Low switching frequency model predictive control of PMSM drives for traction applications

In this paper, direct torque control model predictive algorithm is applied to an interior permanent magnet synchronous motor for traction applications. Models for the electric machine and a 2-level voltage source inverter are presented and their relevant equations are expressed. Model predictive control scheme comprises of using a model of the system to predict the future behavior of the plant variables until a predefined horizon in time, and finally selection of the proper control action which minimizes a cost function. This cost function includes several terms, each representing a specific control objective, which in case of controlling a traction permanent magnet motor drive are torque command tracking, maximum torque per ampere, and complying with system ratings limitations. Also the issue of computation delay, an intrinsic characteristic of model predictive control theory when implemented digitally, is studied and a solution is proposed. Finally, an alternative method to meet the rating limitations and a novel switching frequency reduction technique are introduced. The control strategy is evaluated under standard maneuver scenario of motor starting and a speed reference step followed by load torque application and the results are illustrated.