Direct model predictive control - a new direct predictive control strategy for electrical drives

Since more than 20 years the so-called "field oriented control" is state-of-the-art for electrical drives. Strategies based on this principle fulfil nearly all demands of drive technology today. Due to its basic ideas, the performance of field oriented control cannot be improved without limitations. Furthermore, linear controllers always produce a continuously valued output. Hence, for controlling an inverter additional components, which discretise the output of the controller, have to be inserted. Predictive controllers, which allow to overcome this drawback, are known in drive control for quite a long time. Nevertheless all predictive control schemes proposed for controlling electrical drives until today only precalculate the system behaviour for one single future sampling cycle. Powerful control strategies like model predictive control, well-known in chemical engineering, allow higher prediction horizons, but these strategies have not yet been used for drive control. This paper proposes a new predictive controller for electrical drives based on model predictive control, which allows optimal control of the inverter directly, i.e. without using any kind of modulator. In contrast to all predictive strategies published so far for drive control, the new controller optimises the control action not only for the next sampling period, but for a longer prediction horizon