Iterative learning control for hybrid active power filter

Conventional PI-type iterative learning control method has some defects such as the convergence of control algorithm heavily depending on initial control input and parameters of the iterative controller being constant value, the paper proposes an improved algorithm of PI-type iterative learning control to improve control performance of system. Base on the proposed iterative algorithm, the paper also proposes a feedback-feedforward control strategy for RIHAPF (Resonant Impedance type Hybrid Active Power Filter). The proposed algorithm of PI-type iterative learning control is applied in feedback controller, and it establishes a fuzzy rule corresponding to RIHAPF system to optimize parameters of the iterative controller to improve control precision of the system. In order to improve dynamic performance of the system, it constructs a feedforward link based on derivative learning law of harmonic current error signal as control input. Simulation and experimental results confirm the value of the proposed iterative algorithm and the control strategy.