Using Historic Materials Data from Assurance Testing to Optimise Future Manufacturing Processes of High Integrity Components

Element Materials Technology等公司在为多个行业的一系列高完整性应用提供材料性能独立保证方面发挥着关键作用。因此，Element创建了一个强大的数字平台，用于记录和报告测试结果;这为探索工业4.0背景下此类数据流的理想提供了一个很好的机会，并为设计师和制造商提供了一个强大的反馈回路。通常，将纸质记录转换为数字平台的挑战阻碍了这些数据的使用。与任何大数据项目一样，可以查询现有的数字数据以了解趋势，但将统计过程控制方法应用于此信息，为行业带来真实的好处，生成可供Element Materials Technology及其客户群使用的工具，以监控性能并在失去控制之前提供预防性干预措施。在某些情况下，通过定义真正控制生产的监测参数，符合过程的证据也可能排除未来测试的需要。断裂韧性数据通常代表两种关键类型的机械行为;与结构灾难性失效相关的低电阻穿晶解理和描述韧性断裂的高电阻微孔聚结（见图1）。试样几何形状、材料机械性能和失效模式的许多其他影响也会影响建立适当的性能估计。主曲线方法已成为处理过渡区低合金钢断裂韧性数据的公认工程解决方案，在过渡区，韧性值的记录变化很大（见图2）。这已被纳入国际标准，作为评估方法的支柱（2，3），并取决于假定的材料行为，如概率分布的设定参数所示。失效过程的随机性可能导致即使在单一材料中也会出现较大的变化;主曲线提供了一个框架，用于对稀疏数据进行性能估计。在此过程中，它已被证明在能源行业的应用中非常成功，延长了关键基础设施的使用寿命，该项目将发展有关保证方法、所涉制造过程冶金学的知识，并深入了解可用于正确评估数据的统计方法。纯粹的数据驱动方法可能会导致制造过程的过度规范，通过不必要地拒绝合适的材料来浪费时间，材料和资源。因此，该项目将在材料科学与工程系和数学与统计学院之间建立伙伴关系。

Companies such as Element Materials Technology play a key role in providing independent assurance of materials performance for a range of high integrity applications across several sectors. As such Element has created a strong digital platform for recording and reporting of test results; this presents a great opportunity to explore the ideals of the flow of such data in an Industry 4.0 context and provides a robust feedback loop to the designer and manufacturer if enacted. Often the challenge of converting paper records to a digital platform prevents the use of this data. The existing digital data can be interrogated for trends as with any big data project but there is real benefit to industry in applying a statistical process control approach to this information, generating tools that can be used by Element Materials Technology and their customer base to monitor performance and provide preventative interventions in manufacturing prior to loss of control. Evidence of conforming to process may also preclude the need for future testing in some circumstances by defining the parameters to monitor that truly control manufacture. Fracture toughness data often represents two key types of mechanical behaviour; low resistance trans-granular cleavage associated with catastrophic failure of structures and high resistance micro-void coalescence that describes ductile rupture (see Figure 1). Many other effects of specimen geometry, materials mechanical properties and failure modes can also effect establishing appropriate estimates of performance. Identifying when these have happened is key to providing assurance of future performance.The Master Curve methodology has become the accepted engineering solution for processing fracture toughness data of low alloy steels in the transition region where large variability in recorded toughness values are observed (see Figure 2). This has been adopted into international standards as the backbone of assessment methodologies (2,3) and is dependent on assumed materials behaviour, as exemplified by set parameters for probability distributions. The stochastic nature of the failure process can result in large variations even within a single material; the Master Curve provides a framework for making estimates of performance on sparse data. In doing so, it has proved very successful for energy industry applications, affording life extensions to key infrastructure.This project will develop knowledge of assurance methodologies, metallurgy of the manufacturing processes involved and an in-depth understanding of the statistical methods that can be employed to assess the data correctly. A purely data driven approach could result in over specification of the manufacturing processes, costing time, material and resources through unnecessary rejection of suitable materials. As such, the project will be run in partnership between the Department of Materials Science and Engineering and the School of Mathematics and Statistics.