Cycling and academia team up to explore thousands of design solutions for swingarm
University of Bristol engineering student combines his love of mountain biking with a curiosity about generative design to offer A.I.-powered solutions to a cycling company in the UK.
The story begins with a fortuitous meeting in a garden shed over a cup of tea. Engineering student Ben Boxer contacted Joe McEwan—an accomplished engineer who was in the process of turning his hobby of designing bicycle frames into a new career path—to discuss engineering and design in the mountain bike industry. It happened that McEwan needed help transferring 2D drawings of his bicycles into 3D CAD models for his swingarm design, and Boxer was more than willing to volunteer for the job.
The shed where they met was the workshop that would become custom mountain bike frame manufacturer Starling Cycles. Operating out of Bristol in the UK, Starling Cycles brings together passion for mountain biking and expertise in engineering to create one-of-a-kind, durable bikes rooted in simplicity.
Passion leads to opportunity
“This was an exciting opportunity for me as it helped develop my CAD skills alongside my studies, while getting my foot in the door of an industry in which I hope to pursue my career,” Boxer says.
Boxer, a bright and enthusiastic student who was coincidentally looking to select an area of research for his thesis, was a perfect candidate to help focus on manufacturing methods for Starling’s swingarm parts. He had been using Autodesk® Fusion 360® software and was keen to experiment more with its generative design tools. Boxer had also been exposed to generative design through his engineering program and had been encouraged to further develop these skills by his professor and the Director of Strategic Partnerships at the University of Bristol, Professor Ben Hicks.
This collaboration demonstrates the massive synergies that exist between student, university, and industry. In this case, the result was a swingarm design optimized through part consolidation, time savings, and the freedom to choose which manufacturing option, out of several, worked best. Professor Hicks’ foresight into the value of developing generative design skills in his students paired with Boxer’s knowledge and work ethic led to untapped opportunities for Starling Cycles.
Complex iterations, AI, and cloud-powered solutions with generative design
Joe McEwan had initially considered using topology optimization to modify and improve his swingarm design. However, Boxer convinced him that generative design was a more sophisticated design solution that would ultimately produce better results. With topology optimization, a single part, single material, and single manufacturing method are optimized. But generative design offered the freedom and depth of design to explore multiple solutions with multiple materials and manufacturing methods in the same time it would take to explore one method with topology optimization.
According to Boxer, a central challenge with generative design is that it requires a more thorough understanding of loading requirements than traditional methodologies. Accuracy became a critical component in the initial stages of software inputs. Boxer tackled this challenge by using a combination of methods, including free-body analysis, finite element analysis, and testing to determine requirements for stiffness, all aimed at improving manufacturability, reducing part count, and maintaining (or improving) performance. Using generative design, he could simultaneously generate multiple CAD-ready options based on multiple specified manufacturing constraints and product performance requirements. The outcome? Part count was reduced by eight and total welded length by about 50%. Boxer loved the time he saved due to the built-in simulation, and McEwan was impressed with the reduced waste and more sustainable product. Overall, the time savings enabled more iteration - always a good thing when it comes to design optimization.
“Efficient design conceptualization is one of the main advantages of generative design that I have identified through this project” says Boxer. “The use of the unrestricted algorithm to explore the ideal, most structurally effective solution to a problem for each material being investigated is really useful— even if you don’t necessarily have additive manufacturing in mind. The organic forms produced are unconventional and beautiful and would be very difficult to design without such technology.”
Student-led, professor-supported innovation, and career development
Professor Hicks, who was Boxer’s academic supervisor throughout the project, believes in the power of the student-employer relationship. The engineering program at the University of Bristol offers students the opportunity to be matched with a summer internship to undertake a research project and gain valuable, on-the-job skills. And that’s not all. Professor Hicks and his colleagues in the engineering department are working closely with Autodesk to co-design and co-develop CAD curriculum integrated throughout the engineering programs, including generative design training.
“At Bristol, we believe that provision of industrially relevant projects is essential for students to apply and develop their transferable skills,” says Professors Hicks.
Boxer, Hicks, and McEwan are striking examples of the mutual benefit that exists when industry and academia collaborate. As the world prepares for a growing population and the era of AI-driven automation, educators know industry is looking to hire students with a new set of skills to assimilate to this future of work transformation. Collaboration has a greater collective impact while meeting individual goals — a win-win(-win) for this trio of innovators.
Kristen Pearce contributed to this article.