Improved efficiency of DC-DC converter through modified switched inductor-switched capacitor configuration using ANN optimization for photovoltaic sources

Abstract

In photovoltaic applications, the DC-DC converters are of utmost importance, allowing for the regulation of output voltages to satisfy system needs. A preliminary analysis of several converter topologies revealed that the switched inductor switched capacitor (SISC) configuration provides better performance with lower ripple, reduced stress, and better voltage boosting. The performance of the proposed topology is compared with different SISC topologies, and it is observed that the chosen configuration is suitable for photovoltaic sources. The efficiency of the suggested topology is further enhanced by using artificial neural networks (ANN) for regulating the switching frequency and duty cycle of the switch. In this work, two ANNs are used to train both switching frequency and duty cycle. For the training process, the Levenberg-Marquardt algorithm is used to achieve fast convergence with precise predictions. A prototype model is constructed and tested to validate the simulation results. The results prove that the projected converter achieves considerable efficiency and is suited for photovoltaic (PV) systems.