Appliance-Level Energy Awareness Model for Flexible Load Management in Residential Buildings
Open AccessBuildings are responsible for 40% of primary energy consumption, 72% of total electricity usage, and 40% of total carbon emissions in the United States, with residential buildings accounting for more than half. A significant portion of the total energy consumed in residential buildings is attributed to appliances, commonly known as miscellaneous electric loads (MEL). MEL explored in this research include thermostatically controlled appliance types such as refrigerators, non-thermostatic devices such as dishwashers and clothes washers, and battery-derived energy loads like electric vehicles. This research introduces a novel wavelet-based convolutional neural network architecture called wavelet denoising autoencoder (WDAE) for appliance energy disaggregation in residential buildings. The selected appliances consume more than half the overall electricity usage across households and therefore offer energy-saving opportunities. The WDAE method realized an 11.4% lower mean error rate and a 17.6% improvement in F1-score compared to state-of-the art denoising autoencoders (DAEs). This research aims to provide utility providers in competitive electricity markets with accurate information on nonintrusive energy disaggregation of the MEL footprint. Such information is useful in designing and offering economically efficient incentives to reduce usage during peak periods in residential buildings and thus creates a surplus that can be sold to commercial customers at a premium.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.