Non-Intrusive Appliance Identification for Energy Disaggregation of Indian Households–An Use Case for Energy Informatics

Abstract

The goal in non-intrusive load monitoring (NILM) is to design a system/method which accepts aggregate data of energy consumption measured using smart meters as its input and provide an appliance level breakdown of aggregated energy consumption as its output. For practical usefulness of NILM solution there is an additional mandatory requirement viz. to map its output to appliance name recognized by humans. In this paper we adapt CNN architecture of deep neural nets for non-intrusive appliance identification which is a sub-problem for practically useful NILM. Further energy informatics, NILM and energy disaggregation terms have been heavily used in literature but to the best of our knowledge no concrete distinction between them has been specified. This is first work to synthesize the link between energy informatics, NILM and energy disaggregation. We have proposed NILM as one subject under the research subfield of energy informatics and energy disaggregation as a method for implementing NILM. Further, we also give a representation of method for Energy Disaggregation, which uniquely defines whole method, and is also coherently able to represent ideas of previous work.