A new filter test for assessing crossflow membrane filtration

摘要

IN CONTRAST, the predictive test presented in this article has been developed specifically for crossflow membrane filtration and has the advantage that both filterability as well as backwashing of the membranes can be checked [7]. In the course of time, various possibilities of investigating filterability of beer have been developed. In addition to investigations to identify filtration inhibiting beer components, filter tests were established in order to predict filterability at an early production stage [1-3, 5,6,10,12,15,16].These filter tests are based on the physical principles of industrial filter systems and simulate nitrations on a laboratory or pilot scale. In 1968, Dr. Karl Raible developed one of the first predictive tests for kieselguhr filtration [12]. This test method was aimed at assessing filtration properties of beer, based on calculating the filter cake factor. Higher values indicate inferior filterability. It was effective in investigating interactions between filter aids and beer haze particles and could be used for setting kieselguhr dosage and composition. This method was simplified and standardised later on in that the type of kieselguhr suspension and ap- plication (screen with a pore size of 15 μm), filtration pressure (1 bar) and evaluation were adjusted [13]. In addition to the filter cake factor, the specific filtrate volume (F_(spec)) was proposed as a guide for hourly filtration throughput and filter surface area (m~2) at a constant filter pressure. Kreisz [4] described an assessment of the F_(spec) value [51. In adsdition, practical tests confirmed the comparability with industrial kieselguhr filtration [1, 5, 11, 13]. However, the small test volumes were criticised in terms of homogeneity in large storage tanks. Ziircher et al. [ 16] and Reed et al. [14] thus carried out kieselguhr candle filter tests simulating pressure increase and filter performance on a pilot scale (up to 50 1 sample volume). The tests were much more elaborate as they used a larger sample volume.