基于RMO-BP算法的感应电动机转子断条故障诊断  

Broken Rotor Bar Fault Diagnosis Based on RMO-BP Algorithm in Induction Motors

作　　者：崔京华[1] 苗旭辉 王惠贞[3] CUI Jinghua;MIAO Xuhui;WANG Huizhen(Department of Mechanical and Electrical Engineering,Hebei Chemical&Pharmaceutical College,Shijiazhuang 050000,Hebei,China;Hebei Province Water Conservancy Bureau,Shijiazhuang 050000,Hebei,China;Department of Electrical and Information Engineeing,Hebei Jiaotong Vocational and Technical College,Shijiazhuang 050000,Hebei,China)

机构地区：[1]河北化工医药职业技术学院机电工程系,河北石家庄050000 [2]河北省水利工程局,河北石家庄050000 [3]河北交通职业技术学院电气与信息工程系,河北石家庄050000

出　　处：《电气传动》2019年第10期96-101,共6页Electric Drive

基　　金：河北省科技厅项目(15211039)

摘　　要：为了更加快速准确地识别感应电动机转子断条故障,提出一种径向移动优化算法(RMO)优化BP神经网络的感应电机转子断条故障诊断方法。RMO作为一种新的全局优化算法,与其他优化算法相比,具有搜索速度快、存储空间小、计算精度高等优势。经过RMO优化的BP神经网络具有更好的权值系数和阈值,能够进一步减小预测误差,提高故障诊断正确率。RMO-BP神经网络的训练和测试数据集通过提取定子电流Hilbert模量的小波能量获得。实验验证了基于RMO-BP神经网络的感应电动机转子断条故障诊断方法的有效性。To identify broken rotor bar faults in induction motors accurately and rapidly,a novel method to diagnose broken rotor bar fault on the basis of BP neural network evolved by radial movement optimization(RMO)was proposed.Compared with any other optimization algorithms,RMO,a newly developed global optimization algorithm,has the advantages of faster searching rate,lesser storage space and higher computing precision,and RMO-BP neural network has superior initial weights and threshold,which can further reduce its forecast error and strengthen the classification correctness.The training data and test data of RMO-BP neural network were obtained by extracting the Hilbert modulus of stator current through wavelet energy.The effectiveness and superiority of the proposed method are verified by experiments.

关 键 词：感应电动机 转子断条 径向移动优化算法 BP神经网络 故障诊断 

分 类 号：TM307[电气工程—电机]