AME NDT (Anisotropic Media Evaluation for Non-Destructive Testing)

Completed

Nuclear Fission and Fusion(Nuclear Fission, Other nuclear fission)
Not Energy Related

Basic and strategic applied research

ENGINEERING AND TECHNOLOGY (Mechanical, Aeronautical and Manufacturing Engineering)

Not Cross-cutting

Miss K Tant
Mathematics
University of Strathclyde

Standard

EPSRC

29 June 2018

28 December 2021

42 months

£283,726

Info. & commun. Technol.

Scotland

ISCF - Skills

Miss K Tant, Mathematics, University of Strathclyde

Project Contact, Rolls-Royce PLC
Project Contact, National Physical Laboratory (NPL)
Project Contact, National Nuclear Laboratory
Project Contact, PZFlex Limited (UK)
Project Contact, EDF Energy

Manufacturing is a key activity of the UK economy and accounts for more than half of all UK exports. The ability to reliablytest components at every stage, from manufacture to end of service, is crucial for maximising economic growth, minimisingenvironmental impact and ensuring public safety. End of life inspection is particularly important as much of the UK'sinfrastructure is ageing and, due to global financial pressures, cannot be replaced. Thus, the lifetimes of key UK assets,such as nuclear plants, must be extended. Ultrasonic non-destructive testing presents an economically andenvironmentally desirable solution for detecting damage in such components. Similar to medical ultrasound, ultrasonicwaves can be passed through industrial components and subsequently collected, without damaging their internalcomposition. Large networks of sensors, typically arranged in linear arrays, are deployed to carry out these inspections,resulting in large volume, noisy, time-series data. Mathematical algorithms are then required to decipher the informationencoded within these recorded signals and construct images of the component's interior. Such algorithms are fundamentalenablers of the fourth industrial revolution facilitated by robotics and automated systems, which are largely dependent onaccurate sensing, measurement and imaging systems. In many cases, the component under inspection exhibits ananisotropic, heterogeneous microstructure (that is, the material properties are directionally dependent and vary spatially ina random fashion). This is detrimental to standard imaging methodologies as the ultrasonic wave is bent and scattered bymicrostructural features and the responses from defects are obscured. Examples of such difficult to inspect materialsinclude coarse grained steel welds and carbon-fibre reinforced polymer (CFRP) composites. In fact, materials with complexand highly scattering microstructures are becoming increasingly common as industries continue to invest in thedevelopment of lighter, stronger composite materials. To combat the difficulties in imaging within these materials, thecurrent, cutting-edge imaging research within the NDT community endeavours to map the spatially varying materialproperties using time of flight tomography. However, time-of-flight tomography uses only one data point from each recordedtime series and thus does not fully exploit the wealth of information made available by the inspection. The first objective ofthe proposed research is to develop a material mapping methodology which exploits the full recorded signal, addressingthe non-uniqueness issues faced by time-of-flight tomography. This will be achieved via the development of newmathematical models that capture the varying properties of heterogeneous media using probability theory and stochasticmodels. The resulting material maps will then be incorporated into an advanced imaging system whereby the deviation of theultrasonic wave path in the heterogeneous media can be corrected for so that reliable defect detection can be ensured. The secondobjective of the proposed research is to create an algorithm which can reconstruct complete datasets from incompleteobservations using novel matrix and tensor completion techniques (an emerging area within data-science), facilitating faster inspection times and real-time imaging

14/09/18