Scholarlink

<p>This research paper presents a comprehensive approach leveraging data-driven techniques and customer electric vehicle charging behavior analysis for user behavior-based forecasting and optimal charging demand scheduling in residential microgrid scenarios. The proposed approach combines the empirical ensemble mode decomposition for data decomposition, convolutional long short-term memory for forecasting, and gray wolf optimizer for optimization and fine-tuning of the model. The popular k-means clustering technique is employed to classify user’s EV charging behavior. The demand forecast is then utilized to formulate an optimization problem to determine the optimal charging schedule for each electric vehicle within the microgrid. The proposed approach was validated using data from 200 consumers, encompassing 348 electric vehicle charging records. Results show precise short-term forecasts, achieving a mean absolute percentage error below 5%, meeting real-world applications. This approach empowers electric vehicle owners and microgrid operators to optimize charging scheduling and achieve substantial energy cost savings, approximately $1550 monthly. This work advances energy management in low-carbon microgrids, promoting efficient electric vehicle integration and sustainable energy practices. </p>

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

Forecasting and Optimal Scheduling of Electric Vehicle Charging Demand: A Cluster and Decomposition-Based Optimized Hybrid Approach

  • 期刊论文

  • 发表日期：

    2025年8月29日

查看全文
  • Firuz Ahamed Nahid,
  • Weerakorn Ongsakul,
  • Jai Govind Singh,
  • Joyashree Roy

摘要

This research paper presents a comprehensive approach leveraging data-driven techniques and customer electric vehicle charging behavior analysis for user behavior-based forecasting and optimal charging demand scheduling in residential microgrid scenarios. The proposed approach combines the empirical ensemble mode decomposition for data decomposition, convolutional long short-term memory for forecasting, and gray wolf optimizer for optimization and fine-tuning of the model. The popular k-means clustering technique is employed to classify user’s EV charging behavior. The demand forecast is then utilized to formulate an optimization problem to determine the optimal charging schedule for each electric vehicle within the microgrid. The proposed approach was validated using data from 200 consumers, encompassing 348 electric vehicle charging records. Results show precise short-term forecasts, achieving a mean absolute percentage error below 5%, meeting real-world applications. This approach empowers electric vehicle owners and microgrid operators to optimize charging scheduling and achieve substantial energy cost savings, approximately $1550 monthly. This work advances energy management in low-carbon microgrids, promoting efficient electric vehicle integration and sustainable energy practices.