Scenario-Based and Sensitivity-Driven Energy Management for Carbon-Neutral University Campuses in KMUTT

The planning of carbon-neutral university campuses requires precise forecasting tools that provide interpretable results and handle uncertainties to make effective decisions. The research presents an extensive AI-based methodological framework that merges scenario analysis with sensitivity evaluation and uncertainty quantification to boost the technical reliability of long-term energy forecasting. The hybrid deep learning model consisting of Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BiLSTM) and Transformer-based attention mechanisms received training from five years (2020–2024) of detailed energy and meteorological and occupancy data from King Mongkut’s University of Technology Thonburi (KMUTT) in Thailand. The model demonstrated excellent predictive performance (R2 = 0.96) and received validation through Monte Carlo Dropout for generating probabilistic prediction intervals which captured model-based epistemic uncertainty. Four policy-aligned scenarios were simulated: Business-as-Usual (BAU), Energy Efficiency (EE), Renewable Integration (RE), and Behavioral Change. The EE scenario projected a 20% decrease in energy consumption for 2040 while RE and behavioral interventions led to 15–20% and 8–10% reductions respectively. The most influential variables according to SHAP (Shapley Additive Explanations) analysis were historical energy consumption and ambient temperature and occupancy. The Tornado chart-based sensitivity analysis demonstrated that model outputs remain highly sensitive to ±10–30% variations in these inputs which underscores the importance of maintaining high-quality data and robust monitoring systems. The study provides a replicable technical framework for university energy governance by linking high-resolution forecasting with scenario-driven energy planning under uncertainty. The proposed approach enables data-driven sustainable campus operations that scale toward carbon neutrality throughout ASEAN and additional regions.