Energy Disaggregation with NILM on a Raspberry Pi with Smart-Metering Extension

Our recent work on Energy Disaggregation with Non-Intrusive Load Monitoring (NILM) on a Raspberry Pi with a Smart-Metering Extension was presented at the 2nd International Conference on Power Systems and Electrical Technology (PSET) in Milan, Italy, from August 25th to 27, 2023.

Smart Metering Extension for Raspberry Pi

Non-intrusive load monitoring (NILM) is a promising technology for efficient energy feedback in residential settings, supporting low-cost energy management systems. However, achieving accurate disaggregation necessitates higher sampling frequencies than standard smart meters (15-minute intervals). State-of-the-art methods require a minimum frequency of 1Hz, increasing system costs and privacy concerns. To address this, we propose a cost-effective single-device smart meter utilizing Raspberry Pi and YoMoPie Monitor for efficient and accurate local processing of user data.
Our concept involves a low-cost single-device smart meter that provides direct feedback based on local user data processing. The system’s performance was tested in a laboratory setting under two different scenarios, and promising results were obtained.
Our system demonstrated promising results in disaggregation performance and computational complexity in laboratory tests under two scenarios. This study evaluates implementing NILM on an embedded system with limited resources, achieving satisfactory outcomes for five appliances. The open-source software and hardware enable easy replication and further exploration by the research community and other stakeholders.

Color indicating detected devices by the NILM algorithm

 To learn more, check out the paper

Johannes Winkler, Hafsa Bousbiat, Stefan Jost, and Wilfried Elmenreich. Energy Disaggregation with NILM on a Raspberry Pi with Smart-Metering Extension. In Proc. 2023 2nd International Conference on Power Systems and Electrical Technology (PSET 2023), Milan, Italy, August 25-27, 2023.

or visit our NILM Raspberry Pi project on Github.