Event Detection and Disaggregation Algorithms for NIALM System

Despite the profusion of NIALM researches and products using complex algorithms, addressing the market for low cost, compact, real-time and effective NIALM smart meters is still a challenge. This paper talks about the design of a NIALM smart meter for home appliances, with the ability to self-detect and disaggregate most home appliances. In order to satisfy the compact, real-time, low price requirements and to solve the challenge in slow transient and multi-state appliances, two algorithms are used: the CUSUM to improve the event detection and the Genetic Algorithm (GA) for appliance disaggregation. Evaluation of these algorithms has been done according to public NIALM REDD data set [6]. They are now in first stage of architecture design using Labview FPGA methodology. KeywordsNIALM, CUSUM, Genetic Algorithm, K-mean, classification, smart meter, FPGA.