Predictive Modeling of the Performance of the Hydrocyclone with Different Cone Combination

The gypsum dehydration hydrocyclones, with cone combination of 20-10° (bi-conical), 10° (single-cone) and 8-10 ° (bi-conical), were investigated in this article. Sampled data were obtained from the separation experiments. The MATLAB neural network toolbox was used to build 3-layer BP neural network models. The resultant of separation performance prediction errors show that it is feasible to build a predictive separation model for hydrocyclones with neural network. And the built model can serve as an effective method for the design of cyclones.