Only-Current-Signal-Driven Working Condition Identification Method for Induction Motor

Abstract

Condition identification is the basis of induction motor condition monitoring. However, existing non-invasive methods have different degrees of problems in accuracy, robustness and generalization. To solve this challenge, this paper proposes a new method to identify induction motor working conditions, including speed and load torque. This method is a physical-empirical hybrid model method, which combines the advantages of a clear mechanism of the physical model method and easy implementation of the empirical model method. The proposed method introduces the multi-dimensional empirical information contained in the stator current, and the fitting function adopted has a solid physical basis, so the proposed method has the innate advantages of high identification accuracy and strong robustness. The experimental results show that the working conditions identified by the proposed method are in good agreement with the real values. Moreover, by comparing with other existing condition identification methods, the advantages of the proposed method in condition identification accuracy are further verified.