Design and simulation of model based controllers for quasi resonant converters using neural networks

In this paper, the feasibility of neural modeling and model-based controllers for a non-linear and time varying power converter are investigated. The neural models are developed using dynamic back propagation algorithm. This algorithm is applied to reproduce the dynamic behavior of multi-output flyback ZVS quasi-resonant converter. Based on the developed neural models, inverse control and internal model control (IMC) are developed and their performances are compared through simulation studies. The result reveals that the IMC produces better performance.