Energy Disaggregation via Clustered Regression Models: A Case Study in the Convenience Store

• Main Authors: Hsiao-Hui Chen, Ping-Feng Wang, Ching-Tien Sung, Yi-Ren Yeh, Yuh-Jye Lee
• Format: Conference Proceeding
• Language: English
• Subjects: clustering; Correlation; Electricity; Energy consumption; energy disaggregation; Hidden Markov models; Home appliances; Sensors; support vector regression
• ISSN: 2376-6816; 2376-6824
• DOI: 10.1109/TAAI.2013.21

Global warming and the depletion of natural resources are two of the most difficult problems we have ever faced. To address this problem, people have begun paying more attention to carbon emission reduction and energy saving. For the residential electricity use, many studies have demonstrated that feedbacks, such as energy consumption of each appliance in the home, can help consumers reduce electricity consumption usage. In this article, we propose a novel framework for the disaggregation of energy consumption, which is looking forward to reaching reducing the number of smart meters installed and providing usage statistics as a feedback for consumers to decrease their energy cost. In our proposed framework, we have a chief meter which measures total energy consumption, and install smart meters at few key appliances. Based the energy consumption from these meters, we proposed a clustered regression models for energy disaggregation. More specifically, we first cluster appliances by the correlation between the using behavior of appliances, and select one of them as the key appliance in each cluster. By using the appliance with installed meter, we apply regression model to estimate the energy consumption for other appliances within each cluster. Our experimental results confirmed our proposed framework can achieve high accuracy for energy disaggregation while reducing the number of smart meters.