Inspiring energy conservation through open source metering hardware and embedded real-time load disaggregation

Abstract

Utility companies around the world are replacing electro-mechanical power meters with new smart meters. These digital power meters have enhanced communication capabilities, but they are not actually smart. We present the cognitive power meter (c-meter), a meter that is actually smart. By using load disaggregation intelligence, c-meter is the realization of demand response and other smart grid energy conservation initiatives. Our c-meter is made of two key components: a prototype open source ammeter and an optimized embedded load disaggregation algorithm (μDisagg). Additionally, we provide an open source multi-circuit ammeter array that can build probabilistic appliance (or load) consumption models that are used by the c-meter. μDisagg is the first load disaggregation algorithm to be implemented on an inexpensive low-power embedded processor that runs in real-time using a typical/basic smart meter measurement (current, in A). μDisagg can disaggregate loads with complex power states with a high degree of accuracy.