利用微粒群算法提取的正负序相量检测感应电机定子故障

Positive and negative sequence phasor extraction by particle swarm optimization algorithm for induction motor stator fault detection

针对感应电机电流中转子断条故障特征分量、谐波和噪声影响定子故障诊断结果的问题,提出一种利用骨干微粒群优化(BBPSO)算法提取基波正、负序相量的故障检测方法。该方法利用BBPSO算法提取出三相定子电流的基波幅值和相位,进而直接计算出总的负序电流。由于在实际电机中,供电电压不平衡、电机先天不平衡和负载的变化等都会影响负序电流的大小,因此通过等效负序阻抗和支持向量机来消除这些非故障因素的影响,从而得到仅与定子故障相对应的残余负序电流,实现感应电机的定子故障诊断。实际电机实验结果表明,采用所提方法提取的残余负序电流能够更加可靠地诊断感应电机定子故障。

As the fault characteristic components of broken rotor bar,harmonics and noise in stator current influence the fault diagnosis of induction motor stator,a fault detection method is proposed,which applies BBPSO（Bare-Bones Particle Swarm Optimization） algorithm to extract the positive and negative sequence phasors.The total negative sequence current is calculated directly according to the fundamental amplitude and phase of three phase stator currents extracted by BBPSO. As the negative sequence currents of an actual motor are normally effected by the unbalanced power supply voltages,inherent asymmetry and load variation,equivalent negative sequence impedance and support vector machine are applied to remove these influencing factors and the obtained residual negative sequence current caused by the stator fault alone is used for fault detection. The results of laboratory experiments demonstrate that,the proposed method applying the extracted residual negative sequence current for stator fault detection is more accurate.