Scholarlink

<p>The increasing demand for EV charging can be met sustainably with Fast EV Charging Stations (FEVCSs) that are powered by Renewable Energy Sources (RESs), such as wind turbines (WTs), photovoltaic (PV) panels, and battery storage systems (BSS). Yet, issues such as large amount of installation costs and dynamic energy demands necessitate intelligent energy management (EM) strategies. This study suggests a novel hybrid technique that combines the global optimization capability of the Parrot Optimizer (PO) with the advanced learning capacity of a Dynamic Weighted Hypergraph Convolutional Network (DWHCN). The key innovation lies in the dynamic hypergraph modeling, which effectively captures complex, high-order spatiotemporal dependencies among RESs, storage units, and EV loads, allowing for adaptive and accurate energy forecasting and allocation. Simulation results demonstrate that the suggested PO-DWHCN method significantly enhances prediction accuracy, reduces operational cost, and increases RES utilization compared to existing intelligent methods. Benchmark comparisons with Fuzzy-Neural Network-Improved Particle Swarm Optimization (FNN-IPSO), Dung Beetle Optimizer-Binarized Spiking Neural Networks (DBO-BiS4NN), Multi-Objective Red Kite Optimization Algorithm (MORKOA), Multi-Agent Deep Neural Network (MADNN), and Modified Snake Optimization (MSO) confirm the superiority of the suggested approach. The PO-DWHCN framework achieves an EM efficiency of 98.7%, a total operational cost of 423.56 yuan, and an average charging time of 35&#xa0;min per session. These results validate its potential as a fast, economical, and high-performance solution for managing FEVCSs integrated with RESs.</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

Hybrid Parrot Optimizer and Dynamic Weighted Hypergraph Convolutional Network for Fast Electric Vehicle Charging Stations with Renewable Energy

  • 期刊论文

  • 发表日期：

    2025年7月15日

查看全文
  • G. Soundradevi,
  • Rahila Jawahar,
  • J. Dani Abraham,
  • A. Radhika,
  • A. Lawrence Sahaya Sundar

摘要

The increasing demand for EV charging can be met sustainably with Fast EV Charging Stations (FEVCSs) that are powered by Renewable Energy Sources (RESs), such as wind turbines (WTs), photovoltaic (PV) panels, and battery storage systems (BSS). Yet, issues such as large amount of installation costs and dynamic energy demands necessitate intelligent energy management (EM) strategies. This study suggests a novel hybrid technique that combines the global optimization capability of the Parrot Optimizer (PO) with the advanced learning capacity of a Dynamic Weighted Hypergraph Convolutional Network (DWHCN). The key innovation lies in the dynamic hypergraph modeling, which effectively captures complex, high-order spatiotemporal dependencies among RESs, storage units, and EV loads, allowing for adaptive and accurate energy forecasting and allocation. Simulation results demonstrate that the suggested PO-DWHCN method significantly enhances prediction accuracy, reduces operational cost, and increases RES utilization compared to existing intelligent methods. Benchmark comparisons with Fuzzy-Neural Network-Improved Particle Swarm Optimization (FNN-IPSO), Dung Beetle Optimizer-Binarized Spiking Neural Networks (DBO-BiS4NN), Multi-Objective Red Kite Optimization Algorithm (MORKOA), Multi-Agent Deep Neural Network (MADNN), and Modified Snake Optimization (MSO) confirm the superiority of the suggested approach. The PO-DWHCN framework achieves an EM efficiency of 98.7%, a total operational cost of 423.56 yuan, and an average charging time of 35 min per session. These results validate its potential as a fast, economical, and high-performance solution for managing FEVCSs integrated with RESs.