Identification of induction motor parameters with measurement errors

Abstract

This paper proposes an algorithm for estimating parameters of asynchronous traction motors in the presence of measurement errors. The presence of measurement errors of voltage and current leads to biased estimates of motor parameters. Algorithmic approach to eliminating the influence of noise is used most often as it increases the accuracy of estimates without precision current and voltage meters. Known modifications of algorithms based on the method of total least squares suggest filtering the derivatives of noise, which increases the complexity of the algorithm and also changes the statistical properties of noise. The proposed algorithm is a generalization of the total least-squares technique and does not require knowledge of the laws governing the distribution of measurement errors. The results of modeling show that the parameter estimation is highly accurate than the ordinary least square. The proposed algorithm can be used to create precision control systems, as well as to diagnose faults in electric motors by K parameters.