Adaptive Neuro Fuzzy Inference System Based Sensorless Rotor Position Estimation of Srm

This paper presents sensorless rotor position estimation of Switched Reluctance Motor (SRM) where the position is to be determined by Adaptive Neuro Fuzzy Inference System (ANFIS). The rotor position sensing is very essential for the SRM for its efficient operation. Previously rotor position sensors are used to estimate the position of rotor for SRM. Due to its drawback the sensors have to be replaced by sensorless techniques. So in this paper ANFIS is used to map the nonlinear behavior of the SRM and rotor position is estimated. Mapping is done by the inputs of flux linkage and current to the rotor position as its output. The error between the target and the actual rotor position output is to be calculated. Also the time period of the process, Mean Absolute Error (MAE), Mean Square Error (MSE) and MSEREG are calculated and the comparison is to be made among them. Then comparison of different membership functions, number of epochs and number of membership functions are being carried out for ANFIS. The performance of the ANFIS is analyzed using the error and efficiency. The proposed application will proves the superiority of ANFIS for the rotor position estimation.