Those involved in product development understand it is a deeply iterative process driven by strict timetables, and the path is riddled with trial-and-error. Competition and profitability create extreme pressure to make the design marketable as soon as possible.
As a time-bounded exercise, it is not a matter of getting the design done but instead getting it good enough within the allotted time. When it comes time to manufacture an early revision, it’s like taking a picture of a growing child: a moment frozen in time. The development process continues even as those early revisions are being manufactured. So, when you think about it, it is yesterday’s part made today.
READ MORE: Generative Design, Artificial Intelligence and Manufacturing
If only there were more time, the design could almost always be improved. Excess material could be removed to reduce weight, increase surface area to improve cooling, or improve the flow characteristics of air or liquid. Maybe even just improving aesthetics and adding a bit of style can make a difference.
Designers are often left with mixed emotions, both proud of their work but realizing that not all their ambitions were met.
The promise of generative design is enticing. CAD development that would normally take weeks or months could be condensed into days or hours, allowing tomorrow’s part to be made today.
Generative processes perform hundreds or thousands of tasks to the CAD model automatically based on carefully crafted rules, creating a compliant, manufacturable design. The experience is the CAD designer’s equivalent of transitioning from sewing by hand to using a sewing machine. It can require a shift of mindset that instead of practicing a craft, you are now wielding a productive machine.
Tech Offers More Accessibility to CAD
Technologies like AI, cryptocurrency and video graphics have spurned large advancements in processor and graphics technology in recent years. 3D CAD software is enjoying this tailwind and generative design is more accessible than ever. Whereas the typical computer just a few years ago would slow to a crawl calculating these complex processes, modern computers can perform many of these tasks nearly instantly.
READ MORE: Protolabs Network’s Advanced Software Integration Offers Streamlined Collaboration
Increasingly, canned generative functionality is being introduced into the most common CAD software packages. These allow even novice CAD designers to implement these complex procedural processes into their design without having to understand the complex math under the hood. Voronoi patterns—a variant of tessellation—and minimal surfaces are downright trendy nowadays in 3D printing. In flexible plastics like thermoplastic polyurethane (TPU), Voronoi lattices create pliable structures similar to foam. Minimal surfaces create strong and lightweight structures but are also used in diverse applications such as fluid mixing and heat exchangers. Both make visually striking and futuristic printed parts.
On the other end of the spectrum, generative design allows pioneers to push the needle and explore what is at the edge of possibility. At the frontier is advanced topology optimization and integration with AI and machine learning. Some advanced processes are akin to how ants dig their colonies or how fungus spreads to find a food source. These designs are highly complex and unfeasible for a person to design without the aid of such automation.
Perhaps the most exciting aspect of generative design is that it allows designers to ruthlessly explore the entire solution space for an available part, sometimes having hundreds of potential design solutions to review. This sheer diversity of design would not be possible using human effort alone, and it opens interesting applications for mass customization.
The historic bias in manufacturing is for each part to be as similar as possible to the previous one, but generative design allows the opposite approach to be explored, allowing each part to be as different as possible while still being functional and meeting engineering requirements.
In the further future, generative AI may design not just individual parts but entire complex assemblies.
Not Just for Additive Manufacturing
While additive manufacturing seems to pair naturally with generative design and can most easily leverage the design freedom, generative design is increasingly being used in other manufacturing processes. Recently, NASA and Protolabs showed the world that generative design can create complex and optimized designs for 3-axis CNC machining in an afternoon (and receive the parts the next day). Other exciting avenues are origami-like approaches to folding sheet metal or complex wire bending operations.
Only time will tell if generative design makes good on its promise, but it is exciting to see the future unfold on this iterative technology.
