A Back-EMF Estimation Method for a Switched Reluctance Motor using Model Predictive Control

Switched reluctance machines are simple, robust, fault-tolerant and do not use permanent magnets, which makes them a strong candidate for vehicular propulsion. Despites the advantages they still suffer from high torque pulsation and acoustic noise, which can be reduced by the controller. In this paper the concern is in having an advanced current control, so it is used the model predictive control (MPC). This requires an accurate model to estimate the future behavior of current and the back-electromotive force (emf) signal is essential. As this signal cannot be directly calculated or measured it is proposed a new algorithm to calculate its estimation in real time. The algorithm is easy to implement and the numerical results show the accuracy of the method, which permits a very low current estimation error in the MPC framework.