基于随机游走蜂群算法优化的RBF神经网络电动机故障诊断研究

Fault Diagnosis of Motor Based on Random Walk Artificial Bee Colony Algorithm Evolving RBF Neural Network

为解决目前电动机故障诊断方法收敛慢、易陷入局部极小、准确性不足等缺点,提出了一种电动机故障诊断新方法,即将Levy Flight随机游走机制、人工蜂群算法和径向基神经网络结合的方法。首先通过改进的蜂群算法即随机游走蜂群算法(LFABC)对网络参数进行全局寻优,然后对网络进行监督训练,局部细化网络参数,最终得到故障诊断模型。经实例验证,本文所提方法能更快速、准确地实现电动机故障诊断,取得了较好的效果。

In order to solve the problem of slow convergence,local minimum and insufficient accuracy in motor fault diagnosis,a new method is proposed,which combines Levy Flight,artificial bee colony algorithm and radial basis function network.Firstly,the improved bee colony algorithm called random walk bee colony algorithm is used to optimize the network parameters.Then the network is trained to refine the parameters,and the fault diagnosis model is established.The example proves that the proposed method can realize motor fault diagnosis more quickly and accurately,and has achieved a good effect.