Auto-tuned Model Parameters in Predictive Control of Power Electronics Converters

Model predictive controlled (MPC) power electronics converters (PECs) features numerous benefits including fast transient response, multi-objective control in single-loop, and inclusion of constraints and nonlinearities in a simple manner. However, the control approach requires the accurate knowledge of the model parameters of the system. Potential model parameter mismatches with the actual values will result in performance deterioration of the MPC. Various agents like aging and harsh environmental conditions would yield changes in circuit impedance. This paper proposes an autonomous MPC for PECs subject to the variations and uncertainties in input filter impedance. The proposed controller adaptively tunes the model parameters used in the predictive equations, alleviating mismatch among the predictive model parameters and the actual system. The proposed auto-tuned modeling scheme requires no additional sensors while maintaining the simplicity of the controller. The proposed method is applicable to all types of PECs and filter topologies. The theoretical analysis and results demonstrate that the proposed approach embedded in the MPC makes the control algorithm robust to variations of model parameters of the system.