DARPA Lattice

Design in the Future: Siemens works with DARPA on revolutionary technology

By: Mark Burhop, Principal Investigator

When people ask me what I do at work, the first thing that becomes clear is just how big Siemens is as a company. My friends often think that I work for Siemens Digital Industry Software (formerly Siemens PLM) because I talk a lot about product design. Or they think I work for one of the Siemens Energy companies because of the work I do with additive manufacturing (3D printing).

 

The truth is that I work for Siemens Corporate Technology, essentially the research arm of Siemens. Along with supporting all the Siemens business units with relevant research, we also work closely with universities and the U.S. government. Right now, nothing excites me more than a project called “DARPA TRADES.

 

As you can guess from the name, this is a DARPA-sponsored project called “Transformative Design,” one that, to quote the official language, “aims to advance the foundational mathematics and computational tools required to generate and better manage the enormous complexity of design.” Put simply, DARPA TRADES intends to find revolutionary solutions to designing and manufacturing tomorrow’s ultra-complex products. We have been working with our project partners—Georgia Tech, Michigan State University, and PARC—for three years and have derived some amazing results.

 

One of the first accomplishments was the ability to represent huge amounts of geometry in the design tool. For you CAD designers, go to your system right now and see how many objects (spheres, cylinders, cones, etc.) you can create in your design before it gets too big to be usable. Can you do 100,000? 1,000,000? 10,000,000? We got to 1015 (100,000,000,000,000) with no loss in accuracy.

 

Why is this important? For some time, additive-manufacturing companies have liked to say that customers get “complexity for free” because their hugely flexible machines can create complex shapes (almost) as easily as simple shapes. Designs that were previously not possible with machining, injection molding, sheet-metal bending and other methods have now become physically possible. However, representing such complexity has been a massive challenge for design software because of the incredible amount of geometry involved. DARPA TRADES recognizes that advancing the state of the art is crucial for future manufacturing that is expected to be far more flexible than we have today. 

Addressing this representation problem was just the first step in the research. How do you analyze and optimize such a huge amount of geometry resulting from these new methods? You certainly can’t do a simple finite element mesh. It is just too big.  After analyzing the design, how do you optimize it? It won’t be just three or four design variables that you want to change but maybe hundreds or millions.

 

Despite the marketing around additive manufacturing (AM), any regular user of AM equipment that is printing complex geometry, such as a lattice, knows that the process is hardly “push button.” Design for Additive Manufacturing (DfAM) plays a very important role here. This does not just mean considering things like supports and warping of the AM part in the design. It also means being able to explore the huge design space that is now possible.

 

For Siemens and our customers, the potential of this project means better products. New designs can be more complex but with more and better performance characteristics and reliability. This project can help expand what designers can do without making their job more complicated. That’s why Corporate Technology got involved in the first place and why we see DARPA TRADES as crucial to the evolution of manufacturing. 

So, like I said, this is exciting work. We have other fascinating things going on with both this DARPA project as well as work with other partners. America Makes, MXD, and ARM come to mind.

 

I’d be glad to share a few of our DARPA TRADES models that you can print yourself.