xvEPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
EPA-6 00/2-78-177
August 1978
Research and Development
Demonstration
of Ultrafiltration
and Carbon Adsorption
for Treatment
of Industrial
Laundering Wastewater
-------
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-600/2-78-177
August 1978
DEMONSTRATION OF ULTRAFILTRATION AND CARBON ADSORPTION
FOR TREATMENT OF INDUSTRIAL LAUNDERING WASTEWATER
by
Myles H. Kleper
Robert L. Goldsmith
Arye Z. Gollan
Walden Division of Abcor, Inc.
Wilmington, Massachusetts 01887
Grant No. S-804367-01
Project Officer
Ronald Turner
Industrial Pollution Control Division
Industrial Environmental Research Laboratory
Cincinnati, Ohio 45268
This study was conducted in cooperation with the
Institute of Industrial Launderers
Washington, D.C. 20036
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory-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.
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on
our health often require that new and increasingly more efficient pollution
control methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
The results of a research and demonstration project for treating
industrial laundry wastewaters using ultrafiltration - activated carbon
treatment processes are presented herein. The principal treatment step was
ultrafiltration (UF). The UF process concentrates suspended solids and emul-
sified oils while producing a high quality permeate stream. A portion of the
UF permeate was then treated by activated carbon for removal of residual
organics and color. It is hoped that the results of this study will aid
industrial launderers and municipal treatment authorities in better under-
standing the problems unique to this industry and will encourage further
research in this area.
The Organic Chemicals and Products Branch of the Industrial Pollution
Control Division Industrial Environmental Research Laboratory - Cincinnati
45268 should be contacted for further information on this subject.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
iii
-------
ABSTRACT
The treatment of industrial laundering wastewaters lay ultra-filtration
(UF) and activated carbon adsorption (ACA) has been investigated. Three
program tasks were performed:
Task 1: Pilot-scale testing at Abeor, Inc. facilities
Task 2: Pilot-scale testing at a field demonstration site
Task 3: Economic analysis of full-scale treatment systems.
All experiments were conducted with non-celluliosic spiral-wound UF modules.
This study of industrial laundry wastewater treatment by uTtrafiltratian
and activated carbon adsorption has indicated that a consistently high
quality product water, potentially reusable within the laundry, can be
produced. The operation of the spiral-wound ultraf nitration modules was,
however, hindered by the fouling tendency of the feed stream. Average
module permeate flux was therefore low. This factor, in turn, resulted in
high capital and operating cost estimates for full-scale treatment systems.
Successful feasibility tests with industrial laundering effluents were
performed with the spiral-wound modules during. Task 1. Average flux levels
of 40-50 gal/ft2-day (125°F) and stable membrane performance were realized.
Adsorption isotherms conducted with composite UF permeate samples indicated
that carbon adsorption capacity for color bodies was quite, good, whereas
TOG adsorption was marginal.
In Task 2, field demonstration experiments with a 5000 gpd (nominal
capacity) treatment system were conducted. Initial tests identified the
need for prefiltration to prevent plugging of the spiral-wound modules by
lint. A successful, but temporary and non-commercial, prefiltration step
was developed for lint removal.
Subsequently, four 2-week demonstration tests were conducted. Although
plugging by lint was eliminated, severe membrane surface fouling occurred
and difficulty in recovering membrane flux using standard detergent cleaning
procedures was encountered. UF membrane flux at system conversions ranging
from 67% to 99% averaged 14,gfd (135°F).
-------
A slip-stream of UF permeate was continuously passed through a 2-inch
diameter carbon column. Overall UF/ACA removal efficiencies were 99.2% for
turbidity and 98% for oils and grease. Also, BOD, COD, and TOC removals were
82%, 86%, and 82% respectively.
Capital and operating cost estimates for treatment systems employing
spiral-wound UF modules are, based on the low permeate flux, understandably
high. For a 100,000 gpd treatment system (assuming 14,gfd flux and 98%
water recovery), operating cost for the prefiltration and UF systems is
estimated to be $9.01/1000 gal. Total carbon adsorption costs of $1.07/
1000 gal could be offset by credits gained from reuse of the product water
within the laundry. Reuse of this product water was not demonstrated during
this program.
This report was submitted in fulfillment of the original scope-of-work
for Grant No. S-804367-01 by the Walden Division of Abcor, Inc. under the
sponsorship of the Institute of Industrial Launderers and the U.S. Environ-
mental Protection Agency. The original grant was twice amended to include
sampling and analysis for priority pollutants at industrial laundries and an
assessment of sludge disposal alternatives for industrial laUnderies. Work
on both of these amendments is ongoing and will be reported as addenda.
This report covers the period from March 15, 1976 to March 14, 1977, and work
was completed as of April 8, 1977.
-------
CONTENTS
Foreword i i i
Abstract i v
Figures vii
Tables vi i i
English-Metric Conversion Table x
Acknowledgment xi
1. Introduction 1
2. Conclusions 7
3. Recommendations 9
4, Discussion of Unit Processes 10
5. Program Overview 23
6. Test Systems, Procedures and Analysis 24
7. Results and Discussion 32
A. Pilot-Scale Testing at Abcor's Facilities 32
B. Preliminary Field Demonstration Experiments 43
C. Formal Field Demonstration Program 52
8, Economic Scale-Up Analysis 73
References 85
Appendices
A. Additional Data from In-House Pilot-Scale Experimentation 86
B. Additional Data from Preliminary Field Demonstration
Experimentation 101
C. Detailed Costing Projections for Tubular UF Systems 105
-------
FIGURES
Number Page
1 Generalized Process Flow Schematic 5
2 Cut-away View of Tubular Ultra-filtration Assembly 14
3 Cut-away View of Spiral-Wound Ultrafiltration Module 15
4 Various System Designs for Modular Membrane Equipment 17
5 Hypothetical Isotherms and Breakthrough Curves for
Carbon Adsorption .20
6 Flow Schematic of Field Demonstration Test System 25
7 Flow Schematic of Carbon Column 30
8 Equilibrium Adsorption Isotherms for "Medium"
Loading Waste 42
9 UF Permeate Flux vs. Time for Third Preliminary Field Test ...46
10 UF Permeate Flux vs. Time for 67% Conversion Test 53
11 UF Permeate Flux vs. Time for 90% Conversion Test .54
12 UF Permeate Flux vs. Time for 97% Conversion Test 55
13 UF Permeate Flux vs. Time for 99% Conversion Test 56
14 Carbon Column Breakthrough Curves during 67% and 90%
Conversion Tests 68
15 Carbon Column Breakthrough Curves during 97% and 99%
Conversion Tests 69
16 Flow Schematic of 100,000 gpd Waste Treatment System 74
vm
-------
TABLES
Number Page
1 Laundry Water Contaminant Levels for One Typical
Plant and Its Sewer Discharge Standards 2
2 Ranges of Waste Treatment Efficiencies for a
Conventional Industrial Laundry Waste Treatment
System 4
3 Differences Between Reverse Osmosis and Ultrafiltration 11
4 Comparison of Membrane Configurations of Interest 13
5 Chemical Analyses Routinely Performed 31
6 Average UF Membrane Module Performance during
Experimentation with "Medium" Loading Industrial Laundry
Wastewater Sample 33
7 History of Membrane Module Performance During
Experimentation with "Medium" Loading Industrial
Laundry Wastewater Sample 35
8 Theoretical Power Requirements for UF Spiral-Wound
Modules during In-House Experiments 37
9 Average Contaminant Analyses and Membrane Removal
Efficiencies for UF Batch Concentrations of
Industrial Laundry Wastewaters 39
10 Contaminant Analyses During UF Total Recycle
Experimentation 41
11 Theoretical Power Requirements for UF Spiral-Wound
Modules During Preliminary Field Tests 48
12 Contaminant Analyses and UF and Carbon Removal
Efficiencies During 97% Conversion Period of Test P3 50
13 Contaminant Analyses and UF Removal Efficiencies
During 99.5-99.8% Conversion Period of Test P3 51
14 Flux Recovery and Accumulated Operating Times for
UF Membranes During Fie!d Tests 59
15 Contaminant Analyses and UF and Carbon Removal
Efficiencies during 67% Conversion Test 61
16 Contaminant Analyses and UF and Carbon Removal
Efficiencies during 90% Conversion Test 62
IX
-------
Number Page
17 Contaminant Analyses and UF and Carbon Removal
Efficiencies during 97% Conversion Test 63
18 Contaminant Analyses and UF and Carbon Removal
Efficiencies during 99% Conversion Test 64
19 Average Carbon Removal Efficiency Data 66
20 Summary of Average UF/ACA Product Water Quality and
Removal Efficiencies during Field Demonstration Tests 70
21 Estimated Purchased Equipment Costs for Wastewater
Treatment Systems of Various Capacities 76
22 Estimated Annual Operating Costs for Wastewater
Treatment Systems of Various Capacities 77
23 Estimated Power Requirements for Wastewater
Treatment Systems of Various Capacities 78
24 Estimated Additional Purchased Equipment and
Operating Costs for Complete Water Reuse 80
25 Estimated Annual Credits for Reuse of Treated
Water 81
26 Summary of Estimated Capital and Operating Costs
for Wastewater Treatment Systems Employing Tubular
Ultrafiltration Modules 83
-------
ENGLISH-METRIC CONVERSION TABLE*
To Convert From
To
Multiply by
Inch
Feet
Square inch
Square feet
Cubic feet
Gallon
Pound
Pound per sq. inch
Horsepower
Gallon per day
Gallon per minute
Gallon per sq. ft-day
Gallon per minute per sq. ft.
Meter
Meter
Square meter
Square meter
Cubic meter
Cubic meter
Kilogram
Atmosphere
Watt
Cubic meter per day
Cubic meter per day
Cubic meter per sq. meter-day
Cubic meter per sq. meter-day
2.54x10-2
6.45xlO"4
9.29x10-2
2.83x10-2
3.79xlO-3
4.54X10-1
6.80x10-2
7.46x10 2
3.79xlO'3
5.45
4.10x10-2
5.87x10 !
The units most familiar to the projected readership of this report
have been maintained.
-------
ACKNOWLEDGMENTS
The efforts and dedication of Mr. Charles E. Humphrey, Research
Coordinator of the Institute of Industrial Launderers, in maintaining program
direction and in providing technical assistance to the authors is
acknowledged.
The technical guidance of Mr. Mervyn Sluizer, Jr., and Mr. Manfred Tidor
of the Institute of Industrial Launderers, and Mr. Ron Turner of EPA
throughout the entire program is also gratefully acknowledged.
Roy F. Weston, Inc.,as consultants to the Institute of Industrial
Launderers, served in a review capacity for this program.
The authors thank the management and staff of all the industrial :
laundries that participated in this program and are particularly grateful
to the employees of Standard Uniform Rental Service, Dorchester,
Massachusetts, for their helpfulness and support during the field demonstra-
tion program.
Ms. Cheryl Renaud prepared the draft of this report and Ms. Sharon
Collins typed the final manuscript.
XII
-------
SECTION 1
INTRODUCTION
NATURE OF THE PROBLEM
Industrial laundries handle a variety of items, including uniforms,
shop towels, printers towels, mops, mats and gloves. The wastewaters
generated from the laundering of these items are of significant environmental
impact in terms of both waste volume and contaminant loading. Although
wastewaters from all laundry sources are reported to account for 5% to 10% of
municipal sewer discharges the portion of these wastewaters attributable to
industrial laundries is in the order to 0.5% (8). Industrial laundry waste-
waters can be from 3 to 20 times higher in suspended solids and BOO than
average domestic sewage (1).
Representative wastewater contaminant data for a typical industrial
laundry operation (1) are presented in Table 1, along with the sewer dis-
charge limits set by the City,of Chicago, where this plant is located. In
addition to high suspended solids and BOD loadings, the levels of oil and
grease, lead, and mercury in the plant effluent are in excess of the
Municipal Discharge Standards. Other characteristics of industrial
laundering effluents are:
both flow and composition are highly variable, over
both short-term (minutes to hours) and long-term
(days to months) operations, and
the emulsions are very stable chemically.
It is apparent from these observations that this industry will require a
stable treatment system highly efficient in waste removal to meet sewer
discharge standards.
CURRENT TREATMENT
One of the most thoroughly investigated methods of treating commercial
laundry wastewaters consists of coagulation and flocculation followed by
dissolved air flotation with polishing of the underflow by sand or
diatomaceous earth (DE) filtration and dewatering of the flotation scum by
vacuum filtration (1). Though this sequence was found acceptable for
-------
TABLE 1. LAUNDRY WATER CONTAMINANT LEVELS FOR ONE TYPICAL
PLANT AND ITS SEWER DISCHARGE STANDARDS (1)
Contaminant
Total solids, mg/1
Suspended solids, mg/1
Total dissolved solids,
mg/1
BOD, mg/l'
TOC, mg/1
Hexane extractables, mg/1
pH, units
Alkalinity, ppm CaCOs
Total Chromium, mg/1
Copper, mg/1
Lead, mg/1
Zinc, mg/1
Cadmium, mg/1
Iron, mg/1
Nickel , mg/1
Mercury, mg/1
Range in
Waste Sampled
Over 3 Mos.
4,900-8,600
650-5,000
650-1,300
1,000-6,300
400-3,800
10.2- 11.9
1.0- 3.6
0.2- 9
3.0- 36
0.6- 9
0 - 0.6
3.5-126
1.0- 2.5
1.2- 7.0
Average
Levels in
Waste
6,800
2,800
4,000
830
2,500
1,500
325
2.3
4,0
12.7
3.9
0.24
40
1.6
3.3
Typical Sewer
Discharge
Standards
_ _
300*
300*
100*
4.5-10
25
3.0
0.5
15.0
2.0
50
10
0.0005
*
Surcharge assessed based on mass discharge above acceptable limitation
and not concentration.
** Oil and Grease
-------
uniform and linen laundry wastewater, it could not consistently produce an
effluent meeting municipal sewer standards when processing industrial laundry
wastes.*
The ranges of removal efficiencies by the flotation portion of the
treatment system alone and the flotation/DE filter combination are given in
Table 2 for industrial laundry wastewater processing. Although very good
quality water can be achieved by the dissolved air flotation/filtration
scheme, the inconsistency of this system in treating the highly variable
industrial laundering effluents is evident from these data. Therefore, even
with this method of treatment, industrial launderers might periodically incur
municipal sewer surcharges for suspended solids and BOD above those normally
present in domestic sewage. Heavy metals may have to be removed completely
according to local ordinances and the capacity of municipal treatment
facilities. Annual operating costs without these surcharges were estimated
at $2.70/1000 gal if chemicals were supplied in bulk.
A dissolved air flotation scheme employing an electrolytic process that
involves an electrocoagulation cell followed by an electro-flotation basin
was tested at an industrial laundry for 7 days (8 hours per day ) (2).
Maximum suspended solids, BOD, and hexane extractive reductions of 92%,
86%, and 94%, respectively, were observed. Though reported power and
chemical costs for this system are low ($0.66/1000 gal) no estimation was
made of the cost for operating labor. This cost could be quite high
because of the constant variability of the waste. Waste equalization can
potentially lower the labor cost. A trade off must be made, however, ;
between reduced operating costs and the high space requirement for holding
tanks.
TREATMENT APPROACH SELECTED FOR EVALUATION
The overall goal of this program was to develop an economically viable
wastewater treatment system which could consistently produce an effluent
meeting municipal sewer discharge standards. The method of industrial laundry
wastewater treatment investigated was ultrafiltration (UF) followed by
activated carbon adsorption (ACA). Ultrafiltration is the principal unit
process. The UF system concentrates suspended solids and high-molecular
weight solutes, producing two streams: a purified product water (permeate)
and a concentrate. Typically, the concentrate volume is 1% to 5% of the
influent volume. Further treatment of the UF permeate, principally for
removal of residual low molecular weight organics (i.e., dissolved detergent,
dyes, and solvents), is achieved by adsorption on activated carbon.
As defined in reference (1), an industrial laundry washes mostly shop
towels, printers towels, and dust mops, which results in wastewater
contamination that is abnormally high compared to other laundry types.
-------
TABLE 2. RANGES OF WASTE TREATMENT EFFICIENCIES FOR A CONVENTIONAL
INDUSTRIAL LAUNDRY WASTE TREATMENT SYSTEM (1)
Parameter
BOD
TOC
Suspended Solids
Total Solids
Hexane Solubles
Copper
Lead
Mercury
Cadmi um
Zinc
Total Chromium
Iron
Nickel
Range in Waste
Sampled Over
3 Months (mg/1)
650 -1,300
1,000 -6,300
650 -5,000
4,900 -8,600
400 -3,800
0.2- 9
3.0- 36
1.2- 7.0
0- 0.6
0.6- 9
1.0- 3.6
3.5-126
1.0- 2.5
Range of %
Removal by Flotation
0-67
0-93
38-98
0-54
0-95
0-93
17-99
0-91
0-95
36-99
0-73
62-99
0-70
Range of %
Removal of System
48-73
54795
79-99
5-57
34-99
0-98
40-99
33-91
0-95
89-99
0-88
85-99
0-80
-------
Make up Water
to Final Rinse
en
UHrafiltration System
Water for
Reuse
Discharge
Figure 1. Generalized process flow schematic.
-------
A generalized flow schematic of this treatment process is given in
Figure 1. Laundry wastes flow into a sump and are transferred through a
strainer to the ultrafiltration system. The strainer removes coarse solids
and lint that could plug pumps, valves, and controllers. The strained
feed stream is recirculated within the UF membrane loop. A small concentrate
stream is continually bled off, and a permeate stream, essentially free of
suspended solids, is continuously produced. The permeate is passed
directly into a column packed with granular activated carbon. The carbon-
column effluent is potentially suitable for reuse within the laundry.
Features of the ultrafiltration/activated carbon treatment approach are:
Ultrafiltration is largely insensitive to waste shock
loads. Since the bulk of the contaminant removal is
in the Ultrafiltration step, the overall process will
be shock insensitive.
No chemicals are added. This eliminates both chemical
costs and chemical handling equipment. Also,
operating and maintenance labor requirements associated
with chemical addition are eliminated. Membrane systems
will operate successfully in the pH range 0.5 to 13
and in the presence of free chlorine (up to 50 ppm).
The effluent from the system will be essentially free
of suspended solids and will have a low organic content.
With regard to these contaminants, a reusable water will
be available.
Dissolved inorganics will tend to build up in the system
to a steady state level. This will be established when
the rate of addition of salts in makeup water, detergent
formulations, and soiled articles equals the rate of
removal in the Ultrafiltration concentrate. It may be
desirable to reformulate detergents to have a lower
builder content and/or to demineralize the makeup water.
Heavy metals (except hexavalent chromium and mercury)
may be efficiently removed in the Ultrafiltration
step. This is because the metals may be insolubilized
by reaction with anionic detergents and do not pass
through the Ultrafiltration membrane.
There is no need to cool the waste before treatment. The
hotter the effluent the better the Ultrafiltration system
will perform at the temperature levels encountered in
laundry wastewaters. Temperature will have only a
minimal effect on carbon adsorption efficiency. Thus the
overall process will conserve the sensible heat of the
water for reuse, providing a major process credit.
A total package system will be compact and easy to install.
-------
SECTION 2
CONCLUSIONS
This study of industrial laundry wastewater treatment by ultrafiltration
and activated carbon adsorption has indicated that a consistently high
quality product water, potentially reusable within the laundry, can be
produced. However, the performance of the spiral-wound ultrafiltration
modules was unacceptable due to membrane fouling by oil and grease and the
ineffectiveness of standard membrane cleaning procedures. Average module
permeate flux was therefore low. This factor, in turn, resulted in high
capital and operating cost estimates for full-scale treatment systems.
Specific conclusions reached during this program are:
1. Contaminant Removal Efficiencies
- The overall UF/ACA product water averaged <17 mg/1
suspended solids, 190 mg/1 BOD, 353 mg/1 COD, 123 mg/1
TOC, and <9 mg/1 total freon extractives. An effluent
of this quality indicates average removal efficiencies
for the treatment process of >96% for suspended solids, >97%
for freon extractives (oil and grease) and 82%, 86%, and 82%
for BOD, COD, and TOC, respectively. An effluent of this
quality should be acceptable for discharge to municipal
sewer systems. Based on local ordinances, a surcharge
may be applied if the mass discharge of BOD and
suspended solids exceed acceptable limits.
- Metals removal by the UF/ACA process was, in general,
calculated from the lower detection limit values of the
assays. All metals of interest, except mercury, were
removed to levels below those specified in the sewer
ordinances of the Metropolitan Sanitary District (MSD)
of Greater Chicago.
2. Membrane Flux
- Severe membrane fouling by free oil occurred for
the three types of spiral-wound modules evaluated. A
variety of membrane cleaning techniques was largely
ineffective in restoring membrane flux.
- Of the three spiral-wound module feed-side spacers and
two UF membrane materials investigated, the best
performance was obtained with Abcor, Inc. Type HFM
-------
membranes in open-spacer spiral-wound modules. For
these, average permeate flux was 14 gal/sq. ft.-day
(gfd) [135°F] during eight weeks of tests at system
conversions of 67% to >99%. A flux level of this
magnitude is not economically acceptable for treatment
of industrial laundry wastewaters.
- It is doubtful if a spiral-wound membrane configuration
can withstand prolonged operation in an industrial
laundry. A membrane configuration, which is less
susceptible to fouling and more amenable to cleaning,
e.g., tubular, is clearly required.
3. Estimated Process Costs
- Assuming stable performance could be achieved at an
average flux of 14,gfd, spiral-wound UF system and
pretreatment operating costs are estimated to range
from $11.82/1000 gal when processing 25,000 gpd to
$9.01/1000 gal when processing 100,000 gpd.
- The activated carbon's adsorptive capacity for color
bodies is greater than its adsorptive capacity for
TOC. Breakthrough curves developed for color indicate
a carbon replacement cost of $0.85/1000 gal.
8
-------
SECTION 3
RECOMMENDATIONS
On the basis of the knowledge gained during this program, the following
recommendations for future work are offered:
- No further ultrafiltration work should be conducted
with spiral-wound membrane modules.
- The flux performance of tubular UF assemblies should
be evaluated to verify or amend the capital and operating
cost projections developed for the tubular configuration.
If a flux rate of 40 gfd is assumed for tubular UF
modules rather than the 14,gfd observed for spiral-
wound modules, estimated operating costs become $4.88,
$2.82, and $2.57/1000 gal for'treatment systems of
25,000, 75,000, and 100,000 gpd, respectively. For
the 100,000 gpd system, operating costs for tubular
systems are predicted to be $2.39/1000 gal because of
advances in membrane technology expected in the next
2 to 4.years.
- In conjunction with a tubular UF evaluation, an engineer-
ing and economic survey of UF concentrate disposal
options should be performed. For higher capacity UF
systems, concentrate disposal by contract hauling was
calculated to be as high as 17% of the UF system
operating costs. Significant reductions in this cost
could greatly enhance the economics of the entire
treatment system.
- Should tubular UF prove to be economically viable, the
suitability of the final UF/ACA product water for reuse
must be evaluated.
-------
SECTION 4.
DISCUSSION OF UNIT PROCESSES
The purpose of this section is to set forth certain principles and
definitions which will be used in subsequent sections. Many general
references are available which describe the relevant unit processes in
more detail.
ULTRAFILTRATION
Ultrafiltration and reverse osmosis (RO) are similar processes in that
bqth employ a semi permeable membrane as the separating agent and pressure
as the driving force to achieve separation. There are important differences,
however, which lead to different applications, process conditions and
equipment for each of the two processes. Although the approach in this
program is based on Ultrafiltration, the differences between UF and RO are
presented in Table 3 to aid reader understanding of the subject matter.
In an ul.trafiltration process a feed solution/suspension is introduced
into a membrane unit, where water and certain solutes pass through the
membrane under an applied hydrostatic pressure. Solutes whose sizes are
greater than the pore size of the membrane and all suspended solids are
retained and concentrated. The pore structure of the membrane thus acts
as a molecular filter, passing smaller size solutes and retaining the
larger size solutes. The pore structure of this molecular filter is such
that it does not become plugged because suspended solids are rejected at
the surface and do not penetrate the membrane.
For solutions which have no rejected species, such as water, the flux
through the membrane is given by:
J = AP 0)
0 Rm + Rf
where,
JQ = Flux rate (gal/ft2-day)
AP = Pressure drop across the membrane (pressure
driving force) (psig)
R = Resistance of clean membrane (ft2-day-psig/gal)
in
Rf = Resistance of fouling layer (ft -day-psig/gal)
10
-------
TABLE 3. DIFFERENCES BETWEEN REVERSE OSMOSIS AND ULTRAFILTRATION
I tern
Reverse Osmosis
Ultrafiltration
Size of solute retained
Osmotic pressures of feed
solutions
_ Operating pressures
Nature of membrane retention
Chemical nature of membrane
Molecular weights generally
less than 500
High salt retention
Important, can range to over
1000 psig
Greater than 400 psig, up to
2000 psig
Diffusive transport barrier;
possibly molecular screening
Important in affecting trans-
port properties
Molecular weights generally
over 1000
Nil salt retention
Negligible
10 to 100 psig
Molecular screening
Unimportant in affecting transport
properties so long as proper pore
size and pore size distribution are
obtained
Typical membrane flux levels 2 to 15 gal/day-ft'
20 to 200 gal/day-ft'
-------
No material from the process stream builds up on the membrane surface and,
therefore, for water the flux is pressure dependent and flow independent.
When ultrafiltering solutions having high concentrations of
rejected material, the observed flux levels are much lower than the water
flux of the clean membrane. A gel layer develops and the following
equation applies:
C
J, = AQX In ,£ (2)
1 Lb
where,
J, = Flux rate
A = A constant which is a function of feed channel
dimensions and fluid properties
C = Concentration of rejected species in the gel layer
C. = Concentration of rejected species in the bulk solution
Q = Circulation rate of fluid through the membrane'
modules (gpm)
x = Empirical constant (generally 1
-------
Four module configurations are available commercially to house the
ultrafiltration membranes. These are plate-and-frame, tubular, spiral -
wound, and hollow-fine-fiber (tubeside feed) geometries. For treatment of
industrial laundry wastes the tubular and spiral-wound geometries are
judged to be most suitable in terms of both process reliability and system
expense.
A tubular membrane, as shown in Figure 2, consists of a porous tubular
support with the membrane either cast in place, or inserted into the
support tube. The feed solution is pumped through the tube; the concentrate
is removed downstream; and the permeate passes through the membrane/
porous support composite.
A spiral-wound module is shown in Figure 3. It consists of a large
membrane sheet(s) wound into a compact spiral configuration around a central
permeate collector tube. The feed solution is passed over one side of the
sheet, and the permeate is withdrawn from the other side.
Each module configuration has particular advantages and disadvantages
which are summarized in Table 4, Tubular membranes are desirable in that
they can process feeds containing high suspended solids with minimal
pretreatment, and can be easily cleaned, either chemically or mechanically,
if they become fouled. Spiral-wound modules are less expensive than tubular
modules in dollars/sq. ft of membrane area, generally have lower power
requirements, and are more compact. Spiral-wound modules are, however,
more susceptible to plugging and may be difficult to clean. Because of their
potential cost savings, spiral-wound modules were chosen for testing in
this program.
TABLE 4. COMPARISON OF MEMBRANE CONFIGURATIONS OF INTEREST
FOR INDUSTRIAL LAUNDRY WASTE TREATMENT SYSTEMS
Configuration
Advantages
Disadvantages
Tubular
Spiral-wound
1. easily cleaned chemically
or mechanically if
membranes become fouled
2. can process dirty feeds
with minimal pretreat-
ment
3. individual tubes can
be replaced
1. compact-good membrane
surface to volume ratio
2. less expensive than
tubular membranes
1. high holding volume
required per unit
membrane area
2. moderately expensive
at present
1. susceptible to plugging
by particulates
2. badly fouled membranes
are difficult to clean
13
-------
FEED
CONCENTRATE
MEMBRANE
PERMEATE
PVC END FITTING
DOUBLE O RING
SEAL
EPOXY RE-INFORCED
FIBERGLASS BACKING
Figure 2. Cut-away view of tubular ultrafiltration assembly.
-------
FEED SOLUTION
FEED SOLUTION
MODULE SEAL (SEALS AGAINST THE INSIDE WALL OF A PRESSURE
f VESSEL TO FORCE THE FEED SOLUTION THROUGH THE MODULE)
PERMEATE COLLECTION HOLES
CONCENTRATE
PERMEATE OUT
CONCENTRATE
FEED FLOW
ACROSS FEED
CHANNEL SPACER
FEED CHANNEL
SPACER
MEMBRANE
PERMEATE COLLECTION
MATERIAL
MEMBRANE
FEED CHANNEL
SPACER
PERMEATE FLOW (AFTER PASSAGE
THROUGH MEMBRANE INTO PERMEATE
COLLECTION MATERIAL)
COVERING
'ADHESION LINE
Figure 3. Cut-away view of spiral-wound ultrafiltration module .
-------
Three different types of spiral-wound modules are currently available. The
difference between them is the spacer geometry employed on the feed flow
channel side. The module depicted in Figure 3 utilizes a Vexar spacer- The
Vexar material is a T mm (nominal) turbulence promoting spacer. Corrugated
spacer material consists of 2 mm (nominal) corrugation. Open-spacer feed
channels have a 4,mm (nominal) mesh-like geometry similar to the Vexar
design.
Of the three spacer geometries, the narrow Vexar spacing is the most
susceptible to plugging by suspended solids, while the open-spacer material
is the least susceptible to plugging. Should one passage within a Vexar or
open-spacer module become plugged, the feed solution can bypass that passage
and little effective membrane area is lost. However, with corrugated spacer
modules, once a feed channel becomes plugged the entire membrane area along
that channel is lost.
One significant advantage of Vexar spacer modules over the corrugated
and open-spacer modules is their lower power requirement. Vexarmodules
operate with feed circulation rates of 10 to 20 gpm, while the other spiral-
wound module types require 50 to 100 gpm circulation flow.
The three different type spacers produce different feed flow patterns
within a module thus influencing ultrafiltrate flux. Also, the width of the
spacer material controls the membrane surface area per module. For
commercial-scale modules (4rinch diameter x 36 inches long) the membrane
surface area is 40±2 sq. ft. for the Vexar spacer spirals, 34±2 sq. ft.
for the corrugated spacer spirals and 18 ±2 sq. ft. for the open-spacer
spirals.
SYSTEM DESIGNS FOR ULTRAFILTRATION EQUIPMENT
Three common ultrafiltration system designs are shown in Figure 4, In
the batch concentration mode of operation (Figure 4a), the feed tank is
charged with waste only at the beginning of each concentration cycle.
During operation the permeate is continuously withdrawn while the concentrate
is recycled to the feed tank. As the run proceeds the volume of waste in the
feed tank decreases, and its concentration increases. When the volume of
waste is sufficiently low, it is discharged and a fresh batch of waste is
charged to the feed tank. The degree of volumetric concentration is given
by
Vo
cv= v^nr w
where V0 and Vp are the initial batch volume and the collected permeate
volume, respectively. The degree of volumetric concentration, Cv, is
related to the overall water recovery, by the relationship
Y (%) = (1 - f- ) 100 (5)
16
-------
Intermittent
Addition of
Waste
Intermittent
Blowdown of
Residue
(a) Batch Concentration
Pump
Concentrate
Permeate
UF Membrane
Module
Concentrate
(b) Continuous Feed and Bleed
Concentrate
Waste N
Recycle
/ Fppd
^4
*--
-*
X"
.*-
?
Perm pa tp -.
^
Concentrate
(c) Continuous Staged; Once-Through
Figure 4, Various system designs for modular membrane equipment.
17
-------
Corresponding values of the volumetric feed concentration and the system
water recovery are shown below.
P Equivalent Water
_y_ Recovery (%)
IX (Vp = 0) 0
2X 50
10X 90
20X 95
50X 98
100X 99
There are three advantages to the batch concentration mode of
operation:
1. Feed circulation rate within the modules can be adjusted to
control membrane fouling and/or concentration polarization.
2. High system conversions can be obtained by concentrating to
a very low residual volume.
3. The average feed concentration over the batch concentration is
minimized (compared to other modes of operation) resulting in
a maximum time-averaged module flux and rejection over the
concentration cycle.
The disadvantages of this mode relate to its intermittent nature of
operation. Since it is not continuous, it requires large holding tank
capacity and somewhat more operator time than the other modes.
The continuous feed and bleed mode of operation is shown in Figure 4b.
The advantages of this mode are:
1. It is continuous.
2. Feed circulation can be adjusted to control concentration
polarization.
3. High system conversions can be obtained.
The disadvantages of this mode is that the system is operated at the
concentration level of the concentrate stream. Thus, the average flux and
rejection will be low relative to that of the batch concentration mode.
For sufficiently large systems continuous once-through operation,
shown in Figure 4c, is preferred. This mode combines the advantages of
both the batch and the feed-and-bleed modes of operation. The feed passes
through each module in a single-pass which minimizes the average feed
concentration and achieves maximum utilization of the modules in terms of
flux and rejection. In this mode, operation is continuous and a high over-
all system conversion can be obtained.
18
-------
The preferred mode of operation for any given application may be a
modified form of one of these three more common modes. The operating mode
selection depends upon UF feed flow conditions, membrane flux, water
recovery desired, membrane cleaning frequency, etc.
CARBON ADSORPTION
Adsorption by activated carbon is a surface phenomenon in which
dissolved organics are removed from wastewater and concentrated at the carbon-
liquid interface. The degree of adsorption which occurs is a combination
of solute solubility in the wastewater and the strength of the attractive
forces between the solute and the carbon. The more hydrophilic the organic,
the less likely it is to move toward the carbon-water interface. Thus,
highly soluble organics tend to be poorly adsorbed by carbon; whereas less
solute organics are more highly adsorbed.
Activated carbon is a highly porous material which is characterized by
a typical surface area yield of 1000 m2/gram. Since adsorption is a surface
phenomenon, activated carbon has the potential (depending on the nature of
the dissolved organics) to be a highly-effective, economical unit process
for improving water quality.
There are two important factors in the evaluation of carbon adsorption:
the amount of organic adsorbed at equilibrium and the rate at which
equilibrium is attained. The equilibrium uptake is usually expressed by an
"adsorption isotherm". The isotherm is a plot of the weight of organic
adsorbed per unit weight of carbon (X/M) versus the organic concentration in
the waste (C) when equilibrium is established at a constant temperature.
Three types of hypothetical isotherms are shown in Figure 5a. Type II is a
linear isotherm over the region of interest. Type I is a favorable isotherm
since the adsorptive capacity of the carbon remains high at low feed
concentrations. This results in lower carbon dosages for a given organic
removal. Type III is an unfavorable isotherm since uptake at a given
concentration is low requiring a high carbon dosage for a given organic
removal.
A number of mathematical expressions have been proposed (3) to describe
the shape of the isotherm. The most generally applicable expression is the
Freundlich adsorption equation:
(6)
where,
X = amount of organic adsorbed
M = weight of carbon
k = constant
n = constant
C = concentration of unadsorbed organic in surrounding
solution at equilibrium
19
-------
Concentration of Organic at Equilibrium
a) Types of Adsorption Isotherms
Ul
c
E
o
CJ
10
i.
4->
-------
Restating this equation in logarithmic form,
J) = log k + (1) log C (7)
A plot of X/M vs. C on logarithmic paper will yield a straight line with
slope 1/n if the Freundlich isotherm is followed.
Little progress has been made for liquid systems in predicting the
isotherm from the properties of the carbon and organic. Therefore, iso-
therms must be determined experimentally for each waste-carbon combination.
The Freundlich expression given by Equations (6) and (7) is useful for
correlating the experimental data.
In a practical system, the amount of contaminant removed may be limited
by the rate of adsorption, in addition to the equilibrium uptake. The
transport of a dissolved organic molecule from the bulk solution to attach-
ment at an adsorption site on the carbon can be broken into three
consecutive steps:
-- diffusion through the liquid film surrounding the
carbon particle,
-- diffusion through the carbon pores to an available
adsorption site, and
-- adsorption.
In principle any of these sequential steps could limit the observed rate
of adsorption. However, in practice, adsorption is very fast relative to
diffusion. For fixed bed contactors operated at normal velocities, film
diffusion is generally the rate limiting step (4).
The effect of adsorption rate on the performance of a carbon column
is to govern the shape of the breakthrough curve. The breakthrough curve
is a plot of the concentration of contaminant in the column effluent as a
function of the volume of waste treated. Two hypothetical breakthrough
curves are shown in Figure 5b. The favorable Type A curve makes much
better use of the adsorptive capacity of the carbon before the quality of
the effluent exceeds the maximum allowed concentration (Cmax). A sharp
breakthrough curve is obtained when the rates of film and pore diffusion
are fast relative to the linear velocity of the waste through the column.
Generalized correlations are available (5,6) for the prediction of the
shape of the breakthrough curve. Factors influencing the shape include:
linear velocity of the waste, diameter of carbon particles, film diffusion
coefficient, pore diffusion coefficient, and isotherm shape. For complex
wastes, such as industrial laundry wastes, the breakthrough curve must be
determined experimentally.
21
-------
Both the adsorption rate and capacity of the carbon can be adversely
affected by the presence of participates in the feed. These tend to coat
the carbon particles reducing access to the internal porous structure.
Therefore, pretreatment (usually by depth filtration) is required for
suspended solids removal.
For system capacities of interest to industrial launderers carbon
regeneration would be performed off-site.
22
-------
SECTION 5
PROGRAM OVERVIEW
The program for evaluating ultra-filtration and carbon adsorption
treatment of industrial laundry wastewaters was divided into three program
tasks:
Task 1: Pilot-scale Testing at Abcor, Inc.'s Facilities
Task 2: Pilot-scale Testing at a Field Demonstration
Site
Task 3: Economic Analysis of Full-scale Treatment
Systems
All UF experiments were performed with spiral-wound modules manufactured by
Abcor, Inc. Carbon adsorption testing was conducted with Filtrasorb 400,
a general purpose granular activated carbon, produced by Calgon Corporation.
In Task 1 three actual industrial laundry wastewaters, each from a
different source and each representing a different strength industrial load,
were processed by ultrafiltration modules at Abcor1s pilot laboratory. Two
membrane types and two module feed-side spacers were evaluated to select
the preferred membrane/module combination for field testing. Composite
samples of the ultrafiltrates were further treated with granular activated
carbon in isotherm experiments.
The Task 2 field demonstration was conducted at Standard Uniform
Rental Service, Dorchester, Massachusetts. The average wastewater from
this plant was considered to be of a "light" to "medium" industrial loading^
initial work under Task 2 consisted of test system design, construction,
installation, and shakedown testing. Several preliminary experiments and
four two-week concentration cycles were performed with the 5000 gpd
(nominal) test system. A slip stream of UF permeate was continuously drawn
off and passed through a two-inch diameter column for ACA treatment.
Based on the results of the Task 2 studies, the economics of full-
scale treatment systems were analyzed. This third program task developed
capital and operating cost estimates for industrial laundries discharging
25,000, 75,000 and 100,000 gpd.
23
-------
SECTION 6
TEST SYSTEMS, PROCEDURES, AND ANALYSES
ULTRAFILTRATION
Test System
The basic operation of the UF test system employed during the in-
house tests is similar to that of the field demonstration test system;
therefore, only the latter will be described. The field test system,
including the carbon column, is shown schematically in Figure 6. Plant
wastewater was collected in an existing sump and delivered to the UF feed
tank (Tl) by a transfer pump. The suction line of the pump was connected
to a float and withdrew liquid from ~15 inches below the surface,
minimizing free oil content in the feed. The transfer pump was equipped
with a low pressure switch and a bypass loop. A Y-strainer on the dis-
charge line removed gross solids.
The plant wastewater was fed into the UF feed tank (Tl) as
signalled by a high/low level switch mounted within the tank. The average
volume in Tl was maintained at 75 gallons. A polyurethane foam partition,
3 inches in thickness, was positioned vertically within the feed tank.
This partition was not originally incorporated into the test unit; it was
devised during preliminary testing when lint breakthrough to the UF
modules proved troublesome. Besides effectively trapping lint particles
within its pore structure, it also served as an oil coalescer.
Feed was withdrawn from the side of Tl and passed by a booster pump
through one of two parallel 100 mesh stainless steel basket strainers.
A differential pressure switch measured the pressure drop across the
strainer in use and indicated when the second strainer should be brought
on-line and the first strainer cleaned. A check valve prevented any back
pressure at the strainer outlet.
The feed entered the suction of a centrifugal circulation pump and
was delivered in parallel to the inlet of the two module housings. The
circulation pump was equipped with a low pressure switch/alarm (IPS) to
protect it against running dry and a temperature indicator/alarm to
protect the system against excessive temperatures.
24
-------
Concentrate Bl eed
Plant Sump
UHraflltration
Membranes
TOO Mesh S.S.
Partition Basket Strainers
ro
01
To Sewer
Transfer
Pump
Sample stations
Differential pressure switch/alarm
Drain valve
Flow indicator
Flow totalizer
High/low level controller
Low pressure switch
Pressure gauge
Solenoid Valve
Tank
Temperature indicator/controller
Timer
Carbon
Col umn
Figure 6. Flow schematic of field demonstration test system.
-------
Each module housing was equipped with two spiral-wound UF
cartridges in a series arrangement. The type of spiral-wound modules
within each housing is identified in the Results and Discussion Section.
The feed flow through each housing was indicated and could be controlled
by butterfly valves at the housing inlets and by a ball valve on the by-
pass loop. Pressures before and after each housing were measured. The
main portion of the concentrated waste was returned to the suction of the
circulation pump while the remainder was bled from the system to maintain
the desired conversion. When very high conversions (i.e., very low
concentrate bleed flows) were being investigated a timer controlled the
operation of a solenoid valve to maintain the proper concentrate dis-
charge.
The UF permeate flow rate from each housing was measured before the
permeates were combined. The overall permeate pressure was indicated
and a flow totalizer on the permeate line recorded the cumulative volume
of permeate produced.
The UF permeate flowed into a 50 gallon surge tank (T2) which served
as the carbon column feed reservoir. The overflow from T2 was delivered
to holding tank T3 and pumped to the sewer.
A metering pump provided a constant feed flow through the carbon
columns which were operated in the upflow mode. The feed pressures at
the inlet and outlet of the carbon columns were measured. The carbon
effluent flowed into a four gallon surge tank (T4) which also overflowed
to T3.
Test Procedures
Task 1 Experiments--
Task 1 in-house tests were conducted on samples (250 gallons each)
from the wastewater sumps of three different industrial laundries. Prior
to discharge of wastewater on the day of sampling the waste collection
pits at each site were drained. Also, all accumulated sediment was
removed from the pit before sampling at one of the three sampling sites
("light" loading wastewater). The 250 gallon sample was returned to
Abcor and pumped into a 500 gallon holding tank for next-day processing.
The wastewaters were processed in the batch concentration mode (i.e.,
concentrate returned to the feed tank and permeate continuously withdrawn).
The UF module pressure drop and the UF permeate flux vs. feed concentration
and time were monitored. Initial feed, final concentrate and composite
permeate samples, were analyzed for a number of constituents (see below).
Operating conditions for these tests were:
Feed Circulation Rate
Open-Spacer Modules: 60-95 gpm
Corrugated Spacer Modules: 45-90 gpm
26
-------
Inlet Pressure: 45-50 psig
Feed Temperature: 125°F
Feed pH: Actual
Upon completion of each batch concentration experiment, proportionate
amounts of UF permeate and concentrate were combined to provide a 5X feed
for total recycle processing by the UF system. A 5X feed volumetric
concentration (80% conversion) corresponds to the average feed concentration
within the membrane loop during an entire batch concentration to a
volumetric feed concentration of 50X. In experiments of this type, both
the concentrate and permeate are returned to the feed tank. Thus, the feed
concentration remains the same throughout the test allowing permeate flux
to be monitored as a function of time. The total recycle tests were
conducted for ten day periods and feed and permeate samples were collected
on the first and tenth days.
Task 2 Experiments--
Field experiments were conducted with the UF system operated in the
feed-and-bleed mode. The UF system was maintained at each of four
different system conversions for a period of two weeks. System operation
was continuous, 24-hours/day.
At the start of each experiment, the UF feed tank was filled with
laundry waste and the system was started. Until the desired system
conversion was reached all concentrate was returned to the suction of the
circulation pump. When the concentration within the membrane loop reached
the proper level, the concentrate bleed timer was engaged and concentrate
was periodically ejected through a solenoid valve to drain. The amount of
concentrate removed was controlled by the timer which was set in proportion
to the total permeate flow, maintaining the proper system conversion.
Task 2 UF system operating conditions were:
Feed Circulation Rate
Open-Spacer Modules: 45-90 gpm
Corrugated Spacer Modules: 55-75 gpm
Vexar Spacer Modules: 15-20 gpm
Inlet Pressure: 45-60 psig
Feed Temperature: 135°F
Feed pH: Actual
UF permeate flux and the standard operating data were monitored daily.
During each 247hour period composite samples of the feed, ultrafiltrate,
and carbon effluent were collected using a metering pump. In addition, a
grab sample of UF concentrate was taken each day. Corresponding daily
samples were combined to form a weekly composite sample for analysis.
Additional assays were performed on "two-week" composites.
27
-------
UF Membrane Cleaning--
Various chemical formulations were employed during the UF membrane
detergent cleaning cycles. These formulations and the entire cleaning
operation are discussed in more detail in Section 7. A generalized
procedure for cleaning the spiral-wound UF membranes is presented below.
1. The concentrated waste was drained from the system.
2. Clean water was passed through the system at a low
flowrate (-20 gpm) to flush out the residual
concentrate.
3. A 1%, by weight, solution of Abcor, Inc.'s "Ultra-clean"
was recirculated through the system for 30 minutes
under the following operating conditions:
Recirculation Flowrate: 20-25 gpm
Inlet Pressure: 20-25 psig
Temperature: 115-120°F
The "Ultra-clean" was dissolved in warm water then
added to tap water being circulated through the
system under the proper operating conditions.
4, Clean water was passed through the system for 20-30
minutes at low flow and low pressure to flush out all
traces of the detergent.
5. The water flux of the clean membranes was determined.
CARBON ADSORPTION
Isotherm Procedures
The carbon adsorption isotherm tests were conducted using the following
procedure.
1. Filtrasorb 400 granular activated carbon was ground
with a mortar and pestle and sifted through a 335 mesh
screen.
2. Seven samples of dried carbon were weighted out:
2 mg, 5 mg, 10 mg, 20 mg, 50 mg, 100 mg, and 500 mg.
3. Each sample of dried carbon was placed in a separate
erlenmeyer flask.
4. 100 (±1) ml of ultrafiltrate were added to each
flask.
5. The flasks were stoppered and placed on a Burrel
Wrist Action Shaker for 24-48 hours.
6. The flask contents were filtered through a 0.22 micron
Millipore Filter, and the center portion of filtrate
was collected for analysis.
28
-------
7- The seven carbon treated samples, an original feed
sample taken through all procedures except for carbon
addition, an original feed sample not taken through
the isotherm procedures, and a high purity water sample
are analyzed. Isotherms were developed for both TOC
and color.
Carbon Column Operation
A slip stream of UF permeate was continuously fed into a 2-inch diameter
carbon column during the field demonstration tests. A flow diagram showing
the carbon column's operation is given in Figure 7. The column contained
2400 grams of Filtrasorb 400 (Calgon Corp) granular activated carbon and had
a residence time of 11.3 minutes. The column was operated in an upflow mode
at a flowrate of 6.7 gpm/ft2 (480 cc/min). The carbon effluent flowed into
a 4-gallon surge tank from which a composite sample was continuously with-
drawn. Excess liquid overflowed to drain.
ANALYSES
Table 5 lists the assays routinely performed during both the in-house
and field demonstration experiments. Also indicated are the methods employed
during each analysis. All assays, except'mercury, were performed by Abcor's
Analytical Laboratory. The mercury analyses were conducted by Environmental
Research and Technology, Concord, Massachusetts.
29
-------
Pressure
Indicator
co
o
UF Permeate
Surge
Tank
Overflow to drain
Carbon Columns
Metering
Pump
To
Composite
Sampl er
Figure 7. Flow schematic of carbon column.
-------
TABLE 5. CHEMICAL ANALYSES ROUTINELY PERFORMED DURING
IN-HOUSE AND FIELD DEMONSTRATION EXPERIMENTS
Constituent
Assay Method
Reference
Alkalinity as CaCO,
BOD
Cadmi urn
Chromium
COD
Color
Iron
Lead
Mercury
Nickel
pH
Suspended Solids
TOC
Total Freon Extractives
Total Solids
Turbidity
Zinc
HC1 Titration
5 Day Incubation, Electrode
Atomic Absorption
Atomic Absorption
Dichromate Reflux
Visual Comparison
Atomic Absorption
Atomic Absorption
Atomic Absorption
Atomic Absorption
Meter Reading
Glass Fiber Filtration
Combustion-Methane Detection
Separatory Funnel Extraction
Gravimetric
Meter Reading
Atomic Absorption
**
SM 403
SM 507, 422F, 422B
SM 301A
SM 301A
SM 508; EPA, p. 21
SM 204A
SM 301A
SM 301A
SM 301A VI
SM 301A
Manufacturer's
Manual
SM 208D
EPA, p. 236
SM 502A, EPA, p. 229
SM 208A
SM 214A
SM 301A
SM 403 (etc) refers to procedure number in "Standard Methods for the
Examination of Water and Wastewater," 14th Edition, APHA, 1975.
**
EPA refers to "Manual of Methods for Chemical Analysis of Water and Wastes,"
U.S. EPA, 1974.
31
-------
SECTION 7
RESULTS AND DISCUSSION
PILOT SCALE TESTING AT ABCOR INC.'S FACILITIES
Introduction
The Task 1 tests were designed to determine whether spiral-wound UF
processing of industrial laundry wastewaters is feasible, and if so, to
determine the preferred membrane type and module spacer. Each of the three
laundry wastewaters tested was of a different industrial strength and was
classified by the Institute of Industrial Launderers as either a light,
medium, or heavy industrial load.
Four membrane/spacer combinations were tested during these in-house
experiments. These combinations are summarized below:
Membrane Type Spacer Configuration
Abcor, Inc. Type HFD Corrugated
Abcor, Inc. Type HFD Open-Spacer
Abcor, Inc. Type HFM Corrugated
Abcor, Inc. Type HFM Open-Spacer
Vexar spacer modules were not tested since at this stage of the program it
was felt the very narrow channel Vexar spacer would be plugged by lint.
Washroom production schedules for the time periods during which
sampling occurred are given in Appendix A.
Ultrafiltration
UF Membrane Flux--
Average UF membrane flux levels during the batch pumpdown and total
recycle experiments are given in Table 6 (the permeate flux vs. time curves
for each test are presented in Figures Al through A6 in Appendix A). These
data indicate slightly better flux performance for the Type HFM membrane
as opposed to the Type HFD membrane.
32
-------
TABLE 6.- AVERAGE UF MEMBRANE ^FLUX LEVELS DURING IN-HOUSE
EXPERIMENTS (Values given in gfd @ 125°F)
Membrane/Spacer
Wastewater
Loadi n.g-
Light
Medium
Operating
Mode
Batch
Total Recycle
Batch
Total Recycle
HFD
Corrugated
30
17
35
25
HFD
Open
t
t
t
t
Combination*
HFM
Corrugated
t
t
f
t
HFM
Open
40
38
30
27 tt
Comments
Final Concentration Reached =
5X feed diluted to 3X after 2
due to loss of feed.
Final Concentration Reached =
Modules fouled by free 'oil,
13. 6X
days
17. IX
oo
co
Heavy Batch
t
Total Recycle
45-50
40
45-50
45
test terminated after 2 days
Final Concentration Reached = 11X
Feed Circulation Rates 90-95 gpm
Free oil removed
Test duration 14 days
No module cleaning performed between
batch pumpdown and total recycle
tests.
* Two membrane/spacer combinations used during each experiment.
t No tests made in these instances.
tt As noted in comments, run was of a short duration. See Figure A4 for flux vs. time curve.
-------
The final volumetric concentration achieved was limited by the sample
size and the dead volume of the UF system. For the three batch pumpdowns,
with an average final concentration factor of 13.9X (92.8% water recovery)
the HFM membrane modules averaged -40 gfd (125°F) and the HFD membrane
modules averaged -38 gfd (1250F). During the total recycle experiments,
however, the average flux levels were 37 gfd (125°F) for the HFM modules and
27 gfd (1250F) for the HFD modules.
The duration of the total recycle experiments with the "light" and
"heavy" loading waste streams was 10 days. The total recycle test with the
"medium" loading wastewater was curtailed after two days due to a sharp
decline in the permeate flux for both modules. A review of membrane module
performance before this flux decline and during subsequent operations is
presented in Table 7. As the data of Table 7 indicate, the standard cleaning
cycle with "Ultra-clean" recovered the flux for both membrane modules to
acceptable levels following the batch pumpdown. When the low membrane flux
levels were observed during this total recycle test, a second washing with
"Ultra-clean" was performed, resulting in flux recoveries to 47 gfd (69% of
initial flux) for the HFM module and 21 gfd (29% of initial flux) for the
HFD module. These flux levels were not considered acceptable.
"Dishmate" (Calgon Corporation), a detergent containing free-available-
chlorine, was used during the next cleaning cycle. Because of the
susceptibility of the HFD membrane to chemical attack by free chlorine, only
the HFM membrane module could be cleaned with "Dishmate". The resultant HFM
water flux was 76 gfd (112% of initial flux). As discussed in the following
paragraph, the validity of this flux measurement is in question.
Since the HFD module remained fouled, a 10% kerosene in water mixture
was circulated through the modules. Little change,in water flux occurred
for either membrane type. The flux for the HFM membrane module was
acceptable, and therefore, the extended recycle experiment was reinitiated
with only the HFM open-spacer module. Upon system startup a leak was
detected in the HFM module which caused feed to enter the permeate stream.
Since it is not known precisely when this leak developed, the water flux
measurements for the HFM module following the last two cleaning cycles are
questionable.
The HFM module was removed from its housing, carefully cut open, and
unwound. A coating of free oil was observed on the membrane surface. The
leak in the module occurred at the interior glue seam where the membrane
and spacer material begin to wrap around the permeate collection tube. The
spacer material was inserted slightly askew during manufacture, placing
excess stress on the glue seam and causing it to fail. The module failure
was attributed entirely to this manufacturing defect.
It is suspected that the flux decline observed with the "medium" loading
sample was related to the free oil found in this wastewater. As the
ultrafiltration process proceeds, oil droplets may progressively adhere to
the membrane surface to form a water impervious coating. In a properly
34
-------
TABLE 7. HISTORY OF MEMBRANE MODULE PERFORMANCE DURING EXPERIMENTATION WITH
"MEDIUM" LOADING INDUSTRIAL LAUNDRY WASTEWATER SAMPLE *
Membrane/
Spacer Geometry
Event
Membrane R-ux
at T2:5°FV (gfd)
HFM/
Open-Spacer
co
in
Prior to batch pumpdown
Following 4 hour batch pumpdown, 10 minute "Ultra-clean"
and 50 minute "Ultra-clean"
Total recycle experiment at 5X, flux after 1 hour
Total recycle experiment at 5X, flux after 24 hours
%-hour "Ultra-clean"
Jj-hour "Dishmate"
%-hour 10% kerosene (by volume) in water; 10 minute
"Ultra-clean", 50 minute "Ultra-clean"
Leak developed in HFM membrane upon restart in total recycle,
membrane cut open for inspection
68.0
67.7
38.5
13.2
46.7
75.8
79.3
HFD/
Corrugated Spacer
Prior to batch pumpdown
Following 4 hour batch pumpdown, 10 minute "Ultra-clean" and
50 minute "Ultra-clean"
Total recycle experiment at 5X, flux after 1 hour
Total recycle experiment at 5X, flux after 24 hours
%-hour "Ultra-clean"
, 10% kerosene (by volume) in water; 10 minute
"Ultra-clean", 50 minute "Ultra-clean"
Subsequent cleaning postponed until completion of third
wastewater experimentation
73.0
65.2
36.4
9.9
21.2
29.2
* Similar data is not given for "light" and "heavy" loading wastewater samples
since routine cleaning methods effectively recovered membrane flux.
-------
designed system, free oil droplets would be removed from the feed prior to
its introduction to the membrane system. The treatment of the "light" and
"medium" loading wastewater was performed without the inclusion of oil
removal in an attempt to eliminate the oil separation step. The severe
fouling encountered during the total recycle test with the "medium" loading
wastewater suggests that this would not be a preferred mode of operation.
The removal of free oil from industrial laundry wastewater can be
accomplished by a number of commercially available oil skimmers. In lieu of
an oil skimmer, testing with the "heavy" loading wastewater sample was
conducted in the following manner. The wastewater sample was pumped into the
1500-gallon feed tank, and, one hour prior to processing by the UF system,
was well-agitated for 15 minutes. Feed was drawn into a 55-gallon feed tank
from the bottom of the 1500-gallon tank allowing free oil to remain floating
in the larger tank. The concentrate from the membrane loop was returned to
the 55-gallon tank, preventing any mixing of the free oil into the feed
stream. Comparison of the flux data for this test with the previous
experiments (see Table 6) shows the advantage of free oil flotation (or
skimming). The average flux for both membrane modules was 45-50 gfd (125°F)
while processing the wastewater with the heaviest loading.
Module Pressure Drop--
The two types of spacer materials used in constructing the spiral-wound
modules tested during the Task 1 experiments create different feed flow
patterns within each module. Variations in the feed transport through the
modules results in different pressure drops across each spiral and thus
significantly affects the UF system power requirement.
The power requirement for a UF system is determined almost entirely by
the power input to the feed circulation pump. This power input is directly
proportional to the product of the volumetric output of the pump and the
pressure drop across the membrane system. Table 8 presents the projected
horsepower requirements for each type of spiral-wound module evaluated from
the data obtained during the batch pumpdown and the total recycle experiments
with each industrial laundry wastewater. The horsepower requirement per
module was higher for the corrugated spacer module during all experiments
with the exception of the "medium" loading batch pumpdown. For this test the
open-spacer and corrugated spacer modules had essentially identical horse-
power requirements.
The horsepower requirement, per gallon of product per day, favors the
use of corrugated spacer modules because of their increased membrane surface
area over open-spacer modules. The corrugated flow channels are, however,
more susceptible to plugging than the open-spacer flow channels. Therefore,
preference for one feed-side spacer geometry over the other was riot clearly
demonstrated.
36
-------
co
TABLE 8. PROJECTED POWER REQUIREMENTS FOR UF SPIRAL-WOUND MODULES OPERATING
ON INDUSTRIAL LAUNDRY WASTEWATERS DURING IN-HOUSE EXPERIMENTS
Wastewater Spiral Wound Module Mode of Average Feed
Loading Spacer Geometry Operation Flowrate (gpm)
"Light" Corrugated
Open Mesh
Corrugated
Open Mesh
"Medium" Corrugated
Open Mesh
"Heavy" Corrugated
Open Mesh
Corrugated
Open Mesh
Catch Pumpdown
Batch Pumpdown
Total Recycle
Total Recycle
Batch Pumpdown
Batch Pumpdown
Batch Pumpdown
Batch Pumpdown
Total Recycle
Total Recycle
45
60
45
60
65
75
95
90
95
90
Average Pressure
Drop (ps1)
11
6
15
6
10
9
12
11
13
12
Projected
Horsepower .
per Module (hp)
0.29
0.21
0.39
0.21
0.38
0.39
0.66
0.57
0.71
0.62
Projected Horsepower
.Average Productivity per gallon of product
per Module (gpd) per Day (hp/gpd)
960
720
608
630
1,088
540
1,530
810
1,530
720
3.02 x TO'4
2.92 X 10~4
6.40 x 10"4
3.33 x ID"4
3.49 x 10"4
7.22 x 10"4
4.31 x 10~4
7.04 x TO"4
4.64 x 10"4
8. 61 x 10"4
-------
UF Membrane Flux Recovery--
The measurement of the flux of tap water through UF membranes, under
standardized conditions, indicates the water transport properties of the
membrane and is one means of detecting membrane degradation due to compaction,
plugging, biological fouling and/or chemical attack. This measurement is
always performed after membrane cleaning. Except for the one instance of
severe membrane fouling after exposure to free oil (the "medium" loading
wastewater test), reasonable membrane flux recoveries were obtained after up
to 330 hours of accumulated exposure to industrial laundry wastewater. Thus,
free oil must be removed prior to the ul trafiltration process so membrane
cleaning can be performed effectively using straight-forward, standardized
cleaning procedures.
UF Membrane Removal Efficiency--
Initial feed, composite permeate, and final concentrate samples were
taken during each batch pumpdown and analyzed for a wide range of contam-
inants. The composite permeate analyses were made on a mixed sample of the
total ultrafiltrate produced by both the HFD and HFM membranes. Detailed
analytical results for the "light", "medium", and "heavy" wastewaters are
presented in Tables A4, A5, and A6 of Appendix A. These results are
summarized, in terms of membrane removal efficiency, in Table 9.
Nearly complete removal of suspended solids and freon extractives
(i.e., oil and grease) was obtained for all three wastewaters. Overall
removal efficiencies for BOD ranged from 66.5%-88%, averaging 80.6%.
Composite permeate BOD analyses ranged from 360 mg/1 for the "light"
wastewater to 930 mg/1 for the "heavy" loading sample. In all three cases,
COD and TOC rejections of >80% were noted. It is thus apparent that a
significant contribution to the organic pollutant loading in industrial
laundry wastewater is associated with suspended matter. Based on the
individual test results, the UF rejection of metals (and metallic compounds)
generally ranged from >70 to 98% during the in-house experiments. Only the
rejection for mercury (20% for the "medium" loading waste and 11.1% for
the "heavy" loading sample) was low. The average mercury concentration in
both the initial feed and final composite permeate was, however, <0.001 mg/1.
Rejection of dissolved solids is not characteristic of ultrafiltration
membranes, and therefore, low removal efficiencies for dissolved solids and
alkalinity are expected. Post treatment by carbon absorption can be employed
to remove dissolved organic species.
Feed and individual permeate grab samples were also collected at the
beginning and end of the total recycle experiments. These data are presented
in Table 10. Due to the short duration of the "medium" loading wastewater
recycle experiment, no samples were analyzed for this run.
38
-------
TABLE 9. AVERAGE CONTAMINANT ANALYSES AND MEMBRANE REMOVAL EFFICIENCIES
FOR UF BATCH CONCENTRATIONS OF "LIGHT," "MEDIUM," AND "HEAVY"
LOADING INDUSTRIAL LAUNDRY WASTEWATERS
Assay
Total Solids (mg/1)
Suspended Solids
Average
Initial
Feed
4,910
1 ,960
Average
Final
Concentrate
32,870
19,400
Average
Mixed
Compos i te
Permeate
1,750
<4
Average
Removal
Efficiency, % *t
54.0
>99.6
(mg/1)
Dissolved Solids
(mg/1)
Turbidity (NTU)
BOD (mg/1)
TOC (mg/1)
COD (mg/1)
Freon Extractives
(mg/1)
Alkalinity
(ppm CaC03)
pH (units)
Chromium (mg/1 )
Copper (mg/1 )
Lead (mg/1 )
Zinc (mg/1)
Cadmium (mg/1 )
Iron (mg/1 )
Nickel (mg/1)
Mercury (mg/1)
2,950
ft
4,100
3,030
12,200
3,140
1,320
11.2
<3.3
4.6
9.3
4.8
0.77
37.8
<0.58
O.001
13,470
tt
37,200
22,900
96,400
22,600
2,320
11.2
1,746
4
614
347
1,280
25.2
1,230
11.4
<1.3
<0.7
<1.0
<0.29
<0.008
1.3
<0.5
<0.001
30.6
80.6
85.4
86.4
98.2
5.5
67.0
>73.0
>74.1
>85.7
>83.3
>92.2
>32.4
15.6
Removal Efficiency, r =
Feed Concentration-Composite Permeate Concentration
Feed Concentration
X 100
tAverage removal efficiency is based on individual test results and not on
average feed and permeate concentrations.
ttVery high.
NOTE: See Appendix A for detailed data from each experiment.
39
-------
As shown in Table 10, the permeate quality of the two membrane types
was quite similar. For the "light" loading wastewater the average removal
efficiency for total solids was 59% and for TOC, 84%. Ultra-filtration of
the "heavy" loading wastewater resulted in average removal efficiencies over
the ten day period of 92% and 97% for total solids and TOC, respectively.
Carbon Adsorption
Equilibrium adsorption isotherms at 20°C were determined for each UF
composite permeate for both TOC removal and color removal. Figure 8 presents
the equilibrium isotherms for the "medium" loading industrial laundry waste.
This figure is representative of the data obtained from all waste samples.
The isotherms for the "light" and "heavy" loading wastes can be found in
Appendix A. The points in most curves fall reasonably close (within
experimental error) to straight lines indicating agreement with the
Freundlich isotherm expression. Similar curves for TOC removal were obtained
in all cases. Likewise, the adsorptive capacity for color followed the same
trend for all three permeates.
The steep slope of the TOC isotherms indicates that as the TOC concen-
tration decreases, the loading drops off very rapidly. For a two-fold
decrease in concentration (see Figure 8) the adsorptive capacity of the
carbon decreased by over an order of magnitude. This indicates that the TOC
content of the waste is composed of a small amount of strongly adsorbed
material and a larger amount of weakly adsorbed material. Isotherms of this
nature indicate that a rapid breakthrough of TOC will occur during processing
of industrial laundry ultrafiltrates through a carbon column.
The adsorption isotherms for color removal exhibit much more gradual
slopes than the TOC isotherms. This indicates that the carbon's adsorptive
capacity for color producing compounds remains high even as the color of
the UF permeate becomes reduced.
For the "heavy" loading wastewater, an additional isotherm was performed
to determine the effect of neutralization of the sample pH on the carbon's
adsorptive capacity. This isotherm is also presented in Appendix A. If a
substantial improvement in organic adsorption occurred, then a trade off
study between pH adjustment costs and increased carbon efficiency would be
warranted. However, performing the adsorption isotherm test on a neutralized
industrial laundry UF permeate (pH = 7-1) had little, if any, affect on the
carbon's adsorptive properties for either TOC or color.
From the above discussion, it appears that activated carbon treatment
of industrial laundry UF permeates will prove very beneficial in terms of
color removal, but will only marginally lower the TOC of the ultrafiltrate.
Task 1 Conclusions
Based on the Task 1 experimental results, the following conclusions can
be drawn relative to UF/activated carbon treatment of industrial laundering
effluents.
40
-------
TABLE 10. CONTAMINANT ANALYSES DURING UF TOTAL RECYCLE EXPERIMENTATION
a). "LIGHT" LOADING INDUSTRIAL .LAUNDRY WASTEWATER
Assay
Total Solids
(mg/1 )
TOC (mg/1)
Turbidity
(NTU)
pH (units)
Sampled after 53
HFD
Feed Permeate
4,660 1,820
2,740 362
5.3
10.4
hours
HFM
Permeate
1,570
310
4.7
10.4
Sampled after 239
HFD
Feed Permeate
4,260 2,040
2,280 456
5.2
9.4 9.4
hours
HFM
Permeate
1,820
432
5.5
9.4
b). "HEAVY" LOADING INDUSTRIAL LAUNDRY WASTEWATER
Assay
Total Solids
(mg/1 )
TOC (mg/1)
Turbidity
(NTU)
Sampled
Feed
39,000
32,800
--
after 19.4
HFD
Permeate
2,990
808
2.8
hours
HFM
Permeate
3,060
716
2.7
Sampled after 242
HFD
Feed Permeate
38,900 3,100
36,200 1,070
4.8
hours
HFM
Permeate
3,200
1,120
4.6
pH (units)
11.8
12.0
11.9
9.7
9.7
9.7
41
-------
5000
c
o
.n
i.
1000
S_
O
s-
o
a
c
o
3
CTI
O>
-o
0)
.a
o
o
500
200
100
50
20
10
x/m at C0 = 2570 Color units/g Carbon
x/m at Co - 0.88 TOC/g Carbon
O
n TOC
° Color
sS
Co* 350
color
iunits
|C0=167mg/J
I !
1 I I
I
10
20
50
100
200
500 700
TOC (mg/£) or Color (units) Concentration
Figure 8. Equilibrium adsorption isotherms for TOC and color removal
from "medium loading" industrial laundry waste UF permeate
42
-------
The Type HFM membrane is preferred over the HFD
membrane because of its higher flux level, its
greater resistance to environmental attack, and
its tolerance to free chlorine (for cleaning
purposes).
It appears from the limited test data that average
flux levels of 40-50 gfd (125°F) can be maintained
if the feed is pretreated for free oil removal.
Both membrane spacer types, corrugated and open-
spacer, appear applicable for processing industrial
laundry wastes. Field tests with both feed-side
channel spacers will determine the preferred option.
Membrane rejection for suspended solids was >99%,
and for freon extractives, >98%. BOD, COD, and
TOC rejections were typically >80%. Rejections for
all metals except mercury generally ranged from
>70-98%.
If free oil is removed from the feed prior to UF
treatment, the membrane flux should be recover-
able by standard detergent formulation and cleaning
procedures.
Based on the projected power requirements for the
spiral-wound modules, and a power cost of $0.04/kw-hr,
power costs of $0.25/1000 gal and $0.52/1000 gal are
predicted for operation of the corrugated and open-
spacer modules, respectively. Power costs of the
magnitude are considered acceptable for spiral-wound
module UF systems.
Carbon adsorption treatment is technically feasible
for color removal from the UF permeate; however, only
marginal TOC removal is anticipated.
PRELIMINARY FIELD DEMONSTRATION EXPERIMENTS
Introduction
Four experiments were conducted at the field demonstration site prior
to the initiation of the formal test program. These preliminary experiments
served as shakedown runs for the UF/ACA system and were designed to obtain
data on a number of parameters:
effectiveness of pretreatment in preventing lint
breakthrough to the membrane system circulation
loop;
membrane flux vs. time relationships at low (67%)
and high (99%) conversions;
43
-------
comparative flux levels for open-spacer and corrugated
spacer spiral-wound modules;
feasibility of employing Vexar spacer spiral-wound
modules;
contaminant removal efficiencies for UF membranes
and activated carbon over an extended time period; and
effectiveness of establishing cleaning procedures in
recovering membrane water flux.
The results of these experiments and their implications for the formal test
program are discussed below.
Ultrafiltration
UF Membrane Flux, Tests PI and P2--
The first preliminary test (PI) was conducted with the feed transfer
pump suction line inadvertently placed 15 inches from the bottom of the sump
rather than 15 inches beneath the liquid surface. A 20 mesh Y-strainer was
located on the discharge of the transfer pump and a wooden board was situated
in the UF feed tank to provide a quiescent region (~15 minute residence time)
from which feed to the UF system was withdrawn. The feed was then screened
through 30 mesh basket strainers. Type HFM membrane corrugated spacer and
open-spacer modules were tested in parallel housings. Each housing
contained two membrane cartridges.
With a system conversion of 65-70% (~3X concentration factor) the flux
of the corrugated modules averaged 33 gfd (135°F), and the flux for the
open-spacer modules averaged 36 gfd (135°F). Although no severe membrane
fouling was observed, the test was concluded after four hours due to plugging
of the 30 mesh basket strainers with lint. The permeate flux vs. time curve
for this test is shown in Appendix B.
Prior to the second preliminary experiment (P2) the suction line from
the sump was raised to the proper height. This action had little, if any,
effect on the rate at which the basket strainers became plugged. The second
test also had to be terminated early (after 5 hours operation) and, as
evidenced by the reduced circulation flow through the modules during P2,
plugging of the spiral-wound feed-side spacers was occurring. During PI the
feed circulation rate was 70-75 gpm. In P2 this rate was reduced to 55-69
gpm.
The open-spacer module flux was similar for Tests PI and P2 until the
final hour of operation, at which point it began to decline more rapidly
for Test P2. For the corrugated module the permeate flux during P2 was
lower than in the first test by as much as 50%. These losses in flux are
probably a result of the lower feed velocity through the cartridges, and in
the case of the corrugated module, loss of membrane exposure due to channel
plugging. The flux curve for Test P2 is also given in Appendix B.
44
-------
Improved Pretreatment--*
System operation was clearly limited by the plugging of both basket
strainers and modules with lint particles and threads. To provide improved
lint removal a new tank partition, constructed of reticulated polyurethane
foam, was positioned in the center of the UF feed tank. The foam is highly
porous and had been shown during pilot studies on other programs (7) to be
an effective depth filter. It was also anticipated that the foam partition
would act as an oil coalescer. Once the foam becomes loaded with suspended
solids and/or oil, it can be removed from the tank and regenerated (a
limited number of times) by surface cleaning and squeezing.
Grab samples of the feed from the sump, the feed prior to the foam
partition and the feed after the foam partition were analyzed for suspended
solids content. These assays are shown below.
Feed Feed Prior to Feed After
Assay from Sump Foam Partition Foam Partition
Suspended Solids (mg/1) 1940 2610 976
It is clear from these data that the foam acted mainly as a surface filter,
concentrating the suspended solids in the first portion of the tank.
Regardless of the mechanism of filtration, the foam's potential for
suspended solids removal was evident.
A further demonstration of the effectiveness of lint removal by the
foam partition was the continuous operation of 100 mesh basket strainers in
the ultrafiltration system. With both strainers in line, overnight operation
was successfully achieved.
UF Membrane Flux, Tests P3 and P4T-
After installing the foam partition, two more preliminary experiments
were performed. Run P3 began with testing of open-spacer modules at a 67%
(3X concentration factor) system conversion; however, the conversion
increased during an unattended overnight shift. The average conversion over
the first 70 hours of Test P3 was determined from analytical data (see below)
to be 97% (SOX concentration factor). The 99% (100X) system conversion aimed
for during the final 100 hours of the test was determined by sample analysis
to be from 99.5-99.8% (200-500X concentration factor). Variations in the
concentrate flow with time account for the discrepancies between the planned
and the actual system conversions.
The flux data for Test P3 are plotted in Figure 9. The open-spacer flux
through the first five hours declined from 35 to 22 gfd (1350F). At this
point, the system was left unattended overnight and the feed concentration
within the loop increased beyond a 3X concentration factor. As a result of
the increased concentration, the flux declined to 14 gfd (135QF). When the
system conversion was corrected to 67% the flux returned to 22 gfd and
remained stable for 20 hours.
* This pretreatment method was instituted for these pilot studies only. It is
not a commercially-available process. For full-scale spiral module opera-
tion further investigation of pretreatment alternatives is necessary.
45
-------
60
50
x 30
I 20
10 _
0
67% Conversion
97% Conversion
Conversion
Cleaned
Strainers
Installed
Corrugated
Modules
Concentrate Bleed Flow
Q] Found at Zero/Conversion
Readjusted to 6755
Circulation
Rate
O Corrugated Spacer 55-74 gpm 46-50 psig
D Open Mesh Spacer 45-54 gpm 50 psig
A Vexar ~ 15 gpm 50 psig
System Conversion as Noted
Permeate Flux Temperature Corrected to 135°F
Concentrating
Shutdown to
I Clean Foam
Float Valve Plugged
With Lint/Cleaned
System Shutdown
Flushed With HgO
O -I
Installed
"t)ne Vexar
Module -
10 20
Time (hours)
50
100
200
Figure 9. UF permeate flux vs. time for third preliminary
field demonstration experiment (Test P3) .
-------
Two corrugated modules were operated in parallel with the open-spacer
modules beginning at the start of the third day. The flux for both
module types declined steadily and continued to decline as the system
conversion was increased. As mentioned above, the conversion achieved
inadvertently exceeded 99%.
With about two-thirds of the test completed, the open-spacer modules
were replaced with a single Vexar module. The flux of this module declined
over three days to the level of the corrugated modules. The final process
flux at 135°F and 99.5-99.8% conversion was about 5 gfd for each module type.
The fourth preliminary test was conducted to compare the Vexarand
open-spacer configurations. The laundry wastewater was concentrated to a
100X concentration factor and then the UF system was operated in the feed-
and-bleed mode to maintain this conversion. The flux for the open-spacer
module remained stable from 20 to 75 hours operating time; the flux for the
Vexar module constantly declined with time (see flux curve, Appendix B).
The final flux levels were 20 gfd (135°F) for the open-spacer module and
6 gfd (135QF) for the Vexar module.
UF Membrane Flux Recovery--
Recovery of the UF membrane water flux to acceptable levels was achieved
following each preliminary test; however, the cleaning procedure was, at
times, lengthy. Multiple detergent cleanings were always required, and
frequently, from 5 to 7 cleaning cycles had to be performed. A tabulation of
the flux recovery data is given in Appendix B.
Module Pressure Drop
The installation of the foam partition within the UF feed tank eliminated
the plugging of the feed-side spacer channels of the UF modules with lint.
No increase in pressure drop was observed across any of the module types,
including Vexar-
The projected horsepower requirements for each module type are
presented in Table 11. Clearly the low flow rate required by a Vexar module
enhances its power requirement relative to the corrugated and open-spacer
modules. The higher membrane area per Vexar module (40 sq. ft. vs. 34 sq.
ft. and 18 sq. ft. for the corrugated and open-spacer modules, respectively)
makes its use even more economically attractive. The power cost for
processing 20,000 gallons per day of laundry wastewater at a 67% conversion
is roughly 6 times less for Vexar modules than corrugated modules, and
nearly 13 times less for Vexar modules than open-spacer modules.
UF Membrane Rejection
Twice during Test P3 samples of the feed (from the sump), the UF
concentrate and the UF permeate were analysed. These samples were collected
on a continuous basis each day and then combined into either two-or four-
day composite samples. The first series of samples was taken at a 97%
47
-------
00
TABLE 11. PROJECTED HORSEPOWER REQUIREMENTS FOR UF SPIRAL-WOUND MODULES OPERATING ON
INDUSTRIAL LAUNDRY WASTEWATERS DURING 'PRELIMINARY FIELD DEMONSTRATION TESTS
Spiral Wound
Module
Spacer
Geometry
Open Spacer
Corrugated
Vexar
Average
Feed
Flowrate
70
70
15
Average
Pressure
Drop
(psi)
15
15
15
Projected
Horsepower
per Module
(hp)
0.606
0.606
0.130
Average
Productivity
per Module @
67% Conversion (god)
540
1020
1500
Projected
Horsepower
per Gallon of
Product per Day
(hp/gpd)
1.12 X 10"3
0.59 X 10"3
0.087 X 10"3
Horse Power-
Hours per Day
for Processing
20,000 gal.
537.6
238.2
41.8
Kilowatt-
Hours per
Day for
Process ing
20,000 gal .
«,
178
31.2
Dollars
per day
for Processing
20,000 cal
@ $0.04/KWH
16
7.12
1.25
-------
conversion, a composite sample of the carbon column effluent was also
analyzed.
Table 12 presents the analytical data from the 97% conversion test
period. Essentially qomplete suspended solids and turbidity removals were
achieved. BOD, COD, and TOC removals ranged from 70 to 82% in the UF
permeate and from 88 to 94% in the carbon effluent. The total freon
extractives analyses reported are in error on the high side due to contam-
ination of the freon used in the analysis. Note, the concentration of oils
and grease in the permeate should be closer to the levels reported for the
permeate in the industrial laundry wastewaters (28, 10, and 38 mg/1) processed
during in-house tests. Exact values of metals removals are, again,
generally limited by the lower detection limits of the assays. Color
removal increased from 70% after the UF processing to 98% following carbon
adsorption.
The analytical data from the 99.5-99.8% conversion period of Test P3
are given in table 13. The feed assays are very similar to the data shown
in Table 12 indicating little change in the overall feed composition.
Suspended solids and turbidity removals were not affected by the increased
system conversion. However, BOD, COD, TOC, and color removals all decreased
somewhat, as expected, since the concentration of the feed within the
circulation loop at 99.5-99.8% conversion-was nearly an order of magnitude
higher than at 97% conversion.
Carbon Adsorption^
Partial breakthrough of color occurred in the carbon column effluent
after 200 hours of preliminary testing; however, complete breakthrough was
not achieved. Due to the long duration of the preliminary tests and the high
conversions achieved, fresh carbon was placed in the columns prior to start-
up of the formal tests. As noted in Table 12, the carbon was quite effective
in removing color from the UF permeate, and it also enhanced the overall
system BOD, COD, and TOC removal efficiencies.
Conclusions
On the basis of the preliminary field demonstration tests the following
conclusions are drawn:
Effective pretreatment to remove lint particles has
been demonstrated; however, practical pretreatment
options for full-scale units must be chosen, tested,
and economically evaluated.
Processing with Vexar modules is feasible with
proper feed pretreatment. Comparative testing with
both open-spacer and Vexar modules is warranted to
further detail their flux characteristics at very
high conversions and their amenability to cleaning.
49
-------
CJ1
o
TABLE 12. CONTAMINANT ANALYSES AND UF AND CARBON REMOVAL EFFICIENCIES DURING 97%
CONVERSION PERIOD OF THIRD PRELIMINARY FIELD-EXPERIMENT (TEST P3)
Assay
Total Solids (mj/1 )
Suspended Solids (mg/1)
Dissolved Solids (mg/1)
BOD (mg/1)
COD (mg/1)
TOC (mg/1)
pH (units)
Turbidity (NTU)
Color (units)
Total Freon Extract! bias
(mg/1)
Alkalinity (mg/1 as CaCO,)
Cadmium (mg/1)
Chromium (mg/1)
Iron (mg/1)
Lead (mg/1 )
Nickel (mg/1)
Zinc (mg/1)
* Removal Ef
Feed
From Sump
2,240
336
1,900
1,100
2,680
832
11.8
750
1,000
(1,260) t
930
<0.2
<0.5
9.5
1.4
<0.5
2.2
ficiencv. r =
UF Concentrate
21,100
8,450
12,600
33,000
78,200
19,000
11.3
30,000
...
(ll,000)t
...
Feed Concentration -
UF Permeate
1,640
8
1,630
330
520
148
11.6
6.0
300
(iiejf-
...
...
-__
---
UF Permeate (or Carbon
UF Removal ...
Efficiency, %
26.8
97.6
14.2
70.0
80.6
82.2
99.2
70.0
90.8
Effluent) Concentration
Carbon Effluent
-
132
159
55
11.8
3.0
20
(82) t
880
<0.2
<0.5
<1
<1
<0.5
0.13
Overall Removal ...
Efficiency, %
88.0
94.1
93.4
99.6
98.0
93.5
5.4
>89.5
>28.6
94.1
} Suspected error in analysis
-------
TABLE 13. CONTAMINANT ANALYSES AND UF REMOVAL EFFICIENCIES DURING 99.5-99.!
CONVERSION PERIOD OF THIRD PRELIMINARY FIELD EXPERIMENT (TEST P3\
Assay
Total Solids (mg/1)
Suspended Solids (mg/1)
Dissolved Solids (mg/1)
BOD (mg/1)
COD (mg/1)
TOC (mg/1)
pH (units)
Turbidity (NTU)
Color (units)
Total Freon Extractibles
(mg/1 )
Alkalinity (mg/1 as CaCOg)
Cadmium (mg/1 )
Chromium (mg/1)
Iron (mg/1)
Lead (mg/1 )
Nickel (mg/1)
Zinc (mg/1)
Feed
From Sump
2,500
360
2,140
1,290
3,050
820
11.7
1,000
2,000
(957^
840
<0.2
<0.5
10
2.6
<0.5
2.5
UF Concentrate
145,000
92,300
52,700
71,000
617,000
128,000
11.0
150,000
(103,000)t
__
UF Permeate
2,200
5
2,190
580
929
292
11.7
4.5
1,000
(234)*
UF Removal Efficiency,^
12.0
98.6
55.0
69.5
64.4
99.6
50.0
____
_-__
Removal
_ Feed Concentration - Permeate Concentration
Feed Concentration
X 100
( ) Indicates Suspected Error in Analysis due to Freon Contamination
-------
Testing with corrugated modules is no longer warranted.
Their operation is limited to temperatures <125°F, they
are not as compact as Vexar modules; and they are more
difficult to clean than the open-spacer modules.
Conversions as high as 99.8% (500X) can be achieved
with process flux levels ranging from 4-10 gfd (135°F).
At a conversion of 97% UF removal efficiencies of 70%
for BOD, 80% for COD, and 82% for TOC can be expected.
Optimization of cleaning procedures is a necessary step
before scale-up of the pilot unit.
FORMAL FIELD DEMONSTRATION PROGRAM
Introduction
The formal field demonstration program was designed to provide simulation
of system conversions typical in a four-stage ultrafiltration system. These
system conversions and their corresponding feed volumetric concentrations
are:
Volumetric Feed
System Conversion, % Concentration
67 3X
90 10X
97 SOX
99 100X
These system conversions were skewed toward the higher concentrations because
performance characteristics at these levels will have a major impact on the
process economics. That is, if concentrations of 100X or higher could be
achieved the costs associated with concentrate disposal would be greatly
reduced. Note that since most of the permeate will be produced by the first
stages in a full-scale system, flux data for these stages are as critical as
for the latter stages in determining membrane area requirements.
By operating the UF pilot system in the feed-and-bleed mode, each
system conversion listed above was maintained for a two-week period, and
between each test the membranes were chemically cleaned. Two open-spacer
modules and two Vexar spacer modules, containing Type HFM membranes, were
evaluated simultaneously.
Ultrafiltration
UF Membrane Flux--
The UF permeate flux vs. time curves for the field demonstration tests
conducted at 67%, 90%, 97%, and 99% conversion are presented in Figures 10, «
11, 12, and 13, respectively. Average flux levels are summarized below.
52
-------
40 r
CD Open Spacer
Circulation Rate
60-65 gpm
Inlet Pressure
50 psig
30
Vexar Spacer
20 ,apm
Permeate Flux Temperature Corrected to 135 F
60 psig
CJl
co
X
3
10
01
-------
40 i
~T
T
30
20
10
LJ Open Spacer
/\ Vexar Spacer
Permeate Flux Temperature Corrected to 135 F
New Vexar Spacer Modules were installed prior to test initiation
Circulation Rate
65-75 gpm
20 gpm
Inlet Pressure
50 psig
60 psig
Concentration Within
membrane loop
exceeded 10X over
night.
_L
20
40
60
80
100 120
Time (Hours)
140
160
180
200
220
240 250
Figure 1U UF permeate flux vs. time for 90% conversion field demonstration test.
-------
40
,
cn
en
D
30
Dl
X
20
(LI
CL
10
C~] Open Spacer
A Vexar Spacer
Circulation Rate
60-80 gpm
15 or 20 gpm
Inlet Pressure
50 psig
50 or 60 psig
Permeate Flux Temperature Converted to 135
20
'
Vexar Module
operating conditions
changed to 20 gpm, 60 psig
A
f
A"
i
i
40
60
80
100
120 140
Time (Hours)
160
180
200
220
240
250
Figure 12. UF permeate flux vs. time for 97% conversion field demonstration test.
-------
50
40
^.concen-
trating
30
01
g 20
E
03
O-
10 _
Open Mesh
20
Circulation Rate
75-90 gpm
15 gpm
Inlet Pressure
50 psig
50 psig
99% Conversion
System Shutdown
Over
Weekend
-1-
40
60 80
Time (Hours)
120
140
160
Figure 13.. UF permeate flux vs. time for 99% conversion field demonstration test
-------
Average UF Permeate Flux (gfd @ 135°F)
System Conversion, % Vexar Spacer Open-Spacer
67 3 12
90 15 8
97 3 17.5
99 3.5 17.5
The apparent inconsistencies in the trends for these data are readily
explainable. First, the tests were conducted sequentially from highest
conversion to lowest conversion. Therefore, a reduction in flux recovery
following cleaning, or a gradual degradation of the membrane modules due to
increased exposure to industrial laundry wastes could adversely affect the
flux in the latter (lower conversion) tests. Secondly, for the open spacer
modules, the formation of a thin oil layer on the membrane surface during the
90% conversion test limited the average permeate level to 8 gfd (135°F). The
oil film reduced the passage of water through the membrane, while permeating
oil (see below) into the product stream. The exact nature of this oil film
and why it had a greater affect on the open-spacer, rather than the Vexar
modules, has not been clearly defined.
A third factor influencing the average module flux levels is that after
the 97% conversion experiment the water flux of the Vexar modules could not
be recovered. Thus, new Vexar spacer modules were installed in the test unit
for the 90% conversion experiment. While these new membranes exhibited high
flux initially (see Figure 11), the flux curve has a steep slope throughout
the first 175 hours of the test. At this point the permeate flux stabilizes
within the 2-5 gfd (135°F) range. Once having been exposed to industrial
laundry wastewater, as shown in Figure 10, the Vexar modules exhibit the
same magnitude of permeate flux (2-5 gfd) as observepl in the higher
conversion experiments.
With the exception of the sharp flux decline for the Vexar modules
during the 90% conversion experiment, the spiral-wound modules exhibited very
stable flux performance. The average values of the permeate flux were 14 gfd
for the open-spacer modules and 3 gfd (excluding the initial portion of the
90% conversion test) for the Vexar spacer modules. The open-spacer module
flux during the formal field tests was, however, substantially below the
Task 1, in-house test values. Also, the flux levels for both the open-spacer
and Vexar spacer modules were below what would be predicted from the
preliminary field tests. The reasons for these differences in membrane
module flux performance include:
The in-house tests were conducted with a limited
volume of wastewater, and therefore, the membranes
were exposed to only a relatively low fixed level
of foul ants.
The first two preliminary field tests had a total
duration of only 9 hours; therefore, no long-term
fouling effects were observed.
57
-------
All modules used in the first formal field test had
already accumulated between 100 and 200 hours
exposure to the industrial laundry wastewater during
the preliminary tests.
UF Membrane Flux Recovery--
Table 14,presents the UF membrane flux recovery data during all field
demonstration tests. (The corrugated modules were eliminated from evaluation
after the preliminary testing and are not discussed below.) The same two
open-spacer spiral-wound modules were used throughout the preliminary and
formal test programs. They each accumulated over 1000 hours exposure to
industrial laundry wastes and typically recovered 70% of their initial water
flux after detergent cleaning. The final water flux for these modules was,
however, only 53.5 gfd (135°F). As the tests proceeded, recovery of the
water flux became more difficult for these modules. After 4.hours exposure
to the waste (Test PI), two detergent cleaning cycles were required for
satisfactory flux recovery. After 169 hours accumulated exposure (following
all preliminary tests) 7 detergent cleaning cycles were required. During
the formal field testing the number of detergent washings required to recover
the flux of the open-spacer modules increased progressively from 3 to 6.
When the formal field tests began, a reducing agent for iron foul ant removal
was added to the cleaning formulation. Although the extent of membrane iron
fouling was not precisely determined, this change in the formulation of the
cleaning solution was beneficial.
The Vexar spacer modules proved much more difficult to clean. In fact,
following the second field test (533 hours cumulative exposure) the water
flux of the Vexar modules could only be recovered to 11.3 gfd (135°F). These
modules were discarded and new modules were installed in the test system.
Following the final experiment (488 hours of exposure to the waste) the flux
for these modules was recovered to only 17.6 gfd (135°F).
Clearly, the rate of irreversible flux decline for the Vexar modules is
unacceptable. Also, the number of detergent cycles needed to clean the open-
spacer modules or even partially clean the Vexar modules is excessive.
Optimization of the module cleaning formulations and procedures is therefore
required either before, or in conjunction with, any further testing of
spiral-wound modules for the treatment of industrial laundry wastes.
Preferred UF Module Spacer Geometry--
As discussed above, the flux of the open-spacer module averaged 14,gfd
(135°F) while the flux for the Vexar spacer module averaged 3 gfd (135°F)
throughout the formal field test program. Vexar modules contain 40 sq. ft.
of active membrane area while open-spacer modules contain 18 sq. ft. of
active membrane area. Daily module productivity is calculated by multiplying
square feet of membrane area in a given module type by the average module
flux. Average module productivity, in terms of gallons per day is therefore
252 gpd and 120 gpd for the open-spacer and Vexar configurations,
respectively.
58
-------
TABLE 14. FLUX RECOVERY AND ACCUMULATED OPERATING TIMES FOR UF MEMBRANES OPERATED
ON INDUSTRIAL LAUNDRY WASTEhlATERS. DURING FIELD DEMONSTRATION TESTS
Membrane Accumulated Exposure
Type/Spacer Designation Time (Hours)
HFM/Corrugated MCI 334
MC2 0
MCI and MC2 4
n
MC3 and MC4 0
139
HFM/Open Spacer M02 and M03 0
4
11
im
1 U 1
169
320
566
808
1054
HFM/Vexar MV1 0
MV1 and MV2 68
136
287
533
MV3 and MV4 0
242
488
Water Flux at
1350F (gfd)
75.6
200
54.8
...
45.6
100
79.8
94.9
ftfi A
OO . *r
63.0
75.0
72.0
73.1
53.5
162
74.6
55.1
33.8
11.3
140
31.4
17.6
Comments
Used during Task 1
New Module, Preliminary Testing
2 detergent cleanings
Corrugated modules developed leaks
due to high temperature
Water flux not recorded
5 detergent cleanings
New Modules, Preliminary Testing
2 detergent cleanings
3 detergent cleanings
7 detergent cleanings
Following first field test, 3 detergent
cleanings, iron fouling treatment
Following second field test, 4 detergent
cleanings, iron fouling treatment
Following third field test, 5 detergent
cleanings, iron fouling treatment
Following fourth field test, 6 detergent
cleanings, iron fouling treatment
New Module
5 detergent cleanings, initial water flux
, . . . not recorded for MV2
7 detergent cleanings
Following first field test, 3 detergent
cleanings, iron fouling treatment
Following second field test, 4 detergent
cleanings, iron fouling treatment
New Modules
Following third field test, 10 detergent
cleanings, iron foiling treatment
Following fourth field test, 5 detergent
cleanings, iron fouling treatment
-------
Power considerations for each are also based on the average module flux
levels and active membrane areas just given. The theoretical horsepower
requirements per gallon of product per day are 2.4,x 10~3 hp/gpd for open-
spacer modules and 1.9 x 10-3 hp/gpd for Vexar modules. Thus, the power
requirements for the two types of modules are much closer than observed
during the preliminary field tests (see Table 11). One reason that these
values are closer is that the Vexar module power was calculated with a feed
flowrate of 20 gpm and a pressure drop of 20 psig. This circulation rate
and its resulting pressure drop were used during the final two field tests in
an effort to improve the Vexar module flux performance from that observed at
15 gpm (AP = 15 psig) circulation rate.
Another factor which must be considered before selecting the preferred
module type is the relative response of the modules to cleaning. As noted
above, the open-spacer modules showed considerably greater flux recovery than
the Vexar modules.
Therefore, with regard to process flux, daily module productivity, and
flux recovery, the open-spacer module is preferred to the Vexar spacer
module. Since the power savings for the two module types are reasonably
close in terms of hp/gpd, the open-spacer configuration is clearly preferred.
UF Membrane Removal Efficiency--
Tables 15 through 18 present the entire analytical data sets for the
four field demonstration tests. Any deviations from the standard sampling
procedures are noted in the footnotes to each table. Also, oil permeation
through the membranes of the open-spacer modules resulted in abnormally low
removal efficiencies for suspended solids and total freon extractives during
the 90% conversion test. With the exception of the data obtained during that
particular test, the average UF removal efficiencies are:
Average UF
Assay Removal Efficiency, %
Total Solids 41.8
Suspended Solids >96.2
Dissolved Solids 28.9
BOD 76.4,
COD 82.2
TOC 78.4
Turbidity 95.4.
Total Freon Extractives 95.1
The UF process removed >95% of suspended solids, turbidity, and total
freon extractives. These values are somewhat lower than the corresponding
Task 1 data. This suggests that although severe oil permeation was
experienced in only one instance, minor oil permeation was a regular
occurrence. This was probably the result of small quantities of free oil
in the feed stream continuously adhering to the membrane surface and
diffusing through it. In a full-scale operation this problem can be elim-
inated by installing a free-oil skimmer or coalescer upstream of the
UF modules.
60
-------
CTl
TABLE 15. CONTAMINANT ANALYSES AND UF AND CARBON REMOVAL
.EFFICIENCIES DURING 67%.CONVERSION FIELD TEST
«ar
Tottl Solid! ing/l)
SuSEtr.eetZ Col Im (nQ/t)
Otsioived Saiifit (eg/I)
no :«5«)
C53 l-^/i)
TK !r,A)
t« <--.;;s)
Turbidity (STU)
ToUl Freon Excrccllbles (Mg/1
Aik«l(Mty Cssg/I M C*C03)
CJiro=i^. (.-5/i)
Iron (*o/0
U.d t'G/z)
H«rct.r/ (cs/0
Hlcit'. (=3/1)
Zinc ta/0
FEED FSCH SUMP
Ueelc fl Week 12 Ccnposlte
2,140 3.140
512 »0
1.&36 2.5SO
1 .200 1 ,300
720 1 ,GSO
10.9 9.8
700 850
574 650
_.
..-
3.220
...
1.080
99.2
960
37.5
>85.7
XI. 2
33.3
""
'««»», £f«ci«nW. r - *" a-""""..'" fLPtrj.lri"J'rt"'n """""' C°"c""""°" « 100
-------
ro
TABLE 16. CONTAMINANT ANALYSES AND UF AND CARBON REMOVAL
EFFICIENCIES DURING 90% CONVERSION FIELD TEST
ASSAY
Total Solids (mg/i)
Suspended Solids (mg/A)
Dissolved Solids (mg/i)
BOD (mo/A)
COD (mg/A)
TOC (mg/A)
pH (units)
Turbidity (NTU)
Total Freon Extractives
Color (units)
FEED FROM SUMP -
Week #1 Week K
2,240 1,040
680 324
1,560 716
1,100 (270)2
940
11.0 7.6
1,000 360
(mg/A) 9.16 345
Feed Concentration-UF Permeate
, , UF REMOVAL , ' . OVERALL REMOV1L
UF CONCENTRATE UF PERMEATE ' EFFICIENCY, % CARBON EFFLUENT* EFFICIENCY, %'
Week ft
9,690
5,020
4,670
6,000
...
7,120
8,000
6,1.00
(or Carbon
Week #2 Week fl
2,790 1,360
504 174
2,286 1,126
1 ,500 350
322
1 ,600 TOO
1-.300 363
...
Effluent) Concentration
Keek #2 Week #1 Week #2 Week #1
39.3
74.4
27.8
520 68.2 220
948
251 65.7 102
15 90.0 95.8
36 60.4 89.6 5
400
inn
Week « Week Jl Week f2
100 80.0 (54.5)Z
216
72 89.1
2,.8 99.7
12 99.5 96.b
20
( ) Indicates suspected error 1n analysis.
3Keek tZ UF Permeate taken from vexar modules only, grab sample.
During Week #1 one sample of UF Permeate was Invertantly placed 1n carbon effluent sample bottle and vice versa.
-------
CO
TABLE 17. CONTAMINANT ANALYSES AND UF AND CARBON REMOVAL
EFFICIENCIES DURING 97% CONVERSION FIELD TEST
ASSM
Total Solids (19/1)
Suspended Solids («g/l)
Dissolved Solids (g/1)
MO (.g/1)
COO.t-9/1)
TOC (iq/l)
PH (units)
Turbidity (MTU)
Total Fraon Extractives
Alkalinity (xg/1 » UCO)
CadnliM (*9/l)
ChrMlun (9/1)
Copper (ag/l)
Iran
Heck 11 Week 12
33.2
93.0
13.1
72.2 82.4
l
.... 84 3 ...
71.7 80.7
93.8 97.9
....
....
CARBON EFFLUENT'
Week IT Week 12 Cwposite Meek
230 110 70.9
445
188 94 74.9
582
- < 0.5
< 0.5
< 1
< 1
T-- --- 0.001
... ... < 0 5
0.6
iCarbon Col urn rearing breakthrough
OVERALL REMOVAL
EFFICIENCY. *'
l\ Week 12 Conposlt*
'87.9
82.4
78.6
> 44.4
> 79.6
> 33.3
62.5
I Concentration
ConposlU Samples
-------
TABLE 18. CONTAMINANT ANALYSES AND UF AND CARBON REMOVAL EFFICIENCIES
DURING 99% CONVERSION FIELD EXPERIMENT
Feed from Sump UF Concentrate
Assay Week #12
Total Solids (mg/Jl) 2
Suspended Solids (mg/Jl)
Dissolved Solids (mg/Jl) 2
BOD (mg/Jl)
COD (mg/£) 3
TOC (mg/J.)
pH (units)
Turbidity (NTU)
Total Freon Extractibles
(mg/Jl)
Alkalinity (mg/Jl as CaO>3)l
.Cadmium (mg/£)
Chromium (mg/Jl)
Copper (mg/Jl)
Iron (mg/Jl)
Lead (mg/Jl)
Mercury (mg/Jl)
Nickel (mg/Jl)
Zinc (mg/Jl)
^Rom/Wai Fffir
,840
270
,570
880
,120
675
11.6
700
505
,010
<0.2
<0.5
0.99
5.7
2.2
0.002
<0.5
2.0
Meek #1
73,000
36,100
36,900
87,900
276,000
54,400
10.6
70,000
36,000
_ Feed Concentration
UF Removal
UF Concentrate UF Permeate UF Permeate Efficiency, $
Meek #2
37,700
21,600
16,100
31,000
31,500
11.3
40,000
20,600
...
- UF Permeate
Week #1 Week #2 Week #1 *
2,050
<5
2,045
260
592
198
11.4
7.7
20
ipr Carbon Effluent)
1,740
8
1,732
363
208
10.8
4.0
14
Concentration
27.8
>98.1
20.4
70.5
81.0
70.7
98.9
96.0
-
y inn
Carbon Carbon
Effluent Effluent
Week #1 Week #2
2,350
<5
2,345
260
592
202
11.6
9.3
8
1,040
<0.2
<0.5
<0.5
<1.0
<1.0
0.001
<0.5
0.57
_.
--
363
202
11.2
<5
Overall Removal
Efficiency.%,
Week #1
17.3
>98.1
8.8
70.5
81.0
70.1
98.7
98.4
>49.5
>82.5
>54.5
50.0
71.5
Feed Concentration
No feed sample collected during second week due to pump malfunction.
-------
Thq range of BOD, COD, and TOC removal efficiencies - 76, 82, and 78%
respectively - are consistent with previous results and indicate that no
membrane or module failures occurred. These values would, however, have
been slightly improved if the free-oil had been completely eliminated from
the feed stream. In terms of actual permeate quality, average BOD was
302 mg/1; COD, 634,mg/l; and TOC 194,nig/1.
Mass balance calculations performed with the data presented in Tables 15
through 18 indicate that the system conversion was not maintained at exactly
the desired level throughout each test. For example, during week #1 of the
97% conversion test a system conversion of 96-98% (25-50X concentration
factor) is indicated from the analytical data. For the second week of this
test, conversions of 80-94% (5-18X concentration factor) are indicated. The
lower conversion during the second week is attributed to a slight mismatch
between total permeate flow and possible concentrate bleed settings. The
timer-solenoid valve arrangement used for withdrawing a constant flowrate of
concentrate was only adjustable in increments of 18 ml/min and therefore
precise control at the desired conversion was not always attainable.
Carbon Adsorption
Carbon Removal Efficiency--
Throughout the, four field demonstration experiments a slip-stream of UF
permeate was continuously fed to a 2-inch diameter carbon column. The assays
performed on the carbon effluent composite samples are presented along with
the initial feed, UF concentrate and UF permeate analyses in Tables 15
through 18. Table 19 condenses these data giving average values for the feed,
UF permeate, carbon effluent and carbon removal efficiency based on the UF
permeate quality. Overall UF/ACA removal efficiencies are discussed in a
subsequent section.
As observed in Table 19 the activated carbon reduced the UF permeate or-
ganic loading and oxygen demand by ~40%. Freon extractives were decreased by
over 55%. All of these removal efficiencies are based on weekly or biweekly
Composite samples indicating that the activated carbon consistently improved
the UF permeate quality.
Adsorptive Capacity of Carbon
The adsorptive capacity (maximum loading) of a particular carbon column
for a given contaminant is determined by measuring the volume of liquid
processed before the concentration of that contaminant in the carbon column
effluent approaches or equals its concentration in the feed stream. When
these concentrations approach each other complete "breakthrough" is said to
have occurred. By analyzing periodic grab samples for TOC and color, break-
through curves for these contaminants were developed for the field demon-
stration tests.
65
-------
en
en
TABLE 19. AVERAGE CARBON REMOVAL EFFICIENCY DATA
(UF Permeate as Feed, Basis)
Assay
BOD (mg/1)
COD (mg/1)
TOC (mg/1)
Total Freon
extractives (mg/1)
Average
Feed
Concentration
1030
2960
764
596
Average
UF Permeate
Concentration
302
634
194
20.4
Average
Carbon Effluent
Concentration
190
353
123
<9
Average
Carbon Removal
Efficiency, % *t
37.1
44,3
36.6
>55.9
: Damnwal FffirTOnr*\/ v>
UF Permeate
Concentration - Carbon
Effluent Concentration
Y inn0/
UF Permeate Concentration
t Based on average UF permeate and carbon effluent concentrations
-------
Two sets of breakthrough curves were generated: One set was generated
during the 67 and 90% conversion tests; the other, during the 97 and 99%
conversion tests. These data are shown in Figures 14,and 15, respectively.
New carbon was placed in the columns prior to the collection of each data
set. The average UF permeate color and TOC concentrations were 195 color
units and 143 mg/1 in one test compared to 750 color units and 213 mg/1 TOC
in the test at the higher UF system conversions. The substantial difference
in average ultrafiltrate color concentration, while attributable in part to
changes in system conversion, may also be the result of variations in the mix
of articles laundered.
For TOC, both curves indicate that the carbon effluent concentration
approaches the UF permeate concentration at ~2250 gallons processed. After
this point the concentration of TOC in the carbon effluent increased above
the average UF permeate concentration in one case and returned to ~50% of the
average UF permeate concentration in the other. In the former instance, the
increase can be attributed to desorption of contaminants from the fully
loaded carbon. In the latter case, the low final TOC concentration may be
the result of biological activity within the column. Breakthrough after
2250 gallons were processed through the column corresponds to a carbon re-
placement cost of $1.32/1000 gal (at a carbon cost of 56<£/lb).
A lower carbon replacement cost may be indicated if column operation is
dictated by effluent color concentration only. While color breakthrough
occurred at approximately the same time as TOC breakthrough in the test at
the higher.UF system conversions, complete color breakthrough did not occur
during the test at the lower UF system conversions. At the time when these
UF tests were completed, 3234 gallons of ultrafiltrate had already been
processed through the carbon column. Since color breakthrough was in no way
indicated at this point (see Figure 14), a conservative estimate for the
carbon replacement frequency would be after the processing of 3500 gallons
through the 2-inch column. This corresponds to a carbon replacement cost of
$0.85/1000 gallon.
Without performing similar tests on other laundry wastewaters, it is
difficult to assess the degree to which these costs are representative of
all industrial laundering operations. Furthermore, it is possible that
carbon treatment of UF permeates for color removal may not be necessary for
launderies which wash colored articles.
Overall UF/ACA Removal Efficiency
The overall contaminant removal efficiency of the ultrafiltration/
activated carbon adsorption treatment combination was quite good. These
data were presented for the individual field tests in Tables 15 through 18
and are summarized in Table 20. Turbidity removal averaged 99.2%, and
suspended solids removal, determined by the UF permeate concentration, was
> 96.2%. Most of the suspended solids which did pass the UF process were
elirftinated from the final effluent by the depth filter characteristics of the
carbon column. The average reduction in total freon extractives was>97.7%.
Overall removal efficiencies for BOD, COD, and TOC averaged 82%, 86% and
82%, respectively.
67
-------
300
250
200
150
Average UF Permeate Color Concentration = 195 color units
0 Color
CD TOC'
Average UF Permeate TOC Concentration = 143 mg/1
.,300
.250
. 200 -
150 o
100
50
100 -
50 . H
Carbon Effluent TOC
Concentration +
Carbon Effluent Color Concentration
O
o
500
1000
1500 2000
Volume Processed (gallons)
2500
m
3000
3500
Figure 14. Carbon column breakthrough curves for color and TOC during
67 and 90% conversion field demonstration tests.
-------
300
250 _
200
1200
O Color
D TOC
Average UF Permeate TOC Concentration =213 mg/1
o-"-
01
u
c
o
u
150
100
50
500
Average UF Permeate Color Concentration = 750 color units
Carbon Effluent
TOC Concentration
Carbon Effluent
Color Concentration
-O"
O
_L
1000 1500
Volume Processed (gallons)
2000
2500
1000
800
600
400
200
i.
o
o
u
c
o
fO
S-
OJ
o
c
o
o
J-
o
3000
Figure 15. Carbon column breakthrough curves for color and TOC during
97 and 99% conversion field demonstration test.
-------
TABLE 20. SUMMARY OF AVERAGE UF/ACA PRODUCT WATER QUALITY AND
REMOVAL EFFICIENCIES DURING FIELD DEMONSTRATION TESTS
Average
Average UF/ACA
Feed Effluent
Assay Concentration Concentration
Total Solids
(mg/1)
Suspended Solids
(mg/D
Dissolved Solids
(mg/1)
BOD (mg/1)
COD (mg/1)
TOC (mg/1)
Turbidity (NTU)
Total Freon
Extractives (mg/1)
Alkalinity
(mg/1 as CaCO.0
Cadmium (mg/1 )
Chromium (mg/1 )
Copper (mg/1)
Iron (mg/1 )
Lead (mg/1 )
Mercury (mg/1)
Nickel (mg/1)
Zinc (mg/1)
2230
479
1750
1030
2960
764
685
596
8SO
<0.
<0.
0.
5.
1.
0.
<0.
1.
1480**
<16.8**
1463**
190
353
123
6.1
<9.
867
2 <0.2
5 <0.5
9 <0.5
9 <1
8 <1
0015 0.001
5 <0.5
7 0.42
Average UF/ACA
Removal Efficiency, %*t
41 .8**
>96.2**
28.9**
81.7
86.1
81.8
99.2
>97.7
3.1
>43.8
>82.6
>43.0
41.5
>75.9
Rpmnval pffi n'pnrv.
_ Feed
Concentration - UF/ACA
Effluent Concentration v
Feed Concentration
Average removal efficiency is based on individual test results and not on
average feed and permeate concentrations.
Carbor. effluent sample not analyzed for this constituent. UF permeate concen-
tration and UF system removal efficiency are given.
70
-------
Industrial laundries are typically surcharged by municipal treatment
authorities on the mass of BOD and suspended solids discharged. The minimum
allowable discharge before a surcharge is assessed varies between localities;
however, it is clear from these data that the UF/ACA process will eliminate
or greatly reduce these charges. The effluent oil and grease level of
<9 mg/1 should satisfy all sewer discharge standards.
As discussed previously, the metals removal efficiencies for the
combined treatment process were generally limited by the lower detection
limit of the assays. The MSD of Greater Chicago Sewer Discharge Standards
were easily met for all metals, except mercury. For mercury, an average
removal efficiency of 41.5% was achieved with an average effluent concentra-
tion of 0.001 mg/1. The MSD's standard for mercury, 0.005 mg/1, is quite
low. A survey of 20 municipal sewer discharge standards has indicated a range
of 0.005 - 1.5 mg/1 for the allowable concentration of mercury in
industrial effluents (1).
Conclusions
With the conclusion of the formal field demonstration experiments, the
entire test program was completed. On the basis of these experiments,
conclusions derived relative to UF/ACA treatment of industrial laundry
wastewaters are:
-- Open-spacer, spiral-wound, UF modules are preferred
to Vexar spacer modules since they offer higher permeate
flux rates and higher daily productivities. Open-spacer
modules are also more easily cleaned than Vexar modules.
The small savings in power cost per gallon of product per
day offered by Vexar modules is far outweighed by the
open-spacer modules' advantages.
Average permeate flux for the open-spacer modules was
14,gfd (135°F); for the Vexar modules the average flux
was 3 gfd (135°F). Flux levels of this magnitude are
not economically acceptable for treatment of industrial
laundry wastewaters.
-- Free oil in the waste stream can seriously effect the
performance of the UF membranes. Both permeate flux
and permeate quality deteriorate rapidly if a free oil
layer coats the membrane surface.
Optimization of UF module cleaning formulations and
procedures is required in conjunction with any further
testing of spiral-wound modules for the treatment of
industrial laundry wastewaters. In fact, it is doubtful
if a spiral-wound membrane configuration can withstand
prolonged operation in an industrial laundry. A mem-
brane configuration, which is less susceptible to fouling
e.g., tubular, is clearly required.
71
-------
The overall UF/ACA product water averaged <17 mg/1
suspended solids, 190 mg/1 BOD, 353 mg/1 COD, 123 mg/1
TOC, and <9 mg/1 total freon extractives. An effluent
of this quality indicates average removal efficiencies
for the treatment process of >96% for suspended solids,
>97% for freon extractives and 82%, 86%, and 82% for
BOD, COD, and TOC, respectively. An effluent of this
quality should be acceptable for discharge to municipal
sewer systems. Based on local ordinances, a surcharge
may be applied if the mass discharge of BOD and suspended
solids exceed acceptable limits.
Metals removed by the UF/ACA process was, in general,
calculated from the lower detection limit values of the
assays. All metals of interest, except mercury, were
removed to levels below those specified in the sewer
ordinances of the Metropolitan Sanitary District (MSD) of
greater Chicago.
Activated carbon treatment improved UF permeate quality
substantially. Further reductions in the BOD, COD, and
TOC of the wastewater averaged -40%.
The carbon's adsorptive capacity for color bodies is
greater than its adsorptive capacity for TOC. Break-
through curves developed for color indicate a carbon
replacement cost of $0.85/1000 gal.
72
-------
SECTION 8
ECONOMIC SCALE-UP ANALYSIS
INTRODUCTION
In this section purchased equipment (capital) costs and operating costs
are projected for UF/ACA systems treating industrial laundry wastewaters.
Cases are presented for launderies discharging 25,000, 75,000 and 100,000 gpd
of waste.
A generalized flow schematic for the waste treatment system is shown in
Figure 16. Two basic alternatives have been examined: waste treatment for
discharge to the sewer and waste treatment with water reuse. The latter
alternative requires post-treatment of the UF permeate by carbon adsorption
and the installation of a large product water holding tank.
Figure 16 assumes a 100,000 gpd waste discharge. The waste is trans-
ferred from the plant sump to a 75,000 gal surge tank. This surge capacity
permits 24,hour/day operation of the UF system even though the plant has an
8 hour/day waste generation cycle. Waste pretreatment before the UF system
will consist of three steps:
(1) Microstraining for lint removal (e.g. Sweco
vibrating screen or Bauer Hydrasieve).
(2) Skimming for free oil removal in the surge
tank (e.g. rope skimmer).
(3) Filtration for additional suspended solids
removal. The type of filter to be used has
not yet been specified. This step requires
additional investigation, the degree of which
was not anticipated at the start of this program.
The ultrafiltration system costs are based on Abcor, Inc. Type HFM
membranes in open-spacer, spiral-wound modules. Since the average permeate
flux was not strongly dependent on the system conversions, the use of a
single-stage rather than a multi-stage system is preferred.
Key design bases are outlined below:
73
-------
Plant
Sump
Free Oil
Skimmer
Transfer
Pump
MicroStrainer
for Lint
Removal
75,000 gal
Surge
Tank
Transfer
Pump
Pretreatment
for Suspended
Solids Removal
3-Stage
Ultrafiltration
System
Product
Water
Concentrate
^ Water
Reuse
>Fj
100,000 gal
Holding
Tank
. . .._i .<
Carbon
Adsorption
1
i
i
Reuse |
^ Option i
I
i
Sewer
(Optional)
Figure 16. Flow schematic of 100,000 gpd waste treatment system.
-------
Assumed Membrane Flux, gfd 14
Membrane Life, yrs 1
Membrane System Costs, $/ft2 50
Membrane Replacement Costs, $/ft2 20
Based on the average flux of 14 gfd, membrane area requirements for the three
systems are:
Membrane Area
Capacity, gpd Requirement, ft2
25,000 1,786
75,000 5,357
100,000 7,143
PURCHASED EQUIPMENT COSTS FOR PRETREATMENT AND ULTRAFILTRATION SYSTEMS
Purchased equipment costs for the three system capacities are given in
Table 21. The purchased equipment cost per gallon of wasted treated per
day ranges from $4 to $4.65. The Table 21 costs are based on carbon steel
systems containing all necessary monitors and controllers for unattended
operation, a temperature control loop, a product water turbidity monitor
and a central control panel. These costs exclude costs for an activated
carbon adsorption system, which is addressed separately in a subsequent
section. Also excluded is installation cost which will be highly site
specific. If all utilities (power,water, and sewer connections, etc.) are
in place, and no new facility must be constructed, installation costs might
be as low as $25,000 to $50,000. If not, installation costs may be as high
as 100% of the system cost. Installation costs must be calculated on an
individual, site-by-site basis.
ANNUAL OPERATING COST FOR PRETREATMENT AND ULTRAFILTRATION SYSTEMS
Table 22 contains operating cost projections for the three system
capacities. These do not include capital-related costs such as amortization,
interest, and taxes. Also, carbon adsorption operating costs and water reuse
credits are addressed in a subsequent section. Details of the power cost
estimates are given in Table 23. All UF system costs are based on system
operation six days per week with system cleaning taking place on the seventh
day. Pre-treatment system costs are based on operation six days per week
only. The operating costs range from $11.82/1000 gal when processing
25,000 gpd to $9.01/1000 gal when processing 100,000 gpd. These costs are
clearly impractical and necessitate the evaluation of an alternative UF
approach, presented at the end of this section, which is predicted to
significantly lower the operating expenses.
PURCHASED EQUIPMENT AND ANNUAL OPERATING COSTS FOR ACTIVATED CARBON
TREATMENT AND COMPLETE WATER REUSE
Pretreatment and ultrafiltration system costs have been discussed thus
far with the assumption that all UF permeate would be discharged to municipal
sewers. However, it is possible that some, if not all, of the UF permeate
75
-------
TABLE 21. ESTIMATED PURCHASED EQUIPMENT COSTS FOR WASTEWATER TREATMENT
SYSTEMS OF VARIOUS CAPACITIES (Thousands of Dollars)
UF SYSTEM CAPACITY (gpd)
lltM 25,000 75,000 100,000
PRETREATMENT
1. Holding tank and controls with a capacity
of 75% of system capacity. 10 20 25
2. Prefiltration for lint and suspended
solids removal
3. Oil skimmer for free oil removal
ULTRAFILTRATION SYSTEM
1. UF System including controls, monitors
and engineering design costs
TOTAL ESTIMATED PURCHASED EQUIPMENT COSTS
PURCHASED EQUIPMENT COST PER GALLON OF WASTE
TREATED PER DAY
15
2
27
89.3
89.3
116.3
$4.65
18
3
41
267.9
267.9
308.9
$4.12
20
4
49
357.2
357.2
406.2
$4.06
NOTES: 1. This cost analysis does not include installation costs which will vary considerably from
site to site depending on availability of utility connections, space, etc.
2. This analysis assumes discharge of UF permeate to the sewer and therefore carbon adsorption
and product water storage costs are omitted.
-------
TABLE 22. ESTIMATED ANNUAL OPERATING COSTS FOR WASTEWATER TREATMENT
SYSTEMS OF VARIOUS CAPACITIES (Thousands of dollars)*
ITEM
PRETREATMENT
1. Labor, Operating & Maintenance, 2 hrs/day @
$5.00/hr plus 75% fringe benefits and overhead
2. Supervisory Labor, 1 hr/day @ $10/hr plus 75%
fringe benefits and overhead
3. Power, 5 or 10 hp (27,900 or 55,800 kw-hr @
4
-------
TABLE 23. ESTIMATED POWER REQUIREMENTS FOR WASTEWATER
TREATMENT SYSTEMS OF VARIOUS CAPACITIES
Sys tern
Capacity
25,000
-j 75,000
00
100,000
No. of
Spiral -Wound
Cartridges
99
298
397
No. of
Parallel
Passes
33
99
132
Total
Flow,
gpm
2310
6930
9240
AP per
pass,
psig
45
45
45
Actual
Horsepower,
hp
86
257
343
kw-hr
per 1000
gal
61.4
61.4
61.4
Cost
per year,
$
22,400
67,200
89,600
Notes: 1. Open-spacer module membrane area is 18 sq. ft., 3 modules per parallel pass.
2. Flow rate per pass is 70 gpm.
3. Horsepower requirement assumes 70% pump efficiency.
4. Power costs based on 4<£ per kw-hr.
-------
could be reused directly within a laundry which handles a minimum of white
articles. Nevertheless, because of the questionability of UF permeate
reuse, no projections of water reuse credits were made.
As water discharge standards become more stringent and if "closed-loop"
treatment systems become mandatory, activated carbon treatment of the UF
permeate for color removal will be a logical continuation of the waste treat-
ment sequence. It is anticipated that the additional capital investment
necessary to meet any zero discharge limitation will depend largely on the
nature of the effluent guidelines which may be promulgated for the industrial
launderers. These additional capital costs, associated annual operating
costs, and water reuse credits are discussed below.
Estimated additional purchased equipment costs for complete water reuse
are detailed in Table 24. A carbon adsorption system and a product water
storage tank with a capacity equal to one day's discharge would be required.
The carbon system costs are based on a projected carbon capacity of ~3
color units/mg carbon, an average removal of 500 color units per liter, and
a bed service time of two months. The purchased equipment costs total
$17,300 for the 25,000 gpd system, $40,800 for the 75,000 gpd system, and
$51,000 for the 100,000 gpd system.
Operating costs for the carbon adsorption system are also given in
Table 24, These costs range from $1.09 to $1.68/1000 gal of treated water.
Carbon replacement costs account for 50% to 80% of these annual operating
expense estimates.
Certain credits for water reuse within the laundry may be applied to
the annual system operating costs to arrive at a significantly reduced net
treatment cost. These credits are summarized in Table 25. The exact amount
of credit possible for any particular industrial laundry depends upon the
water consumption and sewer use charges in the city where it is located.
High, low, and average values of potential water reuse credits were
calculated using published data (1). The water reuse credits range from
$0.55 to $2.8/1000 gal of water reused with an average credit of $1.08/1000
gal of water. All of these estimated credits are based on charges pre-
vailing in 1971. Charges for consumption for water have undoubtedly risen
while sewer surcharges are now in effect in more areas. Therefore, these
figures should be considered as conservative estimates of potential water
reuse savings. Even at the 1971 levels, however, these credits offset the
cost of carbon adsorption treatment.
PROJECTED ECONOMICS FOR TUBULAR ULTRAFILTRATION SYSTEMS
Introduction
Two factors contributed most significantly to the capital and operating
costs for open-spacer, spiral-wound-module, UF systems. First, the average
UF permeate flux was low (14,gfd) causing membrane area requirements at any
given capacity to be high. Since UF system capital costs are calculated on
79
-------
TABLE 24. ESTIMATED ADDITIONAL PURCHASED EQUIPMENT AND OPERATING
COSTS FOR COMPLETE WATER REUSE (Thousands of dollars)
CO
o
ITEM
PURCHASED EQUIPMENT COSTS
1. Carbon Column; 70, 210 or 280 ft2 @
$60/ft3
2. Engineering design costs for carbon
column @ 25%
3. Product water storage tank, same capacity
as treatment system
TOTAL ESTIMATED PURCHASED EQUIPMENT COSTS
PURCHASED EQUIPMENT COST PER GALLON OF WASTE
TREATED PER DAY
OPERATING COSTS
1. Labor, Operating and Maintenance 1 hour/day
@ $5/hour, pi us^ 75% for frtnge benefits and
overhead
2. Supervisory Labor, 0.5 hr/day @ $10/hour
plus 75% for finge benefits and overhead
3. Carbon Replacement, @56<£/lb
4. Power, 5 or 10 hp (27,900 or 55,800 kw-hr
@ 4<£/kw-hr)
TOTAL ANNUAL OPERATING EXPENSE
TREATMENT COSTS, $71000 GAL
TREATMENT
25,000
4.2
1.1
12
17.3
0.69
2.7
2.7
6.6
1.1
13.1
1.68
SYSTEM CAPACITY
75,000
12.6
3.2
25
40.8
0.54
2.7
2.7
19.7
2.3
27.4.
1.17
(flP<0,
100,000
16.8
4.2
30
51.0
0.51
2.7
2.7
26.4
2.3
34.1
1.09
NOTES: 1. Purchased equipment costs do not include installation costs, which will vary considerably
from site to site.
2.. Annual operating costs do not include capital related costs such as amortization, interest
and taxes.
-------
TABLE 25. ESTIMATED ANNUAL CREDITS FOR REUSE OF TREATED WATER
Current Laundry Operating Costs * Credits. Dollars/I000 gal Treated
K 3 Low High Typical
1. Influent water 0.12 0.70 0.28
2. Heating requirements (Assumes a
20°F increase in water tempera-
ture resulting from reuse as
opposed to use of waste water
heat in a heat exchanger) (1.66
xlO6 BTU/1000 gal-day, 260 days/
3.
4..
yr operation, @ $3.00/10° BTU)
Sewer use charges
Sewer surcharges
0.41
0.02
0.00
0.55
0.41
1.12
0.60
2.83
0.41
0.22
0.17
1.08
* Items #1, 3 and 4 are given in 1971 dollars, and are based on informa-
tion from: Douglas, Gary, "Modular Wastewater Treatment System
Demonstration for the Textile Maintenance Industry," EPA-600/2-73-037,
January 1974.
81
-------
2
a $/ft basis, these costs became quite substantial. Also, the more modules
in the system, the more pumping power required. As shown in Table 22, for a
100,000 gpd treatment system power costs were >$89,000/yr. A second factor,
affecting operating expense only, is membrane life. The replacement cycle
for the spiral-wound modules is estimated to be one year. At a replacement
cost of $20/ft2, .this amounts to ~$143,000/year for a 100,000 gpd system.
Since only a modest improvement in the performance of spiral-wound modules
can be expected by changing the operating conditions, cost projections have
been developed for an alternative UF module geometry, namely, tubular
assemblies.
The tubular configuration, as discussed in Section 4, is preferred
over the spiral-wound configuration from a technical standpoint: It is
easier to clean (either chemically or mechanically) and would not be
susceptible to plugging by lint. The spiral-wound configuration was chosen
for this program since it was potentially a lower-cost geometry. However,
for this particular application, the disadvantages of spiral-wound modules
predominate their cost advantage, and tubular systems are projected to be
more cost effective as discussed below.
Projected Purchased Equipment and Annual Operating Costs
for Tubular Ultrafiltration Systems
Cost estimates for tubular UF systems were developed for the three
system capacities of interest. For each capacity two system options were
costed: The first option is based on typical current prices of tubular UF
systems from Abcor, Inc. and a projected membrane flux of 40 gfd; the
second option is an optimistic case which assumes that improvements in
tubular membrane technology over the next 2-4,years will increase membrane
flux (from 40 to 50 gfd), increase life (from 2 years to 3 years), reduce
system costs (from $65/ft2 to $45/ft2) and reduce membrane replacement
costs (from $15/ft2 to $10/ft2). All four sources of potential cost
reduction are considered realistic and achievable.
A summary of purchased equipment and annual operating costs for
tubular UF systems is presented in Table 26. A breakdown of these costs
and a listing of the design bases used in their calculation is given in
Appendix C. Capital costs are ~50% lower for tubular UF systems using
costs for current technology than for open-spacer, spiral-wound systems of
comparable capacity. These costs range from $1.97 to $2.22 per gallon of
waste treated per day. Future technological advances should reduce these
costs to $1.31 to $1.62 per gallon of waste treated per day.
Estimated system operating costs are reduced from the $9 to $12/1000 gal
estimated for the spiral module systems to $2.6 to $5/1000 gal when current
technology UF is employed. For the 25,000 gpd treatment capacity today's
costs are $4.88/1000 gal. At this low capacity technological advances are
not predicted to keep pace with inflation and future costs become $5.18/1000
gal. Treatment costs at the 75,000 gal capacity fall from $2.82/gal to
$2.69/1000 gal as lower cost tubular UF systems are developed. The effect of
these technological advances on the 100,000 gpd treatment system are to
lower costs from $2.57 to $2.39/1000 gal.
82
-------
TABLE 26. SUMMARY OF ESTIMATED CAPITAL AND OPERATING COSTS FOR WASTEWATER
TREATMENT SYSTEMS EMPLOYING TUBULAR ULTRAFILTRATION MODULES
UF SYSTEM CAPACITY (gpd)
TODAY'S COSTS
FUTURE COSTS
ITEM
25,000 75,000 100,000 25,000 75,000 100,000
00
CO
Total estimated purchased equipment
costs, $
Purchased equipment cost per gallon
of waste treated per day, $
Total annual operating costs, $
Treatment costs, $/1000 gal
55,600 148,900 196,500 40,500 99,900 130,800
2.22 1.99 1,97 1.62 1.33 1.31
38,100 65,900 80,200 40,400 62,900 74,600
4.88 2.82 2.57 5.18 2.69 2.39
NOTES: 1. Purchased equipment costs do not include installation costs and will vary considerably from
site to site.
2. Operating costs do not include capital related costs such as amortization, interest and taxes,
3. These costs assume discharge of UF permeate to the sewer and therefore carbon adsorption and
product water storage costs are omitted.
-------
The tubular UF configuration will produce an effluent similar in quality
to the spiral-wound configuration since both configurations are available
with Type HFM membranes. Since the treatment costs for the higher capacity
tubular UF systems are competitive with the currently employed dissolved air
flotation systems, tubular UF should be investigated for the treatment of
industrial laundry wastes. Such an investigation would allow the cost
estimates presented herein to be revised on the basis of actual permeate flux
measurements.
84
-------
REFERENCES
1. Douglas, Gary. "Modular Wastewater Treatment System Demonstration
for the Textile Maintenance Industry". EPA-660/2-73-037, January 1974,
2. Ramirez, Ernest. "LectroClear Treatment of Industrial Laundering
Wastewaters". Swift Environmental Systems Company, November 20, 1975.
3. Thomas, J.M. and Thomas, W.S. Introduction to the Principles of
Heterogenous Catalysis. Academic Press, New York, p. 32 (1967).
4. Weber, W.J. Physiochemical Processes for Hater Quality Control.
Wiley-Interscience, New York, p. 213 (1972).
5. Vermeulen, T. "Separation by Adsorption Methods", in Advances in
Chemical Engineering, Vol. II, T.B. Drew and J.W. Hooper, eds.
Academic Press, New York, p. 148 (1958).
6. Hiester, N.K.,, Vermeulen, T., and Klein, G. "Adsorption and Ion
Exchange", in Chemical Engineers' Handbook, J.H. Perry, ed., 4th Ed.
McGraw-Hill, New York, Section 16 (1963).
7. Walden Division of Abcor, Inc. Internal Report. February, 1976.
8. Sluizer, Mervyn Jr., Letter from Institute of Industrial Launderers
to Office of Analyses and Evaluation (WH-586), Environmental
Protection Agency, Washington, D.C., May 16, 1977.
85
-------
-o
TABLE Al. WASHROOM PRODUCTION SCHEDULE ON DAY OF "LIGHT" LOADING WASTEWATER SAMPLING* m
CO
MACHINE # 1
TIME
Start/Finish
.0625
0725
0850
0955
1110
1215
1325
1445
0610
0715
0830
0950
1055
1200
1315
1440
1535
Article
Laundered
Rugs
Cotton shirts
and pants
Roll towels
Wipers (rags)
Wipers
Shirts and
pants
Wipers
Shirts and
pants
Cotton shirts
and pants
MACHINE # 2
TIME Article
Start/Finish Laundered
0615 0700 Shirts
0710 0820 Shirts and
pants
0840 0945 Shirts and
pants
1005 1040 Rugs
1050 1210 Wipers
1220 1330 Shirts and
pants
1340 1430 Shirts and
pants ,
1520 Shirts
MACHINE # 3
TIME
Start/Finish
0620 0705
0720 0805
0810 0905
0910 1010
1015 1100
1115 1200
1205 1305
1335 1425
1435
Article
Laundered
Shirts
'Shirts
Shirts
Shirts
Rugs
Rags
Wipers
Shirts
Shirts
3="
3=
0
O
i i
1
i i
[
O
3»
-rt
§
HH
O
cr
o
)
00
0
£
m
EXPERIME
> i
3=
-a
-o
INDICES
* Wastewater sample collected between 1100 and 1200 hours.
-------
TABLE A2. WASHROOM PRODUCTION SCHEDULE ON DAY OF "MEDIUM LOADING" HASTEWATER SAMPLING*
MACHINE # 1
MACHINE # 2
MACHINE # 3
Time Article
Start/Finish Laundered
Time Article
Start/Finish Laundered
Time Article
Start/Finish Laundered
CO
0603 0712 Colored Garments 0608 0717 Colored Garments 0623 0800 White Garments
0728 0817 Mats
0740 0848 Mops
CGI3 0945 White Garments
0843 1030 Wipers
0916 1100 Wipers
1055 1245 Wipers
1122 1307 Wipers
1258 1415 Rags
1327 1508 Wipers
Wastewater sample collected between 0955 and 1055 hours.
-------
TABLE A3. WASHROOM PRODUCTION SCHEDULE ON DAY OF "HEAVY" LOADING WASTEWATER SAMPLING *
CO
CO
Machine #1 Machine #2 Machine #3 Machine #4
Time
Out
0842
1029
Article Time Article Time Article Time
Laundered Out Laundered Out Laundered Out
Paint Towels 0845 Paint Towels 0748 Machine Towels 0751
Paint Towels 1044 Paint Towels 1050; Machine Towels 0920
1028
Article
Laundered
Machine Towels
Machine Towels
Machine Towels
Machine #5
Time Article
Out Laundered
0828 Rags
1035 Paint Towels
'"Wastewater sample collected between 0920 and 1010 hours
-------
TABLE A4. CONTAMINANT ANALYSES AND MEMBRANE REMOVAL EFFICIENCIES FOR UF
BATCH CONCENTRATION OF "LIGHT LOADING" INDUSTRIAL LAUNDRY VJASTEWATER
Assay
Total Solids (mg/1)
Suspended Solids (mg/1)
Dissolved Solids (mg/1)
Turbidity (NTU)
BOD (mg/1)
TOC (mg/1)
COD (mg/1)
Freon Extractibles (mg/1)
Alkalinity, (ppm CaC03)
pH (units)
Chromium (mg/1)
Copper (mg/1)
Lead (mg/1)
Zinc (mg/1)
Cadmium (mg/1)
Iron (mg/1)
Nickel (mg/1)
Mercury (mg/1)
Initial
Feed
2,610
700
1,910
- t
2,800
1,100
3,780
749
875
11.8
<0.5
1.7
3.9
3.9
0.05
17
<0.5
<0.002
-i 1
Final
Concentrate
20,700
9,150 ,
11,500
-t
20,000
13,700
57,300
4,080
1,500
11.6
-
-
-
-
-
-
-
-
Mixed
Composite
Permeate
1,620
<4
1,616
3.7
360
202
672
27.7
(934)tt
11.7
<0.5
<0.5
<1.0
0.2
<0.005
<1.0
<0.5
<0.002
Removal
Efficiency, %*
37.9
>99.4
15.4
-
87.1
81.6
82.2
96.3
-
-
-
>70.6
>74.4
94.9
>90.0
>94.1
-
-
* Removal efficiency, r = Feed Co"centration - Composite Permeate Concentration x 10Q
, . . Feed Concentration
j. Very high
ft ( ) indicates suspected error in analysis
89
-------
TABLE A5. CONTAMINANT ANALYSES AND MEMBRANE REMOVAL EFFICIENCIES FOR UF BATCH
CONCENTRATION OF "MEDIUM LOADING" INDUSTRIAL LAUNDRY WASTEMATER
Assay
. i '
Total Solids (mg/1)
Suspended Solids (mg/1)
Dissolved Solids (mg/1)
Turbidity (NTU)
BOD (mg/1)
TOC (mg/1)
COD (mg/1)
Freon Extractives (mg/1)
Alkalinity (ppm CaCO,)
pH (units)
Chromium (mg/1)
Copper (mg/1)
Lead (mg/1 )
Zinc (mg/1)
Cadmium (mg/1 )
Iron (mg/1)
Nickel (mg/1)
Mercury (mg/1 )
Initial
Feed
1,920
675
1,240
--t
1,650
1,240
5,480
795
740
10.2
< 0.5
1.2
2.1
1.4
0.03
6.5
< 0.5
0.0005
Final
Concentrate
13,200
7,270
5,970
-t
9,800
9,590
43,800
12,200
1,380
10.7
--
Mixed
Composite
Permeate
954
2.4
952
3.6
553
196
796
10
710
10.9
< 0.5
< 0.5
< 1.0
< 0.5
< 0.01
< 1.0
< 0.5
0.0004
Removal
Efficiency,
**
50.3
99.6
23.5
66.5
84.2
85.5
98.7
> 58.3
> 52.4
> 64.3
> 66.7
> 84.6
..
20.0
* Removal efficiency, r = ^eec* Concentration - Composite Permeate Concentration
Feed Concentration
t Very high
90
-------
TABLE A6. CONTAMINANT ANALYSES AND MEMBRANE REMOVAL EFFICIENCIES FOR UF BATCH
CONCENTRATION OF "HEAVY LOADING" INDUSTRIAL LAUNDRY WASTEWATER
Assay
Total Solids (mg/1)
Suspended Solids (mg/1)
Dissolved Solids (mg/1)
Turbidity (NTU)
BOD (mg/1)
TQC (mg/1)
COD (mg/1)
Freon Extractives (mg/1)
Alkalinity (ppm CaC03)
pH (units)
Chromium (mg/1 )
Copper (mg/1)
Lead (mg/1)
Zinc (mg/1)
Cadmium (mg/1)
Iron (mg/1)
Nickel (mg/1)
Mercury (mg/1 )
* Rpmnval t*ff irionrw i
Initial
Feed
10,200
4,500
5,700
-t
7,850
6,750
27,400
7,890
2,350
11.7
8.8
11
22
9.0
0.15
90
0.74
0.0009
Final
Concentrate
64,700
41 ,800
22,900
"t
81 ,900
45,400
188,000
51 ,650
4,080
11.4
--
--
--
--
Mixed
Composite
Permeate
2,680
<5
2,680
4.7
930
642
2,370
38
2,060
11.5
2.9
1.1
<1
0.18
<0.01
1.8
<0..5
0.0008
_ Feed Concentration - Composite Permeate
Removal
Efficiency,
«*
73.7
99.9
53.0
88.2
90.5
91.4
99.5
12.3
--
67.0
90.0
>95.5
98.0
93.3
98.0
>32.4
11.1
Concentration
Feed Concentration
f Very high
91
-------
60
<£>
ro
50
40
-a
<4-
Dl
30
R3
O)
£
S-
a>
Q.
20
D HFH open-spacer
spiral-wound module
O HFD corrugated-spacer
spiral-wound module
Circulation Rate (gpm):
Inlet Pressure (psig):
HFM
~W
44
10
Initial Feed Volume: 265 gallons
Permeate Flux Temperature Corrected to 125°F
0.1
Figure Al.
0.5 1 2345
Time (hours)
UF permeate flux vs. time for batch concentration of
"light loading" industrial laundry wastewater.
10
-------
60
IO
co
CD
X
o.
50
40
30
20
10
,HFM
fj HFM open-spacer
spiral-wound module
O HFD corrugated-spacer
spiral-wound module
Circulation Rate (gpm):
Inlet Pressure (psig):
HFM
60-65
48-55
5X Feed diluted
with permeate to 3X
.HFD
45-50
43-53
Permeate Flux Temperature Corrected to 125°F
I \ I I
10
20
Time (hours)
50
100
200
500
Figure A2. UF permeate flux vs. time for extended exposure test
with "light loading" industrial laundry wastewater.
-------
10
en
x
-------
vp
en
en
X
o>
4->
-------
UD
cr>
bO
50
,-, 40
o
en
X
r
ji
o) 30
id
a>
a.
Li.
s
20
10;
0
0
:<
1 1 111
^ - m
^^ -.^ m\
^i* / B"ira
v / «k ~
x._x^x ^
. / '
^_ "" ^ -
"" t
nx
_ Q HFM corrugated spacer _
spiral-wound module
HFO open-mesh spacer
spiral -wound module
Circulation Rate
MOi HFD
(gpm): 95 90
Inlet Pressure (psig): 52 50
Pressure Drop (ps1g): 12 11
Initial Feed Volume : 265 gal 1 ons
Permeate Flux Temperature Corrected to 125°F
.1
1 1 III
0.5 1 234
Time (Hours)
Figure A5. UF permeate flux vs. time for batch concentration of "heavy
loading" industrial laundry wastewater.
-------
10
en
x
60
50
40
20
10
B. HFM Corrugated,-Spacer-
Spiral-Wound Module .
HFD Open-Spacer
Spiral-Wound Module'
Circulation Rate (gpnv) :
Inlet Pressure (psig) :
HFM
95
50
HFD
90
49
Permeate Flux Temperature Corrected to, 125 F
10 20 50
Time (Hours)
100
200
500
Figure A6. UF permeate flux vs. time for extended exposure test with
"heavy loading" industrial laundry wastewater.
-------
1500
to
00
X/M at C = 1200 Color units/g Carbon
X/M at C = 0.60 g TOC/g Carbon
5 10 20 50 100
TOC (mg/i) or Color (units) Concentration
Figure A7. Equilibrium adsorption isotherm for TOC and color removal from "light
loading" industrial laundry waste UF permeate
-------
o
J3
O
a>
i~
o
I/I
o
o
-------
10,000
c
o
5000
2000
O
l/l
5
° 1000
o
S-
o
c
o
.a
s_
10
a
-------
TABLE Bl. FLUX RECOVERY AND ACCUMULATED OPERATING TIMES FOR UF MEMBRANES OPERATED
ON INDUSTRIAL LAUNDRY MASTEWATERS DURING PRELIMINARY FIELD TESTS
Membrane
Type/Spacer Designation
HFH/Corrugated MCI
MC2
MCI and MC2
MC3 and MC4
HFM/Open Spacer M02 and M03
HFM/Vexar MV1
Accumulated Exposure
Time (Hours)
334
0
4
11
0
139
0
4
11
169
0
68
136
Water Flux at
135°F (gfd)
75.6
200
54.8
~
45.6
100
79.8
94.9
QC A
63.0
162
74.6
55 a
Comments
Used during Task 1
New Module
2 detergent cleanings
Corrugated modules developed leaks due to
high temperature
Water flux not recorded
5 detergent cleanings
New Modules
2 detergent cleanings
3 detergent cleanings
7 detergent cleanings
New Module
5 detergent cleanings
7 detergent cleanings
-o
-o
m
z
o
ii
X
00
m 3>
x o
-o o
m i i
73 -H
m -z.
^
10
*-i J>
O I
-o
73
O
m
oo
-------
Q
ro
50
20
O Corrugated Spacer
LJ Open Mesh Spacer
Basket Strainer
Cleaned
System Plugging
With Lint/Run Ended
Circulation Rate
70-75 gpm
70-75 gpm
System Conversion - 65-70%
Permeate Flux Temperature Corrected to 135°F
Inlet Pressure
47-50 psig
47-50 psig
.Time (hours)
Figure Bl. UF permeate flux vs. time for first preliminary
field demonstration experiment (Test PI) .
-------
60
50
40
Cleaned Y-Strainer
v
Float Valve Plugged
With Lint/Cleaned
V
o
CO
o
-------
60
Concentrating
992 .Conversion
50
40
o
«*-
cn
-------
TABLE Cl. DESIGN BASES USED FOR COSTING TUBULAR UF SYSTEMS
-p
rn
o
II
x
o
Membrane Area Requirements, H
Assumed ft2 @ Each System Capacity Membrane Membrane 5
Membrane Flux gpd Membrane System ? Replacement m
gal/ft2-day 25,000 75,000 100,000 Life, yrs Costs, $/ft Costs $/ft~ °
Design Option
2. Tubular Modules
(future
technology)
50
500
1500
2000
45
10
o
1.
0
Tubular Modules
(today's
technology) 40 625 1875 2500 2
- - - - ' ! ~ C3
1 1
z.
70
65 15 5
m
o
CO
s
CO
CO
-------
o
en
TABLE C2. ESTIMATED PURCHASED EQUIPMENT COSTS FOR WASTEWATER TREATMENT SYSTEMS
EMPLOYING TUBULAR ULTRAFILTRATIQN MODULES (Thousands of Dollars)
UF SYSTEM CAPACITY (qpd)
TODAY'S COSTS
ITEM
PRETREATMENT
1. Holding tank and controls with a
capacity of 75% system capacity
2. Prefiltration, see Note 4
3. Oil skimmer for free oil removal
ULTRAFILTRATION SYSTEM
1. UF System including controls,
monitors and engineering design
costs
TOTAL ESTIMATED PURCHASED EQUIPMENT COSTS
PURCHASED EQUIPMENT COST PER GALLON OF
WASTE TREATED PER DAY
25,000
10
3
2
15
40.6
40.6
55.6
$2.22
75,000 100,000
20 25
4 5
3 4
27 34
121.9 162.5
121.9 162.5
148.9 196.5
$1.99 $1.97
FUTURE COSTS
25,000
12
3.6
2.4
18.0
22.5
22.5
40.5
$1.62
75,000
24
4.8
3.6
32.4
67.5
67.5
99.9
$1.33
100,000
30
6
4.8
40.8
90
90
130.8
$1.31
NOTES: 1. This cost analysis does not include installation costs which will vary considerably from
site to site depending on availability of utility connections, space, etc.
2. This analysis assumes discharge of UF permeate to the sewer and therefore carbon adsorption
and product water storage costs are omitted.
3. Future costs include 20% inflation factor for pretreatment items.
4. Does not require the same level of sophistication as for the pretreatment to spiral-wound
modules.
-------
TABLE C3. ESTIMATED ANNUAL OPERATING COSTS FOR WASTEWATER.TREATMENT SYSTEMS
EMPLOYING TUBULAR ULTRAFILTRATIQN MODULES (thousands pf dollars/yr)
UF SYSTEM CAPACITY, (gpd)
ITEM
' TODAY 'S "COSTS
2§-,000
75 ,000
TOO, 000
FUTURE COST5~
25,000 '
75,000
100,000
PRETREATMENT
1,
1,
3.
Labor, Operating & Maintenance, 1 hr/day
@ $5.00/hr plu§ n% fringe benefits and
overhead
Supervisors' Labor, 0-5 hr/day @ $10/hr
plus 75% fringe benefits and overhead
Power, 5 hp (27,900 kw-hr @ 3«t/kw-hr)
2.7
2,7
0.8
6.2
2,7
2.7
0.8
6.2
2,7
2.7
0.8
6,2
3.3
3.3
1.0
7.6
3.3
3.3
1.0
7.6
3.3
3.3
1,0
7.6
ULTRAFILTRATION SYSTEM
1,
2.
3.
4.
5,
6,
Labor, Operating & Maintenance, 3 hr/day
@ $5.Qp/hr plus 75% fringe benefits and
overhead
Supervisory Labor, 1 hr/day @ $10/hr plus
75% fringe benefits and overhead
Membrane Replacement, see- Table Cl
Power, see Table C4
Cleaning Chemicals, 1 cleaning per week,
25, 35 or 45 Ibs @'$1.85/lb.
UF Concentrate Disposal, assumes 99.5%
conversion; 125, 375 or 500 gpd, 312
days/yr (contract hauling @ 5
-------
TABLE C4. ESTIMATED POWER REQUIREMENTS FOR WASTEWATER TREATMENT
SYSTEMS EMPLOYING TUBULAR ULTRAFILTRATION MODULES
System
Capacity
TODAY'S TECHNOLOGY
25,000
75,000
100,000
o FUTURE TECHNOLOGY
uu
25,000
75,000
100,000
No. of
Tubular
Assemblies
284
852
1136
227
682
909
No. of
Parallel
Passes
36
106
142
28
85
114
Total
Flow,
gpm
1080
3180
4260
840
2550
3420
AP per
pass,
psig
30
30
30
30
30
30
Actual
Horsepower,
hp
26.6
76.0
102
20.1
60.6
81.6
kw-hr
per 1000
gal
18.5
18.5
18.5
14,5
14,5
14,5
Cost
per year,
$
6,800
20,300
27,100
6,600
19,900
26,500
NOTES: 1. Tubular assembly is 1 inch diameter X 10 foot long; membrane area is 2.2 ft .
2. 8 tubes in series per pass.
3. Flow rate per pass is 30 gpm.
4. Horsepower requirement assumes 70% pump efficiency.
5. Current power costs based on 4<£/kw-hr, future power costs based on 5<£/kw-hr.
-------
TECHNICAL REPORT DATA
(Please rend Iiialructious on the reverse before completing)
_L
1. REPORT NO.
EPA- 600/2- 78- 177
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Demonstration of UHrafil tration and Carbon Adsorption
for Treatment of Industrial Laundering Wastewater
5. REPORT DATE
August 197.8 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
M.H. Kleper, R.L. Goldsmith, A.Z. Gollan
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Wai den Division of Abcor, Inc.
850 Main Street
Wilmington, Massachusetts 01887
10. PROGRAM ELEMENT NO.
IBB610
11. GWW*AC=fVGRANT NO.
S-804367-01
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVEREO
Task Final. 3/76-4/77
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
IERL-Ci project leader for this report is R.J. Turner
513-684-4481
16. ABSTRACT
This study of industrial laundry wastewater treatment by ultrafiltration
and activated carbon adsorption has indicated that a consistently high quality
product water, potentially reusable within the laundry, can be produced. The
operation of the spiral-wound ultr.afil tration modules was, however, hindered by
the fouling tendency of the feed stream. Average module permeate flux Was
therefore low. This factor, in turn, resulted in high capital and operating
cost estimates for full-scale treatment systems.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Wastewater Treatment
Carbon Adsorption
Industrial Laundries
Ultrafiltration
13B
8. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. I
PAGES
20. SECURITY CLASS (Tillspage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
109
ftU.S. GOVERNMENTrailfimG OFFICE 1978757-140/1455
------- |