VALIDATION OF THE DISCRETE ELEMENT METHOD (DEM) BY COMPARING PREDICTED LOAD BEHAVIOUR OF A GRINDING MILL WITH MEASURED DATA

摘要

The Discrete Element Method (DEM) is a model that is used to simulate the motion of the media in a grinding mill. The sum of the energy that is lost in inter media and media-wall collisions represents the power drawn by the mill. The fact that DEM predicts the power drawn by grinding mills over a large range of diameters has been used to show the validity of DEM. However, the power drawn by a mill is a function only of the sum of forces of the load on the shell. A more accurate validation of DEM would involve an assessment of the ability of DEM to predict these forces. DEM should also be able to predict details of the behaviour of the load in the mill (e.g., the tendency of the load to cascade/cataract/centrifuge, the position of the toe and shoulder of the load, etc). An experimental mill has been developed to reproduce exactly what two-dimensional DEM simulates. The 0.55 m diameter mill has a length of 0.023 m, which is slightly larger than to the diameter of the media used in the mill. One of the lifters in the mill has been instrumented so that the forces on the lifter can be measured as the mill rotates. DEM was used to predict these forces. A high-speed video camera was used to record details of the load behaviour. Reasonable qualitative agreement between predictions and experiment was obtained, but substantial adjustment of the model parameters and perhaps even of DEM algorithm itself (for example, the particle collision model) will be required before adequate quantitative agreement is obtained. This will give experimental evidence of the validity of DEM for rotating grinding mills. This work will also serve as an assessment of the accuracy of the generic DEM algorithm itself, so has implications for other uses of DEM, e.g. in the design of bins, silos, chutes, jigs and other processes.