Substation-wise Load Composition Identification Based on Step-wise Regression with Load Templates

Abstract

Component-based load modelling has been widely studied. However, extensive investigations on the consumers, their load devices and their load pattern are needed to determine the substation-wise load composition, which makes it challenging for practical application. In this paper, an integrated data-driven approach is proposed for identification of substation-wise load composition. Firstly, templates of daily load profiles are extracted by semi-supervised clustering. With these load templates, the substation-wise load composition is estimated by using step-wise regression. By fitting the aggregated daily load profile of the substation, the percentage of different kinds of loads, such as industrial load, commercial load and residential load, can be estimated without the need of extensive investigation. Numerical results on practical load data are presented to demonstrate the effectiveness of the proposed approach.