Research project KI-NLT

Artificial Intelligence in Network Control Technology

The increasing use of renewable energy sources is leading to a decentralisation of our energy system on several levels. This not only requires the integration of power generation (e.g., from PV and wind power plants) into the electrical distribution grid, but also the incorporation of new electrical loads, such as heat pumps and electric vehicles. For distribution network operators, this results in increased complexity in network operations at the control rooms. Consequently, local transformer stations are increasingly being equipped with advanced measurement technology, and so-called Smart Meter Gateways are being installed in the grid. This leads to the acquisition of large amounts of data, which can be evaluated using machine learning methods, among other things. However, since the power grid is considered critical infrastructure, the use of these methods is associated with high security requirements. The project Artificial Intelligence in Network Control Technology (KI-NLT), funded by the DLR Transport Programme Directorate, aims to define these requirements with regard to data quality.

Research project KI-NLT

 

Duration

Januar 2025 bis Dezember 2025

Funded by

(internal) DLR Transport Programme Directorate

Project participants

  • Institute of Networked Energy Systems
  • Institute for AI Safety and Security

In the KI-NLT project, the initial focus is on developing a framework that serves as a basic structure for analysing input data for the associated machine learning methods, ensuring high data quality. This framework will be based on specific use cases, for which a risk assessment regarding their deployment in the power grid will be developed, and requirement criteria for the data will be defined.

Within the project, the Institute of Networked Energy Systems focuses on researching and evaluating methods for the application of machine learning in the power grid. Additionally, it aims to provide real or synthetic data for various use cases. For the example application of a network-friendly, intelligent home energy storage management system, researchers utilise the solar power forecast from the Eye2Sky cloud camera network. Another use case being explored is the machine learning-based detection of islanding in medium-voltage grids, which is being developed within the project. These island grids occur when, despite the disconnection of the upstream grid, the medium-voltage grid remains active due to a balance between generation and consumption. With the future deployment of grid-forming inverters in medium-voltage grids, which can compensate for imbalances between generation and consumption, islanding formations are likely to become more frequent, making detection increasingly relevant. The necessary data for this purpose will be obtained from a simulated network environment.

Departments, research groups and challenges related to the KI-NLT project:

Department

Energy Systems Technology

The Energy Systems Technology department focuses on the interaction between system-relevant technologies within decentralised networked structures, particularly at the low-voltage and medium-voltage level.

Research Group

Power Grid Technologies

The Power Grid Technologies research group maps the complex interaction between controlling, controllable and non-controllable grid participants in our laboratories in order to understand and optimise its functioning at software and hardware level.

Challenge

Grid management: Electricity, gas, and heating grids

To ensure that energy is not only available on the balance sheet, but also in the right place at the right time, electricity, gas, and heating grids must be analysed in anticipation of future challenges, adapted in their operational management and modernised and expanded in terms of infrastructure.

Contact

Power Grid Technologies

Research Group
Institute of Networked Energy Systems