Report on the 26th Cologne Solar Colloquium
DLR
"Condition Monitoring of Solar Systems". This was the motto of the 26th Cologne Solar Colloquium, which took place again this year as a hybrid event at the DLR site in Cologne-Porz. Over the course of the day, speakers provided a highly topical insight into the state of development of monitoring technologies that can optimise and increase the efficiency of solar power plants and photovoltaic systems. The use of AI for faster and more accurate data processing and condition predictions is also playing an increasingly important role.
As an annual forum, the Solar Colloquium offers the opportunity to present various highlights from the research programme of the DLR Institute for Solar Research and its partners on a specific topic. This year, 95 participants attended on site and 100 via online stream.
Institute Director Robert Pitz-Paal opened the event. In his welcoming remarks, he emphasised the importance of concentrating solar technology for the global transformation of energy systems.
Peter Heller, head of the qualification department at the Institute for Solar Research, gave the first lecture and spoke about the benefits of condition monitoring, components of interest, methodology, CSP success stories and the relevance of the technology with regard to the energy transition.
Patrick Hilger from Synhelion SA noted in his presentation that monitoring is essential for the development of a reliable commercial technology. Synhelion is a pioneering company in the production of solar fuels. With the help of solar plants, it wants to produce large quantities of solar fuels, especially for aviation. Hilger presented a highly parallel calibration and monitoring system for heliostats, which uses drone images and will be integrated into the plant's control system.
Joachim Krüger, Managing Director of Solarlite, highlighted the importance of integrating and monitoring the state of charge of different storage technologies in solar process heating systems. A high degree of automation and remote monitoring and control are a must. The integration of storage media also increases the complexity. Thus, a development towards fully model-based operation and maintenance is being strived for. Solarlite develops and builds solar thermal parabolic trough power plants for the generation of electricity or heat with the option of cogeneration. Its core competencies are solar field design and the manufacture of parabolic trough collectors.
DLR
Christian Raeder from DLR demonstrated that a differentiable ray tracing environment can be the ideal solution for creating digital heliostat twins. In combination with AI, the twins can then perform extremely accurate aimpoint optimisation. The results will soon be verified in real measurements.
Thomas Schmidt from CSP Services reminded the audience that a wide range of tools are already commercially available to monitor solar installations and support optimisation. Hydrogen mitigation systems, for example, continuously remove hydrogen and thus reduce risks; automation of solar field monitoring enables costs to be saved; collector optimisation and nowcasting increase the performance of CSP power plants.
Another exciting possibility was presented by Laurin Oberkirsch from DLR: it is possible to predict the service life consumption of molten salt receivers and integrate the calculation method into the power plant control system. This involves determining the local stress state, estimating the service life consumption and integrating it into the real-time monitoring.
How AI-powered CSP systems can bring us closer to CO2 neutrality and drive sustainable energy solutions was the message of Andrés Carrancá from Industrial Solar. In the solar industry, AI can be used for predictive maintenance, virtual sensor networks, anomaly detection and much more.
Julian Krauth from DLR presented a rapid method for heliostat calibration, which DLR and CSP Services have developed together. With this method, a 50 MW plant can be calibrated within a few weeks. A commercially available drone with a powerful LED is used for this. The procedure has already been tested at the solar tower in Jülich.
DLR
The volume of data and the need for processing methods in solar power plants is growing strongly. This was described by Inga Miadowicz from DLR in her presentation. As a solution, a data platform could be used here as middleware for linking and data management. A first prototype based on the open source solution FIWARE is already being used to manage this data.
Alex Brenner from DLR explained that hydraulic balancing of the mass flow distribution in parallel parabolic troughs is easily possible. The loop (mass)flow is usually not determined and measurements are only possible at certain points in the field. The use of temperature gradients as tracers, on the other hand, makes it possible to determine the mass flow distribution and optimise it.
Finally, Alexander Schnerring from DLR spoke about the generation of synthetic data in a virtual environment. This helps in the development of UAV-based condition monitoring of solar thermal power plants. Improving this condition monitoring requires large amounts of training data or a safe test environment. AI-supported image processing enables the construction of a vital test environment for training the UAV-based systems.
We would like to thank all speakers for their exciting presentations and all participants for the subsequent professional exchange. We look forward to seeing you again at the next Solar Colloquium!