Nonintrusive load monitoring (NILM) using an Artificial Neural Network in embedded system with low sampling rate

Abstract

A Nonintrusive load monitoring (NILM) system is an energy demand monitoring and load identification system that only uses one instrument installed at main power distribution board. In this paper authors have used low sampling rate of monitored data to detect any change of power signal that obtained a 1 Hz sampling rate of active power from energy meter. Using Artificial Neural Network (ANN) for training steady-state real power and reactive power signatures. This paper point to four appliances including air conditioner television refrigerator and rice cooker. The results showed that in simulation test can disaggregation of appliances in correct detection rate 98% and in the installation test can disaggregation of appliances in correct detection rate 95%.