A critical review on granular dam filter behaviour - from particle sizes to constriction-based design criteria
Over emphasis on tests results involving base soils with increasing coefficient of uniformity (Cu) make the reliability of empirical particle-based criterion for effective filters questionable. Extra measures such as regrading of base soil are then necessary in order to capture the effectiveness of filter constriction sizes in trapping eroding fines. Recent studies focused on the role of the sizes of filter voids on overall filtration behaviour have put forth constriction size approaches that were shown to be equally acceptable methods for distinguishing between effective and ineffective filters. This paper examines the evolution of this approach starting from past empirical and mathematical investigations to eventual development of geometric-probabilistic methods with special reference to embankment dams. Models based on dominant and controlling constriction sizes are introduced, and subsequently, the constriction-based retention criteria for granular filters are presented. The proposed retention criteria were verified based on extensive experimental data taken from small- and large-scale filtration tests carried out by various authors. The integration of filter compaction, porosity and Cu, together with the incorporation of analytical principles, provides alternative and rigorous design approaches that remove most limitations of the conventional particle-based criteria thus making the models essentially more comprehensive and quantifiable.