针对传统诊断技术的局限性，研究了基于ＢＰ模型神经网络的故障诊断推理方法，它只需选择足够的具有代表性的故障样本训练神经网络，将代表故障的信息输入给训练好的神经网络，根据神经网络的输出结果，就可以判断发生故障的类型．神经网络一旦训练好，由于其具有容错性，不仅能诊断出已经出现过的故障，还能在一定范围内诊断出从未出现过的故障，使故障诊断智能化和简单化．仿真结果表明，基于神经网络的故障诊断方法是行之有效的 In view of the limitation of traditional diagnosis technology, the fault diagnosis reasoning method based on BP model neural network is studied. It only needs to select enough representative fault samples to train neural network, inputs the information representing fault into the trained neural network, According to the output of the neural network, you can determine the type of failure. Once the neural network is well trained, it can not only diagnose the fault that has occurred but also diagnose the fault that never occurs before to make the fault diagnosis intelligent and simple. Simulation results show that the fault diagnosis method based on neural network is effective