Open AccessInstantaneous active and reactive load signature applied in non‐intrusive load monitoring systems
Author(s) -
Brito Ricardo,
Wong ManChung,
Zhang Hong Cai,
Da Costa Junior Miguel Gomes,
Lam ChiSeng,
Wong ChiKong
Publication year - 2021
Publication title -
iet smart grid
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.612
H-Index - 11
ISSN - 2515-2947
DOI - 10.1049/stg2.12008
Subject(s) - signature (topology) , trajectory , computer science , active load , feature (linguistics) , feature extraction , uniqueness , data mining , real time computing , pattern recognition (psychology) , artificial intelligence , engineering , mathematics , voltage , physics , linguistics , philosophy , geometry , transistor , astronomy , electrical engineering , mathematical analysis
The performance of non‐intrusive load monitoring (NILM) systems heavily depends on the uniqueness of the load signature extracted from the electrical appliances. Different load signatures have been proposed. Recently, in particular, v – i trajectory feature extraction is attracting more and more attention due to its unique characteristics. Herein, instantaneous p – q load signature (IpqLS) feature extraction is first proposed and applied in NILM, which shows that conventional methods cannot distinguish load signatures under some situations. Applying IpqLS with several machine learning algorithms is not only extracting unique features compared to the overlapping problems of P – Q and v – i trajectory but also improving load classification accuracy. Simulations and experimental results verified the effectiveness of the proposed method.