Energy Storage Management and Techno‐Economic Analysis of Standalone Hybrid Renewable Energy Systems Using Deep Reinforcement Learning

Abstract

ABSTRACT The paper presents optimal microgrid (MGs) configurations using artificial intelligence to minimize net present cost (NPC), cost of energy (COE), and optimal utilization of energy storage. It employs methodologies such as deep reinforcement learning (DRL), spoonbill swarm optimisation algorithm (SSOA), genetic algorithm (GA), and artificial neural networks (ANN) for a techno‐economic‐environmental‐storage analysis. The study looks at different parts of MGs, such as photovoltaic (PV) systems, wind turbine generators (WTGs), biomass generators (BMGs), electric vehicles (EVs), diesel generators (DGs), and battery banks (BBs) for storage purposes. The seven different microgrid setups are tested in different weather and load conditions. The best setup, according to techno‐economic analysis, has NPC (₹32 345 782), COE (₹9.23/kWh), and CE (408 348 kg/year). It has PV (210 kW), WTG (91 kW), BMG (25 kW), BB (265 kWh), EVs (22 kW), and DG (28 kW). The simulation results show that this system is the best of all the MGs in every situation. The DRL method is better for the environment and for techno‐economics‐storage than other methods. The DRL approach is the best for future microgrid designs. It gives energy planners and regulators ideas on how to make autonomous microgrid systems more reliable and efficient.