Optimizing renewable energy-based grid-connected hybrid microgrid for residential applications in Bangladesh: Predictive modeling for renewable energy, grid stability and demand response analysis
Md. Feroz Ali, Md. Rafiqul Islam Sheikh, Diganto Biswas, Abdullah Al Mamun, Md. Jakir Hossen
Abstract
Bangladesh is facing persistent power shortages and frequent grid outages, highlighting the urgent need to integrate renewable energy into the national power system. This study models and optimizes a grid-connected residential microgrid using HOMER Pro, incorporating solar photovoltaic (PV), vertical axis wind turbines (VAWT), and battery energy storage systems (BESS). Multiple configurations were evaluated, with the optimal design demonstrating the lowest lifecycle cost, high renewable energy contribution, and net surplus electricity export to the grid. The system substantially reduces carbon emissions compared to conventional grid-based supply and shows resilience against grid instability. Sensitivity analysis identified solar irradiation, discount rate, and wind turbine hub height as key factors influencing performance. A machine learning-based predictive model, developed using a Random Forest Regressor, accurately forecasts renewable power output based on meteorological variables. MATLAB simulations confirmed voltage and frequency stability under variable renewable generation, while a demand response strategy improved energy efficiency and reduced dependence on non-renewable sources. The proposed approach offers a technically feasible, economically viable, and environmentally sustainable solution for residential electrification in Bangladesh, with potential for replication in other regions facing similar challenges.