Statistical-based method to determine the best hour of the day regarding GHG emissions for a smart home appliance

Smart homes play a critical role in supporting energy management and environmental emissions reduction. Different types of residential loads allow flexible scheduling, thus bringing optimization opportunities regarding pricing and Greenhouse Gases (GHG) emissions reduction. In this work, we present an approach to provide GHG emissions information to the electricity users, based on data available online and using prediction tools. Our aim is to predict which is the best time of the day from the GHG emissions point of view. In the first step, information to calculate emission factors was collected near real time for different Canadian provinces comprising local generation and imports/exports from neighbors. With the data structured, we used LASSO regression analysis and feature selection tools to predict and analyze the data. After determining the time of the day with the smallest emission factor (for a given location, season, if the day is during the week or not, and hour), we applied this information to decide the best time to turn on an appliance, achieving 25% fewer emissions for our case study.