Building archetypes in Urban Energy Models. A comparative case study of deterministic and statistical methods in Andorra

(ENG) In the field of urban building energy models (UBEM), numerous efforts have been made to establish a sound methodology to approach the urban reality of each city. Within this framework, this paper presents a comparative study of two methodologies to determine representative archetypes. We assess these methodologies through a case study in Escaldes-Engordany (the Principality of Andorra), which is representative of medium-sized urban areas in the Pyrenees. We present a workflow for classifying residential buildings using a statistical approach to the available data. We describe the steps followed to construct the archetypes in order to set the bases for a new methodology that can be replicated in other urban contexts. We compare our statistical methodology with a deterministic approach based on local expertise. In a deterministic method, buildings are classified according to existing bibliography and two main variables: building use and year of construction. This first classification is compared with that obtained by defining groups using the metered energy consumption and statistical variables of buildings. The statistical method draws upon the local administration's official databases and is complemented with other information sources such as population and housing censuses, regulatory and technical literature, and the monthly metered electricity consumption of all buildings. Data are processed through statistical methods to group them by similar energy consumption behaviour. Our results show the benefits of the statistical method, as buildings can be characterised not only by their use and constructive technique, but also by their form, dimension location, and occupancy behaviour.