本文简述了煤矿机械齿轮和轴承故障诊断的目的、意义及其诊断技术的研究现状及发展趋势，分析了齿轮及轴承常见的故障形式以及发生的原因，讨论了齿轮的几种典型故障信号特征以及故障诊断常用的信号处理方法。 针对煤矿机械传动系统低速、重载以及环境恶劣等特点，重点研究了齿轮和滚动轴承的振动机理和故障机理，以低速重载下的齿轮和轴承为研究对象，结合时域振动分析、频谱分析、功率谱分析、小波分析等信号处理方法，将信号解调技术、边频带分析法应用于齿轮和轴承的故障诊断中，提取出了故障特征。 研究了神经网络智能学习算法，讨论了神经网络的基本原理以及神经网络建模方法，利用传统的信号分析方法提取齿轮箱故障的敏感参数，作为网络的输入，经过BP神经网络的学习训练，实现了齿轮箱的多故障分类，取得了良好的分类效果，为齿轮故障智能诊断提供了理论依据。研究表明将该方法应用于矿山机械故障诊断领域具有重要的现实意义。

Aiming at the gear and bearing of coal mime machinery, this paper briefly described the purpose and significance of their fault diagnosis, as well as the research status and development trends of diagnostic technology, analyzed the common fault types and causes. Besides, as to gears, the signal features of typical fault and the common signal processing methods used in fault diagnosis have been discussed in detail. Given the problems such as low speed, overloading and poor environment, which are existed in the transmission system of coal mine machinery, this paper mainly studied the vibration mechanism and failure mechanisms of gear and rolling bearings, regarding the gear and bearing under low speed and overloading environment as research objects. Then the signal processing methods, such as vibration analysis in time domain, spectrum analysis, power spectrum analysis and wavelet analysis and so on, have been combined. In addition, the signal demodulation technology and the method of sideband analysis were applied into fault diagnosis of gear and bearing to extract the fault features. Neural Network, which is used in complex systems, has been studied in this paper, discussing the basic principles and the process of modeling. Additionally, the fault sensitive parameters of gear box, as the input of network, were extracted by using traditional signal processing methods. After learning and training of BP network, the fault classification of gear box was realized, and also the results of classification has been well obtained, which Provided a theoretical basis for the intelligent diagnosis of gear fault. Finally, this study showed that applying the method into fault diagnosis of mining machines will have great values.