Select and Optimizate Hydrocyclone Based on Artificial Neural Network

In order to design hydrocyclone comprehensively,this paper establishes three-layer BP neural network model,and it can select right hydrocyclone after giving granularity,production capacity and concentration of underflow.After test of 10 samples,result of selection error: underflow diameter is 10.43%,overflow diameter is 7.51%,the insertion depth is 17.86%,feeding pressure is 20.24%,and precision is higher than that of traditional selection method.The network can not only select right hydrocyclone,but also be used to optimize hydrocyclone parameters on site.Select appropriate hydraulic cyclone to prepare magnet powder,and the coarse and fine product can adapt to wet and dry coal preparation,and it is important to development of coal preparation.