A Factorial Hidden Markov Model for Energy Disaggregation based on Human Behavior Analysis

With the widely installed smart meters and improved data collection technology, nonintrusive load monitoring (NILM) has attracted extensive attention in research. Among different methods, hidden Markov models (HMM) has proved to be one powerful tool in energy disaggregation. A variety of extended HM...

Full description

Saved in:
Bibliographic Details
Published in2018 IEEE Power & Energy Society General Meeting (PESGM) pp. 1 - 5
Main Authors Wang, Xinan, Wang, Jianhui, Shi, Di, Khodayar, Mohammad E.
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.08.2018
Subjects
energy disaggregation
Hidden Markov models
HMMs
Home appliances
human behavior analysis
Mathematical model
NILM
Power demand
Smart meters
supervised learning
Transient analysis
Viterbi algorithm
Online AccessGet full text

Cover

More Information
Summary:With the widely installed smart meters and improved data collection technology, nonintrusive load monitoring (NILM) has attracted extensive attention in research. Among different methods, hidden Markov models (HMM) has proved to be one powerful tool in energy disaggregation. A variety of extended HMM methods has been developed attempting to achieve a better performance. However, HMMs inherently assume a uniform prior distribution over the transition probabilities, which is usually not the case in practice. Also, when similar emission possibilities are shared by different devices, appliances with cyclic power consumption patterns will dominate the non-cyclic appliance due to higher state transition possibilities. In this paper, a novel human behavior analysis based factorial hidden Markov Model (HBA-FHMM) supervised learning algorithm is developed which borrows the connections between human behavior and appliance usage to optimize its performance. Testing result based on the real data demonstrates its effectiveness in addressing the mentioned challenges.
ISSN:1944-9933
DOI:10.1109/PESGM.2018.8586047