基于骨干微粒群算法和支持向量机的电机转子断条故障诊断

Broken Rotor Bar Fault Diagnosis of Induction Motors Based on Bare-Bone Particle Swarm Optimization and Support Vector Machine

为了准确识别感应电机转子断条故障,本文提出一种基于骨干微粒群算法和支持向量机的故障诊断新方法,并给出了可行的诊断步骤和分析。首先根据故障电流信号的特点,提出一种基于骨干微粒群算法的基波滤除方法,用以消除基波分量对故障特征的影响。然后利用小波包频带能量分解技术,将残余电流信号分解到不同频带,形成感应电机运行状态的特征向量,并以此作为支持向量机的输入向量。采用"一对一"向量机进行分类,并利用骨干微粒群算法和交叉检验优化支持向量机模型参数。最后实验结果表明,该方法诊断感应电机转子断条故障能取得良好的效果。

In order to accurately recognize the broken rotor bar fault of induction motors, a novel method for fault diagnosis is proposed based on the bare-bones particle swarm optimization algorithm （BBPSO） and support vector machine（SVM）, and feasible diagnostic steps and analysis are also introduced. Firstly, a fundamental-frequency filtering method based on BBPSO is proposed to eliminate the influence of fundamental wave in fault characteristic components. Then the feature vector of induction motor in different conditions is extracted with wavelet packet, by which the residual current is decomposed to series of frequency bands; and is considered as the input vector of SVM. The ＂one-against-one＂ vector machine is used to solve the multi-class classification problem, and the BBPSO and cross-validation are taken to optimize model parameters. Finally, the experiment shows that the proposed method is effective to diagnose the broken rotor bars fault of induction motors.