Artificial intelligence for the safe and secure control of critical power grids
AI-NLT
Figure 1, symbolic picture, AI-generated
Electrical network system
Credit:
https://www.canva.com
Artificial intelligence for the safe and secure control of critical power grids
The project Artificial Intelligence in Network Control Technology (KI-NLT) investigates the use of machine learning methods to enhance and automate operations in the context of critical energy infrastructures, guaranteeing operational safety and data quality standards. The project will run for one year, from January 2025 to December 2025. Due to the expansion of volatile renewable energies and the associated decentralisation of electricity generation, the stable and secure control of electrical network is becoming increasingly complex. Volatile generation and smart storage technologies with controllable loads require the exchange of large volumes of data between system and grid operators, in order to evaluate and predict the grid status. The increasing demand on decentralised generators, particularly in the areas of grid forming, further complicates the situation, making it necessary to adopt advanced grid control technology with comprehensive networking and grid analysis.
Figure 2
Example of power grid design
Credit:
https://pandapower.readthedocs.io
Contribution of the Institute for AI Safety and Security
Machine learning methods can significantly improve grid control technology, automating processes such as intelligent energy management or preventively detecting system faults. However, the use of advanced methods in critical power infrastructure requires a high degree of attention. For this reason, the quality and reliability of the input and output data must be guaranteed at all times. The Institute for AI Safety and Security plays an active role in the development of data quality standards, enabling the safe and reliable use of AI methods in this context. In this way, the degree of automation in the grid can be increased, regulations for safe and efficient power grid operation can be intelligently adapted, and the status detection and monitoring of systems and grids can be facilitated. Moreover, the manipulation and vulnerability of the AI application must be prevented, avoiding unintended or malicious shutdowns of grid areas or individual generators. Within this framework, the Institute for AI Safety and Security aims to develop approaches to pre-emptively identify system malfunctioning or tampering, to prevent or control grid damage during worst-case scenarios.
