Tomography of Materials and Structures最新文献

Tomographic imaging supports a great number of medical and material science applications. The collected projection data usually has different types of imaging artefacts and noise. Various image pre-processing and reconstruction methods are used to obtain volumetric datasets of high quality for further analysis. In order to minimise reconstruction artefacts, one can apply either filtering and/or data completion/inpainting techniques which can recover the data. Deep learning (DL) methods to remove artefacts and noise have been successfully applied in the past. In this paper, we present a novel approach based on conditional generative adversarial networks (cGANs) to remove stripe artefacts. The novelty of the presented technique is in how the training data for DL is extracted from the same tomographic dataset that needs recovery. We also provide new deterministic stripe detection and inpainting algorithms to support the development. The presented methods are compared with other stripe removal algorithms and applied to 3D and 4D high-resolution X-ray data collected at Diamond Light Source synchrotron, UK. The proposed DL method delivers reconstructed images with minimised ring artefacts while being a parameter-free approach. A similar DL strategy can also be applied to remove other types of artefacts in images.

We investigate appliance of different deep learning models to the problem of semantic segmentation of structural defects in computed tomography images of fiber-reinforced polymer composite material. Specifically, we try to segment porosities and delaminations in a specimen using U-Net and DeepLabv3 neural networks. We find out that complex models struggle to generalize solutions on small data samples that are generally available to individual research teams, whereas smaller models are the right choice for approaching defect segmentation in CT images. Our experiments are based on our own laboratory data, collected with X-ray microtomography and labeled manually for the semantic segmentation task.

Within the past decade, X-ray micro computed tomography (µCT) has become an advanced non-destructive tool to analyse the internal structure of opaque materials. In addition to high spatial resolution, new generations of laboratory-based µCT machines and synchrotron imaging facilities can achieve high temporal resolution. This makes µCT the method of choice to study dynamics processes such as multi-phase fluid flow within porous media at the micro-meter scale. To perform such experiments a system compatible with X-ray imaging is needed. This essentially includes an X-ray transparent flow cell which should both be compatible with the requirements of the experimental study and the constraints of the µCT facilities. So far, most µCT flow cells are custom built and optimised for specific experiments/purposes. This paper reviews the previously published X-ray transparent cell designs, their advantages, and limitations. We present the state-of-the-art in design of X-ray transparent flow systems and discuss the technical challenges around performing µCT-based fluid flow experiments. We also present a review of the main applications which have benefited from µCT imaging studies and discuss the flow cell designs according to applications.

The simplicity of nonclinical x-ray tomography data collection has caused some to overlook the importance of saving additional experimental information or experiment meta-data. Sample meta-data are often saved in the experimenter’s logbook while meta-data about the instrument and experimental conditions are saved by the instrument itself or by the instrument operator. The lack of standardization of this approach has limited the development of automatic tools for data analysis and experiment logging but has also hindered the ability to reproduce the data collection and data analysis under the same conditions. In this paper we introduce tomo-meta, a publicly available repository of laboratory and synchrotron based tomography instrument meta-data files with the aim of presenting how meta-data are currently collected and identify best practices that enable data collection and data analysis repeatability. Structured and machine readable meta-data files, such as HDF, CSV, JSON, XML, etc., are essential for creating automatic processing pipeline. When the tomography meta-data files are structured as machine readable, we also provide a simple python script to automatically load them into a python dictionary.

Magnetic Pulse Welding (MPW) facilitates the permanent joining of dissimilar metallic materials through the sudden impact generated by a magnetic pulsed field. The process can introduce distinct morphological features at the interface of bi-material joints, which subsequently affect the joint’s quality and durability. This article delves into the investigation and quantification of various interfacial morphologies in Aluminum/Copper and Aluminum/Steel joints, using high-energy phase-contrast synchrotron micro-tomography. Surface topography is extracted from 3D tomographic datasets between dissimilar materials, enabling a comprehensive comparison between different material pairings and various locations within the weld. The study analyses and compares the roughness parameters of these surfaces. Moreover, it describes the interface’s waves and vortexes through diverse morphological metrics, encompassing their shape and size. The results provide evidences that vortexes evolve in three dimensions, with lateral growth and collapse. The waves and vortexes shapes promote material interlocking, increasing the contact area between the dissimilar materials by up to 20%. The interface morphology of Al/Cu joints exhibits higher roughness and a greater number of vortexes compared to Al/Fe joints. Lastly, the findings reveal the presence of interface damage in the form of pre-existing discontinuities.

The performance and reactivity of coke in a blast furnace is critically dependent on the accessibility of the coke structure to carbon dioxide (CO₂) gas. We used xenon gas K-edge subtraction in synchrotron micro-CT imaging to probe the extent to which gas could penetrate the microstructure of six different metallurgical cokes made from Australian coals. We compared the distribution of the xenon sorbed by the coke samples before and after reaction with CO₂ at 1100 °C to 20–30% mass loss. Xenon is as strongly sorbed onto surfaces as carbon dioxide and can thus be used as an x-ray-visible analogue of CO₂. Aside from traces of pyrolysis ash, coke comprises two major components; the reactive maceral derived component (RMDC), which passes through a molten state during coke manufacture to form a foam-like structure, and the inertinite maceral derived component (IMDC), which are particles ranging from a few microns to a few millimetres in size, embedded in the RMDC. These components were found to behave very differently in this study. Prior to reaction, the RMDC component sorbed only a small amount of xenon and most of the IMDC sorbed little to no xenon. However, a small fraction of the IMDC took up significant quantities of xenon in high concentration. This suggests that a significant fraction of the surface area of unreacted coke comes from rare, high-surface-area IMDC components.

Imaging of the coke after reaction showed the RMDC still sorbed only small amounts of xenon, indicating that the surface area in these components was largely unchanged. However, the previously xenon-inaccessible IMDC regions sorbed large quantities of xenon after reaction, reaching peak xenon densities many times that seen in the free xenon gas. Thus, surface area is produced by reaction with CO₂ or (more probably) much of the pre-existing surface area is made accessible by reaction. This shows that IMDC provide most of the reacting surface during early stages of reaction of coke with CO₂. This was confirmed by the corresponding loss of mass seen in these IMDC particles relative to the RMDC.

The constitutive behavior of granular materials is highly influenced by the degree of saturation and whether a saturated or unsaturated framework is adopted to model the behavior of granular materials. Conventional axisymmetric triaxial compression (CTC) testing of saturated specimens tends to assume that the specimen remains saturated during shearing. X-ray computed tomography (CT) has allowed for 3D in-situ measurements beyond the global scale to investigate microscale and localized events such as fabric evolution. The literature reported several CT-coupled geotechnical experiments of saturated specimens; however, no study examined how X-ray exposure might affect the specimen, specifically the radiolysis of pore water. In this study, we present the synchrotron micro-computed tomography (SMT) results of experiments performed on sand specimens without shearing (acrylic tubes) and with shearing (CTC experiments) to assess the influence of X-ray exposure on the development of the gas phase of saturated sand. The observed phase changes were dependent on the initial pore water pressure and duration of exposure to the X-ray; the behavior of individual gas bubbles was dependent on the bubble’s surrounding sand grains and pore throat sizes leading to changes in the degree of saturation.

Self-piercing rivet (SPR) joining is a process that has been adopted in the automotive industry, and it is important to be able to characterize SPR joints non-destructively. X-ray computed tomography (CT) allows visualizing the internal structure of the sample whilst preserving the workpiece. In this work, the application of X-ray CT to SPR joint characterization is investigated. Many features of a riveted joint can be observed. This includes the boundary between sheet interfaces and the positions where the substrates and rivet meet, which are important to quality assessment. In the case of similar materials, some of these features become difficult to observe but the presence of adhesive between the sheets enables the features to be more easily observed. Furthermore, substrate fracture and rivet cracks, radial cracks and the asymmetry of SPR joints can be detected due to the difference in greyscale to the surrounding materials. It is also revealed that image resolution plays an important role in defect detectability of the X-ray CT technique.

Tristructural isotropic (TRISO) coated fuel particles are a nuclear fuel form under extensive study for use in advanced nuclear reactor concepts. TRISO fuels are subjected to high temperature neutron irradiations and then examined to assess their performance by determining fission product retention and studying morphological changes. Micro X-ray computed tomography is one method of nondestructively studying the effects of TRISO performance. This work addresses the need for image processing to remove X-ray tomographic reconstruction artifacts that prevent the study of TRISO features, as the TRISO particles’ high Z kernel can introduce metal artifacts that degrade the image quality in the surrounding low Z coating layers. These metal artifacts were reduced by imaging the TRISO particles with both high- and low-energy X-rays and applying a mask to the radiographs obtained with low-energy X-rays to digitally remove the dense fuel kernel region. These masked radiographs were then used to produce a tomographic reconstruction which was combined with the tomographic reconstruction of the high-energy data. This enabled the relatively-low-density TRISO buffer layer to be examined in more detail, providing information on irradiation induced dimensional changes of the coatings. This methodology, which helps see the full picture of a TRISO particle, is not limited to nuclear fuels but can be applied to systems that contain highly attenuating material surrounded by less dense materials.

In a fine-focus geometry, for a given detector, the resolution achievable by conventional tomographic region-of-interest (ROI) imaging is limited by the smallest possible distance of the radiation source from the rotation axis, i.e., the radius of the smallest cylinder about the rotation axis that encloses the object. In situations where the specimen to be imaged is irregularly shaped, or the ROI is off-centre, higher magnification can only be achieved from a limited range of angles, (possibly in a separate scan), and a tomographic reconstruction technique able to incorporate this additional data would be advantageous. Here we present such a technique for imaging planar (or laminar) objects based on a combination of multiple tomography and laminography scans with increasing magnification that employ helical and planar source trajectories respectively. Relative to laminography, this hierarchical combination improves depth resolution, (longitudinal direction), as well as reducing imaging artefacts. Relative to full-field tomography the proposed method increases resolution, particularly in the plane of the specimen, (transverse direction). The foundation of the technique is a generalisation of accelerated multi-grid tomographic reconstruction methods to the case of multiple independent collections of radiographs. Here we first demonstrate the concepts and performance of this technique through a simulated example. We then demonstrate the successful application of the method experimentally to a thin rock section and a printed circuit board.