Dassault Systèmes recently announced the acquisition of FE-DESIGN Group, makers of TOSCA for topology optimization. DS’ SIMULIA brand had already released a TOSCA-based Abaqus product, Abaqus ATOM, and the relationship between SIMULIA and FE-DESIGN was good, so why the acquisition? Why now? What can DS/SIMULIA do that FE-Design on its own couldn’t do? Why isn’t topology optimization used more often?
I spoke with SIMULIA’s Steve Levine and Matt Ladzinski about these and other topics and came away with one very clear impression: SIMULIA believes in topology optimization and that it should be employed far earlier in the design process, by designers, not analysts. DS plans to use its considerable reach to expand the user footprint of TOSCA and push its use into the design workflow, rather than just targeting analysts.
Topology optimization is incredibly cool, a gentle nibbling away of excess material until a design concept meets the design specifications. Here’s a video of Abaqus ATOM in action that shows how the designer sets up optimization of the feet of a clothes washer to use less material, yet meet the vibration and weight loads of a full, rotating washer drum. The optimization yields a mathematically correct result but a designer needs to get in there and figure out how to manufacture or source the optimized components and whether other factors (like aesthetics) should affect the design. Optimization creates a jumping-off point, a concept, and probably not a final design.
TOSCA isn’t the only topology optimization solution on the market. Altair Optistruct, FEMtools Optimization, INTES Permas, LSTC LS-OPT and MSC Software Nastran SOL 200 are among the commercial offerings out there, and there are many more in the open-source domain, too. (ANSYS Topology Optimization and Siemens Femap Optimization are based on FE-DESIGN’s TOSCA.) Then there are fluid flow optimizers, like TOSCA Fluid and RBF Morph. Given so many offerings, why isn’t optimization more common?
Mr. Levine believes that optimization hasn’t been more widely adopted because many simply aren’t yet aware that it’s possible — and relatively easy. It requires a bit of know-how to set up and some serious compute power to run but conference papers by FE-DESIGN customers like Audi, BMW, Daimler, General Motors and Volkswagen (and, now, DS’ marketing machine) will drive awareness. Declining CPU prices/HPC centers/cloud computing makes it more affordable to run.
That leaves the user experience and the cost of software licenses for running optimizations, both of which DS intends to address. DS believes (and I completely agree) that the whole point of optimization is to discover new design concepts early in the design process. How can you build a lighter yet stronger design if you’re simulating failure modes just before you release the design to production? TOSCA/ATOM automates early stage exploration of design concepts so that designers, working mainly in CAD, can explore alternatives before it’s too late.
To do that, SIMULIA needs to rework TOSCA’s focus a little bit, improving the user experience and making it more suitable for designers who have basic FEA experience but are not experts. This means streamlining the optimization setup and making the results easier to interpret. Mr. Levine said that ATOM is a good start at this, but that there is more work to be done. I would imagine that TOSCA could find its way into both DS designer brands, CATIA and SolidWorks, over time. Exactly when and how are TBD at this point.
SIMULIA’s global support organization has spent the last two years learning ATOM and can use that background as a springboard to a wider roll-out of TOSCA, within and alongside of the Abaqus base — Mr. Levine says that DS intends to continue support for all of FE-DESIGN’s partners. But he’s a realist and acknowledges that some of them may not want to work with SIMULIA — as the R&D teams look at ways to integrate TOSCA into 3D Experience bundles.
One of the most common objections to optimization has been its cost, since companies may only use the tools a couple of times a year, not enough to justify a large license fee. Mr. Levine says that DS is looking at ways to bundle TOSCA into its offerings to make it more attractive to occasional users.
Bottom line: We’ve got to work on cost, usability, workflow and capability before optimization becomes more widely used. With DS’ backing, and once partners have announced their intentions, TOSCA should be able to grow in a way that was impossible on its own. Does the acquisition make sense? It’s super-green!
Almost all of the above was written before DS announced its acquisition of SIMPOE for its mold analysis software. But many of the reasons that make the FE-DESIGN acquisition a good one apply to SIMPOE as well: better, earlier integration into the design workflow; the potential of more attractive pricing; improved usability. DS is clearly on a mission to build Experience offers that use simulation as a core component rather than an optional add-on.
Image courtesy of 3ds.com.