CLOSED LOOP CONTROL FOR MULTI LEVEL DC – DC CONVERTER USING NEURAL NETWORKS

Abstract

Multilevel DC - DC converter system is the novel development system which may be used as a DC link where several levels of controlled voltages are needed with unidirectional current flow and self balancing. The concept Multilevel is able to be implemented for both Buck converter and Boost converter. For multiple outputs, multilevel converter topology can be extended. This proposed paper shows the method of neural network controller implementation for the Multilevel DC - DC converters. The purpose of this is to decrease the output voltage ripple content and to vanish peak overshoots due to transients in order to improve the system performance. And finally the output voltage of more accuracy is achieved in this method. The losses existing in the conventional DC-DC converter can be eliminated by using this proposed converter. In this paper the MATLAB simulation of control of multilevel DC - DC Buck boost converter with the help of neuro controller is obtained using the software MATLAB simulink model and then the final results of this converter for neuro controller is compared.