Predicting and Optimizing Industrial Boiler Efficiency via Explainable AI

Optimizing boiler efficiency is crucial for minimizing energy consumption, costs, and emissions in industrial power plants. Traditional models struggle to adapt across fuel types and conditions, necessitating data-driven approaches. This study proposes a machine learning framework using XGBoost and SHAP to predict and optimize boiler efficiency while ensuring interpretability. A high-accuracy predictive model is developed using operational data from a coal-fired power plant, demonstrating applicability across diverse boiler systems. XGBoost provides precise efficiency estimates, while SHAP analysis identifies key influencing factors for informed decision-making. Results highlight energy savings, emission reduction, and the role of explainable AI in industrial optimization. This work advances data-driven, interpretable decision-making in power plant operations, enhancing fuel efficiency and system reliability.