. This defoamer is similar to the one
on the paper machine but has some esters added. The machine
defoamer can be used in this area also, but doesn't have the same
efficiencies. Use at this point amounts to approximately 90.9
kilograms per day (200 Ib/day).
PROCESS WATER SYSTEM ASH BUILDUP
Although ash buildup in a closed paper mill water system is
not easily observed or measured as is slime or soluble solids
buildup, it is real and can be a problem due to increased refiner
plate wear. Following the same pattern as that of soluble solids
with system close-up, ash continued to increase until 1975 when
a means was found to remove the normally recirculated ash from
the system in a concentrated, reasonably dry form. Approximately
one-third of the ash percent is normally silica, and the remain-
der is calcium oxalate.
Figure 19 is a diagram of the complete centrifugal cleaning
and reject refining system. Abrasive material is removed at two
points. The first is relatively coarse sand, gravel, and metal
removed by a MagnaCleaner just before the second stage of
centrifugal cleaners. The second is the reject stream from a
13-centimeter (5-in) centrifugal cleaner operating on the efflu-
ent from a Hydrasieve. The fine recirculated ash in this loop is
at its greatest concentration of any point in the system. After
44
-------
13 PRIMARY CLEANERS
WITH VACUUM TANK
^
TO PAPER
MACHINE
m
• •
• mm
HYDRASIEVE
cn
MAGNACLEANER
4 SECONDARY
CLEANERS WITH
STOCSAVERS
PRIMARY
REFINING
REJECTS
14 TONS
PER DAY
COARSE SAND
AND METAL
400 */ DAY
TO MACHINE
CHEST
TO REFINERS
5 INCH CLEANER
WITH STOCSAVER
FAN PUMP
INERTS 700** DAY
PRESSED CHIPS
FROM WASHING
SCREW PRESS
Figure 19. Grit removal system.
-------
being rejected from the small centrifugal cleaner, the separated
ash is dewatered in a grit chamber and removed by a lift screw
to be hauled away to landfill.
REVERSE OSMOSIS PLANT OPERATING PHILOSOPHY
The reverse osmosis plant has operated 60% of the mill
operating time during the two years since it was installed. The
excess inventory of process water required its use only about 15%
of the mill operating time. The plant was run as continuously as
possible because of the project goal to determine membrane life
and rejection efficiency over an extended period of time.
The 40% downtime was the result of a number of problems—
plant control system and piping revisions, pump maintenance,
module replacement, and plant descaling.
Reverse Osmosis Plant Performance
After a few initial revisions in the instrument control
scheme and hardware, the plant control has been very complete and
maintenance has been minimal. The automatic features of the
plant have resulted in reliable unattended operation.
The pumps have operated well to produce the intermittent
high pressures necessary for plant operation but have required
a high degree of maintenance to keep them efficient. This
aspect is more completely described in the section entitled
"Reverse Osmosis Pump and Pump Seal Problems."
The third component of the plant, the modules, constituted
the main research and development area of the project. This part
of the project was slowed considerably by problems relative to
the porous fiberglass support tubes for which details are also
described under "Reverse Osmosis Module Fiberglass Membrane
Support Tube Failure."
The module membrane used in this plant is of a medium per-
meability grade. The quality of the permeate from this membrane
has been better than specified in the listing of plant bid re-
quirements. This property, as well as the other normal operating
parameters, are listed in Table 10.
Power consumption of 82.0 kilowatt-hours per 3785 liters
(1000 gal) of permeate has been less than other alternate evapor-
ative methods of pure water removal. Comparative costs are
described in the report section "Economic Evaluation of Reverse
Osmosis vs. Evaporation."
The only disappointment has been in the area of reduced
productivity brought about by membrane aging.
46
-------
TABLE 10. TOTAL REVERSE OSMOSIS PLANT
AVERAGE OPERATING PARAMETERS
(See Appendix A for Test Procedures and Calculations)
Item
Soluble solids
Sodium
BOD 5
Color
PH
Item
Soluble solids
Sodium
BOD 5
Color
PH
Feed
(ppm)
54,400
6,500
13,900
112,500
5.8
Permeate
(ppm)
372
62
1,005
65
4.1
Concentrate
(ppm)
66,900
8,000
17,600
146,800
5.8
% Rejection
Specified Actual
98.0 99
none 99
96.0 98
99.0 99
—
.86
.86
.56
.94
-
Item
Design
Actual
Flux gallons
(per day per sq ft)
at 5% feed solids,
600 psi (41.4 bar)
module feed pressure,
and 100°F (3?OC)
6.65
5.13*
Power consumption per 1000 gallons permeate = 82.0 kwh*
*average for two years
47
-------
Figure 20 is a comparison of the flux at 5% dissolved solids
obtained on two different plant complements of modules over a
period of approximately 2000 hours of exposure. This figure is
limited to the operational time data we have available on a full
plant basis as of this date. The original modules supplied with
the plant were replaced due to a material defect, allowing the
evaluation of a second set. In both cases, the flux decreased
gradually over the 2000 hours as the modules aged. The differ-
ence observed in the flux levels shown by the two groups has not
been determined but is believed to be related to the original
membrane characteristics. All attempts to return to starting or
even design flux—11.29 liters per square meter-hr (6.65 GFD) at
5% dissolved solids—have been futile. This rate of drop in flux
does not continue indefinitely but tends to level out after
approximately 2500 hours. This can be observed by the longer
term flux exposure times on the individual module (replacement)
racks installed over the last 14 months (Figure 21). Evidence of
the long-term flux stability after the initial drop can be seen
up to the longest rack exposure time of 7000 hours. The initial
drop in flux which occurs within a 24-hour period after the mem-
brane is installed is recognized in the industry and is known to
be caused by initial compaction of the membrane. Whether the
drop after this is continued slow compaction or some other phe-
nomenon is not known at this time. There has been no deterio-
ration in membrane rejection of soluble solids or color during
this period of 7000 hours.
REVERSE OSMOSIS MODULE FIBERGLASS MEMBRANE SUPPORT TUBE FAILURE
The first six months after plant acceptance saw module
failure, due to tube structural strength loss, reaching an
untenable number, as high as 21 modules per 100 hours of exposure*
Initial failures, generally occurring close to the ends of the
tubes, were thought to be caused by questionable manufacturing
processes or lack of employee care in positioning the tubes for
the delicate tube cut-off operation. When further ruptures
occurred randomly spaced along the entire length of tubes, it was
realized that these new failures were a result of some mysterious
and unknown cause.
An analysis of tube processing and qualification data showed
no correlation of the rupture phenomenon with various manufac-
turing variables. The location of failed tubes within the
module appeared to have no correlation with failure incidence.
The one ominous feature that did correlate was the length of
exposure to the process water feed stream. The history of 4000+
hours of successful piloting with several modules was of little
benefit in the light of this new reality in which tubes were
rupturing in less than 300 hours.
Tube burst strengths had dropped during exposure from 276
bars (4000 psi), as originally manufactured, to as low as
48
-------
•c*.
vo
o
9
Q
U.
I
X
•-ORIGINAL MODULES
A-REPLACEMENT MODULES
2OO
4OO
600 800 1000 1200 I40O
OPERATING EXPOSURE, HOURS
1600
I8OO
Figure 20. Flux performance comparison.
-------
01
o
ID -*
n ,
0
m
fO 7
'
1 0 '
CO
.
3
Li- d
;
t
b
if
n A
i
10
oO°
OO 20
yv
O
3 n
O A RACK
A B RACK
O C RACK
Do
A On
n ^
RACK
o
00 3000 4000 5000 6000 70
00
OPERATING EXPOSURE, HOURS
Figure 21. Cleaned module flux data.
-------
166 bars (2400 psi) and even 124 bars (1800 psi). With all manu-
facturing processes and quality control procedures tightened to
eliminate any possible causes from the manufacturing end, it was
recognized these premature ruptures could only be caused by the
chemistry of the feed stream as it related to and affected the
mechanical integrity of the tube structure itself. Samples of
raw and aged tubes plus samples of feed and permeate liquids were
submitted to Universal Oil Products' Corporate Research Group for
analytical investigation.
Analyses of permeate and feed showed acetic acid present in
both, demonstrating the tube was being exposed to an acidic
environment continuously, albeit in low concentrations. Ele-
mental analysis of the tubes' resin system showed no leaching
of the components. Ash contents of raw and aged tubes were also
similar. Submitting the tube samples to Scanning Electron
Microscopy showed the glass fibers had been corroded and etched
severely, giving them the appearance of "swiss cheese" rather
than a long, smooth, and round filament. Xray analysis indicated
a loss of Mg and a higher ratio of calcium-to-silica content in
the aged fiber. The glass then being used contained basic glass
elements in a certain ratio. It was determined that the acetic
acid of the feed stream and the permeate altered the chemical
composition of the glass itself, allowing the acidic water to
etch the more soluble elemental' portions.
The photographs obtained with the Scanning Electron Micro-
scope are presented in Figure 22. An unexposed fiberglass tube
is shown in the top two photos at different magnifications. The
lower two frames are duplicate magnifications of a tube exposed
to acid permeate.
The conclusion was drawn that failure of the tubes at Green
Bay was due to corrosion degradation of glass fibers by the
acidic permeate. Continued exposure time caused a sudden drop of
strength from 124 bars (1800 psi) to the 41 bars (600 psi) opera-
ting pressure due to the failure of microcracks, introduced by
the leaching process, and the growth of these cracks by flexure
during startup and shutdown cycling.
Action Taken to Correct the Problem
In addition to their Corporate Research's efforts, Universal
Oil Products' Norplex Division, with experience in the production
of composite laminates using resinous glass fiber-filled tech- :
nology, Owens Corning Fiberglas' Technical Center personnel, and
a consultant, experienced in resin/glass fiber structures, were
all engaged to assist in finding a solution. The search was
begun to find a glass and resin combination that would stand up
to Green Bay's feed stream and operating conditions.
51
-------
Figure 22. Photomicrographs—new (top) and aged tubes (bottom)
(furnished through the courtesy of
Universal Oil Products, Corporate Research).
52
-------
A major problem in developing an improved tube structure
centered about finding tests that would provide accelerated data
which could quickly and reliably correlate to long-term exposure
in the commercial application. Selection was finally made of a
water boil test over predetermined time periods from two to as
long as 1200 hours. Additionally, a rapid cycle life test was
developed in which tubes were subjected to an acidic water
solution at 49°C (12QOF) while full modules were pressurized to
41 bars (600 psi), held for 60 seconds at pressure, and then de-
pressurized to atmospheric pressure for 30 seconds to complete
one cycle. This cycle was repeated, usually many thousand of
times, until failure occurred.
The ultimate test, of course, in order to check the effec-
tiveness of any improved tube structure, was in their being
placed in actual operation on the Green Bay unit. Each tube
improvement showing significant results in both static water boil
and dynamic life cycle tests was duplicated to get the required
number of tubes to assemble into a module for placement in ser-
vice. Ultimately, 17 tube improvements were placed in service
after prescreening (by the accelerated tests at Fluid Systems)
suggested their ultimate success in the field.
Eventually, one tube's resin/glass composition emerged as
superior to all others. It consisted of a more chemically inert
glass fiber and new resin combination that overcame an encoun-
tered, secondary problem of loss of old resin to new glass
adhesion. The final new resin/new glass combination provided
an adequate matrix for glass bonding, glass distribution, and
tube poro s i ty.
With on-board accelerated testing continuing to show good
results, and units at Green Bay with several hundred hours ex-
posure to the process water feed, it was decided to retube the
entire plant with the new version. The rebuild of the Green Bay
plant commenced January, 1975, and has continued through the
present date with the final 75 modules, including 26 spares,
scheduled for installation in February, 1976. A tremendous
improvement in support tube life was achieved with the new resin/
glass composition. This improved life is shown graphically in
Figure 23. The originally supplied modules are compared with the
new modules based on cumulative percent failures with exposure
hours. Unfortunately, we have only an exposure time of 6000
hours on the new tubes and cannot predict what the maximum effec-
tive module life will be. To date there have been fewer than two
dozen module failures which have been ascribed to tube ruptures,
and these few have been determined to have been caused by manu-
facturing variances rather than structural deficiencies. When a
failure is caused by a flaw, correction requires only the replace-
ment of one tube of the 18, and the module can be placed back in
service. This was not the case previously with the original
modules, as commonly all the tubes were structurally weakened by
53
-------
m
UJ
o
Q
LU
<3O —
fiO
7A
fiO
crx
2in
OU
20 —
iu —
f\
A
A
A
1 • • '
A '
An,
A
,H • •!
1 " '
A "ORIGINAL" MODULE FAILURES
• "IMPROVED" MODULE FAILURES
• 1
• I
1
1000
2000 3000 4000 5000
OPERATING EXPOSURE, HOURS
6000
7000
Figure 23. "Original" vs. "improved" module failure experience.
-------
etching when the first failure occurred. The oldest tubes of
improved structure have over 8000 hours exposure and show signs
of lasting well into the second year of operation, considerably
beyond the warranty period.
As a result of the problem experienced at Green Bay and its
ultimate solution, Fluid Systems has incorporated the new im-
proved tube as its standard product for all high pressure reverse
osmosis applications, a technical development which has already
paid bonus benefits in providing the answer to another applica-
tion where the older tube version showed shortened life in the
field.
REVERSE OSMOSIS MEMBRANE FOULING
Membrane fouling continues to be an area warranting further
research and development. For a more definitive explanation of
membrane fouling, reference is made to the work done by the
Institute of Paper Chemistry, entitled Reverse Osmosis Concen-
tration of Dilute Pulp and Paper Effluents, USEPA, 12040 EEL,
02/72. (2)
Observed Nature of Membrane Fouling
Membrane fouling by mill process water takes place by
seemingly multiple, complicated, and not well understood mecha-
nisms; however, the fouling observed in the Havens-type tubular
module can be classified into two distinct types.
The first type of fouling is caused by fiber debris and
wood fines. Although the process water passing the membrane con-
tains only a small amount of filterable fines, after prolonged
exposure, debris collecting on the membrane will significantly
reduce the membrane permeability. This debris is always present.
It is that passing the two-step filtering process to make the
water suitable for use on paper machine showers.
The second type of observed fouling is scale buildup. After
an operating exposure of 750 to 1000 hours, a tough scale can
form on the membrane surface, reducing the membrane flux from
1.7 to 5.1 liters per square meter-hr (1 to 3 gal/ft2-day).
Analysis indicates that the process water system is normally
saturated with respect to calcium; therefore, the propensity to
form insoluble calcium salts is extremely great. Calcium oxalate
has been identified as the major foulant material in module scale
examined by Scanning Electron Microscopy. The oxalate anion is
a naturally occurring constituent found in varying amounts in
hardwood bark. Red oak contains a relatively high concentration
of oxalate. Red oak chips are regularly received as a component
of the mixed hardwood wood supply to the mill. Calcium oxalate
will salt out of solution upon cooling to form an extremely
55
-------
insoluble scale. As the process water feed to the reverse
osmosis plant must be cooled from 65°C (150°F) to 38°C (100°F),
the reaction takes place quite readily on heat exchanger, pump,
and piping, as well as membrane surfaces.
In addition to the observed long-term fouling, data analysis
indicates short-term effects can be experienced as well. As an
example, the startup of the fluidized bed reactor requires bed
material (granular Na2S04 and Na2COs) to be transferred into the
reactor. Dust carryover of fines is scrubbed out in the
reactor's Venturi evaporator/scrubber. Normally, this dust
carryover is not a problem as liquor is used as the scrubbing
agent; but during startup, process water and not liquor is pumped
to the Venturi evaporator/scrubber. This process water, too
dilute for combustion, is brought back to the mill, saturated
with soluble sodium sulfate and carbonate. The formation and
deposition of relatively insoluble sulfate, carbonate, and
bicarbonate salts accounts for the blinding of the membrane in a
matter of hours. To avoid this rapid flux drop, the precaution
has been taken not to run the reverse osmosis plant during or
immediately following a reactor startup. The observed coinci-
dence of reactor startup and drastic membrane flux loss has
ceased since this change was made.
Countermeasures Currently in Use to Remove Fouling
It was determined in the pilot plant work that an osmotic
backflush would be necessary for 12 minutes in every 120 minutes
of operation. This osmotic backflush allows for a lifting of
deposited material that has not become physiochemically attached
to the membrane surface. For further information regarding the
development of this technique, consult the previously mentioned
reference and Recycle of Papermill Waste Waters and Application
of Reverse Osmosis, USEPA, 12040 FUB, 01/72.(3)
In addition to maintaining a turbulent flow velocity of
1.2 to 1.5 meters per second (4 to 5 ft/sec) and using the
osmotic backflush, it was determined that a daily high velocity
flush with process feed water was necessary to further prevent
the deposition of fiber debris. This technique involves flushing
one-half the modules at a time for 10 minutes per day using the
full pumping capacity of the plant, directing the discharge back
to the process water recycle system.
Although these practices are valuable in preventing the
buildup of inordinate amounts of fiber debris, they are not
effective in eliminating the fouling due to insoluble precipi-
tated salts. It became evident that a cleaning procedure would
be necessary; thus, recommendations were sought from Universal
Oil Products on chemical cleaning techniques that would remove
the foulants without harming the membrane or its support tubes.
56
-------
The cleaning procedure recommended involves flushing all of
the process water from the modules as a first step. Next, a 2%
solution of E.D.T.A. Na4 prepared with tap water and pH adjusted
to 7.0 with HCl is recycled through the modules for two to four
hours at a flow rate of at least 13.2 liters per minute (3.5 gpm)
per module row. A previously recommended 45-hour quiescent soak
period was eventually deleted as no improvement in flux was ob-
served after the initial 4-hour dynamic treatment. After the
E.D.T.A. treatment, the plant is purged with tap water and flux
checks are made prior to returning the plant to process water.
All discharges of flushing and cleaning waters are returned to
the process water recycle system.
The flux rate of a given group of modules can be improved
from 1.7 to 5.1 liters per square meter-hr (1 to 3 gal/ft2-day),
depending on the condition of the modules prior to the time of
cleaning. In general, this cleaning will return the plant flux
to a level of 8.5 to 10.2 liters per square meter-hr (5 to 6
gal/ft2-day), close to the design capacity.
Because of the cost of E.D.T.A. treatment, other types of
cleaning agents have been investigated. One that appears to
show promise is a combination of sodium tripolyphosphate and
enzymatic cleaning agents. It is commercially available as BIZ,
a home laundry product. The procedure used to date requires
purging the plant of process water, introducing a 1% solution of
BIZ, pH adjusted to 7.0 with HCl, and allowing this solution to
dynamically wash the modules for three to four hours. The plant
is then allowed to soak for 12 to 18 hours prior to being flushed
and put back on line. Limited experience with BIZ indicates the
same type of flux recovery as provided by E.D.T.A., but at
considerably less cost.
Areas of Anomalous Behavior Requiring Further Study (
Four years of pilot study and two and one-half years of
full-scale operation have not provided all of the information
needed to understand some of the anomalous behavior observed in
membrane fouling.
Knowing the many chemical and physical constituents present
in the mill's process water, it is not possible at this time to
explain why on some occasions a flux drop of 3.4 to 5.1 liters
per square meter-hr (2 to 3 gal/ft2-day) will recover without
having changed any operating factors and without cleaning. It
appears that, on some occasions, factors causing the fouling are
countered by other unidentified changes occurring in the process
water system.
Six modules have been set apart in a pilot plant receiving
the same feed, at feed conditions, as the main plant. However,
57
-------
the means are available on the pilot unit to test various
techniques of foulant inhibitors and cleaning agents.
The work to be done in the ensuing year will be aimed at
identifying the nature and mechanisms of fouling and developing
a better understanding of mill operation as it affects reverse
osmosis plant function.
REVERSE OSMOSIS PUMP AND PUMP SEAL PROBLEMS
The pump assembly for the reverse osmosis plant came as a
completely piped unit consisting of six pumps, valving, and
interconnecting piping. There are two feed pumps and four
recycle pumps of multistage centrifugal design. The feed pumps
are Goulds, Type 3933, with 47 stages; and the recycle pumps are
Goulds, Type 3933, with 23 stages.
The first eight months of operation for the pumps seemed to
indicate acceptable performance. Then, in the ninth month, a
critical problem appeared which has been partially resolved at
this time. A dramatic drop in pump amperage indicated a loss of
capacity. Upon disassembling the affected pumps, it was found
that a severe degradation of the plastic impellers had occurred.
The impellers are constructed in two pieces by injection molding
and then assembled using a solvent bond. The nature of the
damage to the impellers indicated that the components had separa-
ted and then had begun to disintegrate. Because of the multi-
staging, the damage was extensive as broken pieces of impeller
were forced through the successive units causing further damage.
Chemical analysis of the process water indicated the presence of
constituents which could adversely affect the life of the bond
between the two impeller components.
To overcome the problem of impeller disintegration, an
impeller of more substantial construction was substituted for the
original type. This has significantly reduced the disintegra-
tion; however, the bond separation continues to be a problem at
the time of this writing.
To further complicate the resolution of the pump problems,
it became evident that the main support bearings for the impeller
drive shaft were deteriorating with use. The feed pumps have
five such bearings and the recycle pumps three each. The problem
appeared to be caused by the swelling of the Buna-N bearing. The
impeller shaft has a stainless steel sleeve which was abraded
away by the swollen bearing. Once there was sufficient reduction
of the sleeve cross section, the shaft was no longer firmly held
in place. As the drive shaft began to travel eccentrically, this
caused the plastic impellers to contact the sides of the stain-
less steel bowls. The result was debonding of the impeller com-
ponents and some physical wear to the bowls themselves.
58
-------
In attempting to overcome the bearing problem, new ceramic-
coated stainless sleeves were used to replace the original sleeve
type. This did not completely eliminate the problem; therefore,
the remaining corrective measure appears to be replacing the
Buna-N bearing material. A composition bearing has been sub-
stituted for the Buna-N type on one of the pumps. The bearing
material is a bronze impregnated fluorocarbon polymer that has a
steel backing. The bearings have functioned for three months in
which time the pump has shown no amp loss or increase in vibra-
tion. An actual operating exposure of a year or more will be
needed to demonstrate any superiority over the original bearing
type.
The motor end pump rotary mechanical seals have also re-
quired attention. Leakage around these seals causes considerable
housekeeping duty as the spray of process water from the leaking
pumps can be extensive. A new supplier's seal is being substi-
tuted for the original with some success to date. In addition,
the shaft sleeve at this location on the drive shaft has been
affixed with cementing agents to reduce leakage between the shaft
and the seal shaft sleeve.
For two years, total plant maintenance cost for parts and
labor (excluding modules) was $22,330.26. Of this cost,
$15,877.00 was for pump repairs. It is hoped the pump repair
costs will be reduced in future years with changes that are
presently being evaluated.
59
-------
SECTION 6
RESULTS AND DISCUSSION
RESUME OF MILL OPERATING EXPERIENCE
The use of intensive recycling has created some new require-
ments and costs, such as replacement of traditional materials
with more expensive resistant materials; but in many ways, re-
cycling has resulted in cost reductions, as well. Production
cost comparisons (see pages 61, 62, and 64) illustrate some of
these cost reductions. This experience contrasts with increases
in operating costs required by end of the pipe biological treat-
ment methods and thus illustrates the cost advantages of in-plant
reuse, where possible.
In obtaining the present point of equilibrium, the process
water soluble solids content rose from 2.5% in 1970 to as high as
5.4% in 1974 (Figure 24). The peak concentration during this
six-year period was 7.5%. The system temperature also increased
from the 54-60°C range (130-140°F) to the 65-71°C range (150-
160°F). Both factors affect the corrosive properties of the
process water. Also, as has been previously described under
"Process Water System Ash Buildup," the circulated ash followed
the same pattern as the soluble solids until a method was found
to remove it in a concentrated form from the system. How much of
the increase in cost of refiner plates (Figure 24) and wires
(Figure 25) can be attributed to wear as a result of ash circula-
tion is open to conjecture, but it is certainly a major factor.
The other factors of increased temperature and solubles were
unchanged in 1975 when costs for both were decreased as ash
decreased. In an effort to make a valid comparison, all costs
related in these graphs are adjusted for inflation at 9% per
year. It is still difficult to make good comparisons for wire
and felt cost because of the changes in materials made during
these years. During the base year, 1971, bronze wires were used;
but from 1972 through 1975, most wires were stainless steel.
A normal stainless wire will run on the machine about 40 to
50 days, as compared to seven to ten days for a bronze wire. We
have tried a few plastic wires with mixed success initially but
with good results so far in 1976. Paper machine wet felt costs
60
-------
PROCESS WATER SOLUBLE
SOLIDS CONCENTRATION
1970
SECONDARY REJECT SYSTEM -ASH
CONTENT, ACTUAL PERCENT
1971 •H^KH 7. f%
1972
1973
1974
1975
1971
REFINER PLATE COST
RELATED TO 1971
Figure 24. Relationship of percent dissolved solids,
secondary reject system percent ash, and refiner plate
cost per short ton product, as percent of 1971 costs.
61
-------
PAPER MACHINE PRODUCTIVITY
ACTUAL PERCENT
1970
1971
WIRES-RELATED TO 1971
WET FELTS - RELATED TO 1971
i97i HiiHBHL I0° %
l^^mi&vz&M.ajaiiajijaauuanimaBm'm
1972
1975
Figure 25. Paper machine productivity and clothing costs
per short ton product, as percent of 1971 costs.
62
-------
are not any easier to compare. In 1971, the construction was
mostly wool; but increasing amounts of synthetics have been used
since then so that presently they are 100% synthetic. As can be
seen, the number of felts used did decrease even though the cost
per ton was slightly higher.
The last factor, and most important, is paper machine pro-
ductivity. This is the ratio of the maximum tonnage the paper
machine could produce with no downtime or paper breaks to the
actual tonnage produced. As can be observed, this aspect was
essentially unchanged by recycling. Also, although there has
been a substantial increase in wire pit suspended solids (from
0.2% to 0.3% as a result of closure), there has not been a loss
of machine speed. Higher process water temperature and higher
average sheet moisture (computer control) have undoubtedly been
compensating factors. Furthermore, the increase in water-
extractables in the final sheet—increased from 20 to 41 kilo-
grams per metric ton (40 to 180 Ib/ton) due to closure—has been
accompanied by a reduction in fiber content at the same basis
weight and, presumably, a corresponding reduction in drainage
resistance.
A number of other important cost factors are graphed in
Figure 26. In all cases adjusted for inflation, improvements
were made on the base year of 1971.
The cost reduction for water and steam are easily recon-
cilable as a result of closing the water system. Some portion of
the cost of electricity may also be related to water recycle, but
most of this reduction is due to a conscious effort to conserve
power instituted as a management program.
The last item, maintenance, is the one that might be antici-
pated to rise dramatically; but this was not the case. Fortu-
nately, most of the mill system in 1971 was stainless steel. The
two principal non-stainless areas were parts of the fourdrinier
and many of the process pumps. Pump cases in cast iron con-
struction that had previously lasted two to three years corroded
in as little as five months. As gradual replacement was neces-
sary, it was made in all stainless pump construction.
Also, some changes and modifications had to be made at the
fourdrinier. Grills, catwalks, and general machine components
required more frequent cleaning. Corrosion of cast iron bearing
housings on wire return rolls was accelerated and required more
frequent replacement until these were replaced in stainless.
Higher temperature greases were required for wire return roll
bearings. The wire return roll doctor blades had to be changed
from micarta to polyethylene because the life had decreased from
six months to one week. Corrosion of the fourdrinier main frame
and soleplate is a problem. There have been recent attempts to
apply an epoxy coating (Trowelon) similar to that used
63
-------
MAINTENANCE, LABOR
AND PARTS
1OO %
ELECTRICITY
KWH/TON
IOO % I
STEAM
IO3#/TON
WATER
GALLONS/TON
100 %
Figure 26. Total mill cost factors per short ton product
expressed as percent of 1971 requirements.
64
-------
successfully on concrete in the wire and seal pits. Only time
will indicate how successful this approach is for long-term pro-
tection of these vulnerable components. With all the problems,
it is a testimonial to the people involved that maintenance costs
have actually decreased during this period of ever-increasing
recycling. This decrease, in most part, is due to the proper
selection of materials and good management by the maintenance
staff.
There have been some process changes relative to recycling
that are continuing problems. Chemical requirements for sizing
can be increased from two to three times over that required for
washed stock; but as these are not costly chemicals, this can be
tolerated. We did find it uneconomical to produce wet strength
grades because chemical requirements for equal wet strength are
three to four times greater than those required for washed NSSC
stock, and this additive is an expensive chemical.
Contaminants from waste paper sources (such as latex, hot
melts, and asphalt) accumulate in the process water and have a
greater chance of being deposited on paper machine wires and
press rolls. We have found ways to minimize the problem but not
to eliminate it.
Any additive to the process water system in a closed loop
becomes a potential hazard in that it may react with other chemi-
cals or with itself at some concentration to produce a chemical
slime. Therefore, it is necessary to have an overall deposit-
control program, and its control merits the sustained attention
of operating mill management. Any new additive or change in rate
of an established additive must be carefully scrutinized before
implementation and watched closely for problems after a change
is made.
Mill shutdowns and subsequent startups can be problems, par-
ticularly if the downtime is of long duration. The large volume
of process water—1134 to 1512 cubic meters (300,000 to 400,000
gal) in our system—becomes increasingly susceptible to bacterial
growth as it cools. Increased dosage of the same slimicides used
for control under operating conditions can deter microbiological
activity up to two days in stagnant storage. If a shutdown is to
be for three days or longer, provision must be made to add addi-
tional amounts of slimicide at two-day intervals and to mix the
stored water by good agitation or by circulation with pumps.
For long shutdowns (up to two weeks), we have run out our
inventory of paper stock completely, even if the last reels are
not salable. The machine system is completely emptied and washed.
All process water is stored in three large tanks, which can be
circulated for slimicide addition. Insulation of these tanks is
65
-------
helpful to retain temperature as long as possible. In our cold
climate, exposed pipelines must be steam traced and insulated
to avoid freezing.
Another mill may choose different options to avoid biolog-
ical deterioration of stored water, as each mill has its own
unique system and problems.
MEDIUM STRENGTH CHARACTERISTICS AND CONVERSION BEHAVIOR
One of the concerns during the progressive closure of the
mill process system was the effect on strength and runnability
as the retained soluble matter in the sheet increased. A number
of approaches were used to try to evaluate the consequences of
process water closure:
(1) Physical testing at the paper machine.
(2) A thorough investigation of all customer complaints,
even minor ones.
(3) Continuous monitoring of our own largest corrugating
plant for changes in strength or runnability.
(4) A study was sponsored under the U. S. Environmental
Protection Agency project to try to determine the
effects of degree of closure on medium properties
by testing it under laboratory conditions at the
Institute of Paper Chemistry.
The changes at the paper machine were so gradual that is was
impossible to show any immediate, clear-cut effects. After an
extended period of observation, it has been concluded that the
soluble solids retained in the sheet do have some effect on fiber
bonding. The change is small up to our normal maximum solubles
level of 6% and affects mullen, tear, and tensile strength but
shows little effect in tests related to stiffness, such as con-
cora and ring crush. The measured reduction in mullen brought
about by increasing solubles in the process water from 1% to 5%
was a drop from 17.7 to 15.9 kilograms (39 to 35 Ib), or approx-
imately 10%.
Reductions in tensile strength would appear to be of the
same magnitude; and tear may be just slightly greater, but our
data are not complete enough to give precise numbers. Mullen is
run on each set off the paper machine as a routine test, so good,
continuous data are available. Tear and tensile are not normal
tests. Based on laboratory testing, there is also some indica-
tion of a reduction in wet web strength with increasing sheet
soluble levels, again of about the same 10% magnitude.
66
-------
The complaints from our customers were not very indicative
of problems related to mill closure. There were some intermit-
tent problems involving two' corrugators using archaic tension
systems, but these problems had existed before closure.
Our Green Bay Corrugated Division is a very modern operation
with an extensive quality control program. This continuous
quality program was well established before 1971, so comparison
figures were readily available for strength. Also, our corruga-
tors are run on an incentive program, as are most in this indus-
try, so that any interference with normal operations would be
indicated immediately. Again, no trends were observed during
closure that could be attributed to the increased solubles level
in the medium. Table 11 is a comparison of medium and corrugated
board strength for 1970 and 1975, representing open system and
almost completely closed system, respectively.
TABLE 11. STRENGTH COMPARISON—OPEN SYSTEM, 1970,
TO CLOSED SYSTEM, 1975
Item
Year
1970
1975
Paper machine
Basis weight
(pounds/1000 sq ft)
Concora (pounds)
Mullen (pounds)
Corrugator
Flat crush
(pounds/sq in)
A Flute
B Flute
C Flute
Standard box, top to
bottom compression
strength
(percent of standard)
26.3 26.1
71.8 70.6
42.0 34.0
30.8
48.1
46.5
100.6
30.1
55.0
40.3
107.4
The summary of the work carried out at the Institute of
Paper Chemistry is reported in Appendix B.
67
-------
The slight reduction in bonding strength has not been a
problem involving the paper machine or corrugator. The problem
has been negated to some degree by the use of slightly more long
fiber in the sheet in recent years. This is added to the system
in the form of corrugator plant clippings and/or recycled corru-
gated boxes, depending on market conditions. The percentage has
increased from less than 20% to approximately 30%. This change
has helped wet web strength, and therefore minimized breaks
occurring at the couch and press sections with no loss of concora
strength.
ECONOMIC EVALUATION OF REVERSE OSMOSIS VS. EVAPORATION
An economic comparison is presented to show the relative
costs of water removal by reverse osmosis and evaporation. Up-
dated capital cost data are provided for Green Bay Packaging * s
reverse osmosis plant and the company's two effect evaporator.
Energy cost and maintenance cost data are also provided.
To provide an effective evaluation, the capital and oper-
ating costs are also provided for a vapor compression evaporative
system. However, maintenance cost data are based on limited
field experience from other users of vapor compression evapora-
tion.
For design comparison, it is assumed that the following con-
ditions would hold true for each system:
(1) Daily water removal capacity—109 cubic meters
per day (28,800 gal/day).
(2) Annual water removal—19,100 cubic meters
(5.05 x 106 gal) for 60% annual operating time.
(3) Mill process water feed would be essentially free
of suspended solids with a dissolved solids concen-
tration of 5% average.
(4) Concentrated product will be approximately 7%
dissolved solids.
(5) Product quality should be equally low in
contaminants.
For such water removal capacity, the purchased reverse osmo-
sis plant has 288 modules, or 445.9 square meters (4800 ft^) of
permeable membrane. It is powered by two 25-Hp feed pumps and
six 10-Hp recycle pumps. There is also a 1%-Hp permeate pump for
ultimate disposal of the permeate.
The comparable two effect evaporator would be essentially
the same size as the existing liquor evaporator at the mill's
68
-------
recovery plant. As the existing evaporator is designed to pro-
duce 150 cubic meters per day (39,700 gal/day) of condensate, a
somewhat smaller tube surface area would be required than
actually is in use in the recovery plant evaporator. The two
effect evaporator is designed to use 123 kilograms per square
centimeter (175 psi) steam and up to 1.66 cubic meters per
minute (440 gpm) condenser cooling water. A normal economy for
this device is 0.68 kilograms (1.5 Ib) condensate per 0.45 kilo-
grams (1 Ib) steam.
The vapor compression evaporator would consist of two units
with combined capacity of 109 cubic meters (28,800 gal) daily.
The necessary heat exchangers would be provided to transfer all
available heat to the feed stream to raise the feed to boiling.
Essentially all energy input is electrical with a small fraction
as steam.
1976 Capital and Installed Costs
Comparable
Vapor
Reverse Two Effect Compression
Osmosis Evaporator Evaporator
Total installed cost $324,572 $200,000 $250,000
1974-1975 Operating Cost Experience
Operating cost—$/3.785 cubic meters (1000 gal) purified water
Energy consumption $1.70 $8.60 $3.34
Depreciation 6.45 3.18 3.71
Maintenance and labor 2.58 .91 .99
Total cost—$/3.785 m3 $10.73 $12.69 $8.04
The total installed costs for reverse osmosis and the two
effect evaporation have been updated from the original costs by
the use of CHEMICAL ENGINEERING plant cost indices. The vapor
compression installed cost estimate was supplied by the manu-
facturer of such equipment.
Energy input costs are given based on electricity and steam
costs at the mill for 1974 and 1975. The vapor compression
energy cost is based on a conservative estimate of 118.8 kilo-
gram-calorie per kilogram (66 Btu/lb) evaporation and is
considered to be a cost of electrical energy. This energy
69
-------
consumption value is based on the experience of an existing evap-
orator used in a pulp mill application.
Depreciation charges are assumed to be for a 12-year useful
life on all hardware. It should be noted that the expected use-
ful module life in reverse osmosis is given at five years—this
may be conservative.
Maintenance and labor costs for reverse osmosis and two
effect evaporation are based on mill experience for 1974 and
1975. These costs for reverse osmosis may be excessive because
of the developmental nature of the changes required in the pump-
in system. The maintenance costs for vapor compression are based
on limited field experience and the assumption that maintenance
will cost 2% per year of the installed cost.
70
-------
REFERENCES
1. Miles, H. Spent Liquor Disposal Via FluoSolids Combustion
at an NSSC Mill. Paper Trade Journal, 154(33):26-31,
(Aug. 17, 1970).
2. Wiley, A. J., G. A. Dubey, and I. K. Bansal. Reverse Osmosis
Concentration of Dilute Pulp and Paper Effluents. U. S.
Environmental Protection Agency, Program #12040 EEL, 02/72.
3. Morris, D. C., W. R. Nelson, and G. 0. Walraven. Recycle
of Papermill Waste Waters and Application of Reverse Osmosis.
U. S. Environmental Protection Agency, Program #12040 FUB,
01/72.
71
-------
APPENDIX A
ANALYTICAL AND TEST PROCEDURES AND EQUATIONS
USED IN REVERSE OSMOSIS PLANT
The analytical samples were collected when the operating
conditions of the units were stable. This required that the
samples be taken at the middle of each pause-operate-pause cycle,
as the conditions close to the depressurization (pause) period
were not representative of the majority of the operating period.
Operating data, e.g. pressures, were recorded at the time of
taking samples.
Frequent measurements were made of the total solids in the
feed and concentrate. The sample is weighed in an oven-dried
tared beaker, dried overnight at 105°C, cooled in a desiccator,
and reweighed.
Soluble solids and suspended solids were determined by
standard filtration and oven drying. A sample is filtered
through a glass fiber pad (Reeve Angel 934 AH, 11 cm), and the
filtrate is collected. The pad is oven-dried, cooled in a
desiccator, and reweighed for suspended solids. The filtrate is
weighed in an oven-dried tared beaker, dried overnight at 105°C,
cooled in a desiccator, and reweighed for soluble solids.
Samples of product water from the reverse osmosis units required
no filtering.
Sodium content was determined with a Model 303 Perkin-Elmer
atomic absorption spectrophotometer. The sample is prepared to
the proper working range using distilled water for dilution. The
spectrophotometer is operated according to Perkin-Elmer1s
Analytical Methods Manual.
Analysis of incubated bottles for five-day biochemical
oxygen demand (BODs) was measured according to procedures out-
lined by the Yellow Springs Instrument Company using their
Model 54 Dissolved Oxygen Meter. Chemical requirements, sample
preparation, apparatus, and calculations described in Standard
Methods for the Examination of Water and Waste Water (APHA-AWWA-
WPCF), 13th edition, 1971, page 489, Method 219, were used for
the remainder of the analysis. Volumetric flasks and pipettes
are utilized for measurement and dilution. A sample, two cubic
centimeters or larger, is pipetted directly into the BOD bottle.
72
-------
Color is determined by spectrophotometric comparison of the
sample with known concentration of colored solutions. The proce-
dure is described in NCASI Stream Improvement Technical Bulletin
No. 253, December, 1971.
EQUATIONS
All flux data (unless indicated otherwise) are reported at
standardized conditions of plant design of 38°C (100°F), a feed
soluble solids of 5%, and a plant feed pressure of 41.4 bars
(600 psi), and have either been obtained at these conditions or
have been corrected to these conditions by the following formula.
Correction is avoided when possible because of additional lab
work and errors in measurement of average module soluble solids.
Permeate flux in gallons per day per sq ft of membrane
PPF x T x 1440 AMFP - 185
2
No. of modules x 16.67 ft per module AMFP - TT
where PPF = plant permeate flow in gallons per minute
AMFP = average module feed pressure in pounds per sq in
"TC = actual osmotic pressure
° = 37* x percent average module soluble solids
T = temperature correction to 38°C (decrease by
c 2.3% above 38OC, or increase by 2.3% below 38Oc)
185 = osmotic pressure at 5% soluble solids
= 37* x 5
1440 = minutes per day
*average measured osmotic pressure of process water at
1% soluble solids, as observed at Green Bay Packaging
Material balances and percent rejections for soluble solids,
BOD, sodium, and color are obtained using plant measurements,
laboratory data, and the following expressions.
73
-------
Flow Balance:
V, = V + V
f c p
where V = volume per unit time of plant flow for feed (f) ,
concentrate ( ), or permeate ( ). These are
c p
obtained from plant instrument measurement and
checked by actual timed volume collection in a
container of known volume.
Mass Balance:
- = W + W
f c p
where W = component weight per unit time in each stream of
sodium, soluble solids, BOD, or color. These are
obtained from analytical determination of concentra-
tion described previously, and the expression for
each is as follows:
Wf = Vf x specific gravity x concentration
Percent rejection for each component is then calculated as
follows:
% Rejection = (1 - W /Wf)100
P •*•
74
-------
Figure A-l. Typical routine reverse osmosis data sheet.
Entry
No.
608
609
610
611
612
613
614
615
616
617
618
619
620
Date
1976
2/06
2/10
2/11
2/12
2/13
2/24
2/25
2/26
3/01
3/05
3/09
3/10
3/11
Clock
Hrs.
10378.9
10459.8
10480.0
10501.1
10521.8
10553.9
10570.3
10594.5
10676.0
10762.7
10848.0
10868.5
10889.5
Raw (Process Water) Feed
Flow
Rate
(gpm)
79.5
82.0
81.5
81.0
81.5
82.0
82.0
82.0
81.5
79.5
80.0
80.0
80.0
Temp.
(°F)
98
98
98
98
100
100
100
105
99
99
101
97
98
No. of
Modules
In Service
261
261
258
258
261
261
261
261
258
258
255
258
258
Dis.
Sol.
(%)
5.15
5.08
4.17
5.04
5.12
5.33
5.46
5.12
4.29
5.23
5.64
6.02
6.29
PH
5.6
5.5
5.4
5.5
5.5
5.6
5.6
5.7
6.0
5.7
5.8
5.8
5.9
Module Feed
Flow
(gpm)
240
245
244
244
246
246
240
245
245
240
220
228
237
Feed,-
Press.
(psi)
600
600
600
600
600
600
600
600
600
600
600
600
600
Back
Press .
(psi)
460
450
460
455
451
461
490
471
452
450
450
445
446
Permeate
Conduc-
tivity
(/*mhos )
410
370
280
345
335
385
330
290
260
370
450
500
570
Flow
(gpm)
17.4
16.8
19.8
16.8
17.5
16.5
16.5
16.8
17.1
12.0
11.4
10.8
9.6
en
-------
APPENDIX B
THE INSTITUTE OF PAPER CHEMISTRY
Appleton, Wisconsin
EFFECT OF MILL WATER SYSTEM CLOSURE ON
CORRUGATING MEDIUM QUALITY
SUMMARY
This study was carried out under the sponsorship of Green
Bay Packaging Inc. with the objective of determining the effect
of the degree of closure of their corrugating medium mill process
water system on selected corrugating medium quality character-
istics. For this purpose, three 30-centimeter (12-in) wide rolls
of corrugating medium were obtained for each of the following
degrees of mill water system closure:
1. "Partially" closed system—process water solubles, 2.9%.
2. "Fully" closed system—process water solubles, 5.6%.
It should be noted that it was originally planned to evalu-
ate a third mill process water condition, namely, an "open"
system, to serve as the reference condition; however, these
samples could not be obtained for the study by the cooperator.
Each roll of medium was fabricated into single-faced board
shortly after arrival at the Institute and after 60 days storage
to determine its conversion quality. Conversion quality was
evaluated in terms of (1) the maximum runnability conditions of
corrugator speed and web tensions which could be employed without
causing fractured flutes, (2) flute height, (3) average differ-
ence in flute height (related to high-low flute formation), and
(4) draw factor. At each time of fabrication, samples of the
corrugating medium were obtained and evaluated for selected
physical characteristics.
The following results were obtained:
1. In general, the mediums representing both the "partial"
and "fully" closed mill water system exhibited
76
-------
satisfactory conversion behavior on the corrugator. The
"partial" closure water system medium generally exhib-
ited somewhat higher physical property levels than the
"fully" closed system mediums. However, the differences
in physical properties may not be of commercial impor-
tance in many cases. A more detailed summary of the
results is as follows:
2. Conversion Behavior.
a. Runnability.
When evaluated on arrival, the maximum web tensions
which could be employed without fracture at a
corrugator speed of 183 meters per minute (600 fpm)
ranged from 272 to 363 grams per centimeter (1.5 to
2.0 lb/in) and averaged 327 grams per centimeter
(1.8 lb/in) for the "partially" closed system
mediums. The corresponding results for the mediums
representing the "fully" closed system ranged from
182 to 363 grams per centimeter (1.0 to 2.0 lb/in)
and averaged 272 grams per centimeter (1.5 lb/in).
The difference in average runnability is not con-
sidered to be significant in view of the variability
from roll-to-roll and the levels appear to fall
within the usual commercial range for good runnabil-
ity. After the 60-day storage period, the maximum
runnabilities of the "partial" system closure medium
rolls decreased about 109 grams per centimeter
(0.6 lb/in) on the average. A decrease of this
magnitude due to aging is not uncommon. The medium
rolls representing the "fully" closed system exhib-
ited about the same maximum runnability tension
after aging as was obtained on arrival.
b. Flute height and high-low flute formation.
The degree of system closure did not have any marked
effect on the flute height of the single-faced
boards or the tendency of the mediums to form
high-low flutes.
3. Physical Characteristics of Medium.
a. The Concora flat crush strengths of the "partially"
closed system mediums averaged 2.7% higher than that
of the "fully" closed system mediums in the "on
arrival" tests but the difference was not statisti-
cally significant.
77
-------
b. The degree of closure did not have a statistically
significant effect on the cross-direction edgewise
compression strength of the medium.
c. The "partially" closed system mediums exhibited a
slightly higher water drop time on arrival which was
statistically significant but the difference may not
be of commercial importance.
d. The "partially" closed system mediums exhibited
significantly higher tearing strengths and tensile
properties than the fully closed system mediums in
the "on arrival" tests, but the differences may not
be of commercial importance in many applications.
78
-------
APPENDIX C
PROJECT VISITORS
MAY, 1972, THROUGH DECEMBER, 1976
Number Organization or Affiliation
U. S. Industrial Companies
2 A & P, National Dairy Division
3 Aqua-Chem, Inc.
1 Butter Manufacturing
1 Chrysler Outboard Corporation
2 Clark & Vicario
2 Continental Can Co., Inc.
2 Dresser Industries, Inc.
2 ' Eastman Chemical International Co.
1 W. D. Ehrke Co.
4 Fabric Research Laboratories, Inc./Albany
International Co.
2 Hammermill Paper Co.
2 Kenics Corporation
1 Kimberly Clark Corporation
1 McGark Strapp & Associates
2 McMaster-Carr
1 Milk Specialties Co.
5 Nicolet Paper Co.
5 Owens-Illinois
2 H. C. Prange Co.
2 Proctor & Gamble
1 Reed Paper Ltd.
1 Rex Chainbelt, Inc.
3 St. Regis Paper Co.
1 L. D. Schreiber Cheese Co., Ltd.
1 Scott Paper
5 Sonoco Products Co.
5 Stone Container
1 3M Co.
2 Universal Foods
1 Ward's Cheese
2 WesCor Corporation
1 Westinghouse
2 Weston Paper and Mfg. Co.
2 Whirlpool Corporation
79
-------
Number Organization or Affiliation
Foreign Concerns
3 ASSI Development Laboratory (Sweden)
1 Australian Paper Manufacturers, Ltd.
2 B.D.M.R. Consultants (Canada)
2 T. W. Beak Consultants Ltd. (Canada)
3 Broby Industrier AB (Sweden)
3 Cellulose Attisholz AG (Switzerland)
3 Daicel, Ltd. (Japan)
7 Domtar, Ltd. (Canada)
2 Emsland Staerke (Germany)
5 Fiskeby AB (Sweden)
1 IVL Consulting Ltd. (Sweden)
1 IVL Swedish Pollution Control Co.
1 IVL Swedish Water and Air Pollution Research
Laboratory
2 Kon. Scholten Honig. (Netherlands)
2 LaRochette-Venize Mill (France)
1 Mitsubishi Heavy Industries, Ltd. (Japan)
1 Mitsui & Co. (Japan)
4 MoDoCell (Sweden)
1 National Environmental Board (Sweden)
1 ttrebro Pappersbruk AB (Sweden)
1 Osaka Municipal Technical Research Institute
(Japan)
1 Papeteries de L'Epte (France)
2 Paterson Candy International, Ltd. (England)
3 Sanyo-Kokusaku Pulp Co., Ltd. (Japan)
3 Sasakura Engineering (Japan)
1 South African Pulp and Paper Industries, Ltd.
1 Tomoegawa Paper Mfg. Co., Ltd. (Japan)
1 Woodall-Duckham Pacific Ltd. (Australia)
Government
4 Environmental Protection Service (Canada)
1 Federal Elected Official
1 Illinois State Water Survey Division,
Department of Registration and Education
3 U. S. Environmental Protection Agency
4 Wisconsin Department of Natural Resources
Tour Groups
9 American Society for Quality Control Sponsored
Tour
5 Bay City Alternative School
35 CESA Environmental Education Trainees
27 East DePere High School
80
-------
Number Organization or Affiliation
Tour Groups (continued)
12 Forest Industry Council Sponsored Tour of News
Personnel
13 Green Bay Chamber of Commerce Environmental
Committee
65 Institute of Paper Chemistry
6 Institute of Paper Chemistry Technical Advisory
Committee
39 Instrument Society of America Sponsored Tour
22 International Fibre Box Association
75 Northeast Wisconsin TAPPI Sponsored Tour
19 Northeastern Wisconsin Dairy Technical Society
10 TAPPI Sulfite and SemiChemical Pulping Committee
19 Third SemiChemical Corrugating Medium
Superintendents' Meeting Tour
108 University of Wisconsin-Green Bay
10 University of Wisconsin-Madison
12 University of Wisconsin-Stevens Point
Educators
13 Institute of Paper Chemistry
1 Yokohama National University (Japan)
2 University of Wisconsin-Madison
Miscellaneous
1 Chamber of Commerce
3 News Media (Press and TV)
1 Pulp and Paper Magazine
81
-------
GLOSSARY
3005: Biochemical oxygen demand based on oxygen requirements
needed during a five-day period for stabilization of
decomposing organic matter in an aerobic biochemical action.
CentriCleaner: Trade name. Free vortex cone type cleaner for
centrifugal separation of shives, dirt, bark, etc., from
normal wood fiber by density, size, and shape.
compaction: Decrease of water permeation rate through a membrane
with time at a fixed pressure and temperature.
concentrate: The solution exiting from the reverse osmosis unit
after removal of a portion of the water through the
membrane.
consistency: The percentage, by weight, of air dry fibrous
material in a stock suspension.
DSM screen: Trade name. Pressurized nozzle feed screen with a
curved screen surface.
fluidized bed reactor: Fluid solids process of burning spent
liquor to obtain a dry ash in the form of inorganic pellets.
flux rate: Rate of permeation or transport of water through the
membrane.
Hydrasieve: Trade name. Stationary, three-slope, stainless
steel, wedge bar screen.
NSSC: Neutral Sulfite SemiChemical pulping, usually with sodium
bisulfite, but other bases such as ammonia are also used.
osmosis: Diffusion through a semipermeable membrane separating
a solvent and a solution that tends to equalize their
concentrations.
osmotic pressure: The hydrostatic pressure required to stop the
diffusion across a semipermeable membrane between two
solutions of dissimilar concentrations.
82
-------
pause: Periodic reduction of the pressure from an operating
level to atmospheric for the purpose of restoring the flux
rate lost by the deposition of a fouling material on the
membrane surface.
process water: A general term for all waters of a paper mill
which have been separated from the stock or pulp suspension,
either on the paper machine or accessory equipment, such as
thickeners, washers, and savealls. In a closed system, the
water which recirculates throughout the mill for pulp
dilution and showers.
reverse osmosis: Osmosis in reverse flow through a semipermeable
membrane when external pressure in excess of the osmotic
pressure is applied.
spent liquor: Waste liquor separated from cooked chips and pulp
containing the residual cooking chemicals and dissolved
constituents from the chemical cooking of wood chips.
83
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
i. REPORT NO.
EPA-600/2-77-241
3. RECIPIENT'S ACCESSIOr+NO.
4. TITLE AND SUBTITLE
CLOSED PROCESS WATER LOOP IN NSSC CORRUGATING MEDIUM
MANUFACTURE
5. REPORT DATE
December 1977 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Gerald 0. Walraven, William R. Nelson, Peter E. DeRossi,
Richard L. Wisneski
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Green Bay Packaging Incorporated
P. 0. Box 1107
Green Bay, WI 54305
10. PROGRAM ELEMENT NO.
1BB610
11. CONTRACT/GRANT NO.
S-800520
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Lab.
Office of Research 6 Development
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
- Gin., OH
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Over the last 5 years, the Green Bay Packaging corrugating medium mill has con-
verted to an essentially closed process water system. The mill is a net consumer of
water. This is due to the greater amount of water carried out of the system with the
sheet compared to the lower water content entering the process system in raw materials
Many small dilute water streams are accepted into the process without upsetting the
water balance.
When extraneous water inputs do upset the system balance, the condition is cor-
rectable by thermal evaporation or reverse osmosis'. The reverse osmosis plant design
operating performance and economics are described. Although many reverse osmosis op-
erating problems have been solved, flux rates are somewhat lower than had been pre-
dicted. Other system additions and revisions for process water entrapment, recycle,
and surge protection are described.
When a spill cannot be prevented, a monitoring system is used by production per-
sonnel for early detection and correction. Included in key areas is redundant equip-
ment to help correct failures quickly.
Levels of BOD loss have been reduced from the 20,000 pounds per day range (9072
kg/day)—1971— to less than 1000 pounds per day (454 kg/day)—monthly average— for
1975. The daily maxima of 4000 pounds per day (1814 kg/day) has not been exceeded in
any mill operating day during 1975.
report covers a period from July 19. 1972. through February 2, 1976.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Circulation, Water Pollution, Color,
Biochemical Oxygen Demand
Pulping, Closed System,
Reverse Osmosis Waste
oncentration, Physical
Separation, Product
Quality,Waste Monitoring
Total Suspended Solids,
Waste Control
13B
3. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
4 U.S. GOVERNMENT PRINTING OFFICE; 1978— 757-140 / 66 49
------- |