User electricity load classification portrait based on multidimensional feature analysis

Abstract

Against the backdrop of the widespread application of advanced metering systems and power Internet of Things technology, the rapid growth of electricity big data and the integration of diverse heterogeneous characteristics from various sources present unprecedented complexity. A precise understanding of user electricity consumption behavior and load characteristics holds significant importance for achieving energy efficiency improvements and personalized services. This paper aims to construct user electricity load classification profiles through multidimensional feature analysis using a non-intrusive load disaggregation method. It utilizes daily average power consumption (P), daily average operating duration (T), and daily average activation count (0) as key feature dimensions to establish the PTO model. By comprehensive assessment, the PTO model reveals characteristics such as energy consumption, operating duration, and frequency of user electricity loads. Experimental validation is conducted using the publicly available REDD dataset.