A Techno-Economic Feasibility Assessment of Rooftop Solar PV Systems Using Machine Learning-Based Energy Forecasting

The increasing demand for reliable and sustainable energy solutions has accelerated the adoption of rooftop solar photovoltaic (PV) systems, particularly in facilities dependent on diesel generators for backup power. This study presents a comprehensive techno-economic feasibility assessment of a rooftop solar PV system using real operational energy data and machine learning-based forecasting techniques. Historical data comprising solar energy generation, diesel generator energy usage, runtime patterns, and specific yield are analyzed to evaluate system performance and energy contribution. Machine learning models, including linear regression, random forest, and time-series forecasting, are employed to predict future solar energy generation and assess potential reductions in diesel generator dependency. The forecasted energy outputs are integrated with economic analysis to compute key financial indicators such as levelized cost of electricity (LCOE), payback period, return on investment (ROI), and diesel cost savings. The results demonstrate that machine learning-enabled forecasting significantly improves the accuracy of energy prediction and enhances decision-making for system planning and investment