Renovation strategy support for building portfolios from a life cycle perspective based on machine learning

The property sector of Sweden faces many challenges, such as rising energy prices, high greenhouse gas emissions and a slow renovation rate. Building owners usually lack relevant data on existing buildings and struggle to develop renovation strategies under rapidly changing boundary conditions.

This project addresses these problems by leveraging machine learning (ML) and digitalization. A previously developed workflow to generate 3D thermal simulation models is extended by a novel workflow to replace time-consuming physic-based energy simulations with ML. Furthermore, the novel workflow will include embodied carbon to ensure the renovation reduces greenhouse gas emissions from a life cycle perspective. This will allow building owners and facility managers to quickly assess many potential renovation strategies under different scenarios in seconds.

The workflow will be developed in close collaboration with stakeholders and operationalized in a web tool to be used by building owners.

A reference group has been established for the project, consisting of representatives from Aranäs Fastigheter, Uddevallahem, Skövde municipality, Familjebostäder, Lågan, CIT Renergy, and NCC