Forecasting household electric appliances consumption and peak demand based on hybrid machine learning approach. (December 2020)

• Record Type: Journal Article
• Authors: Haq, Ejaz Ul; Lyu, Xue; Jia, Youwei; Hua, Mengyuan; Ahmad, Fiaz
• Is Part Of: Energy reports. Volume 6(2020)Supplement 9
• Page Start: 1099
• Page End: 1105
• Publication Date: 2020-12
• Subjects: Clustering -- Demand response -- Smart meter -- Machine learning -- Support vector machine -- Forecasting; Power resources -- Periodicals; Energy industries -- Periodicals; Power resources; Periodicals; Electronic journals; 621.04205
• DOI: 10.1016/j.egyr.2020.11.071
• Languages: English
• ISSNs: 2352-4847

Abstract

Abstract: Machine learning approaches have diverse applications in forecasting electrical energy consumption using smart meter data. Various classification techniques and clustering methods analyze smart meter data for accurately forecasting the electrical appliance consumption and peak demand. Electrical appliance forecasting and peak demand forecasting play a vital and key role in planning, maintenance and automation development for electrical power system. However, there is always a variation between electrical appliance consumption and appliance energy demand due to certain parameters including losses in lines and appliance and mismanagement of appliance energy demand. Detail scrutiny of smart meter data is required to identify the decisive attributes and major cause of variation between electrical appliance consumption and customers' peak demand. This paper proposed a hybrid method based on Machine learning for forecasting appliance consumption and peak demand. We have deployed faster k-medoids clustering, support vector machine and artificial neural network for forecasting appliance consumption and customers' peak demand. The proposed algorithm achieves 99.2% accuracy in forecasting electrical appliance consumption which is much better compared to state-of-the-art in same field. Experimental results validate the effectiveness of the proposed method in forecasting the electrical appliance consumption using smart meter data.