摘要
基于定子电流信号的频谱分析方法诊断电机故障,其检测的精度易受到噪声干扰及频率分辨率的限制。为解决这一问题,提出了一种基于线性混合盲分离模型的电机故障诊断方法。该方法采用固定点算法从电机的定子电流信号中分离出故障特征信号,由观测信号估计出混合矩阵,据此计算故障特征信号的幅值,再根据幅值在电机正常和故障状态下的变化实现对电机故障的诊断。以电机转子故障为例进行了实验,结果表明:该方法可实现转子断条故障的可靠诊断,并且在短数据条件下,也能取得较好的诊断效果。
T he detection precision of fault diagnosis based on the frequency spectral analysis of stator current is easily restricted by noise jamming and frequency resolution .To solve this problem ,this pa-per proposes a fault diagnosis method of induction motor based on the linear mixing model .The method uses the Fast-ICA algorithm to separate the fault characteristic signal from the motor stator current . The amplitude of the signal is calculated with the mixing matrix estimated by the observed signal .The fault diagnosis can be made according to the amplitude varying in a normal state and a fault state of the motor .With the diagnosis of faults in the broken rotor bars as an example ,the experiment result shows that the algorithm can well diagnose the broken-rotor-bar fault ,and also achieve a good result in the condition of short data block .
出处
《海军工程大学学报》
CAS
北大核心
2014年第4期37-41,共5页
Journal of Naval University of Engineering
基金
国家自然科学基金资助项目(50677069)
关键词
电机
盲分离
定子电流
转子断条
故障诊断
motor
blind separation
stator current
broken rotor bar
fault diagnosis
