基于神经网络的齿轮故障诊断专家系统  

Expert System of Gear's Fault Diagnosis based on Neural Network

作　　者：李佳[1] 礼宾[1] 王梦卿[1] 

机构地区：[1]东北大学机械工程与自动化学院,辽宁沈阳110004

出　　处：《机械传动》2007年第5期81-83,共3页Journal of Mechanical Transmission

基　　金：国家自然科学基金项目(50275025)

摘　　要：基于齿轮典型故障机理及其信号特征,采用时域、幅值域与频谱分析相结合的诊断方法,建立了齿轮故障诊断神经网络模型,试验验证模型诊断结果具有较高准确性。基于Windows平台和Visual C++语言,开发了基于神经网络的齿轮故障诊断专家系统,将传统的时频信号分析理论与现代小波分析、神经网络和专家系统技术融入齿轮故障诊断之中,形成一个更加完善基于神经网络的齿轮故障诊断专家系统。On the basis of the typical fault mechanisms and vibration characters,a neural network model is set up for fault diagnosis, the test results show that the higher accuracy of the model. An expert system of gear's fault diagnosis based on neural network is also constructed by applying Windows system and Visual C++ program. traditional method of signal analysis, advanced analysis methods are integrated into gear's fault diagnosis, such as wavelet transform, neural network, and expert system technology, and a multifunction expert system. The software of this system has been researched and programmed.

关 键 词：齿轮 故障诊断 神经网络 专家系统 

分 类 号：TH165.3[机械工程—机械制造及自动化]