Scholarlink

<p>The rise of electric vehicles (EVs) and their integration into smart grids (SG) has necessitated advanced cybersecurity solutions to safeguard the charging infrastructure. Machine learning (ML)-driven approaches offer promising pathways to enhance security in EV charging systems. However, existing methods often suffer from vulnerabilities related to unauthorized access, data breaches, and potential disruptions in grid stability, making them inadequate for the growing complexity of EV networks. This study proposes Secure EV Charging using Machine Learning (SEVC-ML) framework to address existing issues and enhance security and resilience in EV charging stations integrated within smart grids. SEVC-ML applies ML models to identify and mitigate threats in real-time, leveraging anomaly detection and predictive analytics to anticipate potential cyberattacks. The proposed method is deployed within the smart grid infrastructure to monitor and safeguard communication between EVs, charging stations, and the grid. The system can detect unusual behaviors indicative of potential security breaches by analyzing data traffic patterns and energy consumption. Findings demonstrate that SEVC-ML improves the detection accuracy of cybersecurity threats, reducing false positives and response time. This approach significantly improves the overall security of the charging network, guaranteeing the secure and effective operation of EVs while maintaining grid stability. The successful integration of SEVC-ML into smart grids offers a robust and scalable cybersecurity solution for the future of EV charging infrastructure. The analysis reveals an accuracy rate of 98.51%, with a false positive rate of 25.38%, response time measured at 18.47%, unauthorized access incidence at 93.91%, and security and reliability at 98.44%.</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

Advancing machine learning-driven cybersecurity solutions for secure electric vehicle charging in smart grids

  • 期刊论文

  • 发表日期：

    2025年5月26日

查看全文
  • S. Poorani,
  • S. S. Sivaraju,
  • G. Arthy,
  • M. Manjusha

摘要

The rise of electric vehicles (EVs) and their integration into smart grids (SG) has necessitated advanced cybersecurity solutions to safeguard the charging infrastructure. Machine learning (ML)-driven approaches offer promising pathways to enhance security in EV charging systems. However, existing methods often suffer from vulnerabilities related to unauthorized access, data breaches, and potential disruptions in grid stability, making them inadequate for the growing complexity of EV networks. This study proposes Secure EV Charging using Machine Learning (SEVC-ML) framework to address existing issues and enhance security and resilience in EV charging stations integrated within smart grids. SEVC-ML applies ML models to identify and mitigate threats in real-time, leveraging anomaly detection and predictive analytics to anticipate potential cyberattacks. The proposed method is deployed within the smart grid infrastructure to monitor and safeguard communication between EVs, charging stations, and the grid. The system can detect unusual behaviors indicative of potential security breaches by analyzing data traffic patterns and energy consumption. Findings demonstrate that SEVC-ML improves the detection accuracy of cybersecurity threats, reducing false positives and response time. This approach significantly improves the overall security of the charging network, guaranteeing the secure and effective operation of EVs while maintaining grid stability. The successful integration of SEVC-ML into smart grids offers a robust and scalable cybersecurity solution for the future of EV charging infrastructure. The analysis reveals an accuracy rate of 98.51%, with a false positive rate of 25.38%, response time measured at 18.47%, unauthorized access incidence at 93.91%, and security and reliability at 98.44%.