Adaptive Multitimescale Event Detection in Nonintrusive Load Monitoring Based on Minimum Description Length Principle
Bo Liu, Jianfeng Zhang, Wenpeng Luan, Bochao Zhao, Zishuai Liu, Yixin Yu
Abstract
Load event detection is the fundamental step for the event-based nonintrusive load monitoring (NILM). However, existing event detection methods with fixed parameters may fail in coping with the inherent multitimescale characteristics of events and their detection accuracy is easily affected by the load fluctuations. In this regard, this article extends our previous two-stage event detection framework, and proposes a novel multitimescale event detection method based on the minimum description length (MDL) principle. Following the completion of step-like event detection, a long-transient event detection scheme with variable length sliding window is designed, which is intended to observe and characterize the same event at different timescales. In that, the context information in the aggregated load data is mined by motif discovery, then based on the MDL principle, the proper observation scales are selected for different events and the corresponding detection results are determined. For the postprocessing, a load fluctuation segment location method based on voice activity detection (VAD) algorithm is proposed to identify and remove the unreasonable events caused by load fluctuations. Based on new evaluation metrics, the comparison tests on public and private datasets demonstrate that our method achieves higher detection accuracy and integrity for events of various appliances across different scenarios.