基于改进粒子群优化算法的电机故障诊断研究  

Research for Motor Failure Diagnosis Based on Improvement Particle Swarm Optimization Algorithm

作　　者：付光杰[1] 李云鹏[1] 杨秀菊[1] 

机构地区：[1]大庆石油学院电气信息工程学院,大庆163318

出　　处：《科学技术与工程》2010年第4期1001-1004,1009,共5页Science Technology and Engineering

摘　　要：针对电机转子故障,利用神经网络方法进行故障诊断研究。将基本粒子群优化(PSO)算法进行改进,并用其训练反向传播(BP)神经网络,对电机转子进行故障诊断。选用电机转子振动频谱分量作为神经网络的训练样本,将故障信息数据作为输入量代入已训练好的神经网络,通过输出结果即可诊断故障类型。仿真结果表明,基于改进PSO算法的BP神经网络可以有效地识别电机常见故障,具有较快的收敛速度和较高的诊断精度。For the failures of the motor, the method of failures diagnosis is studied by using the neural network. Making the improvement to the basic Particle Swarm Optimization （PSO） algorithm, and then it is used in training on （BP） neural network to carry out failure diagnosis for the motor rotor. By withdrawing motor rotor＇ s vibration frequency spectrum component as the neural network＇ s training sample, inputs the failure information data to the neural network trained well already, then diagnoses the failure type through the output result. The simulation result indicates that, the improved PSO algorithm used in training BP neural network can distinguish the common motor failures effectively with quick convergence rate and high diagnosis orecision.

关 键 词：粒子群优化算法 BP神经网络 异步电机 故障诊断 

分 类 号：TM307.1[电气工程—电机]