短时数据下利用双HTLS参数估计的异步电机转子断条故障诊断  

Broken rotor bar diagnosis by double HTLS parameter estimation based on short-time data of induction motor

作　　者：王臻[1,2] 李承[1,2] 陈旭[2] 李惠章[2] 王蕾[2] WANG Zhen LI Cheng CHEN Xu LI Huizhang WANG Lei(State Key Laboratory of Advanced Electromagnetic Engineering and Teehnology,Huazhong University of Science and Teehnology,Wuhan 430074, China School of Electrical and Electronic Engineering, Huazhong University of Science and Teehnology,Wuhan 430074,China)

机构地区：[1]华中科技大学强电磁工程与新技术国家重点实验室,湖北武汉430074 [2]华中科技大学电气与电子工程学院,湖北武汉430074

出　　处：《电力自动化设备》2017年第1期182-186,共5页Electric Power Automation Equipment

摘　　要：传统基于傅里叶分析的算法,其频谱分辨率受限于数据长度,对此,提出一种在短时数据下利用双HTLS算法进行异步电机转子断条故障诊断的方法。利用HTLS算法中的Hankel矩阵左奇异子矩阵的平移不变性质,准确估算出基波幅值、频率、相位参数;然后利用基波参数构造基波表达式,将其从定子电流原始信号中剔除,从而得到残余电流;对残余电流再次进行HTLS参数估计,辨识出故障成分的参数,并有效抑制噪声的影响。对不同数量的转子断条故障进行实验验证,结果表明双HTLS算法具有高鲁棒性的优点,且适用于噪声背景及短时数据下的转子断条故障检测。Since the spectral resolution of traditional Fourier analysis algorithm is limited by the data length, the double HTLS （Hankel Total Least Squares） algorithm is applied to diagnose the broken rotor bar fault based on the short-time data of induction motor is proposed. The shift-invariance of left singular sub-matrix of Hankel matrix in HTLS algorithm is utilized to accurately estimate the fundamental parameters of amplitude,frequency and phase,which are used to construct the fundamental harmonic expression. The fundamental harmonic is removed from the original stator current to obtain the residual current. In order to effectively suppress the noise,HTLS algorithm is utilized again to estimate the residual current parameters for recognizing the fault component parameters. Experimental verifications are carried out for different quantities of broken rotor bar,which indicate the double HTLS algorithm has high accuracy and high robustness, suitable for the broken rotor bar diagnosis based on short-time data with noise background.

关 键 词：异步电机 转子断条 故障诊断 双HTLS 参数估计 总体最小二乘法 短时数据 基波 

分 类 号：TM307[电气工程—电机]