Data-drive decision support system for selecting building retrofit strategies

The building sector in EU countries is primarily comprised of outdated and inefficient structures, which are of high energy consumption and seismic vulnerability. As a result, building retrofit is being stressed as a viable option for addressing existing energy and seismic issues in the construction industry, particularly in residential properties. For this purpose, strategic decisions should be made about the retrofit strategies, which require great time, effort, resources, and expertise. While traditional case-based retrofit scenarios fail to provide rapid and objective solutions, data-driven methods such as Artificial Intelligence (AI) technologies can serve as an effective and efficient decision support system for selecting retrofit strategies. This research offers a clustering of residential properties in the CENED database (Lombardia 2007)(comprising over 1 million energy labels of residential properties), based on the construction year and U-value. These clusters are associated with the type of material and building technique using the National scientific report on the TABULA activities (Corrado, Ballarini, and Corgnati 2012), and the probability distribution of EHP values. Therefore considering a given U-value and an energy class, the most optimum retrofit strategy is suggested to obtain a particular energy label. This research benefits from AI technologies to enhance strategic decision-making for building retrofit by connecting the current dispersed databases. It also helps increase energy-saving on an urban level.