EPSRC Strategic Equipment - High Speed CT

X射线计算机层析成像(XCT)是一种扫描技术，能够实现内部和外部几何图形的全3D可视化和询问。它在工业(尤其是制造业)和学术研究中变得流行起来，因为它使我们能够在各种长度范围内看到比以往任何时候都多的东西，并且完全是非破坏性的。获取如此丰富的数据所花费的时间对于许多应用程序来说是令人望而却步的，单次扫描需要数十分钟，甚至几个小时。这项提案中概述的设备将在数十秒内实现高分辨率扫描，如果进行一些基础研究，扫描速度会更快。这是英国首个将产生大量科学进步的实验。已经有少量使用基于实验室的XCT扫描仪的“动态”实验，在每个加载步骤中顺序加载(例如压缩)和扫描诸如新材料的样品。在这里，人们可以观察材料随时间的变化，确定失效机制，突出潜在的制造改进和帮助确定材料特性。研究如此之少的原因是，所需的扫描次数可能导致采集时间为数天。该设备在速度上的大幅提升将使连续采集数据集的总扫描时间减少到几分钟左右。然后，可以在离散的时间点对样本进行评估，并集中在观察到的故障的关键时刻。这种独特的表征能力将大大加快新聚合物、陶瓷和金属合金开发的科学进步。由于扫描所需的时间，制造应用通常仅限于少数高价值部件。速度的显著变化将允许在生产线内进行所需的高通量扫描。这是一场重大革命的第一步，这场革命将需要由机器学习算法支持的大数据分析，以在合理的时间范围内提供接受/拒绝决策。总而言之，这将是在高分辨率和相关周期内实现大批量部件100%检验的变化的驱动力。

X-ray Computed Tomography (XCT) is a scanning technique that enables full 3D visualisation and interrogation of internal and external geometries. It has become popular within industry (particularly manufacturing) and academic research as it enables us to see more than ever before at a variety of length scales and is completely non-destructive. The time taken to obtain this wealth of data is prohibitive to a number of applications with a single scan taking tens of minutes, up to a few hours. The equipment outlined in this proposal will enable high-resolution scans in tens of seconds, and even faster with some fundamental research. This is a UK first that will generate a wealth of scientific advancement.There have been a small countable number of "dynamic" experiments using lab based XCT scanners where a sample such as a novel material is sequentially loaded (e.g. compression) and scanned at each loading step. Here one can observe the changes in the material through time, identifying failure mechanisms, highlighting potential manufacturing improvements and aids in determining material properties. The reason for so few studies is that the number of scans required can lead to acquisition time of days. The substantial gain in speed with this equipment will reduce the total scan time to a matter of minutes with a continuously acquired dataset. The sample can then be evaluated at discrete points in time, and concentrate around the critical onset of failure observed. Scientific advancement in the development of new polymers, ceramics and metal alloys will be considerably accelerated with this unique characterisation capability.Manufacturing applications are often limited to a few high-value components because of the time taken to scan. The significant step change in speed will allow for high-throughput scanning that is desirable within the manufacturing line. This is the first step in a major revolution that will require big data analytics powered by machine learning algorithms to deliver accept/reject decisions in a reasonable time scale. Together this will be a driver for change in achieving 100% inspection of large component batches, at high resolution and at relevant cycle times.