Title:Event Detection for Non-intrusive Load Monitoring using Tukey s Fences

Abstract:The primary objective of non-intrusive load monitoring (NILM) techniques is to monitor and track power consumption within residential buildings. This is achieved by approximating the consumption of each individual appliance from the aggregate energy measurements. Event-based NILM solutions are generally more accurate than other methods. Our paper introduces a novel event detection algorithm called Tukey's Fences-based event detector (TFED). This algorithm uses the fast Fourier transform in conjunction with the Tukey fences rule to detect variations in the aggregated current signal. The primary benefit of TFED is its superior ability to accurately pinpoint the start times of events, as demonstrated through simulations. Our findings reveal that the proposed algorithm boasts an impressive 99% accuracy rate, surpassing the accuracy of other recent works in the literature such as the Cepstrum and $\chi ^2$ GOF statistic-based analyses, which only achieved 98% accuracy.