LATODA presents AI solutions at workshop of BAM (a senior scientific and technical Federal institute of Germany) in Berlin

The expansion of wind energy is projected to increase to 30 GW for offshore installations and 115 GW for onshore installations by 2030. At the workshop on technology transfer, “From the Lab to the Wind Farm of Tomorrow,” held at the Federal Institute for Materials Research and Testing (BAM) in Berlin, it became clear that the industry faces enormous challenges. Already, most wind turbines produce less than their optimal capacity. This is primarily due to the manufacturing quality of rotor blades and their pitch angles on the tower.

Christoph Pudenz from the technical inspection service DEKRA provided an overview of potential damages to a turbine. During periodic inspections, the condition of the foundation, tower, and machinery is assessed at regular intervals.

For example, the foundation is inspected for cracks, while the tower is checked for bolt tightness and corrosion on attachments. In the machinery, large components like the gearbox and transformer are examined to detect early signs of damage. Rotor blades are also thoroughly inspected during periodic inspections, including internal and external visual examinations. Inspectors look for cracks, erosion, chipping, or delaminations, and lightning protection systems are tested. Traditional rotor blade inspections are labor-intensive, involving industrial climbers who assess the blades up close. Through the conventional tapping of fiberglass-reinforced plastic (GFRP), chipping and its extent can be quickly identified. However, hidden defects are generally not visible to the naked eye and require other methods.

Jochen Aderhold, a researcher at the Fraunhofer Institute, discussed the available methods for detecting hidden defects and potential damage hotspots. He highlighted the use of active and passive thermography in rotor blade inspections. Current thermographic methods are time-consuming and require turbines to be offline for extended periods. Insights gained from investigations over the past 15 years are relatively clear: approximately 90% of rotor blade damages are production-related. Additionally, damages occur during the transport and installation of turbines, and optimal performance is often not achieved.

Lars Osterbrink, Technical Director at the Marburg-based start-up LATODA, explained in his presentation how a novel method using thermography and artificial intelligence to assess material fatigue in rotor blades can sustainably improve the energy yield of wind turbines.

By detecting surface damages on rotor blades early, operators can enhance the efficiency of their wind turbines. Osterbrink reported that under BAM's leadership, the novel inspection method was successfully tested. Thermographic images captured with an infrared camera made erosion damages on rotor blades visible. The feasibility study for the automatic detection of thermographic turbulence patterns was based on an AI prototype model that already reliably and accurately identifies turbulence patterns. Building on these positive results, further development is planned, including the automatic prediction of turbine efficiency losses caused by turbulence patterns and other factors (e.g., location, weather conditions, turbine type).

Compared to conventional inspection methods using industrial climbers, which typically take 2-3 days, a thermographic inspection of rotor blades takes about ten minutes. During this short period, the turbine does not even need to be shut down. The thermograms are then analyzed using complex image processing and AI algorithms, identifying and marking even the smallest temperature differences on the blade surfaces, which indicate damage events. Automated data evaluation enables significantly more efficient planning of maintenance and repair work than before. Additionally, the method is cost-effective and could even be conducted more frequently to identify and address early-stage damages near the surface before extensive repairs are required.

The subsequent discussion focused on the economic conditions for operating wind turbines. A high price sensitivity among operators was noted, along with the fact that downtime is compensated under the Renewable Energy Sources Act (EEG), providing little incentive for technical optimization of turbines. The growing number of offshore turbines will pose logistical challenges for technical monitoring. According to Michael Stamm, project manager at BAM, the workshop initiated a knowledge transfer from science to future turbine designs. Significant improvements can already be achieved during the production of rotor blades through stringent quality control during acceptance. Insurance companies should also play a stronger role in networking manufacturers, operators, scientists, and technical inspection services. They possess comprehensive datasets on damages that could be of great importance for scientific research. Another workshop on knowledge transfer is being planned.