Use of Granular Flow Modelling to Investigate Possible Bias of Sample Cutters

Three-dimensional granular flow models have been run to simulate three types of sample cutters: a cross-stream cutter, a vezin, and a cross-belt cutter. This type of modelling involves solving the equations of motion for large numbers of individual particles as they travel along a conveyor belt and interact with a sample cutter. It allows flows of bulk materials to be visualised and understood. Here we have used spherical particles. These results illustrate some advantages that modelling has compared to physical bias testing: By taking reference samples which are very close to the actual samples, the estimation of bias can be achieve quite precisely with very few runs. (Many Standards about bias testing recommend a minimum of 20 runs, but we have found that five runs generally provide adequate precision.) Potential bias can be divided into two components: the particles which are missed but should have been sampled and the particles which are extra but which should not have been sampled. This more detailed information can provide useful insight. Exactly the same material can be sampled using different sample cutters, thereby avoiding a source of variation that complicates physical testing. Much more detailed visualisation is possible, allowing better understanding of the bulk material flow in various circumstances.