Non-intrusive load identification for smart outlets

摘要: An increasing interest in energy-efficiency combined with the decreasing cost of embedded networked sensors is lowering outlet-level metering. If these trends continue, new buildings near future will be able to install “smart” outlets, which monitor and transmit an outlets power usage real time, for nearly same as conventional outlets. One problem pervasive deployment smart that users must currently identify specific device plugged into each meter, then manually update meta-data software whenever a outlet. Correct important both interpreting historical outlet energy data using building management. To address this problem, we propose Non-Intrusive Load Identification (NILI), automatically identifies attached without any human intervention. In particular, our approach NILI, intuitive simple-to-compute set features from time-series employ well-known classifiers. Our results achieve accuracy over 90% across 15 types on traces collected multiple homes.