Deadlock avoidance in model predictive direct torque control

For medium voltage AC drives, model predictive direct torque control (MPDTC) shows excellent performance characteristics regarding the switching frequency and the harmonic distortion of the torque and the stator currents, outperforming classic control schemes, such as direct torque control and field oriented control. Besides these advantages, the MPDTC algorithm runs occasionally into so called deadlocks, in which no suitable voltage vector exists. Even though an exit strategy is available to resolve these situations, deadlocks tend to cause spikes in the instantaneous switching frequency and impact the overall performance of MPDTC. This paper focuses on new methods to avoid such deadlocks, using terminal constraints and terminal weights. The proposed methods greatly reduce-and in many cases completely avoid-deadlocks. Moreover, a significant reduction of the switching frequency and the harmonic distortion is observable in the case of a five-level topology.