改进BP算法在旋转机械故障诊断中的应用

针对BP算法在实际诊断中表现出来的“学习收敛速度慢”、“易于陷入局部最小点”等局限性 ,改进BP算法的神经网络方法已经被提出 ,并且应用于对设备的故障诊断和分析研究。用转子试验台模拟了几种常见的故障 ,对其进行诊断研究 ,从中得出结论 :将代表故障的信息输入训练好的神经网络后 ,由输出的结果 ,便可以判断发生故障的类型 ,此外 ,改进后的BP算法大大地提高了网络收敛速度和稳定性 ,更能满足实时在线诊断的要求

Application of the improved BP network to rotating machinery fault diagnosis

Because of weakness of BP algorithm during the actual diagnosis,such as,slow convergence rate and easily to fall into part minimums in network studying of perceptions,the method of improved BP algorithm's neural network has been adopted and been applied when diagnosing and analytic researching the faults of equipments.In this paper,imitate several kinds of typical faults on the test-bed-system on rotor and research them.We can get the conclusions:after input representative the fault's information into the neural network that has been trained,according to the output.We can judge fault's type,moreover,the improved BP algorithm can improve network convergence rate and stability very fast satisfy the request of on-line and real time.