Artificial Intelligence in Product Design
Imagine creating boundary condition and letting machine do rest of the work. Well, this is 2022 and this is a reality now! Generative design is an AI (Artificial Intelligence) driven method where you define an engineering problem and the machine generates a wide range of solutions. You can create a design from scratch as well as optimize an existing design.
Let’s take a closer look. In the below example, I am optimizing an existing design. I wish to reduce its overall mass by taking advantage of machine’s capability. You can see that I have added few boundary conditions like which areas the machine should not change (red) and which bodies are to be preserved (green). I have specified what kind of loads are expected, the manufacturing method I am considering is Additive. This means that the machine will offer me suggestion that are compatible with Additive Manufacturing technologies such as Metal 3D printing.
What actually happens behind the scene?
The generative model is an algorithm for constructing a generator that learns the probability distribution of training data and generates new data based on learned probability distribution. In particular, variational auto encoder (VAE) and generative adversarial network (GAN) are popular generative models used in design optimization, where high-dimensional design variables are encoded in low-dimensional design space. In addition, these models are utilized in the design exploration and shape parameterization [8].
Basically, computer uses the evolutionary approach to offer suggestions to optimize the whole part, in order to achieve desired weight reduction while keeping same strength and functional end points.
Broadly speaking, generative design is more suited for application where:
a. Large quantities is not required - Non-additive fabrication methods like casting, forging, machining are still a lot quicker as of today. Producing components with a simple geometry at scale is more economical. Additive Manufacturing of complex geometry parts is time taking.
b. Mass optimization is critical - as in case of aerospace, think of saving gallons of fuels by reducing 500 grams on a flight to Mars. Aviation and automative companies have achieved reduction in mass for several components like GM did for safety belt lock mechanism in their new EV.
c. Personalization is useful - as in case of biomedical devices. Like customized tooth-braces, fabricated rib bone or replacement of amputated limbs. Machine generated design are featherlight as compared to traditional approach.
d. Mission Critical delivery - as in case of Military, you can’t have a multi-million dollar Aircraft or Tank out of action in thick of a battle thousands of miles away from your home base in want of replacing a crucial component.
In this collaborative research project, Autodesk and NASA’s Jet Propulsion Laboratory (JPL) tested the use of generative design to make a concept space lander that is 35% lighter than a traditional design. The software allowed the team to leverage three different manufacturing methods (casting, additive and subtractive manufacturing) to produce the lightest and strongest design, which will be critical for future space travel. In another example, machine can generate designs which are virtually impossible for humans to imagine. See the first image of this blog post, a rocket engine developed using GD and printed on a metal printer via additive methods.
Recently, Netherlands unveiled world’s first pedestrian+bicycle bridge in Amsterdam. Mind you, this is not just a prototype, it is an actual installation that people are using on a daily basis. Since the total quantity of raw material consumed in such generative designed and additive printed components is less, they are inherently sustainable and eco-friendly.
As I see, the possibilities are virtually limitless. Generative design has application in automative, space, aerospace, industrial machinery, medical equipments etc. However, due to current limitation of Additive Manufacturing, making generative design parts at a large scale is not economical yet. But, this will change soon!
Let me know how are you going to take benefit of this emerging technology?
Regards,
Abhinav
Reference:
1. https://redshift.autodesk.com/what-is-generative-design/
2. https://www.sculpteo.com/en/3d-learning-hub/create-3d-file/generative-design/
3. https://www.ptc.com/en/blogs/cad/beginner-guide-generative-design
4. https://all3dp.com/2/generative-design-simply-explained/
5. https://www.danieldavis.com/generative-design-doomed-to-fail/
6. https://www.hyperganic.com/magazine/works/a-new-dimension-in-space/
7. https://manufactur3dmag.com/queen-maxima-of-the-netherlands-unveiled-the-worlds-first-3d-printed-steel-footbridge-in-amsterdam/
8. https://arxiv.org/pdf/1903.01548.pdf
9. https://www.assemblymag.com/articles/94984-generative-design-software-is-transforming-the-design-process
Disclaimer: I am an Autodesk employee and my views can be biased in favour of Fusion 360 software which is a leading provider of generative design capabilities to designers around the world.
