Exploratory Analysis of Building Stock: A Case Study for the City of Esch-sur-Alzette (Luxembourg)

Abstract

AbstractOne of the main steps in developing urban building energy models (UBEM) is the classification of the building stock according to building archetypes. Different approaches have been proposed to accomplish this task, some based on the application of clustering techniques, or a combination of expert knowledge, deterministic classification, and data driven approaches. This paper proposes the utilization of a hybrid approach where exploratory data analysis is combined with feature extraction and feature selection to support clustering. The proposed methodology was applied to the building stock of the city of Esch-sur-Alzette (Grand Duchy of Luxembourg). The used data set includes buildings’ geometrical and physical characteristics, preassigned occupancy estimates, and final energy use simulated with a quasi-steady-state model. According to the variables’ combination and deterministic building stock fragmentation schemes used, the number of archetypes identified varied between 12 and 89. The paper shows the potential of clustering techniques for the development of archetypes, even though this must be combined with other (deterministic) fragmentation methods because clustering alone does not allow for the differentiation of building use typologies and construction periods, both of which must be considered to characterize buildings properly.