Akademik Veri Yönetim Sistemi
IEEE 13th International Conference on Smart Grid (icSmartGrid), Glasgow, İngiltere, 27 - 29 Mayıs 2025, ss.641-649, (Tam Metin Bildiri)
Focusing on anomaly detection in industrial control systems and critical infrastructures, this study discusses the potential integration of digital twin (DT) and explainable artificial intelligence (XAI) technologies. While digital twins create virtual copies of physical systems, allowing monitoring and simulating system behaviors; XAI makes the decision processes of anomaly detection systems transparent, making operator interventions more effective. In the literature, it is stated that digital twins are generally used in technically focused applications, but there are deficiencies in cybersecurity and transparency. In addition, hybrid approaches developed by combining supervised and unsupervised learning methods provide high accuracy rates in anomaly detection. This study focuses on how XAI methods are used to increase security, especially in critical infrastructures, and how these technologies can be integrated with cyber resilience in the future. In addition, the diversity of data sets and models used emphasizes the importance of different methodological approaches for each application area. Finally, it is stated that this study aims to provide a resource that will guide new research areas and technological integrations in anomaly detection and security strategies.