AQUADA-GO: Automated blade damage detection and near real-time evaluation for operational offshore wind turbines

AQUADA-GO

Summary

The AQUADA-GO project develops a methodology for automated, noncontact, near real-time blade damage detection and risk evaluation in a single step using thermography and computer vision without stopping the normal operation of wind turbines. The project will take the AQUADA technology developed in the laboratory of DTU Wind Energy and apply it to operational offshore wind turbines.

The idea for the project originated at DTU Wind and Energy Systems, which has already published scientific articles on the so-called AQUADA technology. The GO in the project title has been added as the partners are now ready to transfer the work from the lab at DTU into the environments in which the industry operates.

The project will offer an innovative, market-ready solution to disrupt the labor-intensive and multi-step blade inspection paradigm. AQUADA-GO will develop and demonstrate software implementation and hardware integration in an all-in-one drone platform for offshore applications. A full-scale prototype system will be tested and demonstrated on RWE-owned offshore wind farms together with the commercial partner Quali Drone and the research partner DTU Wind Energy. The project aims to reduce blade inspection costs by at least 50% and offer a significant contribution to drive down the Levelized cost of energy (LCOE) of wind power by 2-3% over 25-30 years’ project lifetime of offshore wind farms compared to the solutions currently available on the market.

Main activities include (1) Thermal features of blade damages in operational conditions considering environmental effects; (2) Physics-based data-driven computer vision algorithms for automated damage detection and near real-time damage evaluation; (3) Hardware and software integration in an all-in-one drone platform; (4) Onshore testing and offshore demonstration without stopping the normal operation of wind turbines; and (5) Validation of LCOE reduction to verify the economic impact of the developed solutions.

Quali Drone will commercialize the project solution and scale up the business rapidly thanks to the unparalleled competitive advantages of the AQUADA-GO technology and the unique access to hundreds of wind farm sites owned by the project partner RWE Renewables globally.

The AQUADA-GO technology is expected to reduce CO2 emission by 30-50% per turbine inspection compared to commercially available solutions. In total, the AQUADA-GO project is estimated to increase the combined annual turnover of the two involved companies by 125 to 230 million DKK and create 33 to 55 full-time new jobs in 3 to 5 years after project completion.

The project will mature from TRL 5-6 to 7-8 from the beginning of 2023 to the end of 2025.

Research Outcome

Relevant peer-reviewed research articles:

1. Chen, X., Shihavuddin, ASM., Madsen, S. H., Thomsen, K., Rasmussen, S., & Branner, K. (2021). AQUADA: Automated quantification of damages in composite wind turbine blades for LCOE reduction. Wind Energy, 24(6), 535-548. https://doi.org/10.1002/we.2587

2. Chen, X., Semenov, S., McGugan, M., Madsen, S. H., Yeniceli, S. C., Berring, P., & Branner, K. (2021). Fatigue testing of a 14.3 m composite blade embedded with artificial defects – damage growth and structural health monitoring. Composites - Part A: Applied Science and Manufacturing, 140, [106189]. https://doi.org/10.1016/j.compositesa.2020.106189

3. Chen, X., Sheiati, S., & Shihavuddin, ASM. (2023). AQUADA PLUS: Automated Damage Inspection of Cyclic-loaded Large-scale Composite Structures using Thermal Imagery and Computer Vision. Composite Structures, 318, [117085]. https://doi.org/10.1016/j.compstruct.2023.117085

4. Chen, X., Janeliukstis, R., & Sarhadi, A. (2022). Thermographic data analytics-based damage characterization in a large-scale composite structure under cyclic loading. Composite Structures, [115525]. https://doi.org/10.1016/j.compstruct.2022.115525

5. Sheiati, S., & Chen, X. (2023). Deep learning-based fatigue damage segmentation of wind turbine blades under complex dynamic thermal backgrounds. Structural Health Monitoring, [14759217231174377]. https://doi.org/10.1177/14759217231174377

6. Spencer, M., Sheiati, S., & Chen, X. (2023). AQUADA GUI: A graphical user interface for automated quantification of damages in composite structures under fatigue loading using computer vision and thermography. SoftwareX, 22, [101392]. https://doi.org/10.1016/j.softx.2023.101392

7. Eder, M. A., Sarhadi, A., & Chen, X. (2021). A novel and robust method to quantify fatigue damage in fibre composite materials using thermal imagine analysis. International Journal of Fatigue, 150, [106326]. https://doi.org/10.1016/j.ijfatigue.2021.106326

8. Jia, X., & Chen, X. (2024). AI-based optical-thermal video data fusion for near real-time blade segmentation in normal wind turbine operation. Engineering Applications of Artificial Intelligence, 127, [107325]. https://doi.org/10.1016/j.engappai.2023.107325

9. Jia, X., & Chen, X. (2024). Unsupervised Wind Turbine Blade Damage Detection With Memory-Aided Denoising Reconstruction. IEEE Transactions on Industrial Informatics. https://doi.org/10.1109/TII.2024.3459612

10. Sheiati, S., Jia, X., McGugan, M., Branner, K., & Chen, X. (2024). Artificial intelligence-based blade identification in operational wind turbines through similarity analysis aided drone inspection. Engineering Applications of Artificial Intelligence, 137, 109234. https://doi.org/10.1016/j.engappai.2024.109234

Dataset: Drone-based optical and thermal videos of rotor blades taken in normal wind turbine operation

In AQUADA-GO project, we collected a large-scale optical-thermal wind turbine blade video dataset. This dataset contains 36 optical-thermal video pairs and over 20,778 images. Moreover, we make it publicly available to facilitate future studies. It can be accessed at DOI: 10.21227/yzs5-1067.

AQUADA-GO Introduction

We developed a large AI model to detect blade damage without stopping the normal operation of wind turbines. We use paired RGB and thermal videos taken by drone in various weather conditions. The large AI model is trained with over 100,000 images.

AQUADA-Seg: Near real-time optical-thermal wind turbine blade segmentation

In AQUADA-GO, we propose an AI-based optical-thermal blade video segmentation model. This model achieves near real-time optical-thermal wind turbine blade segmentation without stopping turbines. https://doi.org/10.1016/j.engappai.2023.107325

Blade surface damage detection results

A novel AI model has been developed to detect blade surface damage automatically. This model achieves near real-time blade surface damage identification without stopping the normal operation of wind turbine turbines, paving the way for automated surface and underneath blade damage detection.

AQUADA-Det: AI-based thermographic wind turbine blade structural damage detection

A novel AI-based thermographic approach that achieves in-field blade structural damage detection without stopping the normal operations of wind turbines.

Outreach & Media Coverage