Hollow Fiber UF Pilot

Testing Procedures & Case

Studies
By Dave Christophersen, CWT

Several membrane technologies are available to use for industrial water

preparation and wastewater recycling. Typical membrane processes for
solids removal include microfiltration (MF) and ultrafiltration (UF). Hollow
fiber membranes are one relatively cost effective MF / UF design. Other
designs include spiral wound and tubular. Hollow fiber design for UF and
MF is a popular configuration used today on surface water and water reuse
applications due to the ability to handle relatively tough waters at a
reasonable cost and with a small footprint. Tubular designs can handle higher
solids and higher emulsified oil concentrations.
Membrane pore sizes are commonly expressed as nominal which means that
a membrane with a specified nominal pore size or Molecular Weight Cut-off
(MWCO) would be expected to remove 90% of material of that size. The
largest pores in these membranes are likely larger than the nominal pore size.
Variations in pore size occur depending upon the membrane type and
manufacturer.
Microfiltration ranges in size from approximately 0.05 m to 1.0 m.
Ultrafiltration is a tighter membrane and is in the micron range of
approximately
0.005 - 0.1 m. UF is typically expressed in terms of molecular weight cutoff and ranges from 1,000 Daltons for a very tight UF membrane to
approximately 500,000 Daltons for a very open UF membrane. By
convention for MWCO, it is assumed that the molecules are a polysaccharide
of that molecular weight; so actual filtration effectiveness varies according to
the specific chemistry of the molecule. There is some overlap between the
stated ranges for MF and UF, so a membrane with a pore size that might be
considered to be a loose UF membrane might also be considered to be a tight
MF membrane depending on the industry or manufacturer.

Wastewater

346

Hollow Fiber UF and MF

Hollow fiber membrane systems for filtration have gained wide acceptance
in surface water treatment for potable water production. For potable water
applications, hollow fiber membrane systems can guarantee removal of
bacteria such as giardia and cryptosporidium because the integrity of the
membrane system can be verified with integrity tests of the membranes in the
field. This is done by applying air pressure to the membrane and monitoring
for pressure loss and visually looking for air bubbles. Extensive application
of hollow fiber UF and MF for potable water production has led to the costs
of this technology coming down to the point where it is cost competitive with
conventional water treatment and spiral wound UF systems for RO
pretreatment.
Industrial applications include wastewater recycle after primary or secondary
treatment, metals removal, raw water clarification, and RO pretreatment,
among others.
Hollow fiber membranes used for filtration in industrial water treatment may
either be UF or MF membranes. The I.D. of fibers is typically 0.5 mm (0.02
inch) 1.2 mm (0.047 inch) diameter. Up to several thousand hollow fibers
are bundled into a membrane element. At either one or both ends of the
membrane element, the fibers are cast in epoxy. Based upon membrane
design, feedwater can either be fed to the inside of the fibers, with filtrate
leaving from the outside of the fibers (inside-out), or else feedwater can be
directed to the outside of the fibers with filtrate leaving from the inside of the
fibers (outside-in). Membranes are manufactured from several different
materials, depending on the membrane supplier. Typical membrane
materials are polymeric, are naturally hydrophobic and include polysulfone,
polyethersulfone (PES), PVDF, polypropylene, polyacrylonitrile (PAN), and
polyethylene.

A UF system can be set up with membrane modules arranged in parallel or

blocks as shown above.
Wastewater

347

Modes of Operation: Dead-End or Cross-Flow

Dead-end Flow
Hollow fiber systems are commonly operated in a dead-end mode. All the
feedwater is directed across the membrane, leaving the filtered particles
behind on the membrane. Particulates are removed from the membrane
surface by means of a physical backwash that forces the particulates out of
the membrane pores and away from the surface of the membranes. The
backwash may occur every 20 minutes to every few hours, depending on the
system and the feedwater source. With the system operating in a dead-end
mode, operating pressures are generally low (commonly around 10-25 psi),
and there is no recirculation stream that would require extra pumping power.
Over time, the physical backwash will not remove some of the membrane
fouling. Most membrane systems allow the feed pressure to gradually
increase to around 20 - 30 psi and then perform a clean-in-placed (CIP). CIP
frequency might vary from 7 days to several months. A good target is every
one to two months. Another approach is to use a Chemically Enhanced
Backwash (CEB), where, on a frequent basis, chemicals are injected with the
backwash water to clean the membrane and maintain system performance at
low pressure without going off-line for a CIP. CEB chemicals are usually
sodium hypochlorite, caustic, or acid. Much of the success of the hollow
fiber filtration process is establishing an effective backwash and CEB
program.
The backwash and CEB strategy should minimize backwash water losses
while effectively returning the trans membrane pressure back to where it was
at the start of the previous cycle. A good target recovery rate calculated as
(filtrate volume/total volume including flushes)*100, is 92 95% with deadend flow operation.
Flushing and backwash commonly use filtrate water. Sometimes it may be
necessary to use other or better water sources for this operation. The flushing
and backwash cycles allow the following options:
a. Drain downs: This is used to evacuate the water and can be assisted
with air pressure.
b. Air pressurization: This allows a slight expansion of the membranes
and flaking of the deposit to induce better cleaning.
c. Forward flush: Under this step, the filtrate valve is closed and water
is pushed through the feed side of the module to clear accumulated
solids.
d. Bottom backwash: Backwash water enters at the top of the module
through the filtrate line on the filtrate side, through the membrane
Wastewater

348

and out of the module at the bottom and to drain. The backwash flux
will be about three times the processing flux rate.
e. Top backwash: This is the same as bottom backwash, however, the
water exits out of the top feedwater port and then to drain.
f. Chemically enhanced backwash: Provisions for feed of two or three
chemicals are made to inject into the backwash water in front of the
membranes. Typically citric acid, caustic, or sodium hypochlorite is
used. They can be pumped in during the backwash step. Also soak
times can be programmed in to allow longer contact time to help
clean the membrane.
g. A second drain down and air pressurization followed by a second top
backwash may be included.
h. A final rinse puts the unit back into operation mode, but wastes the
filtrate for a short period of time.
An example of a hollow fiber single module pilot unit set-up may be as
follows:
1. Processing time: 45 minutes.
2. Filtrate flow set point: 12 gpm.
3. First drain down: 8 seconds.
4. First air pressurization: 15 seconds.
5. Forward flush: 0
6. Forward flush flow: 0
7. Bottom backwash: 12 seconds.
8. First top backwash: 14 seconds.
9. Chlorine soak: 200 seconds.
10. Chemical soak: 200 seconds.
11. Second drain down: 10 seconds.
12. Second air pressurization: 0
13. Second top backwash: 10 seconds.
14. Final rinse: 10 seconds.
15. Chlorine inject: 4 cycles.
16. Chemical inject: 16 cycles.

Cross-flow
For higher solids waters, the membrane may be set up to operate in a crossflow mode.
Wastewater

349

What could be a rapid buildup of solids at the membrane surface is overcome

by continuously removing a small portion of the flow from the dirty
feedwater side of the membrane. This wastewater along with the water lost in
flushes and backwashes lowers the overall recovery rate down to the 80
90% range. In both the dead-end flow or the cross-flow, strategies can be
incorporated to capture, treat, and reapply the backwash water to the front
end of the membrane filtration process to improve actual overall system
recovery rate.
Forward flush, backwash, and chemically enhanced backwash strategies are
similarly applied whether the system is operated in the dead-end or crossflow mode.

The performance and economics of filtration depend upon the rate at which
water flows through the membrane. This is the flux rate and is expressed in
gallons per square foot of membrane surface area per day (GFD). The system
is set up to operate at a fixed flux rate or filtrate flow rate by the use of a
VFD on the supply pump. Any accumulation of retained material at the
surface will reduce the effective filtration rate and create the need for higher
supply pressure to maintain the set flux rate. Concentration of solids at the
surface occurs in a dynamic state but its effect is similar to the filter cake
build-up at the separation surface in conventional filtration. The trans
membrane pressure builds and the unit will go into the backwash cleaning
cycle. This is established by service cycle time. Excessive trans membrane
pressures must be avoided to prevent damage to the membranes.
Actual desirable flux rates depend upon the membrane manufacturer, flow
path, membrane material, water characteristics, and water temperature. Some
starting point guidelines are shown below. The high-end flux rates are for
relatively clean feedwaters with turbidities of less than 1 NTU or total
Wastewater

350

suspended solids of less than 1. The low-end flux rates are for dirtier waters
with turbidities of 15 NTU or greater, and total suspended solids of 20 ppm
or greater.

Table 1
Water Source:

Hollow Fiber Flux Rate (GFD)

City
water
or
pretreated
surface
water

70 90

Well water

65 90

Raw surface water

58 70

Sea water

45 70

Tertiary waste water

32 50

Treated
industrial
waste water

35 55

Filtrate Water Quality

Both UF and MF will remove suspended particles, algae, and bacteria. UF
will also remove viruses. Table 2 shows the expected filtrate water quality
from MF and UF systems on raw water clarification and compares with
multi-media filtration. Microfiltration will generally provide a 15-minute silt
density index (SDI15) of less than 3 for surface or well waters. Ultrafiltration
commonly achieves an SDI15 below 2. The lower SDI indicates a diminished
potential for downstream reverse osmosis fouling where an RO machine is
part of the treatment process. The removal of suspended solids prevents
fouling and blockage of the RO brine spacer. Biofouling of RO membranes
will be less likely with the removal of bacteria by a MF or UF membrane
process used in front of the RO. Waste waters containing oils and surfactants
and high TOCs that are filtered by MF or UF will commonly result in higher
turbidities and higher SDIs than surface or well waters. Generally,
emulsified oils and certainly free oils are not desirable for most hollow fiber
UF or MF membranes.

Wastewater

351

Table 2. Filtrate Water Quality on Surface or Well Waters

Water
Quality

Multi-media

MF

UF

Turbidity

0.1 2 NTU

< 0.1 NTU

< 0.1 NTU

SDI15

3 Filter
Blinding

<3

<2

Crown Solutions designed and built a pilot UF system to allow dead-end and
cross-flow testing of hollow fiber UF membrane technology for industrial
applications. The purpose for pilot testing includes the following:
1.
2.
3.
4.

Evaluate the effectiveness and reliability of the technology.

Determine chemistry requirements, if any, of the UF feedwater.
Determine reliable flux rates.
Determine necessary backwash frequencies, flow rates, and
durations.
5. Determine chemically enhanced backwash strategies (CEB).
6. Develop budgetary costs for full-scale operation.
7. Test CIP effectiveness.

There are many challenges faced with this technology including variations of
influent water quality, temperature changes, high TOC water,
microbiological effects, and others. Conducting a UF or MF pilot test
requires monitoring of several parameters and making effective adjustments
to determine if a full-scale system will be effective and to determine how to
design it properly.
The pilot system has supervisory control and data acquisition (SCADA)
capability where operating data is automatically acquired and stored. The
data is stored in the RS View software, and then periodically downloaded
into an Excel spreadsheet. Manual data logging is also advisable to backup
the electronically stored data and to check the sensor transmitters.

Hollow Fiber Data Logging and Key Performance Indicators:

The following data should be collected. Manual logging should occur at least
once per day and once per shift if possible. A data set should include data
two minutes before backwash and two minutes after the backwash sequence
is complete. Automatic data entry should be every 2 10 minutes during
pilot testing:

Date.
Time.
Machine run time, hours.

Wastewater

352

Feedwater temperature, C or F.
Screen filter inlet pressure and out pressure, psi.
Membrane module top feedwater pressure, psi.
Membrane module bottom feedwater pressure, psi.
Filtrate pressure, psi.
Feed turbidity, NTU.
Filtrate turbidity, NTU.
Filtrate flow, gpm.
Bleed flow (if in cross-flow mode), gpm.
Recycle flow, gpm.
Drain flow and totalizer.

Filtrate Flux
Filtrate flux is expressed in gallons per square foot of filter area per day
(gfd). Therefore the average filtrate flux is the flow of product water
divided by the surface area of the filter. Filtrate flux is expressed according
to the following equation:
F = Q/S
Where F = filtrate flux (gfd).
Q= filtrate flow in gallons per day (gpd).
S = filter surface area (ft2).

Trans Membrane Pressure

This term describes the average pressure across the filter. It is the net driving
pressure on the membrane and is calculated as follows:
TMP = ((PT + PB)/2) - PF
Where TMP = Trans membrane pressure (psi).
PT = Pressure of feedwater at the top of the filter (psi).
PB = Pressure of feedwater at the bottom of the filter (psi).
PF = Filtrate pressure (psi).

Wastewater

353

Temperature Compensation Factor for Trans Membrane

Pressure (TCF)
The temperature compensation factor is used to normalize the TMP to some
temperature such as 20 C to account for the influence on membrane
expansion or contraction and water viscosity. If the temperature were above
20 C, the TMP would be lower than it would be at 20 C since the pores in
the membrane will be slightly larger and the water will be less viscous. To
normalize to 20 C the TCF will be a number greater than 1.
Likewise, for a temperature below 20 C, the TCF will be less than 1 to
normalize to
20 C, since the membrane will become somewhat tighter and the water will
become more viscous.
The TCF will vary by the manufacturer and the membrane material.
One example for calculating TCF is as follows:
TCF = e 0.031(T-20)
Where TCF = Temperature Correction Factor
T = Water Temperature in C.

Temperature Compensated Trans Membrane Pressure TMP20C

This term normalizes the TMP to account for variations in water temperature.
TMP20C = TMP * TCF
Where TMP20C = Temperature compensated TMP.
TMP = Trans membrane pressure (psi).
TCF = Temperature correction factor.

Wastewater

354

Specific Flux
The term specific flux is used to refer to filtrate flux that has been
normalized for the trans membrane pressure. The equation for specific flux
is as follows:
FTM = F/ TMP20C
Where FTM = Specific flux (gfd/psi).
F = filtrate flux (gfd).
TMP20C = Temperature compensated trans membrane
pressure (psi).

Differential Pressure or P
The pressure difference between the top and the bottom of the UF module of
the feedwater during processing is used to monitor feedwater side fouling.
(Differential pressure of the prefilters should also be monitored).
P = PFeedbottom PFeedtop

(for bottom feed setup)

Where PFeedbottom = Bottom feed pressure.

PFeedtop = Top feed pressure.

Instantaneous Recovery
This is the recovery of water from the system for one processing/backwash
cycle. This only looks at the efficiency based upon backwashes and does not
include chemically enhanced backwashes, CIPs, or bleedoff if in a crossflow mode.
RI =

[ 1V

Backwash

/ VTotalCycle *100, (%).

Where RI = Instantaneous Recovery

VBackwash = Total volume of filtrate used during the single
backwash sequence.
VTotalCycle = Total amount a filtrate produced from a single
processing cycle.
Wastewater

355

Total UF System Recovery

This looks at the percent of filtrate made from the total feedwater flow and
includes water loss to backwash, CEBs, CIPs, and bleedoff.

RT =

[V

Filtrate to Process

/ VTotal Feedwater

] *100, (%).

Where RT = Total System Recovery

VFiltrate to Process = Total volume of filtrate available for use
after accounting for filtrate used for backwashes, CEBs,
CIPs, and bleed-off.
VTotal Feedwater = Total amount a feedwater sent to the UF for
a sufficient period of time that captures all water
consumptions.

Case Studies:
1) Pharmaceutical Plant SBR Wastewater for Recycle.
This pilot test used an extremely challenging water. The plants wastewater
is high in organics so it is treated biologically with four sequential batch
reactors (SBR). The effluent from this secondary treatment process is then
sent through a clarifier followed by multi-media filters prior to RO and then
recycled back for plant cooling. Even after biological treatment and
clarification, the water was still high in TOC and TSS. It was the plants
desire to eliminate the clarifiers and go directly from the SBRs to multimedia, and then UF. A multi-media filter was included as part of the pilot
equipment to be able to test the water without it first going through the
clarifier.
Water after the plants SBR, clarifier, and the plants multi-media filters was
first used as UF feedwater and baseline data gathered.
Next, we tested water directly from the SBRs to the pilot multi-media filter
prior to UF. Even by operating in a cross-flow mode this water was too high
in TSS and TOC for the UF. Fouling occurred rapidly even at a low flux rate
of 30 GFD.
Most of the test period was then dedicated to clarifier effluent water filtered
by the pilot multi-media filter prior to UF. The pilot required the addition of
ferric chloride ahead of the filter to reduce TOC, because TOC can blind UF
membranes rather quickly. The pilot extended several months to establish an
effective strategy. Some key points were as follows:
Wastewater

356

a. Ferric chloride dosage and mix time were important.

b. Backwashing with UF filtrate water created problems when caustic
was used with the CEB cycle. The water is relatively high in calcium
and very high in alkalinity, so the caustic caused scaling. The
backwash water was changed to RO permeate which eliminated the
scaling problem.
c. Developing adequate backwash cycle times, sodium hypochlorite
and caustic dosages, and feed frequencies were important to combat
organic fouling.
d. With variable influent levels of TSS, operation in cross-flow was
necessary.

Typical Water Analyses

SBR
Decant

MMF
Effluent

UF
Filtrate

TSS, mg/L

160

BDL

Turbidity,
NTU

36

<0.1

TOC, mg/L

70-260

70-260

70-190

<3

COD, mg/L

210-320

190-220

180-200

<1

Hardness,

RO
Permeate

200

CaCO3

Alkalinity,

600

CaCO3

pH
Conductivity,
S

SDI 15

7.4
4300
<1

The final trial ran successfully on 45 minutes cycles and a stable specific flux
was managed with an operating flux of 50 GFD. The chemically enhanced
backwash schedule showed to be effective with caustic addition every four
cycles and sodium hypochlorite additions every six cycles. Total system
recovery was 90%.

Wastewater

357

2) Copper Removal in Cooling Tower Blowdown.

A large industrial air separation plant required copper removal from their
cooling tower blowdown water to be less than 17 ppb. Benchtop testing using
an organic metal precipitator followed by ultrafiltration showed that singe
digit Cu ppb numbers could be achieved. The pilot included a mix tank and
chemical precipitator feed along with the UF pilot unit. Very consistent and
reliable results were shown in dead-end mode and 94% recovery.

Average
NTU

Feedwater

39

Turbidity,

Average Filtrate Turbidity, NTU

0.08

Average Flux, GFD

53.8

Average
Pump
Pressure, PSIG

Discharge

18

Average Water Temperature, F

74

System Recovery

94.1%

Total Copper in Influent, ppb

270

Copper in UF Filtrate, ppb

Specific Flux vs Time

12
Chlorine Backwash

Specific Flux (GFD/psi)

10

2
Specific Flux

864.3

833.3

802.3

771.3

740.3

709.3

678.3

647.3

616.3

585.3

554.3

523.3

492.3

461.3

430.3

399.3

368.3

337.3

306.3

275.3

244.3

213.3

182.3

151.3

89.3

120.3

58.3

27.3

Operating Time (hrs)

Wastewater

358

3) Polishing Cold Lime Softened River Water for HRSG

Makeup.
A treatment strategy to provide several thousand gallons per minute of high
purity makeup water for new heat recovery steam generators (HRSGs)
needed to be determined at a large oil refinery. The source water was from a
river notorious for presenting bacteria problems and organic loading. The
water was first clarified with conventional cold lime clarifier, then a pilot UF
showed that it could operate at high recovery and high flux rate and act as a
good pretreatment strategy for RO and demineralization.

Cold Lime Softened Effluent (UF Feedwater)

Parameter

Average Result

Total Organic Carbon (TOC),

mg/L
Ammonia, mg/L
pH, S.U.
Conductivity, S
Silica, mg/L
Total Hardness, mg/L as CaCO3
Calcium Hardness, mg/L as CaCO3
Total Alkalinity, mg/L as CaCO3
Silt Density Index (SDI 15)

3.92
0.20
8.2-8.4
450
6.5
75
50
25
>6

The UF Filtrate quality data that was collected showed excellent results:
Parameter
Product Flow Rate
Water Temperature
Trans Membrane Pressure
Feed Turbidity
Product Water Turbidity
Filtrate Pressure
Silt Density Index (SDI)

UF Filtrate, Average Result

23.8-25.2 gpm; 90 GFD
63.4 deg F
15.8 psig
0.97 NTU
0.05 NTU
8.46 psig
0.80

The process cycle was set for 40 minutes; that is, product water was
produced for 40 minutes before the unit automatically switched over to the
Wastewater

359

backwash sequence. The pilot ran for over 30 days and the TMP never
increased to the 20-psig point where an out of service chemical clean was
actually needed. A cleaning with citric acid was performed followed by
caustic after the pilot and the TMP was fully recovered.
A pilot softener and RO were successfully operated using the UF filtrate for
feedwater as part of this test.

4) Lake Water Clarification.

The city of Chicago uses clarified Lake Michigan water to supply a large
industrial complex. A UF pilot was conducted to determine if UF could be
used for the site to produce their own water. Removal of bacteria and
pathogens was necessary. Operation of the UF to changing TSS levels and
water temperature was important from a design basis.
The UF system was operated in the dead-end mode for the duration of the
study. The average turbidity of the lake water was typically less than 10
NTU. That value ranged from 40 to 100 NTU twice a year when the lake
turns over. Hardness ranged from 132 mg/L to 140 mg/L as CaCO3 and
alkalinity ranged from 108 mg/L to 120 mg/L as CaCO3. Conductivity
ranged from 301 to 319 S/cm, and the pH ranged from 8.0 to 8.1. The pilot
was initiated during the winter months when lake water temperatures were at
the lowest. The temperature range was from 40 to 76F.
The pilot study allowed us to determine the effectiveness of the selected
technology to generate a water quality that is similar to potable water. At the
same time, a budgetary cost estimate was to be developed. This cost estimate
was based on the data and operating conditions experienced during the pilot.
This information was used to project the costs of operating a 1000-gallon per
minute system. The data evaluation included determining the following:

Characterization of Filter Flux and Recovery

Evaluation of Cleaning Strategy and Efficiency

Evaluation of Filtered Water Quality

Filter Integrity Testing

Microbial Removal

During start-up and prior to the UF system running continually, lake water
turbidity was averaging 60 NTU, with surges as high as 100 NTU. While the
system did not run continuously during those high readings, when in
operation, the filtrate turbidity was 0.030 NTU or less.

Wastewater

360

The following testing parameters were conducted throughout the pilot study:

Total Hardness

Total Alkalinity

Silt Density Index

Total Suspended Solids

AMC

Total Organic Carbon

Total Coliforms

Pyrogens

Particle Size Count

Two different UF filter modules were used during the pilot. They differed in
the diameter of the lumens and total surface area.
A multi-media filter was installed in front of the UF unit to act as a prefilter
for lake upset conditions. During the initial portion of the study, the filtration
system ran at varying target flux rates due to a variation in feed water
temperature, which ranged from 40 F to 70 F. The resulting flux rate from
these temperatures ranged from 40 GFD to 51 GFD. Filtrate cycle length,
which is the time between backwashes, was 30 minutes. Each filter cycle run
was followed by a 250-second backwash sequence. Every fourth backwash
cycle, the backwash water was chlorinated at 5 to 10 ppm free chlorine.
Increases in TMP were recovered with extended chlorine soaks.
Initially, the water temperature was 41 F and the operating flux rate was 40
GFD. These operating conditions were experienced until April 14, 2003
when the feedwater temperature increased to approximately 70 F. During
that time period, the flux rate was 46.6 GFD. During the study, a filtration
cycle consisted of 30 minutes. System recovery was 93%.
Silt Density Index
Date

Filtrate

4/18/03

1.05

4/21/03

1.05

5/1/03

1.00

5/22/03*

1.01

5/23/03*

0.78

* Denotes phase II testing

Wastewater

361

Bacteria analysis showed some low levels in the filtrate. The pilot unit used
for this test allowed some cross contamination of backwash water with
filtrate at the sample point and this was thought to be the major source of
bacteria found in the filtrate.

Date
3/25/03
3/31/03
4/14/03
4/21/03
4/28/03
5/1/03
Date
3/5/03
3/21/03
3/25/03
3/26/03
3/28/03
3/31/03
4/2/03
4/4/03
4/14/03
4/16/03
4/18/03
4/21/03
4/23/03
4/25/03
4/28/03
4/30/03
5/2/03
5/5/03
5/20/02*
5/21/03*
5/22/03*
5/23/03*

Total Organic Carbon (ppm)

Lake Water
2.9
<2.4
<2.4

AMC (cfu/100 ml)

Lake Water
194,333
743
1,237
38
300
27
35
82
223
4,600
150,000
179,333
66,333
198,333
199,000
53,000
>200,000
>200,000
92,000
194,000
181,667
64,000

Filtrate
2.03
1.73
1.54
1.73
2.03

Filtrate
12,133
957
136
29
36
4
1
5
3
25
31
1
15
101
88
29
15
74
47
77
56
101

* Denoted Phase II testing

Wastewater

362

Pyrogens (EU/ml)
Date

Lake Water

Filtrate

3/25/03

>32.0

> 2.0, <4.0

3/31/03

>32.0

>0.5, <1.0

4/14/03

>32.0

>1.0, <2.0

4/21/03

>32.0

>4.0, <8.0

3/28/03

>32.0, <64.0

>1.0, <2.0

5/5/03

>128.0, <256.0

>0.5, <1.0

Filtrate Flux vs. Feedwater Temperature

F ilt r a te F lu x v s F e e d w a t e r T e m p e r a t u r e

80
75
70
65

D e n o te s th e e n d o f P h a s e I

60
55
I n c r e a s e in f e e d w a t e r t e m p e r a t u r e

GFD

50
45
40
35
30

F e e d w a te r te m p e ra tu re d e c re a s e

25
20
15
10
73

70

67

64

61

58

55

52

49

46

43

40

37

34

31

28

25

22

19

16

13

10

5
D a y s o f T ria l
F lu x

Tem p

Transmembrane Pressure
17
16
H ig h T M P d u e to in a p p ro p r ia te flu x r a te

15

H ig h fe e d w a te r tu r b id it y p ro c c e d in g t m p s p ik e

14
13
12
11
Pressure

10
9

T M P a f te r c h lo r in e a n d c a u s tic s o a k

8
7

D e n o te s th e e n d o f P h a s e I

6
5
4
3
2
1
73

70

67

64

61

58

55

52

49

46

43

40

37

34

31

28

25

22

19

16

13

10

0
D a y s o f T ria l

Wastewater

363

Feedwater Turbidity
60
55
50
45
D e n o te s e n d o f P h a s e I

40

NTU

35
30
25
20
15
10
5

73

70

67

64

61

58

55

52

49

46

43

40

37

34

31

28

25

22

19

16

13

10

0
D a y s o f tria l

Filtrate Turbidity
0 .1 0 0
0 .0 9 0
0 .0 8 0
R e a d ing o b t a in e d im m e d ia t e ly a f te r s y s t e m re s ta rt
S y s te m s h u t d o w n p re v io u s d a y 4 / 2 7 / 0 3

0 .0 7 0

NTU

0 .0 6 0
0 .0 5 0
D e n o t e s th e e n d o f P h a s e I

0 .0 4 0
0 .0 3 0
0 .0 2 0
0 .0 1 0

73

70

67

64

61

58

55

52

49

46

43

40

37

34

31

28

25

22

19

16

13

10

0 .0 0 0
D a y s o f T r ia l

Wastewater

364

Hollow Fiber Design

Inside-Out
Cross-flow
Top Feed

Filtrate
Feedwater

Concentrate
Wastewater

365

Hollow Fiber Design

Inside-Out
Dead-end Flow

Filtrate
Feedwater

Feedwater
Wastewater

366

Hollow Fiber Design

Inside-Out
Backwash
Bottom

Filtrate

Dirty Water
to Drain
Wastewater

367

Pilot Hollow Fiber UF System and Pilot Reverse Osmosis System

Wastewater

368