Household behavior by load curve analysis with machine learning techniques
Résumé
With the recent deployment of the smart meters in a French context, energy efficiency improvement in residential sector can be approach by the households load curve analysis completed by a detailed survey of the household’s characteristics and habits.
As the buildings are becoming more and more efficient, the energy consumption tends to be more and more behaviorally influenced. It is important to be able to quantify this part to give the pertinent advices to the users to improve their relationship to energy consumption.
We present the first result of the obtained by machine learning technique application on the coupled information given by their load curve and their status and habits. This exploration gives some patterns that can be coupled with some population categories that may help an energy provider having a new customer similar to a category to advise him in improving its energy consumption performance.