Algorithmic Detection of Home Appliances from Smart Meter Data

Reducing the overall energy waste is one of the most pressing challenges of mankind. The energy consumption of individuals can be reduced by providing them with information about the consumption of single appliances in their household. The field of Non-intrusive Appliance Load Monitoring or Energy Disaggregation detects single devices from aggregated loads. Smart meters provide an easy solution to extract momentary values of the device-aggregated energy consumption for further processing. This publication summarizes a proof-of-concept implementation from data extraction via standard smart meters to the detection of appliances of interest (AOIs). Data extraction is based on a low cost hardware with an extraction computer script. The developed disaggregation algorithms were trained with device parameters to detect three AOIs: freezer, dishwasher, and dryer. Through the generality of the concept, the algorithms could be trained to detect other appliance models or classes. Leveraging standard interfaces, the implementation could be reproduced in different households with an installed standard smart meter.