Abstract
In response to the need to improve energy-saving processes in older buildings, especially residential ones, this paper describes the potential of a novel method of disaggregating loads in light of the load patterns of household appliances determined in residential buildings. Experiments were designed to be applicable to general residential buildings and four types of commonly used appliances were selected to verify the method. The method assumes that loads are disaggregated and measured by a single primary meter. Following the metering of household appliances and an analysis of the usage patterns of each type, values of electric current were entered into a Hidden Markov Model (HMM) to formulate predictions. Thereafter, the HMM repeatedly performed to output the predicted data close to the measured data, while errors between predicted and the measured data were evaluated to determine whether they met tolerance. When the method was examined for 4 days, matching rates in accordance with the load disaggregation outcomes of the household appliances (i.e., laptop, refrigerator, TV, and microwave) were 0.994, 0.992, 0.982, and 0.988, respectively. The proposed method can provide insights into how and where within such buildings energy is consumed. As a result, effective and systematic energy saving measures can be derived even in buildings in which monitoring sensors and measurement equipment are not installed.
Highlights
MethodsFour types of commonly used major appliances (i.e., laptops, refrigerators, TVs, and microwaves) were selected regarding their use in general residential buildings
Realistic and effective measures for energy savings are needed for 2030’s national business-asusual greenhouse gas (GHG) emissions goals to reduce GHG in South Korea
To verify the identification accuracy of 10 types of household appliances, electrical signals were extracted from the laboratory and tested, and the results showed an identification accuracy of about 90%, which indicates suitability for residential Nonintrusive load monitoring (NILM) systems
Summary
Four types of commonly used major appliances (i.e., laptops, refrigerators, TVs, and microwaves) were selected regarding their use in general residential buildings. The experiment was thereafter divided into three stages: Data were measured when the appliances were in operation. The electric current of the four types of appliances was measured using a device, and their patterns and characteristics were analyzed, classified, and determined for entry into a HMM. The determined pattern data for each appliance, which estimated the initial values of matrices of state transition probabilities ( A) and emission probabilities ( B), were used as HMM input variables. The HMM was completed by being trained with the Baum–Welch algorithm, and the electric current was predicted in the completed HMM for each appliance. All predicted electric current data for each appliance were combined, the sum of which was compared with that of data measured for the appliance used over the course of a day. A matching rate that met the tolerance level of the measured and predicted data was calculated for all times
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
