Aller au contenu principal

Context

  • Buildings are in charge of more than 40% of power consumption demand and greenhouse gas emissions

  • Need for a better understanding, efficient characterization and prediction of the energy building performance at different scales for informed decision-making to reduce energy costs and consumption

  • Multi-source data (sensors, smart meters, numerical simulations, weather data…)

What's the Objective ?

  • Better understand  and monitor the energy performance of buildings at different scales in order to help stakeholders make better decisions in terms of energy

    • Building sector and real-estate markets, Distribution system operators, Public authorities, Policymakers, …

What's at stake ?

  • Need for a transition towards a low carbon economy to limit the effect of climate change
    • Not only increase the use of renewable energies
    • Also a more efficient use of the energy that we have avoiding waste
  • Better managing the energy of our buildings using AI goes exactly in this direction
  • Building energy modelling & forecasting is a key tool to inform decision-making for energy planning, building design/renovation, establishment of energy performance contract, predictive maintenance, advice to occupants, energy policies definition, …

What are the challenges ?

  • The cost of equipping and monitoring buildings with appropriate sensors and meters is not negligible
  • Historical data are difficult to obtain (monitoring equipment relatively recent)
  • Numerical simulations used to model energy consumption are also costly and are not always sufficiently accurate

     -> Obtaining sufficient and accurate data can be challenging

What are the possible solutions ?

  • Times series forecasting
  • Fault Detection and Diagnosis 
  • Transfer learning is a promising technique to develop accurate and reliable building energy prediction models. 
    • Use the knowledge about energy performance learned from a set of information-rich buildings to buildings with limited data
    • Reduce the computational time and cost, while also improving the accuracy of predictions  
  • Pay attention to the explainability of the provided predictions to make the latter more trusted by end-users

Responsable(s) du grand défi