Load curve analysis at households scale for energy efficiency
Résumé
All stakeholders in the energy field acknowledge a growing need for higher consumers’ involvement in energy use. This will help them be more aware while bringing them more serenity with respect to their bill, their comfort, and their ecological footprint. The residential sector accounts for 40% of the total energy consumption in the world and over 35% of the electricity consumption. Moreover, distribution network operators, have to anticipate more and more local production and consumption peaks. These issues require improvements in ways to conserve energy and to adapt demand response. Many studies show that the knowledge for each device of real-time consumption would be an essential tool to help households to control their consumption. The deployment of smart meters gives the opportunity to collect real-time power consumption data, the second step is to disaggregate load curve between each use. The startup WattGo, based on consumption data from a panel of hundreds of households, develops algorithms of disaggregation with machine learning methods. They recognize patterns of device use, « lineaments », and can thus detect appliance consumption in real-time. WattGo has also collected socio-economic data about this panel. They can be crossed with energy use data, to understand better determinants of appliance use. From this information, a protocol will be implemented to study the potential of several kinds of incentives to promote energy conservation and load management