Abstract

The model-free predictive control (MF-PC) employs a data-driven predicting system instead of the conventional prediction model. So, MF-PC remarkably has a better performance against uncertainties and disturbance than the traditional model predictive control (MPC). Except for the predicting system, MF-PC and MPC utilize the same predictive algorithm in power electronics applications. This paper introduces an alternative predictive algorithm, which dramatically simplifies the implementation of MF-PC compared to MPC. The proposed predictive algorithm utilizes the ABC coordinates instead of the common stationary or rotary frames. The frame transformation is employed to avoid complicated and coupled equations of the ABC coordinates. However, the model-free control does not face this complexity issue to need a frame transformation. This paper presents a general MF-PC strategy in the ABC frame for multiphase converters and electrical drives. This method performs the predictive algorithm independently for each phase of the converter or electrical motor. Therefore, the switching of the multiphase converters, which was extremely complicated in the standard predictive algorithm, can be easily applied. Consequently, the computational burden is decreased significantly. Moreover, this predictive technique does not require a weighting factor. The proposed ABC frame MF-PC is validated by a five-phase induction motor simulation. The experimental verification is made by a three-phase induction motor.