Stator Flux Estimator Using Feed-Forward Neural Network for Evaluating Hysteresis Loss Curve in Three Phase Induction Motor

The operation of induction motors with high performance contributes significantly to the global energy savings but hysteresis loss is one of the factors causing decreased performance. Stator flux density (B) and magnetic field intensity (H) must be plotted to know hysteresis loss quantity. Unfortunately, since the rotor rotates in time series, the stator flux density is unmeasurable quantities, it’s hard to direct sensored this properties because of limited airgap space and costly to install additional instrument. The purpose of this paper is to evaluate the hysteresis loss quantity in induction motor using a novel method of multilayer perceptron feed forward neural network as stator flux estimator and magnetizing current model as magnetic field intensity properties. This method is effective, because it’s non-destructive method, without an additional instrument, low cost, and suitable for real-time motor drive systems. The FFNN estimator response is satisfying because accurately estimate stator flux density for evaluating hysteresis loss quantity including its magnitude and phase angle. By using the proposed model, the stator flux density and magnetizing current can be plotted become hysteresis loss curve. The performance of flux response, speed response, torque response and error deviation of stator flux estimator has been presented, investigated, compared and verified in Simulink Matlab