Ice measurement on the wind turbines' blades by close-range photogrammetry methods

• Main Author: Majlesi, Mohammadreza
• Format: Article in Journal/Newspaper
• Language: English
• Published: École de technologie supérieure 2015
• Subjects: Glace Prévention; Éoliennes; Turbines Aubes; Photogrammétrie numérique; Turbines Aubes Modèles mathématiques; Analyse géométrique; photogrammétrie à courte portée; photogrammétrie rapprochée; méthode de détection de glace; modélisation 3D à base d'images; structure du mouvement; transformation de similarité en 3D; Arctic; Rime
• Online Access: https://espace.etsmtl.ca/id/eprint/1542/; https://espace.etsmtl.ca/id/eprint/1542/1/MAJLESI_Mohammadreza.pdf; https://espace.etsmtl.ca/id/eprint/1542/2/MAJLESI_Mohammadreza-web.pdf

Photogrammetric scanning for measuring the thickness of ice is quite new in the wind energy field. When the rotor blades were mounted on the wind turbine tower, ice thickness inspections became nearly impossible to be performed in the past but by image base modeling methods, geometry extraction of ice of frosted blades are possible in the form of mesh models. In this research thesis, 3D model of wind turbine blades without ice is designed as reference Master CAD data. A prototype of this CAD model is made by a 3D printer machine which after painting, a pattern of retro targets is added to blades surfaces. Digital Single-Lens Reflex (SLR) camera is used to capture several images of objects with different modes of lighting and illumination. A series of point clouds is produced by using the Structure from Motion (SfM) method and VisualSfM software. Each point cloud has an arbitrary coordinate system that is scaled and transferred to Global Coordinate System by 3D transform on CATIA software based on similar features between 3D point cloud and master CAD data. When the ice has sufficient texture, such as rime ice and the texture is captured in the images, a point cloud corresponding to the captured region will also be detected. The position of ice on the blades is detected by coded retro targets on blade’s surface and the order of image acquisition. The generated point cloud and its mesh are compared with the Master CAD or reference geometric model and finally shows in 3D format, the shape, thickness and approximate mass of ice on the blades. This study focused on the feasibility of reconstruction of 3D polygonal mesh models by images captured from a wind turbine blade with a configuration of digital SLR cameras. The main objective of this reconstruction is to identify the volume and shape of ice-accumulated areas on the wind turbine blades in arctic regions. Geometrical comparison of these triangular mesh models with the non-iced blade CAD data could reveal the geometry of accumulated ice.