Generative Design rewrites the engineering process

Computer Aided Design (CAD) software replaced the drafting table and made engineering designs digital, but for years, CAD has been stuck in a rut. There hasn’t been much automation beyond streamlining the drafting-modeling experience. The workflow is a back-and-forth experience, proposing ideas, testing them with simulation and analysis software, and then tweaking the original proposal. The process repeats as needed until a successful product design is considered ready.

Now, a new technology is reimagining the product design workflow. Generative Design is a software technology that brings new automation to CAD by exploring thousands of design options. The product designer provides several goals including shape, purpose, and other required engineering features. The software then uses a combination of intelligent algorithms and brute force calculation to create and compare hundreds or hundreds of thousands of design alternatives. The best ideas are presented to the designer.

Generative Design began in architectural design in the late 1990’s as a way to explore specific shapes for large signature projects, but it was limited to evaluating geometry. The software has since matured significantly, to the point that it can be used for more complicated processes.

How important is Generative Design to the future of product design? All the leading CAD software companies are either buying startups working on Generative Design software or developing their own technology. Leading CAD vendor PTC recently paid $75 million to acquire Frustum, one of the new small companies bringing next-generation Generative Design to product design.

Generative Design is transformative — perhaps even disruptive — because it changes the nature of the design process. Until now design has been a deterministic process where one specific idea at a time was explored, expanded upon, and submitted to engineering review. Generative Design means the software can examine a vast variety of alternatives, then give the designer the best choices that match stated goals and parameters. CAD moves from being an electronic drawing board to a co-designer, giving engineers and designers new freedom to explore alternatives. This becomes even more important with the rise of 3D printing and its ability to create “unbuildable” parts. Generative Design can provide novel solutions to design problems that break existing paradigms by helping engineers think outside the box.

Finite Element Analysis (FEA) is physics software engineering. It has always been used in product development separate from initial design. Generative Design researchers are using Artificial Intelligence (AI), FEA and optimization algorithms to “train” software to use physics as part of design exploration. 

“In mechanical design, we have the privilege of having very accurate physical models described by differential equations which are solved by FEA,” notes Dr. Michael Bogomolny, Co-Founder and CTO of ParaMatters, a leading Generative Design software developer. The data from this rich heritage can now be the source for training design software. “The Generative Design which we do at ParaMatters is a combination of FEA and optimization algorithms which converge to extremely effective designs, fully automated and valid. CAD was tailored for traditional manufacturing, hence the ability to create very complex geometries is limited. ParaMatters Generative Design is “free-form,” meaning we are not bound to a specific way of modeling, so our algorithms are able to generate extremely complex and physically efficient designs.”

Movies about Artificial Intelligence show computers replacing humans. By contrast, Generative Design is an example of AI assisting the human by extending and enhancing work. As Generative Design matures, it will move from geometry and physics into other design criteria including aesthetics, manufacturing processes, and materials selection. Such new features won’t replace the designers, but will instead enhance and augment existing work processes.

« »