Wastewater 346

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 µm to 1.0 µm. Ultrafiltration is a tighter membrane and is in the micron range of approximately

0.005 - 0.1 µm. UF is typically expressed in terms of molecular weight cut-off 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 347

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 348

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 dead-end 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 349

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 cross-flow mode.

Wastewater 350

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 cross-flow 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 351

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 TOC’s that are filtered by MF or UF will commonly result in higher turbidities and higher SDI’s than surface or well waters. Generally, emulsified oils and certainly free oils are not desirable for most hollow fiber UF or MF membranes.

Wastewater 352

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. Evaluate the effectiveness and reliability of the technology. 2. Determine chemistry requirements, if any, of the UF feedwater. 3. Determine reliable flux rates. 4. 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 353

• 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 354

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 TMP20°C

This term normalizes the TMP to account for variations in water temperature.

TMP20°C = TMP * TCF

Where TMP20°C = Temperature compensated TMP. TMP = Trans membrane pressure (psi). TCF = Temperature correction factor.

Wastewater 355

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/ TMP20°C

Where FTM = Specific flux (gfd/psi).

F = filtrate flux (gfd).

TMP20°C = 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, CIP’s, or bleedoff if in a cross-flow mode.

RI = [ 1 – VBackwash / 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 356

Total UF System Recovery This looks at the percent of filtrate made from the total feedwater flow and includes water loss to backwash, CEB’s, CIP’s, and bleedoff.

RT = [ VFiltrate 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, CEB’s, CIP’s, and bleed-off. VTotal Feedwater = Total amount a feedwater sent to the UF for a sufficient period of time that captures all water consumption’s.

Case Studies:

1) Pharmaceutical Plant SBR Wastewater for Recycle. This pilot test used an extremely challenging water. The plant’s 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 plant’s desire to eliminate the clarifiers and go directly from the SBR’s to multi-media, 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 plant’s SBR, clarifier, and the plant’s multi-media filters was first used as UF feedwater and baseline data gathered. Next, we tested water directly from the SBR’s 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 357

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

RO Permeate

TSS, mg/L 160 5 BDL Turbidity, NTU

36 2 <0.1

TOC, mg/L 70-260 70-260 70-190 <3 COD, mg/L 210-320 190-220 180-200 <1 Hardness, CaCO3

200

Alkalinity, CaCO3

600

pH 7.4 Conductivity, µS

4300

SDI 15 <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 358

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 Feedwater Turbidity, NTU

39

Average Filtrate Turbidity, NTU 0.08 Average Flux, GFD 53.8 Average Pump Discharge Pressure, PSIG

18

Average Water Temperature, °F 74

System Recovery 94.1% Total Copper in Influent, ppb 270 Copper in UF Filtrate, ppb 7

Specific Flux vs Time

0

2

4

6

8

10

12

27.3

58.3

89.3

120.

3

151.

3

182.

3

213.

3

244.

3

275.

3

306.

3

337.

3

368.

3

399.

3

430.

3

461.

3

492.

3

523.

3

554.

3

585.

3

616.

3

647.

3

678.

3

709.

3

740.

3

771.

3

802.

3

833.

3

864.

3

Operating Time (hrs)

Spec

ific

Flux

(GFD

/psi

)

Specific Flux

Chlorine Backwash

Wastewater 359

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 (HRSG’s) 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

3.92

Ammonia, mg/L 0.20 pH, S.U. 8.2-8.4 Conductivity, µS 450 Silica, mg/L 6.5 Total Hardness, mg/L as CaCO3 75 Calcium Hardness, mg/L as CaCO3 50 Total Alkalinity, mg/L as CaCO3 25 Silt Density Index (SDI 15) >6 The UF Filtrate quality data that was collected showed excellent results:

Parameter UF Filtrate, Average Result

Product Flow Rate 23.8-25.2 gpm; 90 GFD Water Temperature 63.4 deg F Trans Membrane Pressure 15.8 psig Feed Turbidity 0.97 NTU Product Water Turbidity 0.05 NTU Filtrate Pressure 8.46 psig Silt Density Index (SDI) 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 360

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 µS/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 76°F. 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 361

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 362

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.

Total Organic Carbon (ppm) Date Lake Water Filtrate

3/25/03 2.9 2.03 3/31/03 <2.4 1.73 4/14/03 <2.4 1.54 4/21/03 1.73 4/28/03 2.03 5/1/03

AMC (cfu/100 ml) Date Lake Water Filtrate

3/5/03 194,333 12,133 3/21/03 743 957 3/25/03 1,237 136 3/26/03 38 29 3/28/03 300 36 3/31/03 27 4 4/2/03 35 1 4/4/03 82 5 4/14/03 223 3 4/16/03 4,600 25 4/18/03 150,000 31 4/21/03 179,333 1 4/23/03 66,333 15 4/25/03 198,333 101 4/28/03 199,000 88 4/30/03 53,000 29 5/2/03 >200,000 15 5/5/03 >200,000 74

5/20/02* 92,000 47 5/21/03* 194,000 77 5/22/03* 181,667 56 5/23/03* 64,000 101

* Denoted Phase II testing

Wastewater 363

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

F i l t r a t e F l u x v s F e e d w a t e r T e m p e r a t u r e

5

1 0

1 5

2 0

2 5

3 0

3 5

4 0

4 5

5 0

5 5

6 0

6 5

7 0

7 5

8 0

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73

D a y s o f T r i a l

GFD

F l u x T e m p

D e n o t e s t h e e n d o f P h a s e I

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

F e e d w a t e r t e m p e r a t u r e d e c r e a s e

0

1

2

3

4

5

6

7

8

9

1 0

1 1

1 2

1 3

1 4

1 5

1 6

1 7

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73

D a y s o f T r i a l

Pres

sure

D e n o t e s t h e e n d o f P h a s e I

H i g h f e e d w a t e r t u r b i d i t y p r o c c e d i n g t m p s p i k e

T M P a f t e r c h l o r i n e a n d c a u s t i c s o a k

H i g h T M P d u e t o i n a p p r o p r i a t e f l u x r a t e

Transmembrane Pressure

Filtrate Flux vs. Feedwater Temperature

Wastewater 364

0

5

1 0

1 5

2 0

2 5

3 0

3 5

4 0

4 5

5 0

5 5

6 0

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73

D a y s o f t r i a l

NTU

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

Feedwater Turbidity

0 .0 0 0

0 .0 1 0

0 .0 2 0

0 .0 3 0

0 .0 4 0

0 .0 5 0

0 .0 6 0

0 .0 7 0

0 .0 8 0

0 .0 9 0

0 .1 0 0

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64 67 70 73

D a y s o f T ria l

NTU

D eno tes the e nd o f P ha se I

R e a d ing ob ta ine d im m e d ia te ly a f te r sys te m re s ta rtS yste m shu t do w n p re v io us d a y 4 /2 7 /0 3

Filtrate Turbidity

Wastewater 365

Filtrate

Feedwater

Concentrate

Hollow Fiber Design Inside-Out Cross-flow Top Feed

Wastewater 366

Filtrate

Feedwater

Feedwater

Hollow Fiber Design Inside-Out

Dead-end Flow

Wastewater 367

Filtrate

Dirty Water to Drain

Hollow Fiber Design Inside-Out Backwash

Bottom

Wastewater 368

Pilot Hollow Fiber UF System and Pilot Reverse Osmosis System