Nia Cason
Nia Cason: Could you introduce yourself in a few words?
Daniele Grandi: I’m a design engineer in Autodesk Research’s generative design team. I’ve been there for 3 years, before which I was working for an additive manufacturing start-up as a design engineer. I’m from Italy and I’m now living in London – our team is split between the UK and the US. We’ve been researching ways to bring generative design to a wider audience, as well as to make it easier and more effective to use.
Daniele Grandi - Design engineer at Autodesk
Andreas Bastian: I’m a researcher, designer, and mechanical engineer for Autodesk at the US lab, and I study design and manufacturing methods. I have a background in engineering and studio arts – specifically, Shigaraki clayware, a very specific flavour of traditional Japanese folk pottery. (As every designer makes chairs, every potter has a fling with making tea bowls.) I studied a mix of mechanical engineering, electrical engineering, computer science, and material science – it turned out to be very good preparation for working with 3D printing.
Andreas Bastian - Researcher, designer, and mechanical engineer at Autodesk
NC: What inspired you to choose this research topic?
DG: I guess what initially drove me to mechanical engineering was the problem-solving aspect. That was true for any hobby I ever got into, like biking – I was always excited to work on my bicycle, so I guess fixing things was always something I was attracted to. I started looking at software optimisation and joined the Autodesk team, where I could be closer to the design aspect and have more access to cutting edge research.
AB: From an early age, I’ve been interested in tool making – much of my childhood was spent making a blacksmiths forge, making a foundry, and that progressed to working with milling machines. It became clear that one of the most capable tools out there was 3D printing. So, 8 years ago, I started designing and manufacturing 3D printers, and that really started my career. I even spent some time in a lab printing livers.
NC: What’s a typical day at work like in your respective US and UK offices?
AB: I’ll take an early train, grab a coffee, walk along the waterfront here in San Francisco to our office – it’s on one of the piers on the water, so it’s quite nice to watch the sun rise in the morning. The applied robotics lab is where we do all the assembly and do a lot of physical prototyping, and if we need anything we pop into our machine shop.
DG: I take the research prototypes that my team develops and apply them to customer problems. For that I’m in the office, but the other aspect of my work involves going out and researching what those customer problems are. The ‘customers’ are people and companies that manufacture anything from consumer products to cars to airplanes, and the ‘problems’ are essentially about designing better products for their customers.
NC: Do you follow any online scientific personalities?
DG: There’s many science-related YouTube channels that I find interesting- the SmarterEveryDay channel is pretty inspiring.
AB: I continue to follow some of the folks in tissue engineering, in particular Dr Jordan Miller’s lab at Rice University, and then people in the machine learning space, such as Jeremy Howard.
NC: Do you recommend any scientific events?
DG: As well as Paris Space Week, Andreas and I are doing a sort of euro-tour to another conference, the Hanover Messe Conference, as well as to see some of our customers in Germany and the UK. We also just attended the Additive Manufacturing for Space and Aerospace in the UK.
AB: And as a researcher in industry, the AMUG conference and Laser Additive Manufacturing Conference are both great venues for new work in additive manufacturing. While I’ve never been, I enjoy the proceedings of the Solid Freeform Fabrication Symposium.
NC: Paris Space Week is approaching! Could you tell me about your collaboration with the NASA Jet Propulsion Lab?
AB: The Jet Propulsion Lab (JPL) is a NASA facility operated by Caltech – JPL is tasked specifically to the development of robotic space exploration equipment. They’re the guys that make the exploration space craft, mars rovers, a lot of those spacecraft that are way out in the solar system. We started exploring and structuring a research problem with them, and what features of a design problem would be interesting to tackle. About 18 months ago, we started working on the concept for interplanetary landing craft – so landers for the icy moons of planets like Jupiter and Saturn.
DG: The collaboration was a research collaboration, so we’re still far from putting our parts into space. It was more for us to understand the problems that the JPL might have. We were learning about how generative design can be applied to these structures, so it was always a concept –it was never meant to fly.
AB: That’s how our team within Autodesk works; we partner with a company or institution to work on a hard problem, and basically further develop the tools that would be necessary to solve that problem. For example, all the technology we used to make this lander didn’t necessary exist when we started the collaboration.
NC: Tell me a little more about generative design.
DG: Generative design is a design method that where the designer leverages optimisation methods that computers can run a lot faster than humans in order to explore the design space more effectively, and so in other words – it’s a design exploration method. It allows users to more quickly explore different design options and optimise them for a particular design issue. We’ve been applying this to the manufacturing and design of mechanical structures at Autodesk for several industries, including the automotive and aerospace industries.
AB: There’s also a video to explain more here, featuring Daniele himself. Space craft are some of the highest-performing structures that we’ve ever developed. This was a really interesting application of our technology, as it can manage tremendous amounts of complexity in design problems while also producing very high performance structures.
NC: What are your specific roles in the Autodesk-JPL collaboration?
DG: My role was to lead the design efforts and showcase to NASA’s JPL what our research software can do. And as far as what we’ll be talking about at the Paris Space Week conference, we’ll tell the collaboration story, go into what we learned, and what was surprising.
AB: I think we’ll talk about how spacecraft are traditionally designed and manufactured, and then paint a picture of this collaboration, which has illustrated how spacecraft can be designed and manufactured in a very different way that could be faster and much more agile than traditional methods.
DG: We’ve had some pretty intense sprints during this project as we’re in two different time zones. There was a couple of weeks where I’d work during my day and hand it off at the 12-hour mark to Andreas. So the work would essentially keep going for 24 hours. In theory, my focus is more about the design and simulation aspect, whereas Andreas is more focussed on the design and manufacturing aspect, but the line gets a little bit blurred when you start leveraging generative design. Because of that, we were working very much in parallel rather than using a typical design workflow.
AB: Yes, Dani and I clearly overlap on the design side. We’re kind of like two lobes of the same brain operating on the problem, in a continuous 24-hour cycle. Part of the generative design workflow is you’re creating a description of the problem rather than of the solution. Then you’re using a bunch of smart computers to propose different solutions to the problem. A little like if you’re taking a sculpture class, you’re learning about positive and negative space and inverting the way you think about it. As a result, instead of designing the end piece, you’re designing the process.
NC: So see you in Paris?
DG: I’ve only been to Paris for work before so this time I might have a free weekend, so I’m looking forward to exploring the city. There’s some really big structures that I look forward to exploring.
AB: There’s one in particular with a nice hierarchical truss structure.
