An improved meta-heuristic algorithm for energy optimization in smart grids considering photovoltaic, storage battery, and demand response

Abstract

Residential load scheduling in smart power grids (SPGs), especially those incorporating renewable energy sources (RESs), storage battery, and demand response (DR) faces significant challenges due to the limitations of traditional optimization algorithms. These challenges include premature convergence, high computational costs, imbalanced exploration and exploitation, lack of adaptability, and sensitivity to parameters. Such issues make it difficult to effectively manage energy consumption, alleviate peak loads, and reduce energy costs while maintaining user comfort. To address these challenges, we propose an improved particle swarm optimization (IPSO) algorithm that enhances exploration and exploitation balance through inertia weight adjustment, velocity damping, and the inclusion of crossover and mutation strategies. These enhancements prevent premature convergence and allow for faster, more accurate convergence to optimal solutions. The proposed IPSO is integrated into a power usage scheduler (PUS) for optimal residential load scheduling under an adaptive pricing scheme considering photovoltaic (PV) and storage battery, focusing on reducing peak energy usage, rebound peaks, energy costs, and user discomfort. The effectiveness of the IPSO-based PUS is demonstrated through a comparison with other optimization algorithms such as genetic optimization algorithm (GOA), particle swarm optimization (PSO), and wind-driven optimization (WDO). Results show that the IPSO algorithm consistently outperforms these alternatives in terms of energy consumption, peak energy alleviation, cost reduction, and grid stability, while also achieving faster execution times and superior convergence rates. This work provides a robust solution for residential load scheduling, offering significant insights and practical benefits for energy optimization in SPGs.