分析了影响齿轮早期故障诊断准确性的主要因素,针对齿轮振动信号噪声干扰大的特点,采用小波分析、神经网络滤波、相关滤波等方法进行信号预处理,可大大提高信噪比。现场应用情况表明:该方法可在强噪声环境下提取出齿轮振动中的微弱故障信息,及时发现齿轮早期故障隐患。并在此基础上,开发了一套齿轮早期故障诊断软件系统。 The main factors affecting the accuracy of gear fault diagnosis are analyzed. According to the characteristics of large noise of gear vibration signal, the signal preprocessing using wavelet analysis, neural network filtering and correlation filtering can greatly improve the signal to noise ratio. The field application shows that this method can extract the weak fault information in the gear vibration in the strong noise environment and detect the hidden trouble of the gear early failure in time. On this basis, a set of gear fault diagnosis software system was developed.