For many years, the multi-pass efficiency test according to the standard ISO 16889 has been a well-established and widely accepted procedure to evaluate lifetime and (fractional) efficiency of filter elements. However, the continuous ingression of test dust at relatively high concentrations and the constant flow rate have been subject to some criticism. A major argument for this is that under real operating conditions, both the concentration of contaminants and the flow rate will not be constant in time and therefore, a static multi-pass test with high ingression rates can underestimate relevant effects such as the resuspension of captured particles into the liquid.

Several alternative multi-pass test procedures were proposed to resolve this issue by prescribing conditions such as a variable flow rate for (at least part of) the test duration, a reduced base upstream gravimetric level and so-called stabilization phases during which no test dust is injected at all. Examples include initiatives by companies and industrial standards. A common feature of these test methods is a periodic alternation of the flow rate between two values with relatively short transition stages (see Fig. 1, left). Therefore, the evaluation of the filter performance includes the efficiency properties for two different flow rates and the filtration behavior during rapid changes of the flow.

Specialized computer simulation tools have proven to be very useful for the innovation and optimization of filter element designs. However, filtration models developed and validated for static flow cannot be expected to describe the dynamics in the filtration process under cyclic flow conditions. For instance, it was observed that filter media (or elements) with almost the same performance in a static test can display a quite different efficiency behavior when tested under variable flow conditions (see Fig. 1, right).

The present work is devoted to the extension of filtration models to the case of unsteady flow rates. In the case of media flat sheets and for the sake of quick identification of parameters, a one-dimensional model is sufficient to test the extended models. In particular, the influence of the cyclic flow (transition stages between the two flow rates) and the stabilization phases on the efficiency performance factors are investigated. The modeling approach is presented and illustrated by examples...