BP-ANN model to optimize the structural parameter setting of transfer point

Abstract

In order to optimize the structural parameter setting at the transfer point, the principle of CFD-DEM coupling is used to simulate the production conditions of the material mass flow rate: 150kg/s, height difference: 3m, material density: 1400kg/cm3, and the operation of different structural parameter settings at the transfer point. Based on the pumping volume, equipment wear degree and dust concentration in the exhaust pipe, the weight coefficients of the three indicators given by the comprehensive analytic hierarchy process and the CRITIC method are 0.49, 0.32 and 0.16 respectively, and then the structure parameter settings obtained by the orthogonal test are simulated and optimized by using the BP neural network model. The optimized parameters of orthogonal test are the chute angle of 45 degrees, the number of suction ports of 3, the belt speed of 1.5m/s and bandwidth of 1000mm; the optimized parameters of BP neural network model simulation are chute angle of 45.53 degrees, the number of suction ports of 2.86, the belt speed of 1.48m/s and the bandwidth of 996mm. The results show that the BP-ANN model has the advantages of simplification, high efficiency and time saving, and can provide a new method for enterprises to set the structural parameters at the transfer points reasonably according to their own production conditions.