Niu X., Wang J., Ren Z., 2015, A Fault Diagnosis Based On Combination Model Of VPRS And PSO-BP Neural Network, Chemical Engineering Transactions, 46, 319-324.
Against the disadvantage of slow convergence and falling into local optimal illustrated by the traditional fault diagnosis based on the BP model, the paper proposes a new diagnosis method by combining the variable precision rough set with the PSO-BP neural network. The proposed model covers the new design of heuristic ß simplification algorithm based on the classification quality, the idea of which could dissolve the input redundancy of the network so as to get the simplify the network and decrease the time consumption oftraining; on the other hand, the improved PSO-BP model could better conquer the problem of local optimalfor BP neural network by taking good advantage of the global data search ability of PSO algorithm and local quick search ability of BP algorithm. In the experiment section, the paper conducts the fault diagnosis model on the antifriction bearing to illustrate the efficiency of the proposed model.
