A secure edge monitoring approach to unsupervised energy disaggregation using mean shift algorithm in residential buildings

Liu, Qi; Nakoty, Francis Mawuli; Wu, Xueyan; Anaadumba, Raphael; Liu, Xiaodong; Zhang, Yonghong; Qi, Lianyong

Abstract

Compared to Intrusive Load Monitoring which uses smart power meters at each level to be monitored, Non-Intrusive Load Monitoring (NILM) is an ingenious way that relies on signal readings at a single point to deduce the share of the devices that have contributed to the overall load. This reliable technique that guarantees the safety and privacy of individual users has recently become an increasingly popular topic, as it turns out to be a major solution to assist household users in the process of obtaining details of their electricity consumption. The detailed consumption promotes better management of the electrical power on the consumer side by helping to eliminate any waste of energy. In this paper, an edge gateway has been implemented to safely monitor the overall load in a smart energy system. A load separation method has been introduced based on events detected on a low-frequency power signal, which allows the consumption profile of On/Off and multi-state devices to be generated without relying on the knowledge of the cardinality of these devices Following the extraction of significant features contained in the aggregate signal, an appliance profile recognition approach is presented based on the non-parametric Mean Shift algorithm. The ability of the proposed method to learn and deduce devices profile is validated using the Reference Energy Disaggregation Dataset (REDD). The experimental results show that the proposed approach is efficient in detecting events of binary state and finite state appliances.

• Journal Article Type: Article
• Acceptance Date: Aug 30, 2020
• Online Publication Date: Sep 2, 2020
• Publication Date: 2020-10
• Deposit Date: Oct 20, 2020
• Publicly Available Date: Sep 3, 2021
• Journal: Computer Communications
• Print ISSN: 0140-3664
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 162
• Pages: 187-195
• DOI: https://doi.org/10.1016/j.comcom.2020.08.024
• Keywords: Home energy management, Non-intrusive load monitoring (NILM), Unsupervised learning, Non-parametric algorithm
• Public URL: http://researchrepository.napier.ac.uk/Output/2694758