Design of a fuzzy PI controller for peak-to-average reduction in output current of LED drivers

This paper presents an optimal gain-scheduling fuzzy PI controller applied to an AC-DC discontinuous conduction mode (DCM) flyback converter to achieve a high performance electrolytic capacitor-less LED driver with high efficiency and high power factor. The proposed system consists of seven PI controllers for seven different load currents which are designed based on Particle Swarm Optimization (PSO) algorithm. By using the obtained data, a fuzzy system is trained to give suitable PI gains for any load condition. Employing the proposed controller significantly improves the peak-to-average ratio of the output current to 1.05 while the LED driver reaches to a high power factor of 90.8% and its input current THD is equal to 11.88% meeting the IEC 61000-3-2 class C standard.