Scientist uses maths theory to keep planes flying safely

Defence scientist Dr Nick Armstrong is using probability theory to help keep defence aircraft safe and ready to fly.

Dr Nick Armstrong is researching the time it takes for cirtical aircraft engine components to acquire the sort of wear and tear that can lead to cracks.

Dr Nick Armstrong is using probability theory to help keep defence aircraft safe and ready to fly.

The defence scientist is researching the time it takes for critical aircraft engine components to acquire the sort of wear and tear that can lead to cracks.

He is combining information about the potential damage to the materials that make up the components with probability theory to help determine when an aircraft engine needs maintenance checks and overhauls.

“If we can bring that physical information about the time it takes to acquire damage before crack initiation, we can then use that to say, ‘It’s time for overhauls to commence. This is the best time to undergo overhauls and those inspections to make sure that the components are working’,” he says.

Armstrong uses a branch of probability called Bayes’ theorem. Named after mathematician Thomas Bayes, it describes the probability of an event, based on prior knowledge of conditions.

“It’s the idea that if I have a proposition A, how true is that property, this proposition? What is the probability of that proposition being true conditional on my background data, my assumptions, my hypotheses about that proposition?” Armstrong explains.

A simple example is the probability of rain conditional on clouds being present.

There are currently large uncertainties about the amount of flying time it takes for an aircraft component to initiate a crack and Armstrong is working to provide better information.

“Given the distress profile and the loading profiles of the components, we want to be able to say, ‘it’s reaching a certain life, and it’s more likely to initiate a crack’,” Armstrong says.

“It’s about being risk averse, but also understanding how we can use our components in the fleet.”

Armstrong did his PhD in physics at the University of Technology Sydney in how to extract information from X-ray diffraction data and in particular the shape and size of nanoparticles. In order to understand these problems, it’s best to think of them as probability problems, which is how he came to be combining Bayesian theory with research into materials properties.

An employee of the Defence Science and Technology Group, Armstrong is the recipient of a Chief Defence Scientist Fellowship to further his research.

It is one of the ways Defence is helping researchers create capability for the defence force of the future through science and technology research.

Armstrong describes the fellowship as a great opportunity to collaborate more widely with external scientists and other universities. He is undertaking his research into the sustainment of aero-propulsion systems in conjunction with Associate Professor Peter Lynch at Deakin University, with the Centre of Excellence for Biosecurity Risk Analysis at the University of Melbourne and with students at the Australian Synchrotron.

The Australian Nuclear Science and Technology Organisation’s synchrotron uses electricity to produce intense beams of light more than a million times brighter than the sun, which can be used to examine the molecular and atomic details of a wide range of materials. The research at the synchrotron helps Armstrong understand the physical process and the time it takes for the damage to accumulate in the metal and the time for that damage to reach a critical state in which it is more probable to initiate a crack.

The synchrotron can mimic the types of loads that a component would experience in an aircraft engine and monitor the damage that has been accumulated by it. From there, Armstrong can build a model of the statistical distribution of the time it takes to initiate a crack.

“Once you understand that piece of information, then you can make decisions, and that’s where the Bayesian statistics and Bayesian theory comes into it,” he says.

The research will have broader application than just aircraft, as it can also apply to ships and vehicles, Armstrong says.