Probabilistic Design Framework for Granular Filters

Abstract

A granular filter is required to satisfy two requirements of retention and hydraulic conductivity. The current design approaches are based on the representative grain sizes for retention and hydraulic conductivity requirements. The importance of consideration of the grain-size distribution (GSD) in satisfying the filter requirements is highlighted in this paper. An assessment criterion is presented for retention and hydraulic conductivity requirements based on GSD. A probabilistic assessment criterion is developed for retention requirements considering the grain size and constriction size as random variables. For hydraulic conductivity requirements, a probabilistic assessment criterion is defined by considering hydraulic conductivity’s variability and a semianalytical model for saturated hydraulic conductivity. The limit states for developed criteria are established based on published experimental data. The proposed approach is demonstrated with examples, and it is illustrated that the filter bandwidth is governed by the convexity of the soil GSD.