Deeptrack: Increasing solar efficiency with AI-optimized photovoltaic tracking systems in Germany

Photovoltaic (PV) systems equipped with solar trackers have been shown to increase energy yields by 20-30 percent compared to stationary ground-mounted systems. These systems not only maximize energy production, but also take into account other factors in their design and adjustment, such as the light requirements of specific plant varieties in agrivoltaic (APV) and biodiversity photovoltaic systems, as well as the timing of feeding the grid at different times. Times a day.

In the “DeepTrack” research project, Zimmermann PV-Tracker GmbH, the Zimmermann PV-Steel Group and the Fraunhofer Institute for Solar Energy Systems ISE are collaborating to improve tracking algorithms. They use a deep learning-based digital twin to develop optimized control strategies. This digital twin learns from data collected by its “real” counterpart, a PV tracker built by Zimmermann PV-Tracker, which is located at the Fraunhofer ISE Outdoor Performance Lab in Merdingen near Freiburg, Germany.

According to the International Photovoltaic Technology Roadmap published by the German Federation of Engineers (VDMA), it is expected that 60 percent of all photovoltaic power plants in the world will use tracking systems in the future. In countries with high levels of solar radiation, such as Spain, systems of this type already dominate newly built ground-mounted photovoltaic installations. Germany is also expected to see a significant increase in the number of APV systems with tracking devices following the inclusion of “Photovoltaic Package I” in the German Renewable Energy Sources Act (EEG). Hannes Elsen, product manager at Zimmermann PV, highlights the potential of optimized tracking algorithms in APV systems, taking into account the diversity of crops and systems.

As part of the “DeepTrack” project, Zimmermann PV-Tracker GmbH installed one of its photovoltaic tracking systems on the outdoor Fraunhofer ISE test field to collect real-world data. Using these measurements, the project consortium developed a digital twin that integrates PV monitoring and modeling tools with weather forecasts through deep learning. This allows the calculation of optimal tracking positions for photovoltaic modules in various scenarios.

Dr. Matthew Berwind, team leader at Fraunhofer ISE, explains that the initial control sequences were aimed at maximizing the electricity yield from bifacial solar modules or creating the best conditions for the power plants underneath the APV system. The next challenge is to combine these two approaches to optimize both aspects simultaneously. Achieving this balance is complex but doable with the AI-based approach being developed.

The “DeepTrack” research project is supported by the InvestBW funding program of the Ministry of Economic Affairs, Labor and Tourism of Baden-Württemberg and is scheduled to run until early 2025. Throughout this period, researchers will continue to refine and validate the digital twin model by consistently comparing it with real performance data . This constant comparison ensures the reliability and effectiveness of this innovative technology.

The importance of such advances cannot be overstated. As the demand for renewable energy sources grows around the world, optimizing photovoltaic systems becomes crucial. The ability to accurately track and adjust solar modules based on real-time data and forecasts significantly increases energy efficiency and performance. This, in turn, supports broader goals of sustainability and energy independence.

Moreover, the integration of these advanced tracking systems in APV settings creates a unique opportunity to harmoniously combine agriculture and energy production. By providing crops with adequate sunlight while generating solar energy, these systems promote more sustainable and efficient land use. This double benefit is particularly important in regions where land resources are limited and the need for food and energy security is high.

The DeepTrack project represents a significant step forward in the development of smarter and more efficient photovoltaic systems. Using the power of deep learning and digital twins, researchers and industry partners are paving the way to a future that maximizes solar energy production and minimizes environmental impact. As the project progresses, it promises to provide valuable insights and technologies that can be deployed on a broader scale, contributing to the global transition to clean and renewable energy sources.

Related