On Continuous-Set Model Predictive Control of Permanent Magnet Synchronous Machines

This article presents a real-time realization of a continuous-control-set model predictive current controller for the two types of permanent magnet synchronous machines: 1) surface-mounted permanent magnet synchronous machine (SMPMSM) and 2) interior permanent magnet synchronous machine (IPMSM). The constrained optimization problem is solved online using a slack formulation of the primal-dual interior-point method. The proposed controller is tested on a 14.5 kW SMPMSM based on the linear time-invariant (LTI) model of the machine and on a 0.5 kW IPMSM. For the latter, we present in detail how the nonlinear first-principles modeling yields the fastest possible transient as well as an offset-free steady-state performance. The experimental results were obtained at sampling times typically used in the electrical drive applications (125 and 100
$\mu$
s for the two machines, respectively).