Data Driven Urban Building Energy Modeling with Machine Learning in Satom CH

This article delves into the integration of district heating systems into urban planning for sustainable development in regions with moderate to cold climates. The study introduces the Data-driven Urban Energy modeling framework, which aims to bridge the gap between conventional engineering-based energy simulation models and emerging data-driven machine learning (ML) models. By doing so, it provides accurate and comprehensive insights into urban energy demand (ED) patterns. The methodology involves evaluating engineering and ML model's generalization power, revealing its ability to predict energy demand accurately at both building and urban scales. Machine learning algorithms, including LightGBM (LGBM) and Random Forest (RF) regression, are employed to fine-tune the energy-use model for future energy demand predictions. The results demonstrate the model's exceptional accuracy and suitability for diverse urban scenarios. Incorporating a more straightforward approach like Multiple Linear Regression (MLR) into the methodology also highlights its capability to predict energy demand in less complex research scenarios and offer valuable insights for effective urban energy planning. Overall, this article emphasizes the significance of data-driven approaches and machine learning techniques in optimizing energy demand, promoting sustainable urban development, and guiding informed decision-making for energy-efficient cities. The findings have implications for urban planners, policymakers, and energy analysts seeking to enhance energy efficiency and contribute to a greener and more sustainable future for urban communities.