Membrane integrity monitoring

Most of the commercial MF and UF membranes used for water treatment have small enough pore sizes to remove bacteria and protozoa including Giardia and Cryptosporidium. Since majority of viruses are included in floc and the effective pore sizes of the dynamic membrane (or cake layer) formed on membrane surface are much smaller than the membrane pore sizes, MF and UF are capable of removing vast majority of viruses.

Maintaining rigorous membrane integrity is generally not the highest priority in MBR since the effluent can be disinfected by UV and/or chlorine before being discharged, if necessary. In addition, MBR effluent is not recycled for direct human contact. As a result, regular integrity tests beyond turbidity and/or particle count monitoring are not required in typical situations.

Membrane integrity test methods can be divided by direct methods and indirect methods as summarized in Table 1. The existing direct integrity test methods are based on air pressure decay, diffusive air flow monitoring, sonic test, and bubble point (or visual bubble) test. Pressure decay test (PDT) is most commonly used off line test method. Sonic and bubble point tests (BPT) also can be used offline to locate the leak points when PDT indicates potential leaks (Crozes, 2002). Membrane integrity can be also checked indirectly by testing permeate quality, where turbidity measurement, particle analysis, and microbial analysis are most commonly used.

Table 1. Direct and indirect membrane integrity monitoring methods (modified from USEPA, 2001)

 Direct MethodsIndirect Methods
Definition– Physical method directly applied to membrane unit–  Water quality monitoring to correlated with membrane integrity
Advantages/ Disadvantages–  Mostly simple and accurate

–  Typically in situ measurement

–  Not for continuous monitoring

–  Filtration must pause during the test

–  Continuous monitoring with low accuracy (turbidity/particle monitoring)

–  Intermittent monitoring with high accuracy (bacterial/viral count)

Examples–  Pressure decay test (PDT)

–  Vacuum decay test (VDT)

–  Diffusive air flow test (DAF)

–  Bubble point test (BPT)

–  Sonic test

–  Turbidity monitoring

–  Particle count

–  Bacterial/viral count

The detection ranges of various integrity test methods are compared in Fig. 1. The log removals in y-axis are based on the rejection efficiencies of protozoa such as Giardia and Cryptosporidium at the time of detection. For example, if any method can detect the integrity breach when protozoa removal is 99%, log-removal is calculated at 2 (= -log(1-0.99) ). Turbidity and particle counting are generally less accurate than PDT and DAF.

The accuracies of various integrity test methods are compared in Fig. 2, where the title of y-axis means the probability of missing the leak events. Pressure decay tests (direct method) and microbial challenge tests (indirect method) have virtually no chance of missing membrane integrity breach, but turbidity and particle indexing/counting have 33-67% of chance of missing it.

Details of each integrity test method are found here.
image002 (2)gh Fig. 1. Relative sensitivity of various monitoring methods in detecting protozoa (Johnson, 1998).

image004 (1)hjkFig. 2. Comparison of the accuracy of some integrity test methods. False negative means the case where leaks are conceived intact.