The Role of Campus Buildings in Energy Conservation and Carbon Reduction of China: A Case Study Based on Hybrid Data-Driven Modeling

Campus buildings, particularly in China, contribute significantly to overall energy consumption and carbon emissions, with educational institutions often facing high energy demands. This study aimed to evaluate and optimize campus building, focus on the accurate prediction of energy consumption and the balance between energy consumption levels and indoor comfort. Physical simulation and data-driven machine learning modeling were integrated to predict energy consumption more accurately. It was proved effective during high-energy consumption months like winter, reducing prediction errors by up to 30%. Additionally, multi-objective optimization (MOO) using the SPEA2 algorithm was applied to improve energy efficiency, reducing energy use intensity (EUI) by 39%, from 78.43 kWh/m² to 47.71 kWh/m², while enhancing percentage of thermal comfortable hours (PTC) by 39%. This demonstrated the ability of MOO to balance energy savings with occupant comfort. A comparative analysis with similar campus in other countries was conducted, and revealed the studied campus in Beijing achieved competitive energy consumption levels but still had potentials for improvement. The hybrid model and optimization of this study provided a more reliable and accurate methods for campus building energy forecasting and application.