基于粒子群算法优化BP神经网络的轴承故障诊断  

Bearing Fault Diagnosis Based on PSO-BP Neural Network

作　　者：樊怀聪 田禾[1,2] 冯明文 曹冉冉 FAN Huaicong;TIAN He;FENG Mingwen;CAO Ranran(National Demonstration Center for Experimental Mechanical and Electrical Engineering Education,Tianjin University of Technology,Tianjin 300384,China;Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control,Tianjin University of Technology,Tianjin 300384,China)

机构地区：[1]天津理工大学机电工程国家级实验教学示范中心,天津300384 [2]天津理工大学天津市先进机电系统设计与智能控制重点实验室,天津300384

出　　处：《机械制造与自动化》2024年第3期45-49,共5页Machine Building & Automation

基　　金：国网天津市电力公司科技项目(KJ21-1-21)。

摘　　要：通过PSO优化BP神经网络的权值和阈值,采用此算法对滚动轴承进行故障诊断,以驱动端加速度数据和风扇端加速度数据作为输入,通过训练网络输出轴承3种不同状态,实现对轴承的故障诊断。仿真结果表明:此网络模型能够准确识别出轴承运行状态和故障类型,正常样本测试准确率达到98%,并且相对于BP神经网络来说测试精度和准确性都有较大提升,泛化能力更强,可行性高。PSO algorithm is applied to optimize the weight and threshold of BP neural network and conduct the fault diagnosis of rolling.The acceleration data of driving end and the acceleration data of fan end are taken as input to ouput three different states of bearing by training network,so as to realize the fault diagnosis of bearing.The simulation results show that the network model can accurately identify the running state and fault type of bearings,and the test accuracy of normal samples reaches 98%.Compared with BP neural network,the test accuracy is greatly improve with stronger generalization ability and higher feasibility.

关 键 词：轴承 故障诊断 BP神经网络 粒子群算法 

分 类 号：TH133.3[机械工程—机械制造及自动化] TP206.3[自动化与计算机技术—检测技术与自动化装置]