【摘要】 阐述了基于Levenberg-Marquardt算法训练BP网络的原理及其结构,借助于MATLAB6.x中NeuralNetwork人工神经网络工具箱对旋转机械的故障诊断进行了研究,试验表明利用MATLAB将神经网络引入旋转设备故障诊断,诊断快速,结果精确,并且灵活性强,能够满足工程技术人员的需要,并在此基础上开发了旋转机械诊断系统.更多还原

【Abstract】 In this paper,the authors elaborate the principle and structure of training BP network based on Levenberg-Marquardt.They also study the fault diagnosis of rorating machines with the help of artificial neutral network tootbox in MATLAB 6.x.The experiment indicates that the method of introducing neural network into the fault diagnosis of rotating equipments with the help of MATLAB possesses the characteristics of rapid diagnosis,accurate results and strong feasibility,which can meet the needs of engineers and technicians.And a kind of diagnosis system of rotating machines is developed on the above basis.更多还原