摘要
针对电动机故障诊断的准确性问题,研究一种基于自适应神经模糊推理系统的电动机故障诊断新技术.利用己知的输入和输出数据,构造一个模糊推理系统,系统内部调节使用的隶属度函数是基于最小二乘法和反向传播法相结合的混合学习算法.测试电动机的四个主要状态(包括正常状态,定子短路故障,转子断裂和转子偏心故障)的实验数据结果显示:所测试的输出与期望输出基本吻合.所提算法用于异步电动机的故障诊断可以克服传统电动机故障诊断方法具有的误差大,时间长等缺点.
The paper researches motor fault diagnosis of new technologies based on an adaptive neuro-fuzzy inference system for better accuracy. By using known input and output data, a fuzzy inference system is constructed. The membership function for internal system adjustment is based on a hybrid learning algorithm using a combination of a least square method and back propagation method. The data from testing the four major states of motor ( including the normal state, the stator short-circuit fault, the rotor rupture and the rotor eccentricity fault ) show that the test output and expected output are basically in agreement. The proposed algorithm for fault diagnosis of asynchronous motor can overcome shortcomings of big error, time – consuming process by using traditional motor fault diagnosis method.
出处
《龙岩学院学报》
2015年第5期57-60,共4页
Journal of Longyan University
