Research on charging patterns of electric taxis based on high-dimensional cluster analysis: a case study of Hangzhou, China

Abstract

The promotion of electric vehicles (EVs) poses challenges to the power grid due to the large-scale and disordered charging behaviors. While previous studies have investigated the charging patterns of EVs, little attention has been paid to electric taxis (ETs). To address this gap, this study proposes a novel combinatorial clustering model to investigate the charging patterns of ETs. This model employs Principle Component Analysis (PCA) for dimensionality reduction, an Canopy + to determine the optimal number of clusters, and concludes with K-means for rapid clustering. It exploits the rich information from the high-dimensional features, such as battery status, time, driving range, and environmental conditions, and enables fast and accurate analysis of large-scale ET charging behavior. The model analyzed a year of charging data from 164 ETs in Hangzhou, identifying six typical patterns. The impact of 20,000 ET charging loads on the power grid was further simulated. The results indicate that increasing the proportion of three types of fast-charging patterns can alleviate the peak and standard deviation of the power load of the grid. This study contributes to a better understanding of the charging behaviors of ETs and provides insights for managing the power demand in the context of urban transportation.