This site is dedicated to NILM and disaggregation research done by Stephen Makonin.
Nonintrusive (Appliance) Load Monitoring (NILM or NIALM), sometimes called load disaggregation, is an area in computational sustainability that tries to discern what electrical loads (e.i. appliance) are running within a physical area where power is supplied to from the main power meter. Such areas can include communities, industrial sites, office towers, buildings, homes, and even with in an appliance. By knowing what loads are running, when they are running, and how much power they are consuming we can begin to make informed choices that can lead to a reduction in power consumption.
It is common knowledge that our increase in consumption is neither economically nor environmentally sustainable. Power utilities around the world are or will be bringing in time-of-day usage tariffs These tariffs try to discourage power consumption during peak hours of usage and defer that consumption to off-peak hours (a.k.a. peak shaving). Peak shaving is a response to the rolling blackouts caused by demand exceeding supply. When this happens, the power grid experiences blackouts which is an inconvenience to all. This is only compounded as the demand for power increases. Residential homes consume about 34% of the total power consumption in the USA and are projected to increase to 39% by 2030. Economically this means consumers will be charged higher prices and environmentally this means producing more supply using unclean sources (e.g. coal).
Talks & Presentations
The playlist from NILM 2016 that I organized and hosted In Vancouver, Canada, May 14-15
I gave and inveited talk at the 2nd EU NILM Workshop On July 8, 2015 at Imperial College London.
Open Source Projects
The following are my NILM projects that have been released as open source on GitHub.
A shield for the Arduino Due to mesure current using current transformers (CT) for up to 4 power lines. This ammeter was used to run my PhD thesis disaggregation algorithm called µDisagg. If you use this code please cite my APPEEC 2013 paper.
The following are my NILM datasets that have been released as open to the public.
A dataset that captures smart meter and sub-meter data. Houses are located in and around Vancouver, Canada. The Rainforest Automation Energy (RAE) dataset to help smart grid researchers test their algorithms which make use of smart meter data. RAE contains 72 days of 1Hz data from a residential house's mains and 24 sub-meters resulting in 6.2 million samples for each sub-meter. In addition to power data, environmental and sensor data from the house's thermostat is included. Sub-meter data includes heat pump and rental suite captures which is of interest to power utilities. Read more.
Version 2 of the AMPds dataset has been release to help load disaggregation/NILM and eco-feedback researcher test their algorithms, models, systems, and prototypes. This dataset is intended to be multi-year capture of the consumption of my house. This dataset contains electricity, water, and natural gas measurements at one minute intervals. This dataset contains a total of 1,051,200 readings for 2 years of monitoring (from April/2012 to March/2014) per meter. There are a total of 21 power meters, 2 water meters (with additional appliance usage annotations), and 2 natural gas meters. Weather data from Environment Canada's YVR weather station has also been added. This hourly weather data covers the same period of time as AMPds and includes a summary of climate normals observed from the years between 1981-2010. Billing data from utility companies is also included for cost/benefit analysis.
The AMPds dataset has been released to help load disaggregation/NILM and eco-feedback researcher test their algorithms, models, systems, and prototypes. This is the original version, Release 2013 (R2013). This dataset contains electricity, water, and natural gas measurements at one-minute intervals. A total of 525,600 readings per meter.
This dataset contains power meter, ambient light, and ambient temperature sensors readings; including weather and daylight data. Sensor readings are at 15 minute intervals for about 8 months. This dataset was originally used in the my conference publication and the later expanded journal paper. This dataset monitors the same house that AMPds and should have some limited use for testing very low-frequency load disaggregation (nonintrusive load monitoring or NILM) algorithms and activity inference algorithms.
Published Research Papers
The following are my NILM peer-reviewed journal and conference papers.