Engineers and designers for companies like General Electric, Airbus and General Motors are using a new tool called generative design to explore thousands of design iterations with geometries that are automatically generated within their software.
Although generative design is most applicable to additive manufacturing processes (aka 3D printing), developers are quickly integrating the technique traditional manufacturing processes as well. In April 2018, Autodesk commercially introduced its generative design tool on the Autodesk Fusion 360 platform. This example on their websites shows how they optimized a skateboard truck design with this technique.
In contrast to other automated manufacturing techniques, generative design requires no preliminary CAD instructions.
Engineers input a set of design parameters – special constraints, loading, materials and manufacturing methods. Then, using cloud-based software, hundreds or even thousands of possible designs that meet these criteria are created. Many of the geometries are other-worldly looking with swooping curves and intricate lattices that human designers are not accustomed to thinking up and are very difficult to model parametrically.
Generative design techniques are still in the early stages of development and are not yet practical for large scale production. However, there have been some notable success stories. The Adidas Futurecraft 4D high-performance sneakers, for example, were generatively designed and are set to go into production this year. Meanwhile, Airbus has been exploring generative design techniques to create lightweight interior panels for its planes as well.
The current potential of generative design is still somewhat restricted by hardware limitations, but this novel approach is sure to become more useful and applicable to real-world design challenges with time. In essence, generative design mimics natural evolutionary processes to find optimal solutions through rapid-fire iteration. If and when engineers are able to overcome generative design’s current limitations, this technique could have a transformative effect on the structure and appearance of tomorrow’s industrial design.