Research on an Analysis Method for Electric Vehicle Charging Load Characteristics Based on Adaptive Density Clustering

Read full text Summarise with VIRAT
Authors: Xiaofeng Da, Zheng Xiao, Leixiong Wang, Hang Li, Huiyao Mi, Hanjun Zhai, Ziyan Zhou
Cited by: 0
References: 10

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Abstract To enhance the accuracy of electric vehicle charging behavior recognition, this paper proposes a load characteristic analysis method based on the comprehensive Load Morphology Index (LMI) and adaptive density clustering. First, an LMI indicator system is constructed from three dimensions: power fluctuation, peak occurrence time, and initial growth trend. Data cleaning, interpolation, and smoothing steps are then applied to improve the continuity and stability of charging curves. Second, to address the difficulty of traditional DBSCAN in adapting to multi-density structures, the study introduces an adaptive neighborhood radius based on local k-nearest neighbors. Combined with density consistency constraints, this approach forms the Adaptive-DBSCAN clustering model, enabling precise classification of user charging behaviors. Finally, case studies using real-world data from DC fast-charging stations in a specific region validate the proposed method. Results demonstrate its ability to effectively distinguish three typical user behavior categories, with superior clustering stability and noise robustness compared to traditional approaches. This research provides technical support for charging station planning and load control strategies.

Fields of Study & Subjects

Electric Vehicles and Infrastructure Smart Grid Energy Management Advanced Clustering Algorithms Research

Publication details

Journal / source
Journal of Physics Conference Series
Volume
3198
Issue
1
Pages
012008–012008
Publication date
1 March 2026
Language
EN