July 2011 - Issue #13
Plane safety climbs with smart inspection system
Story by Alexandra Roginski
View articles in related topics: Engineering, Industry Collaboration
Key points
- Swinburne engineers have developed an inspection system based on artificial intelligence
to detect and characterise internal flaws in composite materials - The technology has the potential to increase aeronautical safety and speed up component safety checks
Aircraft made mostly from composite materials such as carbon-fibre-reinforced polymers are already on the drawing boards of major aeronautical manufacturers, which seek lighter planes able to carry more passengers, cargo and fuel.
While these ultralight materials are currently available, their widespread use is problematic because the process of scanning for potential flaws is expensive, time-consuming and gives less confidence than similar processes used for checking and certifying metals.
It’s a challenge that researchers from Swinburne University of Technology have been asked to help solve by developing an automated approach to processing data from scans of composite materials. The goal is a process based on artificial intelligence (AI) technology that would enable analysis to be carried out with much greater speed and accuracy than a human technician could achieve.
The project is being led by the Defence Materials Technology Centre (DMTC), in collaboration with Swinburne, industry partner GKN Aerospace, and technical collaborator the Defence Science Technology Organisation (DSTO).
Dr Mark Hodge, CEO of the DMTC (which is based at Swinburne’s Hawthorn campus), explains the limiting human factors involved when checking composite panels. “There is a lot of pressure on the technicians who analyse the scans of composite materials for certification. Getting it wrong could cost lives and a lot of money.
The risk of those consequences means there is a tendency for the technician to be conservative and not certify parts that have any potentially threatening flaw,” Dr Hodge says.
Detecting hidden flaws and defects
The challenge is such: drop a spanner on a sheet of metal and a mark will be visible to indicate potential damage; drop a spanner on carbon fibre and there probably won’t be a mark, even though a structural flaw may now run through the sheet.
Defects can be introduced into a composite material during manufacture or while the plane is in service. The difficulty in detecting these is one reason why composite parts are currently used only on non-load-bearing aircraft parts such as aerolons (the flaps that descend from the wing to control side-to-side movement), and even then only after a rigorous and time-consuming certification process. If more composite parts were used, these processes would result in production bottlenecks.
Tim Barry, a research engineer at the Industrial Research Institute Swinburne (IRIS), is developing the AI technology for the project. He explains that composite materials have multiple layers and during inspection their ultrasonic signals interfere and interact with each other. “You have to identify any unusual signals among all that activity, and then determine why they are unusual,” he says.
Mr Barry, who is part of the Robotics and Non-Contact Inspection research team led by Professor Romesh Nagarajah, says interiors of composite panels are examined using non-destructive inspection (NDI) technologies, such as ultrasound.
The panel is scanned with an ultrasonic probe attached to a robot that then sends the thousands of signals to a computer for storage – raw data telling the story of the panel’s internal structure. This data is displayed as squiggles akin to those produced by an electrocardiogram.
A technician studies these signals, looking for clues to flaws – too many air bubbles, a foreign object, or delamination (separation of layers). Only a handful of people in Australia are experienced enough to read the data at this level and to sign off on component safety. The issue from a commercial perspective is that it can take a day to scan and assess one wing panel of just two square metres. And interpretation is subjective.
This is why industry needs an automated, objective process that is cost-effective and trusted by end users.
Professor Romesh Nagarajah’s research group has over two decades’ experience in developing intelligent sensing techniques to check component integrity for clients from the automotive, aerospace, defence and food industries, with the Ford Motor Company a prominent example. The research team’s work over the years has led to the development of a number of AI tools, some of which were used in this project.
“When you use sensor technology in an industrial environment, the signals you get are noisy, making interpretation difficult,” Professor Nagarajah explains. “AI mimics human intelligence and allows you to examine a sensor signal and draw out valuable information. The signal information can then lead to identification of any defects existing in the component.”
Boost in details and inspection speed
Mr Barry spent two years writing more than 16,000 lines of code that are the heart of the intelligent inspection system. He applied the resulting software to ultrasonic signal data supplied by the NDI division at DSTO. The software’s algorithms processed the signals at eight times the rate of a human inspector.
Human inspection catalogues only two properties: irregular echoes in the signal and where those echoes occur. The automated method is able to go into much more detail, much faster.
“We start by calculating 91 different properties that describe key features of the ultrasonic signal. We have this larger feature set to not only say that we have found something, but to describe what we’ve seen, what we think it is, and what our confidence of this decision is,” says Mr Barry, who in February was honoured as an Outstanding Young Researcher at the DMTC’s annual technical conference.
There are many other benefits to this AI-based inspection system. Without any specialist support, it can be instructed to look for new types of defects. It also stores the data from each scan by date, a necessary tool when adhering to the aerospace industry’s long audit trails.
Because of the stringent quality assurance requirements of aeronautical engineering, a technician will still be required to sign off on the inspection results provided by the ‘smart’ inspection system. And although the project is still in development, the technology passed a major milestone recently when a top-level technician who carries out contract work in Melbourne trialled the software and affirmed its value for industry.
Now that this stage of development is complete, the next step is for software coding engineers to refine the system for commercial use.
Professor Nagarajah and Mr Barry say there is also research potential in scanning the same panels over time to see how the composite degrades, providing crucial information about how an aircraft holds up with use.
Competitive edge
For industry collaborator GKN Aerospace – a multinational manufacturer of composites and provider of engineering services – the revolution in engineering capability promised by the technology is seen as providing a competitive edge with which to secure major contracts.
Adam Groszek, head of technical development with GKN Aerospace, says that when considering market competitiveness, the key question came down to efficiency: “The conclusion was that we would need smarter engineering tools to help automate our work.”
An AI system for decoding scan data also opens up cost-saving possibilities. A technician based in Australia could review scans from anywhere in the world, making GKN an attractive candidate for maintaining aircraft such as the F–35 Lightning II, a model being developed as part of the Joint Strike Fighter (JSF) program funded by the US, the UK and other partners.
“We are still talking to Lockheed Martin about the JSF sustainment program. This is one of our business goals, to be part of the support team for JSF operations. We want to increase our capabilities. It’s something that will differentiate us from other companies,” Mr Groszek says.
More opportunities, fewer human factors
Although it has been adapted to crunch ultrasonic signals, this AI-based inspection system can be used to process signals from any inspection system, including those using computer-vision, capacitive-sensing and light-scattering technologies.
The resulting decrease in time and costs compared with current certification methods may also open up opportunities to manufacture a greater number of aircraft components out of composite materials, says Dr Viktor Verijenko, operations and education manager at the DMTC, and project leader for the technology’s development program.
“The large-scale production of composite parts is a driver for this technology. Currently some aircraft parts can be uneconomical to manufacture with composites because they require the same amount of time to evaluate as large structural components. This new automated analysis technique that we are developing will enable the analysis of test data to occur very quickly, opening up the opportunity to process hundreds of components a day rather than just a few,” Dr Verijenko says.
