Loading...
Publication
Inspeção autónoma de turbinas eólicas offshore com UAVs residentes
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 10.02 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A crescente adoção de turbinas eólicas offshore exige soluções para a sua inspeção, capazes de reduzir custos e riscos associados a métodos convencionais. Esta dissertação apresenta um pipeline de inspeção autónoma com UAVs residentes, estruturado em várias etapas que vão desde a aproximação à turbina até ao retorno seguro, incluindo deteção de obstáculos, seguimento das pás e coordenação multi-UAV. O sistema integra sensores LiDAR e câmara RGB com algoritmos de visão computacional e segmentação, permitindo planear trajetórias seguras e garantir cobertura completa das pás. A abordagem foi validada em ambiente simulado, recorrendo
às ferramentas Gazebo, ROS e PX4, onde foram avaliadas métricas de desempenho como tempo de missão, distância percorrida e robustez em diferentes cenários operacionais. Os resultados demonstraram que a solução proposta possibilita inspeções completas e seguras, mesmo perante restrições impostas pela geometria da turbina, e que a utilização de múltiplos UAVs contribui para ganhos significativos de eficiência na execução da missão.
The increasing adoption of offshore wind turbines demands innovative inspection solutions that can reduce costs and risks associated with conventional methods. This dissertation presents an autonomous inspection pipeline with resident UAVs, structured in several stages ranging from the approach to the turbine to the safe return, including obstacle detection, blade tracking, and multi-UAV coordination. The system integrates LiDAR and RGB camera sensors with computer vision and segmentation algorithms, enabling safe trajectory planning and ensuring complete blade coverage. The approach was validated in a simulated environment using Gazebo, ROS, and PX4, where performance metrics such as mission time, traveled distance, and robustness under different operational scenarios were evaluated. The results demonstrated that the proposed solution enables complete and safe inspections, even under geometric constraints imposed by the turbine, and that the use of multiple UAVs provides significant efficiency gains in mission execution.
The increasing adoption of offshore wind turbines demands innovative inspection solutions that can reduce costs and risks associated with conventional methods. This dissertation presents an autonomous inspection pipeline with resident UAVs, structured in several stages ranging from the approach to the turbine to the safe return, including obstacle detection, blade tracking, and multi-UAV coordination. The system integrates LiDAR and RGB camera sensors with computer vision and segmentation algorithms, enabling safe trajectory planning and ensuring complete blade coverage. The approach was validated in a simulated environment using Gazebo, ROS, and PX4, where performance metrics such as mission time, traveled distance, and robustness under different operational scenarios were evaluated. The results demonstrated that the proposed solution enables complete and safe inspections, even under geometric constraints imposed by the turbine, and that the use of multiple UAVs provides significant efficiency gains in mission execution.
Description
Keywords
UAVs turbinas eólicas offshore LiDAR visão computacional Inspeção autónoma Autonomous inspection Offshore wind turbines Computer vision
Pedagogical Context
Citation
Publisher
CC License
Without CC licence