Modeling and optimization of pressure drop in cyclone separators using artificial neural networks and genetic algorithm

In order to build complex relationships between cyclone pressure drop coefficient (PDC) and geometrical dimensions, representative artificial neural networks (ANNs), including back propagation neural network (BPNN), radial basic functions neural network (RBFNN) and generalized regression neural network (GRNN), are developed. Multi Step Search Method (MSSM) configures the optimal parameters for designing ANNs. To choose the network with best performance of prediction, mean square error and correlation coefficient of three networks is compared. It is found that, the RBFNN provides higher generalization performance than the BPNN and GRNN. And BPNN has the least precision. Comparing proposed RBFNN simulation results with three semi-empirical models shows the great vantages for RBFNN. In next step, using RBFNN and genetic algorithm, optimized geometrical parameters for minimizing pressure drop of cyclone are found. A comparison between the new design and the standard Stairmand design shows that the new cyclone design results in about 80% of the PDC obtained by the old Stairmand design at almost the same cut-off diameter and the same volume flow rate.