Signal Injection Based Sensorless Online Monitoring of Induction Motor Temperature

Abstract

This research work presents an accurate sensorless temperature monitoring for the commercial induction motor (IM) operated by soft starters (SSs). This technique is relying on calculating stator winding temperature (SWT) of the IM. The methodology is to inject or translate a DC signal into the motor at its normal operation without interrupting the motor running. The research work employs the artificial neural network (ANN) techniques in the asynchronous thermal monitoring scheme for supplying efficient and active temperature calculation of the stator winding, also for decreasing the estimation error. The value of Stator Winding Resistance (SWR) and SWT that computed from signal injection is processed by the ANN with other parameters such the current of stator winding to obtain an accurate value of SWT. A comparison between the results of the proposed ANN scheme and other techniques is performed and the obtained results are inventively.