An ANN operated P-I controller for asymmetrical AC-DC converterbased battery charger to mitigate power quality issues

Abstract This article discusses an asymmetrical lithium cobalt oxide (LiCoO 2 ) battery charger that employs an artificial neural network (ANN) controlled adaptive proportional-integral (P-I) controller to mitigate power quality issues. The converter omits the use of rectifier by using two different converters in bridgeless network. Two different converters used are buckboost and Canonical cell (CC) converter for negative 1/2 cycle and positive 1/2 cycle respectively. The usage of these two converters provides a possibility of utilizing a single inductor with center tapping in place of two separate inductors and thus need for one inductor is omitted. These converters are designed to convert AC-DC in null current switching mode (NCSM) which eliminates the necessity of sensing grid voltage and current and in this manner it omits the need for two extra sensors as compared to regular current switching mode. NCSM also require the energy storage element of lower value which means lesser cost and volume of EA charger. Thus elimination of certain costly and heavy components such as extra sensors, an extra inductor and two network completing diodes are removed from the circuit, making it lighter, more compact, and less expensive. Additionally, an ANN-operated adaptive P-I (ANN-API) controller is suggested in this work. The ANN-API controller’s adaptive control method enables real-time system performance optimization. In-depth mathematical modeling and a stability analysis of the given AC-DC converter using a bode plot and poles-zeros (P-Z) map are included in the paper. The findings under both dynamic and steady-state conditions are shown and discussed in order to validate the performance of the LiCoO 2 battery charger that is being presented.