This paper focuses on identifying which appliance is currently operating by analyzing electrical load signature for home energy monitoring system. The identification framework is comprised of three steps. Firstly, specific appliance features, or signatures, were chosen, which are DC (Duty Cycle), SO (Slope of On-state), VO (Variance of On-state), and ZC (Zero Crossing) by reviewing observations of appliances from 13 houses for 3 days. Five appliances of electrical rice cooker, kimchi-refrigerator, PC, refrigerator, and TV were chosen for the identification with high penetration rate and total operation-time in Korea. Secondly, K-NN and Naive Bayesian classifiers, which are commonly used in many applications, are employed to estimate from which appliance the signatures are obtained. Lastly, one of candidates is selected as final identification result by majority voting. The proposed identification frame showed identification success rate of 94.23%.
[1]
Jian Liang,et al.
Load Signature Study—Part I: Basic Concept, Structure, and Methodology
,
2010,
IEEE Transactions on Power Delivery.
[2]
Jian Liang,et al.
Load Signature Study—Part II: Disaggregation Framework, Simulation, and Applications
,
2010,
IEEE Transactions on Power Delivery.
[3]
Bo-Hyeun Wang,et al.
A Study on Electric Power Monitoring System per Appliance
,
2010
.
[4]
Bo-Hyeun Wang,et al.
Development of Home Electrical Power Monitoring System and Device Identification Algorithm
,
2011
.
[5]
G. W. Hart,et al.
Nonintrusive appliance load monitoring
,
1992,
Proc. IEEE.