A new DEM calibration method for wet and stick materials based on the BP neural network

The assessment of flow characteristics and the Discrete Element Method (DEM) calibrations for Wet and Sticky Materials (WSMs) are significantly demanded by the mining and transportation industries. However, the angle of repose (AoR), as a common DEM calibration method, has been facing challenges in adopting for WSMs. Due to the relatively high adhesion and cohesion of WSMs, it is difficult to form repeatable material heaps, resulting in inaccurate measurements of the AoRs. Therefore, this study aims to calibrate DEM parameters for WSMs under various conditions, utilizing a modified AoR test method and the Back Propagation-Neural Network (BP-NN). In the modified AoR test method, WSMs are discharged from a funnel using a screw feeder and fall closely against a vertical baffle to form a heap in a constrained direction. The new method combines the features of the traditional AoR test and the slump plane test to accurately measure the AoRs for different WSMs, including wet white sand, bulk coal, and corn powders. Test results are analysed to show the advantage of the modified AoR test. Furthermore, an innovative DEM calibration method based on the BP-NN is developed. By training the BP-NN with the database established according to the relationship between the measured and simulated AoRs, the DEM calibration parameters can be predicted rapidly for WSMs with different particle sizes and moisture contents. Validation is conducted for the wet white sand, from which the deviations and relative mean deviations between the experiment AoRs (Ae) A e ) and the calibration AoRs (Ac) A c ) are less than 2 degrees and degrees and 4 % respectively, indicating the capability of the developed method for WSMs DEM calibrations.