MONTE CARLO SIMULATION OF THE NO-EROSION FILTER TEST

摘要

Much laboratory testing has been done on the performance of filters for a given base soil, and guidelines exist that result in adequate filter performance. The analysis of filters, in particular, existing filters that do not meet existing filter criteria, lags behind. The authors have developed a simple simulation of filter performance utilizing the entire grain size distribution of the base (core) soil and the filter soil. The simulation was developed to improve understanding of the phenomenon of the base-filter interaction, and also to provide guidance in situations where existing filters do not meet filter criteria. The simulation randomly selects a single particle from the base soil, assumes it gets transported via a crack or slot to the face of the filter soil, and then randomly selects three particles from the filter and calculates the size of the opening formed by the three particles. If the particle passes through the opening, the new position within the filter layer is calculated, and the process is repeated. If the particle becomes lodged in the filter zone, the gradation of the filter is adjusted. If the particle became lodged on the first collision, then it becomes added to a "filter cake" layer forming on the upstream face of the filter. The program presented uses computer graphics to help convey the "lodging" of transported particles from the core into the filter material, and how that affects the development and performance of the filter cake layer, the gradation of the filter and the filter performance. Foster and Fell (2001) have presented an excellent description of how filters that don't meet traditional criteria perform, and this simulation attempts to model the performance described.