Sensorless Speed Control of Synchronous Reluctance Motors Using Model Predictive control associated with Model Reference Adaptive System

摘要

This paper presents a predictive strategy and Model Reference Adaptive System (MRAS) technique for controlling the synchronous reluctance motor (SynRM). A number of sensorless control methods have been developed such as extended Kalman filter (EKF), high-frequency signal injection methods and extended electromotive force (EEMF) models considering magnetic saturation. The aim of our approach is to estimate the rotor speed by the MRAS estimator and then submit this estimate to the Model predictive control. Thus, to obtain a predictive control, the most idea is to extract the variables to be controlled from the model of the machine to calculate their next behavior and to choose the best optimization criterion. In order to feed the predictive control model and reach the optimal speed, it is necessary to subject it to the model reference adaptive system, which employs the state observer model with current error feedback and the rotor current model as two models for current estimation. Thus, to extract the estimated speed, the difference between the outputs of the state observer model and the rotor current model is used to obtain a zero error.