Finite set model predictive control with on-line parameter estimation for active frond-end converters

This paper proposes a finite set model predictive control (FS-MPC) system for active front-end (AFE) converters in combination with an extended Kalman filter (EKF) for on-line parameter estimation to overcome the issues of parameter uncertainty and measurement noise. The FS-MPC performance of such converters is largely affected by variations in the model impedance (filter impedance and grid impedance), especially for systems with low short circuit ratios. Therefore, an EKF is used to estimate these model parameters on-line. Moreover, the EKF is used to filter out measurement noise in the feedback variables. For implementation of the FS-MPC, the discrete-time models of the AFE converter and the filter are derived using two discretization methods (forward Euler method and Taylor series expansion). The observability matrix of the linearized model is computed, and the observability is checked for each time instant to verify that the EKF is working properly. Moreover, the delay time due to the digital calculation is compensated. The performance of the proposed method is illustrated via simulation results.