Where to place questions, suggestions for Rigid Body Physics Sim Schemes?

We hope that USD can be used in the future to fully describe the mechanical behaviour of assets as “digital twins”.

But looking at the white paper for Rigid Body Physics Simulation schemes and NVIDIAs use of PhysX related attributes, I could not see a way to describe a product like shown in this YouTube product video.

Here’s our keyframe animated version as USDZ/GLB landing page (without bending).

  • The elements interlock like screws but still the moving joints have a range of 3D rotation.

  • The product has an expander rope in its core, fixed on both ends that force all elements in its collapsed / rotated state, while it still can bend in all directions like a spine.

  • In addition, the material itself is injection moulded plastic and therefore flexible.

  • And how to handle/simulate the cable movements (orange in the animated USDZ)?

My questions:

  • Would the Rigid Body Physics schemes in the current white paper be sufficient to describe this ?

  • If not, who to talk to, to make this possible? The working group for this topic seems to be a private/closed group.

  • The company (igus) would be happy to provide all CAD geometry and related information. This seems to be their most tricky energy chain asset to be described as digital twin to be used in next gen factory planning software. There are thousands of other digital twins to be described.

Hi Thomas,
The existing support for physics does not cover deformable bodies yet. We are currently working on a collection of schemas to support simulation of tetrahedral meshes, triangle surface meshes and curves. To be clear though, that doesn’t mean the simulator is provided by Usd, just a way to represent the data. PhysX in Omniverse might be one of the adopters of these schemas once they have been gone through the review process and are part of a release. If the simulator can handle contact coupling between FEM deformable meshes for the cables and rigid bodies for the housing, then I think simulating something like this would be feasible. An alternative and more adhoc way now would be to segment the cables into many rigid bodies and have constraints between them. Depending on the discretization and the simulator used, you might be able to achieve pleasing results for a visualization purposes.