A digital twin framework for optimized energy harvesting IN PV charging systems under fluctuating irradiation

Abstract

Conventional solar-powered battery charging systems tend to operate at their rated capacity; however, at low solar insolation conditions, efficiency reduces. To ensure maximum energy extraction, an energy optimization system is implemented in our research, comprising an impedance-matching mechanism, a battery model by means of a digital twin, and an offset transformer. Furthermore, the machine learning-based method is used to precisely predict the battery state of charge (SoC), State of Health (SoH), and internal resistance, thus enhancing the accuracy of the real-time charging control. An experimental validation was provided by developing a prototype at a laboratory scale. However, the findings reveal a remarkable energy extraction with a gain of 16.68% in the conditions of low irradiance and 11.25% in conditions of moderate irradiance. The suggested system is a good option to incorporate renewable energy sources into the infrastructure of the electric vehicle (EV) battery charging process, significantly increase the efficiency of the energy expenditure, and minimize the total time of charge.