Paper Conference

Proceedings of eSim 2018: 10th Conference of IBPSA-Canada


Generator: a stochastic virtual smart meter data generation model for residential building stock characterization

Adam Neale, Michaël Kummert, Michel Bernier

Abstract: Smart meters are becoming prevalent for the purpose of electricity consumption billing. In the province of Québec alone there are over 3.5 million smart meters collecting data at fifteen-minute intervals. In this paper a series of building energy simulations are used to produce virtual smart meter (VSM) data sets. This represents a first step towards using real smart meter data for building stock characterization, by developing an inverse model and training methodology in a controlled environment. The inputs for the VSM model are based on single family home building stock characteristics found in the province of Québec. Parameters for each simulation are randomly selected based on probability mass functions, which allow for realistic, statistically representative VSM data sets to be produced using building energy simulations. A data set of 100,000 VSM profiles is targeted, which will require big data analysis techniques for visualization and processing.
Keywords: residential, smart meter data, building stock modeling, building energy modeling briefly how current building archetype development
Pages: 65 - 74