Identification and control of power converters by means of neural networks

Abstract

This paper investigates the use of neural networks for identification and control of power converters. A nonparametric model of a dc-to-dc switching converter implemented by means of a neural network emulator identifies the converter dynamics in cases of uncertainty in the load parameter. A pseudo-linearization control technique resulting in converter regulation and closed-loop linear dynamic behavior is also implemented by means of a neural controller. Simulation results in a PWM boost converter under large-signal operation illustrate both applications This paper investigates the use of neural networks for identification and control of power converters. A nonparametric model of a dc-to-dc switching converter implemented by means of a neural network emulator identifies the converter dynamics in cases of uncertainty in the load parameter. A pseudo-linearization control technique resulting in converter regulation and closed-loop linear dynamic behavior is also implemented by means of a neural controller. Simulation results in a PWM boost converter under large-signal operation illustrate both applications.