MCPSO与RBFNN在滚动轴承故障诊断中的应用

将多群体协同粒子群优化算法应用于RBF神经网络优化设计,不仅拓宽了算法本身的应用范围,而且在一定程度上提高了神经网络的泛化能力,为进一步利用神经网络解决实际工程问题提供了便利。利用优化后的RBF神经网络建立从滚动轴承故障特征向量到故障模式之间的映射,达到了滚动轴承故障模式识别的目的,具有重要的理论和实际意义。

Application of MCPSO and RBFNN in Rolling Bearings Fault Diagnosis

Applying multi-group cooperative particle swarm optimization algorithm(MCPSO) to the optimization design of RBF neural network(RBFNN) not only widened the application scope of the algorithm itself,in a certain extent,and improved the neural network generalization ability for the utilization of neural network to solve practical engineering problems.Using the optimized RBFNN to establish the mapping from fault feature vector to failure pattern achieved the purpose of fault pattern recognition of rolling bearing and had important theoretical and practical significance.