Pretreatment & Membrane Performance

The Guadalupe Blanco River Authority (GBRA) has developed a new water supply system that withdraws water from Canyon Lake in central Texas, treats it at the Western Canyon Regional Water Treatment Plant and distributes potable water to wholesale customers in three counties.

The treatment process is direct filtration and consists of coagulation, flocculation, membrane filtration and disinfection. Chlorine is used for primary and secondary disinfection. The current capacity of the plant is 10 million gal per day (mgd), and its ultimate capacity after expansion will be 15 mgd. The membrane filtration process uses Pall’s Aria microfiltration system. The plant has eight racks, each of 1.43-mgd capacity. The design membrane flux is 34.3 gal per sq ft per day (gfd).

Since the plant startup, GBRA has been able to consistently meet customers’ requirements in terms of plant production and water quality; however, a few operational challenges related to coagulation using ferric sulfate were observed. Therefore, GBRA has evaluated the use of alternative coagulants such as polyaluminum chlorides (PACls) and aluminum chlorohydrates (ACHs).

Startup Challenges

GBRA started the full-scale plant in April 2006 with ferric sulfate as the coagulant. Ferric sulfate was added at dosages of 30 to 60 mg/L. Two rounds of acceptance testing were conducted. The goals of acceptance testing were to operate the membrane racks at their design flux and demonstrate 30-day operational cycles between clean-in-places (CIPs). Transmembrane pressures (TMPs) for the test racks exceeded the threshold TMP for CIP of 35 psi in approximately 25 days. Due to excessive fouling of membranes from iron-organic complexes, the run cycles were shorter than 30 days.

Also, the CIPs were not able to restore the specific fluxes or the TMPs. Normal CIPs consisted of citric acid, sodium hydroxide and sodium hypochlorite. GBRA had to conduct aggressive CIPs with hydrochloric or sulfuric acids (supplementing the recipes for normal CIPs) to recover fluxes/TMPs. Sometimes back-to-back CIPs were needed to remove the membrane foulants. These operational measures to recover fluxes/TMPs resulted in increased downtime, lower recoveries and increased waste generation.

Alternative Coagulants

To overcome the operational challenges, GBRA initiated bench and pilot scale testing of alternative coagulants that included PACls and ACHs. The following are key observations from the alternative coagulant tests:

• PACls/ACHs were able to sufficiently remove the disinfection byproduct precursor materials and allow GBRA to continue to use free chlorine for primary/ secondary disinfection.
• PACls/ACHs resulted in negligible pH/alkalinity suppression, which allows GBRA to reduce its post- filtration caustic dosage; caustic is being added to raise the pH of treated water and reduce the corrosion potential.
• No dissolved aluminum was observed in filtered samples after the PACl/ACH addition.
• Use of PACl/ACH resulted in lower TMPs, easier recovery of fluxes/TMPs from enhanced flux maintenance (EFM) and CIP procedures.

PACl/ACH Conversion

Based on the promising results from bench and pilot studies, GBRA decided to convert to PACl/ACH from ferric sulfate. GBRA discussed the conversion plan with the state regulatory agency, the Texas Commission of Environmental Quality. GBRA obtained approval from the state to demonstrate the performance of PACl/ACH for a few months at the full-scale plant. The coagulant conversion included the following steps:

• Cleaning of backwash wastewater recovery basins to minimize interference of metal hydroxides from different coagulants.
• Purging of coagulant supply lines to minimize chance of coagulant blending.
• A contingency plan to return the plant to ferric sulfate in the event of any issues using PACl/ACH.
• Modifications to the membrane chemical cleaning regimes; normal cleaning regimes were used.

Full-Scale Plant Results

The TMPs over the 30-day test period were well below the threshold TMP for CIP of 35 psi. This plot also shows the steady recovery of TMPs with weekly EFMs and complete recovery of initial TMP after CIP.

Since conversion to PACl/ACH, GBRA has been operating the plant with fewer operational issues. The benefits that GBRA has been deriving from conversion to PACl/ACH are:

• Improved plant operation and performance. PACl/ACH are more compatible with the membranes compared to ferric sulfate.
• Lowered usage of other chemicals—eliminated the use of sulfuric or hydrochloric acid in CIPs, extended EFMs interval from every four days to weekly and reduced the caustic dose at post- filter location.
• Decreased the amount of CIP waste generation by eliminating back-to-back CIPs and maintaining a CIP interval of 30 days.
• Decreased energy costs from lower TMP and decreased power for feed pumps.

The conversion to PACl/ACH needs to account for some issues. PACl/ACH solids are more voluminous and less concentrated compared to the ferric sulfate solids. The PACl/ACH solids are difficult to dewater. This translates to increased solids handling and disposal costs. On a unit cost basis (dollars per pound), PACls/ACHs are generally more expensive than the ferric sulfate or ferric chloride. However, the cost comparison for PACl/ACH should account for cost savings that would be realized from lower usage of other chemicals.