Tag Archives: jet engine

Thinking out of the box leads to digital image correlation through space

This is the third in a short series of posts on recent engineering research published by my research group.  Actually, two have already been published: ‘Salt increases nanoparticle diffusion‘ on April 22nd, 2020; and ‘Spatio-temporal damage maps for composite materials‘ on May 6th, 2020 and then I got distracted.  This third one arose from the same project as the time-damage maps which was sponsored by the United States Air Force.  The time-damage maps allow us to explore the evolution of failure in complex materials; however, we already know that damage tends to initiate from imperfections or flaws in the microstructure in the material.  New continuous fibre reinforced composite (CFRC) materials based on ceramics are very sensitive to defects or anomalies in their microstructure, such as misalignment of fibres.  However, they are capable of withstanding temperatures in excess of 1500 degrees Centigrade, which offers the opportunity to use them in jet engines or nuclear power plants to help generate energy more efficiently.  Therefore, it is worthwhile investigating effective methods of inspecting their microstructure which we can do either destructively by repetitively polishing away the surface of a sample and viewing it in a microscope, or non-destructively using x-ray tomography.  In both cases, the result is hundreds of ‘images’ containing millions of data values from which it is challenging to extract useful information.  In our work, we have used a little lateral thinking, to show how digital image correlation, usually used to track deformation of structures using multiple images collected over time [see ‘256 shades of grey‘ on January 22nd, 2014] , can be used to track fibres through the multiple images of the layers of the microstructure.  The result is the sort of ‘stick’ diagram in the image showing the orientation of fibres through the sample.  We have demonstrated that our new algorithm was more reliable and 30 times faster than its nearest rival.

The image shows, at the top, a typical stack of images from the microscope of a ceramic matrix composite; and, at the bottom, a plot of 3d profiles of the fibres tracked using the DIC-based method with the fibres orientated nominally at ±45° from the sectioning (x-y) plane shown in red and green colours.


Amjad K, Christian WJR, Dvurecenska K, Chapman MG, Uchic MD, Przybyla CP & Patterson EA, Computationally efficient method of tracking fibres in composite materials using digital image correlation, Composites Part A, 129:105683, 2020.