Lagrange Multipliers Based Predictive Power Control for Grid-Tied Power Converters

Abstract

Model predictive control has already become an attractive control technique for power electronics and drives. Recently, it has been increasingly widely adopted to the control of renewable generations. To conquer the shortcomings of the large variable-ripple problems, caused by the one-voltage-vector-per-interval principle of the classical predictive methods, non-optimal multi-vector based solutions have been presented and achieved reduced control variable ripples. In this paper, we analyze the problem of the existing multi-vector based predictive control technique, and thereafter propose a new Lagrange Multipliers based predictive power control scheme with globally optimal vector design process. The proposal has been applied to a two-level grid-tied power converter, demonstrating considerably enhanced steady-state control performances in comparison with a recently reported multi-vector based predictive control technique.