Silo pressure predictions using discrete-element and finite-element analyses

The storage of granular solids in silos provides many interesting problems concerning pressures and flow. It is difficult to obtain repeatable and tru stworthy results from either experimental studies or theoretical modelling. Comparisons of the best computational models with experiments are, at best , weak, and provide little assurance of the accuracy of any existing predic tive model. The study described here was undertaken to explore the predicti ons of different models on a set of simplified exercise silo problems. For these problems, no experimental results exist, but simpler tests for truth can be used. This paper reports briefly on an international collaborative study into the predictive capacity of current discrete-element and finite-element calcula tions for the behaviour of granular solids in silos. The predictions of one research group, however eminent, are often not regarded as authoritative b y others, so a commonly agreed theoretical solution of simple silo exercise s, using different computational models from research groups around the wor ld, is a valuable goal. Further, by setting the same unbiased exercise for both finite elements and discrete elements, a better understanding was soug ht of the relationships between the two methods and of the strengths of eac h method in practical silo modelling. The key findings are outlined here from three of the challenge problems: fi lling a silo; discharge of granular solid from a flat-bottomed silo; and di scharge from a silo with a tapered hopper. Both computational methods displ ay considerable shortcomings for these difficult exercises. Different resea rch groups make widely different predictions, even when the problem stateme nt is very detailed. There is much scope for further comparative studies to identify the reasons why different models based on comparable assumptions can produce such varied predictions.