Skip to content Skip to navigation

Disaggregation Survey Paper

 (Figure from Hart, 1992)
Disaggregation refers to the extraction of appliance level energy use data from an aggregate, or whole-building, energy signal, using statistical approaches.
(Figure from Hart, 1992)

How much of your electric bill is going for your refrigerator? For your entertainment center or air conditioning? Consumer awareness of electricity consumption for specific appliances and devices in homes and small businesses may be the most effective type of energy feedback. However, traditional approaches for getting appliance-specific information require installing expensive hardware throughout the home or office. Another approach is software that elicits energy consumption information about individual appliances and devices from changes in whole-home electricity use at the meter.

This study, "Is Disaggregation the Holy Grail of Energy Efficiency? The Case of Electricity," explained how appliance-level data can provide numerous benefits, such as automated personalized recommendations, and why using algorithms in conjunction with smart meters is the most cost-effective and scalable solution for getting this data. This survey of published studies, supplemented with expert interviews, explained the research, technological and policy requirements for using disaggregation with smart meters successfully and relatively quickly. The study guided a significant amount of the research of the Stanford Energy Behavior Initiative. In addition, the paper was viewed on the Precourt Energy Efficiency Center's website more than 60,000 times over 18 months beginning with the posting of a final draft. The published paper was been cited by more than a dozen other published studies in one year.

Additional research to develop disaggregation algorithms, this study found, is needed to improve accuracy and increase the number of appliances the algorithms can identify. A common data set that captures variability over appliances as well as operating conditions would hasten development and enable comparison of algorithms. Electric current must be measured more frequently—every second or even more often—for the algorithms to identify all appliances and major devices. Smart meters can do this, though data compression software must be developed to communicate the data. Additional research ought to determine which appliances are most important to target with disaggregation, how often or quickly feedback is required, and whether periodic snapshots of energy use are sufficient in lieu of complete records.

Disaggregation refers to the extraction of appliance level energy use data from an aggregate, or whole-building, energy signal. A set of statistical approaches that extract patterns characteristic of a given appliance are applied to accomplish this. (Figure from Hart, 1992)
Disaggregation refers to the extraction of appliance level energy use
data from an aggregate, or whole-building, energy signal. A set of
statistical approaches that extract patterns characteristic of a given
appliance are applied to accomplish this. (Figure from Hart, 1992)

As for the technology, the firmware on smart meters already installed should be upgraded to measure reactive power in addition to real power, so that algorithms could identify more household devices. The firmware on such meters should also be upgraded to support data compression. Transmitting events and transitions instead of raw load profiles could significantly improve the frequency of data available to networked home devices, as bandwidth is currently the bottleneck. Future smart meters need to be capable of 10-15 kHz frequency. This would raise meter costs by $5 to $10, but would likely enable a jump in accuracy and the number of appliances recognized. The 802.15.4 standard radio should be replaced with 802.11 (WiFi or low power WiFi) so that meters can communicate directly with the broadband routers, rather than require additional hardware.

In terms of policies, this survey recommends that disaggregation developers should contribute use case specifications and requirements to the standards process so that forthcoming communications technologies are better suited for disaggregation. Also, state utility regulators should institute policies to ensure that utilities enable the home area network (HAN) communication interface soon, at a minimum beginning with pilots. Until then, consumers will benefit little from the billions of dollars spent on smart meters. Rule makers should also ensure that utilities share anonymous data collected during HAN pilots with research institutes and companies not large enough to participate in the pilots. This would facilitate algorithm development and other HAN applications. Rebates should make HAN gateways effectively free to consumers, this study recommends, so they can get real-time data from their smart meter. And utilities should select HAN devices for pilots that allow consumers to access or share their data with any third party, not just their local utility. Finally, both federal and state regulators should expediently approve guidelines for addressing privacy issues, because delays prevent individuals from sharing data they own, and limit third parties from helping to realize consumer benefits.

Publications and Presentations

Is Disaggregation the Holy Grail of Energy Efficiency? The Case of Electricity
Energy Policy, 52, 213–234 (2013)
Armel, K. C., Gupta, A., Shrimali, G., & Albert, A.

  • Without journal access, you may download the paper here: Technical Paper (0.9MB PDF)
  • Video Connectivity Week, Santa Clara, CA, May 22-24, 2012
  • Slide Set (2.25MB PDF) This paper describes how disaggregation can address two timely energy problems - achieving significant low-cost energy reductions in the residential and commercial sectors, and achieving the energy saving potential of the smart grid.

Real-Time Feedback and Electricity Consumption: A Field Experiment Assessing the Potential for Savings and Persistence
The Energy Journal, 34, 87-102 (2013)
Houde, S. Sudarshan, A., Todd, A. Armel, C. & Flora, J.A.

Appliance-specific electricity feedback: Implications for energy conservation programs and policies
Behavior Energy & Climate Change conference (2010)
Houde, S., Todd, A., Sudarshan, A., Flora, J., & Armel, C.

A Randomized Controlled Trial to Evaluate the Power of Information
Behavior Energy & Climate Change conference (2010)
Houde, S., Sudarshan, A., Todd, A., Armel, K.C., Flora, J.A.

Energy and the Environment: Conventional and Unconventional Solutions
The 29th USAEE/IAEE North American Conference, Canada (2010)
Sudarshan, A., Houde, S., Todd, A., Flora, J., & Armel, C.