期刊
IEEE TRANSACTIONS ON SMART GRID
卷 8, 期 2, 页码 812-821出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TSG.2015.2462719
关键词
Energy management; home automation; pattern recognition; substractive clustering; unsupervised learning
资金
- Laboratoire des technologies de l' energie d'Hydro-Quebec
- Natural Science and Engineering Research Council of Canada
- Foundation of Universite du Quebec a Trois-Rivieres
In this paper, a low-sampling-rate and nonintrusive appliance loads monitoring (NIALM), which required only few tuning parameters and which is little sensitive to the grid power noise, is presented. Using transformed active power transitions as features, the proposed approach is based on the subtractive clustering and the maximum likelihood classifier. In order to validate the NIALM, a 1 Hz sampling rate experimental data from the reference energy disaggregation dataset is selected. The validation results with six commonly found ON/OFF residential appliances indicate that the proposed approach is effective. In addition, the obtained results from a Monte Carlo simulation suggest that this approach is less sensitive to power grid noise than a K-mean-based NIALM method.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据