Predictive torque control for AC drives: Improvement of parametric robustness using two-degree-of-freedom control

Predictive torque control (PTC) is amongst the most dynamical control methods for electrical drives. However, the steady-state performance is negatively affected by parametric uncertainties. For instance, magnetic saturation corrupts the inductance values and results in excessive switching and high current ripples. An extension of the PTC controller is proposed to make it more robust while not compromising its performance. By using two-degree-of-freedom (2DoF) control, a part of the controller can be put into a feedforward path. Thereby, a considerable robustness improvement is obtained that does not reduce the strong reference tracking capability of PTC. While the two-degree-of-freedom control structure is well established in continuous control systems, it is a novelty for systems with switched inputs. Experimental results on an industrial servo system with a permanent magnet synchronous machine demonstrate that, under uncertainties, both the excessive switching and higher ripples disappear while the dynamic response remains identical.