采煤机是煤矿开采的关键设备之一，其运行环境复杂，一旦检修不及时就会出现故障，将会造成巨大的经济损失。开发具有机载微机控制功能的多窗口人机交互式，集数字和文字、图形、曲线、图片显示于一体，具备在线监测功能的采煤机是今后采煤机设计和制造的发展趋势。在采煤机的故障诊断方面，采用小波分析理论与人工神经网络相结合的智能故障诊断方法将是以后应用的热点。 本文在深入分析采煤机结构、工况及运行故障的基础上, 针对采煤机关键部件轴承和齿轮的故障机理，设计了采煤机故障诊断系统的总体方案，确定了故障监测参数；在采煤机故障诊断的信号处理方法方面，通过深入研究和比较经典信号处理与小波分析方法的特点，确定采用小波分析方法进行采煤机故障诊断的故障特征的提取，并对故障特征提取的具体方法进行了深入研究；本文还将人工神经网络技术应用于采煤机的故障诊断特征信号的模式识别中，重点研究了PSO-BP粒子群优化算法进行特征模式识别的具体方法，编制了相应的算法程序；在上述工作的基础上，开发了基于虚拟仪器技术的采煤机运行状态监测与故障诊断系统的软件，该系统能实时监测采煤机工作状态中的关键参数，实现采煤机数据采集、数据分类、设备管理、故障报警、结果输出、打印报表等功能。

Coal-mining machine is one of the key equipment which is used for coal mining. Running environment of coal-mining machine is complex, once the examination and repairment is not in time it will appear fault and cause huge economic loss. In the future, the developing trend of designing and manufacturing is developing a kind of coal–mining machine which can display digital, character, graph, bight and picture, and the mode of the coal–mining machine is mult-windows with human and machine interactive, meanwhile it has the skyborne microcomputer control function. In fault diagnosis of coal–mining machine, the brainpower fault diagnosis means which combines wavelet analyses theory and artificial neural network will become the applied hotspot in future. On the base of further analyzing the structure, behaviour and operation disturbance of coal-mining machine, aiming at the failure mechanism of bearing and gear which are the critical components of coal-mining machine, we designe a general planning of fault diagnosis system and confirmes the detection parameters. In signal processing technique about fault diagnosis of coal-mining machine, through futher researching and comparing the classical signal processing technology with wavelet analytical theory, it has confirmed that abstracting a fault signature of failure diagnose by utilizing wavelet analytical theoty and thorough analyzing the specific method on abstraction of fault signature. The text applied artificial neural netrok technology to the pattern recognition on fault diagnosis of coal-mining machine. It primarily studies the concrete means on using PSO-BP algorithm for pattern recognition, and the corresponding program has been developed. Based on the above mentioned work, it has been developed a kind of software which is applied to the fault diagnosis and operation monitoring system, the system is based upon virtual instrument technology, and can realtime monitor the key parameters in working state, as a result realizing these functions about data acquisition, data qualification, facility management, malfunction alarm, outcome output and print accounting and so on.