Global optimization approach for the parameter synthesis of a Pole Restraining controller on the example vienna rectifier

Abstract:
The Pole-Restraining approach is a powerful method to control power electronic converters. One of the main challenges is the synthesis of the multivariable controller parameters. The well-known LQ-approach is a practicable way to design multivariable controllers. Nevertheless, the parameters of the weighting matrices have to be chosen. They have a direct impact on the final control matrix and thereby on the dynamics of the closed control loop and the manipulate variable demand. In this paper a simulative experiment is presented to parametrize the weighting matrices with a genetic algorithm. By defining the fitness function of the genetic algorithm multiple control goals can be achieved. Different configurations are investigated for the vienna rectifier. The approach is transferable to other power electronic converters and controller designs.