Model Predictive Direct Power Control of Doubly Fed Induction Generators Under Balanced and Unbalanced Network Conditions

Model predictive direct power control (MPDPC) has been widely studied for the control of doubly fed induction generator (DFIG) systems because of its conceptual simplicity and multivariable control ability. However, conventional MPDPC suffers from the problems of high power ripples and intensive computational effort. Furthermore, this approach presents highly distorted currents under unbalanced networks. To address the problems mentioned above, this article proposes a universal and low-complexity MPDPC, which can work effectively under both balanced and unbalanced networks. On one hand, the proposed method unifies conventional MPDPC and multiple-vector-based MPDPC under a common framework with lower complexity. The optimal vectors and their respective durations in the proposed MPDPC are obtained in a substantially more efficient manner than conventional enumeration-based MPDPC. On the other hand, a flexible power control method with a universal power compensation expression is proposed. By adding the calculated power compensation value to the prior power reference value, the proposed universal MPDPC method can be applied to unbalanced networks. Three control objects under unbalanced network conditions can be realized. Current distortion and power ripple can vary smoothly among the three objects by regulating the coefficient determining the universal power compensation value. The presented experimental results confirm the effectiveness of the proposed method.