• Elberfeld, Tim

Abstract

Fiber-reinforced composites are an important part of modern material science. They are used in every area of our lives in some shape or form. Car manufacturers use them to make the car body lighter and give it better compression properties if there is an accident. Concrete mixed with fibers is now a widely used material not only in construction, but also in private home renovation and improvement. Wind turbines have blades made of plastic, reinforced with longer strands of glass fibers to make them lighter and stronger against the forces of the wind. The more the field progresses, the more the composites can be improved and their applications diversified. A fiber-reinforced composite consists of a base or matrix material that embeds some sort of fibrous material, which is the reinforcement component. This can be glass or carbon fibers, steel bars or natural fibers like bamboo. They can be added in layers, bundles, weaves or loosely dispersed within the material before its set. A preferred method for investigating composites is X-ray computed tomography. This technique can generate a fully 3-dimensional image, that reveals the inside structure of the analyzed samples. As X-rays penetrate most materials easily, it is possible to acquire images of the attenuation of the incoming radiation when placing such a material sample between the source and the imaging device. Several of those so-called projections are collected over a large angular range, preferably 180 degrees or more. Using mathematical methods the volumetric image, the reconstruction, can then be computed from the projections. The thesis you are reading is dealing with the problem of the quantitative and statistical analysis of such composites, using X-ray computed tomography. More precisely, it shows approaches to detect the fibers in volumetric images of glass fiber-reinforced polymers and analyze their structure to give insight into their geometrical properties. After introducing the foundations for the work, a first framework for the extraction and analysis of straight fibers, called PARE, is introduced. This framework is then extended to also deal with fibers with arbitrary curvature. Parametric models for both straight and curved fibers are presented and it is shown how those parametric models can be improved using numerical optimization based on the information in the X-ray projection images.

Topics

• impedance spectroscopy
• polymer
• Carbon
• extraction
• tomography
• glass
• glass
• steel
• fiber-reinforced composite