Scholarlink

The construction and maintenance of charging stations have increasingly become critical components of the supporting infrastructure for electric vehicles. Despite their importance, there exists a paucity of research concerning the operational characteristics of charging stations, hindered the development and operational efficiency of charging station. Consequently, analyzing and exploring the intrinsic operational features of existing charging stations through daily operating data has emerged as a significant issue, which could prompt the development of electric vehicle industry. This paper aims to analyze charging demand utilizing actual charging data from various charging stations. Besides, peak charging periods and the correlations between charging demand across different intervals are investigated. Based above analysis, we establish a charging demand prediction model for charging stations using Long Short-Term Memory (LSTM) neural network, which average prediction error is less than 0.25.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Data-Driven Based Electric Vehicle Charging Station Demand Prediction Using a Deep Learning Model

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yu Duan,
  • Hanchen Ke,
  • Weihang Bi

摘要

The construction and maintenance of charging stations have increasingly become critical components of the supporting infrastructure for electric vehicles. Despite their importance, there exists a paucity of research concerning the operational characteristics of charging stations, hindered the development and operational efficiency of charging station. Consequently, analyzing and exploring the intrinsic operational features of existing charging stations through daily operating data has emerged as a significant issue, which could prompt the development of electric vehicle industry. This paper aims to analyze charging demand utilizing actual charging data from various charging stations. Besides, peak charging periods and the correlations between charging demand across different intervals are investigated. Based above analysis, we establish a charging demand prediction model for charging stations using Long Short-Term Memory (LSTM) neural network, which average prediction error is less than 0.25.