研究基于小波分析的神经网络故障诊断方法根据齿轮驱动装置在工作时通常可能出现的故障类型，人为地在该装置上制造了类似的故障供诊断．并对实验结果进行了分析和讨论．实验表明，利用小波进行诊断信号的预处理，并将其形成向量供给神经网络进行故障诊断具有较好的效果． Research on Neural Network Fault Diagnosis Based on Wavelet Analysis  According to the types of faults that may occur during operation of gear drives, similar faults are artificially created on the device for diagnosis. The experimental results are analyzed and discussed. Experiments show that the use of wavelet for diagnosis of signal preprocessing, and its vector formation to the neural network for fault diagnosis has a good effect.