Engineering, Rail Supply, Research & Development, Rolling stock & Rail Vehicle Design

Harnessing processing power to develop smarter train assembly

Terence Vu sees a future where robots can work out the best way to assemble rollingstock components. The PhD student spoke with Rail Express about his project.

With research co-funded by the University of Wollongong and the Rail Manufacturing CRC, Terence Vu is working to advance the field of automated assembly for rollingstock fabrication.

Put simply, Vu wants machines to determine the best way to assemble components by learning and adapting to new processes on the fly.

“Traditionally in manufacturing, people program robots the step-by-step process to manufacture a part,” Vu explains. “But any time they have a new component, they have to program that additional component again.”

In Vu’s version of the ‘factory of the future’, that engineer would simply provide inputs and a desired result – a 3D model – and the robot would establish the steps required to get there.

“My PhD project is aiming to develop a certain brain for the robot to ‘think for itself ’, where it can determine how to assemble a product from many individual parts,” he says.

“When the parts are brought together, we want the robot to consider lots of constraints from the assembly configurations. We also want other requirements such as the assembly stability and the part’s tolerance to be considered, before making a decision regarding the sequence, and the path to bring those parts together into the final product.”

For every new part or assembly, the machine would be able to adapt and adjust to ensure its process was the most efficient possible, given the constraints.

Vu’s PhD research combines physics, mathematics, engineering and artificial intelligence. Given the large number of potential assembly routes to take, running software simulations helps to select the quickest, safest and most tolerant route when assembling components. The algorithms Vu builds currently work to assemble virtual components, but in the future may be trialled and adapted to be a real-world rail scenario.

Vu says this kind of capability is thanks to recent developments in computing power, and research, specifically within the field of artificial intelligence.

“In recent years, people started to think of many ways to present the different constraints and data efficiently, so the robot can make use of that information,” he says. “And certain algorithms have made the process faster as well.”

With an undergraduate degree in Mechanical Engineering and a Masters in Design Engineering, Vu gained industry experience working in robotics at the Rolls-Royce@NTU Corporate Lab – a collaboration between Rolls-Royce and Singapore’s Nanyang Technological University, before relocating to Australia to pursue his PhD studies and continue progressing his passion in robotics.

Vu believes technology like this has a future not only in manufacturing, but in the maintenance and repair side of the rail industry. He says more students should be encouraged to enter the rail sector.

“I think the rail industry is very lively and very colourful,” he says. “It serves and employs people from all walks of life, and I believe that people of any disciplines can participate and contribute to this industry.”

The Rail Manufacturing CRC (Cooperative Research Centre) sponsors and directs collaborative research and commercialisation partnerships in rail manufacturing. It pairs researchers with commercial businesses, with the aim of furthering the rail manufacturing sector within Australia.

The Rail Manufacturing CRC also supports the next generation of rail leaders, with more than 30 students supported across industry projects, scholarships and internships run by the Centre.

For more information, visit: www.rmcrc.com.au