Abstract: To improve the accuracy of induction motor state evaluation，a model based on correlation algorithm ( CF) and self-organizing map minimum quantization error ( SOM-MQE) was proposed to solve the problem of fundamental current signal interference and lack of fault data． Firstly，the autocorrelation algorithm and cross-correlation algorithm theory were briefly introduced． The characteristic harmonic components in the stator current were analyzed and input into the SOM neural network as performance degradation indicators． Based on this，the minimum quantization error ( MQE) value was calculated． MQE was used as a measure of condition monitoring of induction motors． The example shows that the proposed model can accurately estimate the health status of induction motors，so it has strong engineering application value and versatility．

Key words: induction motor, correlation fundamental wave cancellation, minimum quantization error, fault prediction and health management