Image-based numerical modelling of heterogeneous materials

Sammanfattning: In science there has always been a desire to visualise the invisible. Since the discovery of X-rays in 1895, imaging research has made remarkable progress. Nowadays, state-of-the-art technology allows to visualise the micro-structure of objects in three dimensions. However, merely visualising the structure is often insufficient. The quantitative information regarding morphology and structure is of great interest. Therefore, in addition to significant advancements in X-ray image acquisition and three-dimensional reconstruction, image analysis has become an active research field in recent years. Modern image analysis methods enable to extract even invisible information from image data. The heterogeneous micro-structure of composites imposes advanced material characterisation as even for the largest composite structures, such as wind turbine blades or airplane wings, the material properties are dictated on the micro-scale. Image-based modelling offers exceptional capabilities in analysing the micro-structure at the fibre level and numerically predicting material behaviour even at larger scales. However, image-based modelling is a complex process and all work-steps must be in line with the final modelling goal. Therefore, X-ray computed tomography aided engineering has been introduced to emphasise the importance of a holistic point of view on the image-based modelling process. The developed X-ray computed tomography aided engineering methodology has been developed based on micro X-ray computed tomography scans for non-crimp fabric glass-fibre reinforced composites. It is demonstrated that local fibre orientations and fibre volume fractions can be accurately imaged and transferred onto a finite element model. Thereby, the tensile modulus of the scanned samples can be accurately predicted and possible stress concentration regions detected. However, conventional micro X-ray computed tomography presents a major drawback. Achieving the required high resolutions to visualise carbon or glass fibres, typically ranging between 5 to 20 μm, limits the scanning field of view, which remains in the millimetre range. This drawback is overcome with new approaches in image-based modelling involving advances in imaging and image analysis. Therefore, targeted approaches for accurate image-based modelling are presented which increase the possible scanning field-of-view of fibrous composites by up to three to six orders of magnitude.