Mechanical engineers shift gears with AI
Research commercialisation from AUT is set to revolutionise the process of gear system design.
Despite gears being essential to everyday technologies include car transmissions, commercial airplanes and heavy machinery, gear design has hit an innovation roadblock.
Currently, the mechanical engineering industry relies on highly sophisticated software to test and refine gear system designs. However, these state-of-the-art tools share a fundamental limitation: they are evaluative, not creative. They are great at predicting the performance of a gear system that a human engineer has already dreamed up, but they can’t generate a completely new concept from scratch.
Tim Hart of Compact Motorsport, an industry leader who has spent the past decade researching and commercialising "direct gear design" methodologies, has partnered with AUT's Associate Professor in Mechanical Engineering Michael Gschwendtner and master’s student Andre Paynter to build a tool that could unlock truly novel gear systems.
Instead of an engineer manually drawing a gear and testing if it works, they simply input their performance goals into the new desktop application. Paynter’s AI framework then takes over, using generative design algorithms to explore countless geometric possibilities and discover superior gear systems that humans might never have conceptualised.
Crucially, the AI is completely unconstrained by traditional manufacturing methods, meaning it can generate highly complex, optimised 3D gear geometries ready for immediate fabrication.
“For decades, gear design has been limited by what a human can conceptualise and what traditional machines can physically cut,” says Associate Professor Michael Gschwendtner. “By leveraging AI and generative design, we are completely decoupling the design process from those historical constraints. We’re giving engineers a tool that doesn't just test their ideas but actively discovers better ones.”
To bring this ambitious software from theory to reality, the project is being backed by a $40,000 AUT Ventures Tīmatanga Initiate Grant. Designed to accelerate promising AUT research, this Innovation Fund initiative will support the project over the next year to create a validated, user-friendly prototype.
The funding will help the team navigate complex technical challenges, ensuring the AI algorithms can consistently find optimal solutions and generate designs that are actually possible to reliably manufacture.
“This project represents a critical leap forward for mechanical design,” says Professor Robert Niven, the Head of Department – Mechanical Engineering. “Associate Professor Gschwendtner and his team's research is perfectly aligned with the future of advanced manufacturing. The Tīmatanga Grant provides the essential support needed to build a working prototype, helping to translate this complex academic research and deep industry knowledge into a highly valuable commercial tool.”
The commercial potential of this AI framework is immense. Pathway to market is already being established through Compact Motorsport's manufacturing and marketing base.
By empowering engineers to create superior, highly optimised mechanical systems, the software offers a tangible competitive advantage for industries reliant on high-performance machinery. Whether it's creating lighter parts for the aerospace sector, more efficient transmissions for automotive manufacturers, or more durable components for industrial machinery, this Kiwi collaboration is preparing to shift the global manufacturing industry into a higher gear.
Useful Links
- Michael Gschwendtner academic staff profile
- Robert K Niven academic staff profile
- Bachelor of Engineering (Honours) in Mechanical Engineering
- Bachelor of Engineering Technology in Mechanical Engineering
- Compact Motorsport
AUT supports the role our academics play as critic and conscience of society, as set out by the Education and Training Act 2020. Our approach to academic freedom and freedom of expression is set out in our Charter of Academic Freedom, which was developed and endorsed by the AUT academic community.