An ANN based switching network for optimally selected photovoltaic array with battery and supercapacitor to mitigate the effect of intermittent solar irradiance

Abstract

ABSTRACT Power fluctuation at the PV panel output is caused by variation of input irradiance. Several optimization techniques have been used for selecting the Global Maximum Power Point (GMPP) from different Local Maximum Power Point (LMPP) that are caused due to variable irradiance. Reconfiguring the PV array by various combination based on the available irradiance is used to overcome the effects of partial shading. In these methods, system complexity increases and efficiency reduces as the size of PV array increases. Hence, in this paper Artificial Neural Network (ANN) based switching network of optimum number of PV panel with battery and supercapacitor is formed. Genetic Algorithm (GA) is used to optimize the number of PV panels, battery and supercapacitor. Battery and supercapacitors are used as energy back up to support the load during low power generation from PV panels. Size of supercapacitors is calculated to support the system during instant power fluctuation caused either due to sudden increase in load or unexpected variation in irradiance. Power enhancement of optimized PV array is performed by ANN based switching and numerically analyzed, and compared with other reconfiguration methods. Optimization is carried out for a system having 245 kW load which led to 1281 number of PV panels, 161 number of batteries and 8 supercapacitors for minimum cost of energy sources. Results of optimization are used for switching of 1281 PV array to form 16 × 81 strings. Out of which 16 × 9 strings are selected for Month Ahead Scheduling (MAS) and string of 16 × 1 PV panels for Real Time Scheduling (RTS). Considering variations at the input irradiance while load is constant. For the selected value of irradiance 16 × 45 strings remain ON which is selected as the base value.