Developing a Top-Down Statistical Urban Building Energy Model for Space Heating in Mendoza, Argentina

Energy represents a fundamental requirement for the development process, while buildings are becoming essential consumers in urban landscapes. Given the global need to optimize energy use, urban planners and decision-makers are placing increasing emphasis on energy efficiency in urban planning. The focus of this project is on developing predictive energy consumption models that are tailored to different building characteristics and thus offer differentiated insights into energy dynamics. Using a data-centric approach, the study draws on a comprehensive dataset of building characteristics and weather information, with a focus on residential space heating and domestic hot water consumption. The key to this research is the use of district-level gas consumption data as the dependent variable for modeling purposes. Given the different scales of other datasets, a top-down modeling technique was used. In addition, to strengthen the robustness of the analysis and improve understanding of energy consumption patterns, dependent variables were normalized based on factors that have the greatest influence on gas consumption. These factors include urban altitude, population density, number of families, building area, and volume. Another critical aspect of this study concerns the various independent variables that reflect the quality and characteristics of buildings in Mendoza City. Given the heterogeneous nature of the districts, efforts have been made to delineate homogeneous districts with similar consumption patterns through K-means clustering. Examining the relationship between dependent and independent variables required the use of correlation analysis and then applying multi-linear regression to create the model.