Silicon carbide membrane are submerged ultrafiltration flat-plate ceramic membranes with a pore size of 0.silicon carbide membrane 1 microns that act as a physical barrier, blocking solids and pathogens but allowing clean water to pass through. They are known for their chemical resistance and long service life, offering a cost-effective alternative to conventional polymeric membranes. They also operate with lower energy consumption and require fewer chemical cleans, decreasing downtime and operator intervention. In this case study, SA Water piloted Cembrane silicon carbide (ceramic) membranes at its Mount Pleasant WTP in South Australia, to explore their performance in varying River Murray source water quality and upstream aeration conditions. The Mount Pleasant plant operates two process streams – a traditional green sand filter stream and a polymeric membrane one. Variability in river water quality can cause issues with odour, taste and algal concentrations which, coupled with constraints on chemical dosing upstream of the polymeric membranes, often force a trade-off between capacity and premature irreversible fouling.

Coagulant dosing is an important aspect of preventing fouling, and is a common component in the operation of many WTPs.Silicon carbide membrane to understand how a silicon carbide membrane would perform at the plant, a trial was conducted to test its ability to cope with a brief failure in coagulant dosing. The results showed that the ceramic membranes were able to continue operating at the normal flux rate and recovery, with only a slight decline in permeability compared to when coagulant dosing was in full effect. In fact, the membranes were able to recover from an ACH overdose in less than a day.

In addition to the permeability data, back-pulse pressures and tank drain cycles were recorded. This allowed a comparison of how different reductions in coagulant doses impacted on turbidity buildup, initial fouling rates, recovery and TMP increase over time. In general, the higher the coagulant dose used, the slower the TMP increase and the longer the membrane could be expected to run before needing a chemical clean (Figure 4).

The sensitivity of the ceramic membranes to varying levels of coagulant dosing was evaluated using grab samples taken from the membranes following each back-pulse, between tank drain cycles and after a chemical clean. A sensitivity matrix was developed that offered insight into which combination and concentration of chemical cleans are needed to achieve the same result. This information can then be used to establish a protocol for cleaning and de-dosing. TMP increases over time provided an indication of when a membrane could be expected to experience irreversible fouling. As a result, the frequency of back-pulses and tank drain cycles can be optimised.