粒子群优化融合随机森林的电机故障诊断方法

Motor Fault Diagnosis Method Based on Particle Swarm Optimization and Random Forest

针对三相电机实际识别准确率较低的问题,研究了一种智能的电机故障诊断方法。以三相电机振动数据为研究对象,结合粒子群优化算法和随机森林算法,建立了优化的随机森林算法模型对电机故障状态进行模式识别。提出一种融合K均值聚类算法和随机森林重要性选择方法的敏感特征提取算法,用以对故障敏感特征进行提取。对电机的八种运行状态进行实验验证,实验结果显示该方法能准确和高效地识别出电机故障状态。

Aiming at the problem of low recognition accuracy of three-phase motor,an intelligent fault diagnosis method was studied.Taking the three-phase motor vibration data as the research object,combining particle swarm optimization algorithm with random forest algorithm,an optimized random forest algorithm model was established to identify the motor fault state.A fusion K-means clustering algorithm and random forest algorithm were combined,a sensitive feature extraction algorithm was proposed to extract the fault-sensitive features.The eight operating states of the motor are experimentally verified.The experimental results show that the method can identify the motor fault state accurately and efficiently.