Acoustic inspection system with unmanned aerial vehicles for offshore wind turbines: A real case study
Wind energy has become fundamental in the global transition towards renewable energies, with the deployment of larger and more complex wind turbines. CMS play a crucial role in early fault detection, enhancing productivity while decreasing downtimes and maintenance costs to ensure the optimal perfor...
| Autores: | , , , |
|---|---|
| Tipo de documento: | artigo |
| Data de publicação: | 2025 |
| País: | España |
| Recursos: | Universidad de Castilla-La Mancha |
| Repositório: | RUIdeRA. Repositorio Institucional de la UCLM |
| OAI Identifier: | oai:ruidera.uclm.es:10578/47320 |
| Acesso em linha: | https://hdl.handle.net/10578/47320 |
| Access Level: | Acceso aberto |
| Palavra-chave: | Offshore wind turbines Acoustic analysis Maintenance management Unmanned aerial vehicle Structural heal monitoring |
| Resumo: | Wind energy has become fundamental in the global transition towards renewable energies, with the deployment of larger and more complex wind turbines. CMS play a crucial role in early fault detection, enhancing productivity while decreasing downtimes and maintenance costs to ensure the optimal performance and viability of the wind energy industry. This paper presents a novel non-destructive testing system embedded in an unmanned aerial vehicle designed to acquire acoustic data from rotating wind turbine components. This approach develops pre-processing and filtering methodologies based on wavelet transform, Fast Fourier or energy transformation to avoid undesired noise sources, e.g., the rotor of the drones or the environment, and to obtain patterns associated with the real state of the wind turbine. The implementation of acoustic monitoring in wind turbines is a novelty in the current state of the art, and this methodology is tested in an operating offshore wind turbine. The experiments incorporate an external condition monitoring system and introduce noise records from simulated mechanical faults. The results demonstrate that all the noise sources and faulty and healthy scenarios can be differentiated, proving the reliability of the methodology and the robustness of the fault detection approach. |
|---|