Multi-objective Constrained for Predictive Controller Design of Grid-inverter and Online Algorithm

LIU Bin,XIA Longqing,LI Jun,WU Jiaju,LI Dewei

DOI: https://doi.org/10.13334/j.0258-8013.pcsee.2014.30.003

2014-01-01

Abstract:In order to tackle with the grid-inverter system control and efficiency problems arising from using LCL filter device, based on derivation, analysis and transformation of its mathematical model, this paper designed a discrete current model predictive control (DCMPC) strategy with multi-objective constrained in the stationary coordinate system, which makes the trace ability of current and efficiency optimization can be embodied in the controller. While as for constrained optimization problems, due to the complexity and huge computation of traditional optimization methods, this paper proposed a modified branch and bound method which combined with predictive control model. The method sought the control vector sequence by complying with the switching principles of the two adjacent input vector among the 27 vectors, and considered that the comprehensive performance index value which contains the system trace ability, switching losses and large current conditions for switching losses is extremely small, so optimal control can be achieved. The method can solve the resonance problem of a LCL filter, guarantee the stability of the system and trace ability, reduce the system loss and the switching frequency at a certain extent. In addition, the CPU’s computational burden can be reduced compared with other constrained controls. Simulation and experimental results are proved the correctness and effectiveness of the proposed control strategy.