Forecasting of Global Solar Irradiance and Optimal Tilt Angle: Machine Learning Approaches

There is a pressing social need for accurate renewable energy forecasts in light of the current environmental movement's emphasis on these sources of power. The study investigates the optimal tilt angle for solar panel installations in Chennai, India, focusing on Global Horizontal Irradiance (GHI) prediction. Using historical weather data and modern predictive tools, the study found that a <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$13^{\circ}$</tex> tilt angle maximizes solar energy capture, outperforming other angles <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$(10^{\circ},20^{\circ})$</tex> during crucial daylight hours. The optimal tilt enhances energy generation efficiency, increasing the return on investment in solar technology under Chennai's unique climatic conditions. Three machine learning models (ML) such as SVM (support vector machine), Neural Network (NN) and Gaussian Process Regression (GPR) are used to predict the performance on the basis of performance metrices (R2, RMSE, MAE). The NN model demonstrated superior performance in power generation predictions, achieving <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{R}^{2}$</tex>. RMSE, and MAE values of 98.3453, 3.4456, and 98.3455, respectively. These findings highlight the importance of tailored solar panel installation strategies for renewable energy companies in the region