Forecasting and optimization of a residential off-grid solar photovoltaic-battery energy storage system

Achieving full energy independence in residential settings remains a challenge with conventional grid-connected solar photovoltaic (PV) systems.This study presents a data-driven approach to designing a grid-independent solar PV and battery energy storage system (BESS) using machine learning techniques.Actual residential load profiles and solar generation data, combined with weather and calendar-based inputs, were analyzed using four regression models: Polynomial Regression, Support Vector Regression, Gradient Boosting, and Random Forest.These models were evaluated based on R, mean absolute error (MAE), and root mean square error (RMSE) to determine the most accurate predictor.The Random Forest model delivered the highest accuracy, achieving an R of 0.92 for generation and 0.90 for load forecasting.The optimized forecast supported the expansion of the PV system from 11 kW to 30 kW and the integration of a 120 kWh BESS, ensuring complete residential grid independence.This framework offers a reliable and adaptable methodology for intelligent energy management and autonomous PV-BESS system design.