What they are doing is called “load disaggregation” and it is a very broad and hot research topic in energy policy academia. In nearly all cases, appliances are not monitored directly, and usage is estimated using algorithms. The methods they use for this estimation are non-trivial and there are many competing methods. These look at appliance signatures (unique characteristics of each appliance), the times at which your loads are used, the temperature on that day, etc. Such data is very useful for “demand response” markets and it will likely play a big role in renewable energy adoption.
Latest Answers