Kordowich G, Bouchkati S, Conrad T, Kubis A, Mitrentsis G, Lutat P, Dauer M, Vogt M, Stiasny J, Jäger J, Mirz M, Ulbig A (2026)
Publication Type: Conference contribution, Conference Contribution
Publication year: 2026
Series: Proceedings of the CIGRE Paris Session 2026
The integration of AI-based systems offers significant potential to alleviate the challenges of increasing grid complexity and decreasing operational safety margins. However, in contrast to the rapid progress of AI in research and development, the deployment in critical infrastructure faces substantial barriers. The goal of this paper is to address limitations of the regulatory and technical frameworks and to offer guidance for implementing trustworthy AI. For this purpose, we provide an in-depth analysis of the current technical and regulatory landscape to identify significant barriers. Then, we propose a process driven approach, which consists of three stages: requirement specifications, offline development and supervised online operation. Each of these stages is described in detail and criteria are developed that allow or prohibit transition from one stage to the next. The process driven approach is exemplified using a case study in which an AI-based solution determines optimal protection settings for given power systems. Finally, the decision support tools and techniques in managing the increasing complexity of power grids results of the analysis are discussed, and recommendations are made for the implementation of AI in power systems.
APA:
Kordowich, G., Bouchkati, S., Conrad, T., Kubis, A., Mitrentsis, G., Lutat, P.,... Ulbig, A. (2026). Towards Trustworthy Artificial Intelligence in Grid Control. In Proceedings of the CIGRE 2026 Paris. Paris, FR.
MLA:
Kordowich, Georg, et al. "Towards Trustworthy Artificial Intelligence in Grid Control." Proceedings of the CIGRE 2026 Paris, Paris 2026.
BibTeX: Download