Image processing based detection of the fibre orientation during depth-controlled laser ablation of CFRP monitored by optical coherence tomography

Materials & Design

Matthias Buser¹  Daniel Holder²  Thomas Graf³  Steffen Boley³  Rudolf Weber³ 

[1] Institut für Strahlwerkzeuge (IFSW), University of Stuttgart, Germany;Corresponding author at: Institut für Strahlwerkzeuge (IFSW), University of Stuttgart, Pfaffenwaldring 43, 70569 Stuttgart, Germany.;Institut für Strahlwerkzeuge (IFSW), University of Stuttgart, Germany;

关键词: Repairing of composites;    Layer-accurate laser ablation;    Fibre orientation;    Image processing;    Optical coherence tomography;   

DOI  :  

来源: DOAJ

【 摘 要 】

Layer-accurate and precisely stepped removal of damaged areas is required for the repairing process of parts made of carbon fibre-reinforced plastic (CFRP) to obtain a high bonding strength of the repair plies in the rebuilding process. Conventional techniques are characterized by a high degree of manual labour and mechanical load of the processed part. Here, a novel laser-based and controlled approach for automated and layer-accurate removal of damaged areas is presented. In-line depth measurements by optical coherence tomography during nanosecond laser ablation enable depth-controlled homogeneous material removal with minimum damage <10 μm and low surface roughness in the range of 7 μm to 13 μm. The processed layer must be detected for an unknown or varying layer thickness to stop the ablation process at the interface between two layers and not somewhere within one ply. The image processing-based determination of the fibre orientation from the depth measurements allow for a reliable online detection of the processed layer with a maximum error of 4° and root mean square error of 1.1°. As a result, layer-accurate damage removal is demonstrated for a complex repairing geometry. The suitability of the automated laser-based approach as preparation for repairing is proven by cross-sections and scanning-electron-microscopy.

【 授权许可】

Unknown