Neural network approaches for generation forecasting and conceptual design in hybrid renewable energy systems

• AI Forecasting: Neural nets predict solar, wind, biomass generation & grid load for Hybrid Renewable Energy Systems (HRES). • CO 2 Reduction: Raising renewable share from 10% to 90% cuts emissions up to 80%; TensorFlow shows fast drop, PyTorch predicts linear decline. • Model Performance: TensorFlow and PyTorch provide stable HRES forecasts, with minor differences in solar energy generation. • Random Forest achieves ∼80% accuracy but struggles with complex nonlinear relationships, making it less suitable for HRES design. • Conceptual HRES design: TensorFlow favors balanced mix, while PyTorch emphasizes renewable share, aligned with local wind and solar potential. This study investigates energy generation, load forecasting, and CO 2 emission reduction in Hybrid Renewable Energy Systems (HRES) using advanced artificial intelligence (AI) models. Deep learning frameworks, TensorFlow and PyTorch, alongside the Random Forest algorithm, were employed to develop neural models for forecasting solar, wind, and biomass energy production and grid load. Historical data were analyzed to train Long Short-Term Memory (LSTM), Multilayer Perceptron (MLP), and ensemble models, with their accuracy rigorously evaluated. TensorFlow demonstrated superior stability and precision in time series forecasting, while PyTorch offered robust generalization. This research advances prior studies by integrating precise HRES forecasting with location-specific conceptual designs, optimizing renewable energy deployment for a specific region (52.2297°N, 21.0122°E). Results indicate that increasing the renewable energy share from 10 to 90% could reduce CO 2 emissions by up to 80%. TensorFlow models showed the highest CO 2 reduction efficiency when scaling renewable energy from 10 to 50%, with diminishing returns thereafter, while PyTorch predicted a linear emission reduction. These AI-driven models enhance HRES management and provide tailored solutions for local energy planning, contributing to sustainable energy transitions.