Finite Control Set Model Predictive Control with Secondary Problem Formulation for Power Loss and Thermal Stress Reductions

Abstract

Thermal stress has been identified as one of the major failure causes in power modules. Generated from the power loss, thermal stress accelerates the degradation of semiconductor devices and downgrades the system reliability. This article presents a finite control set model predictive control (FCS-MPC) oriented to reduce the power loss over the mission profile and relieve the thermal stress in power modules. Conventional control approaches including the switching frequency regulation, the reactive power injection, and the dc-bus voltage adaption show an effective progress. However, the increased control loops and complicated modulation schemes limit the system performance and practical implementation. In the proposed FCS-MPC, a secondary problem formulation is defined to reduce the power loss for the thermal stress reduction in power modules. It is simply integrated with the primary problem formulation in order to achieve the power flow control and power loss reduction simultaneously. An energy-based loss model is proposed for the loss prediction. The impact of the weightings between primary and secondary problem formulations is investigated and a most efficient weighting curve with a weighting-zones strategy is presented to design the proposed FCS-MPC. The proposed FCS-MPC is validated in the simulations and experiments. A 2.5-kW grid-tied inverter prototype is developed for the hardware testing and validation.