Skip to main content
EU Science Hub

PINN FLOED

PINN FLOED aims to harness the potential of data-driven machine learning approaches to provide foresight on the effect of building renovation.

Given that the EU building stock is ageing, the Commission’s Renovation Wave initiative promotes increasing building renovation rates to ensure that climate-neutrality targets for 2050 can be achieved.

To inform policy-making processes, modelling approaches that are applicable at large-scale, yet adequately consider building physics parameters, are needed.

Detailed building energy models are valuable, but too computationally intensive to be applied at EU-level. Instead, PINN FLOED explores for the first time a novel physics-informed neural network (PINN) approach for the EU building stock, with the aim of carrying out foresight studies on the renovation of the EU building stock.

This will allow exploring future scenarios, such as the effect of climate change, changes in energy-carrier mix and different renovation rates. This novel bottom-up approach will allow to integrate results with the effect of applying seismic retrofitting to existing vulnerable buildings and hence include insights in terms of reductions in energy consumption, CO2 emissions and potential monetary losses from seismic damage.