Performance Enhancement of a Dual Active Bridge by a Genetic Algorithm Based Routine for Optimal Parameters of a Cascade Control

An online update routine (OU-Routine) is presented to determine the optimal parameters of a PID-PI cascade control using a genetic algorithm to improve the transient response of an isolated and bidirectional dc-dc converter. The effects of the cascade control parameters on the behavior of the converter are analysed in this study. The parameters are determined using Ziegler-Nichols (ZN), Takahashi (TK) and the genetic algorithm (GA) methods. The performance of the converter with the respective controller is compared on the basis of the time curve of the output voltage. Optimal parameters of PID-PI cascade control for different loads are found using an OU-Routine based on a special cost function. The obtained optimal parameters reduce overvoltage and oscillations while improving the dynamics of the system. The concept is validated by measurements taken from a hardware test setup of an isolated bidirectional dual active bridge dc-dc converter.