The overall objective of the project is to facilitate an increase in water recycling through the use of membrane bioreactor (MBR) technology. This is being accomplished over two phases. Phase I consists of a side-by-side pilot test of six different MBR systems at the Honouliuli WWTP in Ewa Beach Hawaii to investigate the consistency of water quality, the reliability and the operability of the technology for three different waste streams. Phase II consists of pilot testing some of the MBRs for procurement-based design at actual full-scale application sites and/or at permanent locations. Phase II includes a comprehensive study of potential applications for MBRs on Oahu including satellite reclamation, plant expansions, plant upgrades to facilitate recycling, and decentralized treatment for proposed/new developments.

The City and County of Honolulu (Division of Wastewater Treatment and Board of Water Supply) have a series of wastewater treatment applications for which MBRs may be feasible. These include 1) treatment of raw wastewater at pump stations for nearby water recycling applications, 2) treatment of primary effluent to upgrade existing wastewater treatment plants for water recycling, 3) treatment of primary effluent for concurrent nitrogen and phosphorus removal for discharge in environmentally sensitive areas, and 4) treatment of a high-strength solids handling recycle stream for organic and color removal. Because there was no experience with MBRs in Hawaii, a side-by-side pilot study with five MBR vendors was conceived and is currently underway at the Honouliuli Wastewater Treatment Plant in Honolulu, Hawaii. The study is being conducted in two phases. In the first phase, three different waste streams were treated one at a time by the MBRs side-by-side: raw wastewater, high-strength solids handling recycling stream (centrate), and primary effluent. In the second phase, some of the MBRs will be moved to other locations for pilot testing associated with full-scale design and a countywide MBR application feasibility study will be conducted. Under each operating condition in Phase I, the following parameters are being evaluated for each MBR wastewater treatment system: 1) basic performance, 2) operation and maintenance requirements, and 3) life cycle treatment costs. One of the key features of the study is that City operations staff are operating and maintaining the systems and keeping detailed records three shifts per day.

US Filter	Ionics	Zenon	Enviroquip	Koch	Huber

METHODS:

The investigation began in April 2003 with one MBR (Enviroquip) in place. The Ionics unit began operation in June 2003, and the Zenon unit got started in August 2003. US Filter and Huber pilot units came on-line in April 2004. Most recently a unit from Koch Membrane Systems was installed. Each vendor employs somewhat different technologies including membrane configuration and pore size:

• US Filter – microfiber membranes, 0.4 µm pore size, vertical arrangement in off-line tank, air scour and backpulsing;


• Ionics – microfiber membranes, 0.4 µm pore size, horizontal arrangement in aeration tank, air scour and relaxation;


• Zenon – microfiber membranes, 0.04 µm pore size, vertical arrangement in aeration tank, air scour and relaxation and backpulsing);


• Enviroquip – flat panel membranes, 0.4 µm pore size, vertical arrangement in aeration tank, air scour and relaxation;


• Koch – Hollow fiber membranes, 0.1 µm pores, vertical arrangement in membrane tank, intermittent air scour, backflushing;


• Huber – flat panel membranes, 0.025 mm pore size, vertical arrangement on rotating shaft in aeration tank, air scour and spray wash

Four of the pilot units (all except Huber) include anoxic tankage for biological nitrogen removal. Phase IA began in September 2003, Phase IB began in April 2004, and Phase IC began in August 2004.

In Phase IA the MBR units treated raw wastewater with the goal of demonstrating suitability for water recycling and for nitrogen removal. Raw wastewater was conveyed to a headworks structure with two alternative fine screens. One of the screens was a static slotted-type screen with 0.5 mm openings and the other screen was a brush-type punched-hole screen with 3 mm openings. In Phase IA, all three MBRs used the 0.5 mm screen. For Phase IB, all five of the MBRs were switched to the 3 mm screen and treated a high-strength solids handling stream (centrate). In Phase IC, the feed was switched to primary effluent and the 3mm screen was utilized.

24-hr composite samples are collected and analyzed 5-days-per-week for BOD5, TSS, and color; 3- days-per-week for TOC, nitrogen and phosphorus species, TDS, turbidity, UV transmittance, and pH: one-day-per-week for anions, cations, and oil and grease. Grab samples are also collected 5-days-perweek and analyzed for turbidity, UV transmittance (UVT), fecal coliform, and F-specific coliphage. Standard methods are used for analyses where applicable.

RESULTS AND DISCUSSION:

MBRs are generally operated by maintaining a target mixed liquor suspended solids concentration (MLSS) rather than a target solids retention time (SRT). The MBR pilot units have been operated at MLSS between 6,000 and 16,000 mg/L. We have found that the operation is optimal between 10,000- 12,000 mg/L. We have found that when MLSS was very high, the pilot units were dissolved oxygen limited and nitrification was inhibited. The MBR units were able to operate at their advertised permeate flux rates of either 10 or 15 gallons/ft2-day. The TMPs of each unit were monitored and were a good indicator of needed cleaning (generally when TMP > -4 psi). Each of the MBR pilots was cleaned in place several times using either a dilute chlorine solution or a dilute acid solution in the event that the chlorine cleaning was inadequate. Cleanings would normally only be required on an annual or semi-annual basis. However, during this pilot test there were initially several incidents in which polymer from the main treatment plant were allowed to contaminate the MBRs. Practices were modified to alleviate this problem. Also, during Phase IA, there were several sustained power outages and several large storm events which caused shutdowns and/or greatly fluctuating influent conditions that necessitated membrane cleanings.

Many of the water quality parameters were analyzed five days per week resulting in hundreds of data points for each phase. Overall average values and selected removal efficiencies for each phase are shown in Tables 1-3. Selected data are presented in the form of distributions rather than time-course plots (Figures 1-12) in which the data sets are analyzed to determine the percentage of the data points that are smaller than a given numerical value. This allows the reader to easily see the overall distribution of the data as well as get a feel for the maximum, minimum and average values. Figures 1-3 show permeate BOD5 data for the pilot units during Phase IA, IB, and IC, respectively. These figures show the very low values of effluent BOD5 that are typical for MBR systems. Figure 1 indicates that greater than 90% of the BOD5 values are less than 5 mg/L for each of the MBRs and that the performance of the three different types of MBRs are very comparable. Figures 2 and 3 show that in Phase IB some of the BOD values were higher, and in Phase IC the values were the lowest. Figures 4-6 show permeate TSS data for Phases IA, IB, and IC, respectively. Again, these figures show the very low values of effluent TSS that are typical for MBR systems. For both raw wastewater and primary effluent, greater than 90% of the TSS values are less than 3 mg/L for most of the MBRs. Values were slightly higher for the difficult centrate stream.

Figure 7 shows that the MBR permeates contain low yet significant amounts of organic carbon. Comparison of Figures 1, 2 and 7 indicates that effluent TOC is larger than effluent BOD5. This means that a small amount of soluble organic matter which is not readily degradable as BOD5 passes through the MBR systems. This is often denoted as soluble microbial products (SMP) which can be fractionated into carbohydrates, proteins and lipids. Figure 7 shows that the MBR permeates contained greater amounts of presumably recalcitrant organic carbon during Phase IB than IA. Figure 7 indicates that permeate TOC concentrations increased by 10 to 20 mg/L across the whole distribution. The permeate TOC is most likely indicative of additional SMP present in the centrate waste stream (the centrate results from a heat-treat sludge process that lyses all bacterial cells). This additional SMP and or fine organic colloidal material was responsible for enhanced fouling rates in this phase.

Figure 8 shows total nitrogen removal efficiency during Phase IA. Because all three of the MBR pilots used in Phase IA were equipped with anoxic zones and mixed liquor recycle systems, these units are capable of significant nitrogen removal. However, the degree of nitrogen removal is dependent upon achievement of nitrification prior to denitrification. At various times there was insufficient dissolved oxygen present in the MBRs to allow complete nitrification, and under these conditions, denitrification-based nitrogen removal was reduced. Figure 9 shows residual orthophosphate during Phase IB. Effluent data for turbidity for Phase IA are shown in Figure 10. The data show the very low values of permeate turbidity that are typical for MBR systems. The figure indicates that greater than 90% of the turbidity values are less than 0.1 NTU for each of the MBRs and that the performance of the three different types of MBRs are somewhat different. These data are typical of Phases IB and IC. Figure 11 indicates that greater than 85 to 95% of the UVT values are greater than 65% for each of the MBRs and that the performance of the three different types of MBRs are very comparable. Figure 12 shows typical oil and grease data from the study.

CONCLUSIONS:

The conclusions thus far from the pilot study are that all six MBR technologies produce excellent quality permeate suitable for water recycling. There are differences in permeation cycle times, nitrification/denitrification capabilities, required amount of operator attention, membrane cleaning frequency, power requirements, and robustness of the systems. It is apparent that there are many factors other than just water quality that are important in the selection of an MBR system.