A speed and current cascade Continuous Control Set Model Predictive Control architecture for synchronous motor drives

Model predictive control represents a quite mature technology for current control of electric drives. Continuous control set MPC has been widely investigated for the current loop of permanent magnet synchronous motor. Standard PI controllers are often preferred for the speed loop. An innovative speed and current cascade model predictive control architecture for synchronous motor drives is proposed in this paper, to comply with the different requirements of the two loops. Parameter uncertainties represent a critical issue in the implementation of predictive controllers. Considering continuous control set MPC, these mismatches induce bias errors in the desired reference tracking. On one hand, a MPC with intrinsic integral action is adopted in the current loop. The integral action avoids the error induced by electric parameters uncertainties. On the other hand, a separate continuous control set MPC is adopted for the speed loop. A moving horizon estimator supports the latter MPC algorithm, compensating mismatches in the mechanical model of the system. This estimator is used, for instance, to compensate external loading torques. The proposed architecture is validated through experiments on an interior permanent magnet motor. Performances of the proposed solution are compared to the ones obtained using a cascade of PI controllers. Robustness of the architecture against mechanical and electric model uncertainties is investigated too.