Good morning, Thank you very much for your replies and the examples you gave me. I had to reinstall the entire simulator because some modules (PardisoMKL) were missing. Now I am able to run FEA/ANCF/Reissner models on tire test rigs.
However, the simulation is extremely slow, even when using all my CPU cores. I was wondering if there are any plans to implement FEA computation on NVIDIA GPUs in the future. For now, I will stick with rigid tires, since the simulation time is close to real time and works well when the terrain is softer than the tire. I noticed that several rigid tires are available, such as HMMWV_RigidMeshTire_CoarseClosed, HMMWV_RigidMeshTire_Coarse, HMMWV_RigidMeshTire, and HMMWV_RigidTire. Could you please explain the differences between them? Thank you for your help! 😊 Le mercredi 6 août 2025 à 13:02:40 UTC+2, Radu Serban a écrit : > Some clarifications and comments: > > > > You are correct. As indicated in the documentation, handling tire models > (aka force-element) tire models cannot be used with deformable terrain. > > For that, you need a tire model that carries collision shapes. That can be > an FEA-based tire model or, as already pointed out, a rigid tire model. > The latter is appropriate in soft soils (when the soil strength is below > the tire inflation pressure). > > > > So, you should first decide if you indeed need flexible tires or if a > rigid tire model will be good enough. Not that, should you use an FEA tire > model, you also need to come up with appropriate parameters for the tire > geometry and materials, data that is not easily obtainable. > > > > If you stick with FEA (ANCF) tires, whether you use the SCM terrain (as > you first intended) or the CRM terrain (as Dan suggested), the bottleneck > will be the FEA calculation. The computationally cost will be that much > higher if you simulate all tires on a single node. I do not see any crashes > with demo_VEH_SCMTerrain_WheeledVehicle if I switch the tire model to an > FEA-based one, but simulation is indeed very slow. Note that the code that > Ganesh shared with you also uses a single CPU node to simulate all 4 tires. > That simulation ran about 500 times slower than real time (on a relatively > potent CPU). > > > > The solution to this is to use the vehicle co-simulation framework > <https://github.com/projectchrono/chrono/tree/main/src/chrono_vehicle/cosim> > which would allow you to run individual tires on their own node. Of course, > to really benefit from that, you now need access to a cluster. Note that I > am aware of an issue with the co-simulation code when using FEA tires and > CRM terrain which I plan on looking at soon. I don’t know if that same > issue shows up when using SCM terrain and FEA tires. > > > > We’ll also make some changes to the HHT integrator that will help with > performance of integrating large FEA problems. That should also come in > Chrono in the near future. > > > > --Radu > > > > > > > > *From:* [email protected] <[email protected]> *On > Behalf Of *juju > > > *Sent:* Friday, August 1, 2025 3:05 PM > *To:* ProjectChrono <[email protected]> > > *Subject:* Re: [chrono] HMMWV with ANCF tire > > > > Hello, > > I would like to study the interactions between a tire and various terrain > obstacles while the vehicle is in motion. I saw that several tire models > are available, such as TMeasy, Pacejka02, ANCF, and Reissner. > > According to the documentation, force-based tire models (e.g., Pac02, > TMeasy) cannot be used with soft obstacles or deformable terrain (such as > sand, snow, or grass). On the other hand, the ANCF tire model is > specifically designed to accurately capture soil–tire interactions. > > Therefore, for my use case, the ANCF tire model seems to be the most > appropriate choice. > > I tried using it with the demo script > demo_VEH_SCMTerrain_WheeledVehicle.cpp, but after just a few milliseconds > of simulation, the tire explodes, and the simulation is extremely slow > (even with SetNumThreads set to the maximum number of available threads). > > I’m wondering if anyone has successfully run the ANCF tire model with the > HMMWV (or other vehicles) in a simulation that utilizes hardware resources > efficiently (CPU, GPU, etc.). > > Thank you for your help! 😊 > > > > Le jeudi 31 juillet 2025 à 19:43:55 UTC+2, Dan Negrut a écrit : > > Hi there - can you please describe what you try to accomplish? > > We might be able to help with the ANCF tire model, but before that, please > tell us what you want to accomplish since there might be easier paths > forward. ANCF is heavy duty. > > Dan > > --------------------------------------------- > > Bernard A. and Frances M. Weideman Professor > > NVIDIA CUDA Fellow > > Department of Mechanical Engineering > > Department of Computer Science > > University of Wisconsin - Madison > > 4150ME, 1513 University Avenue > > Madison, WI 53706-1572 > > 608 772 0914 <(608)%20772-0914> > > http://sbel.wisc.edu/ > > http://projectchrono.org/ > <https://urldefense.com/v3/__http:/projectchrono.org/__;!!Mak6IKo!MOF05q2wXkkfBDHMnzYQlr8F1XTyRnV9WN2zouxf4gzBc2kCjT162b9jv-2okTnEQyt42oYsuhHiLTd7$> > > > --------------------------------------------- > > > > *From:* [email protected] <[email protected]> *On > Behalf Of *juju > *Sent:* Thursday, July 31, 2025 11:41 AM > *To:* ProjectChrono <[email protected]> > *Subject:* [chrono] HMMWV with ANCF tire > > > > Good morning, > > I am a PhD student working on off-road vehicle navigation, and I found > ProjectChrono very interesting for testing my algorithms. > > I would like to use the HMMWV vehicle with ANCF tires for more realistic > behavior. However, when I run the demo script > demo_VEH_SCMTerrain_WheeledVehicle.cpp with TireType::FEA, the simulation > becomes unstable and extremely slow. It runs fine with other tire types. > > Does anyone have a demo script using HMMWV with ANCF tires that is both > stable and performs well (accelerated and reliable)? > > Thank you very much! > > -- > You received this message because you are subscribed to the Google Groups > "ProjectChrono" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion visit > https://groups.google.com/d/msgid/projectchrono/18ae5673-3e2f-40fe-80a4-ceac72d16a63n%40googlegroups.com > > <https://urldefense.com/v3/__https:/groups.google.com/d/msgid/projectchrono/18ae5673-3e2f-40fe-80a4-ceac72d16a63n*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!Mak6IKo!JNvqHAR3wRVRusJdp9o7q-21GOJFi79d_DrmD3hkYcRW_7Zt7MXQJVR3DC8iuN4-hzbB98xTJXyflWry$> > . > > -- > You received this message because you are subscribed to the Google Groups > "ProjectChrono" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > > To view this discussion visit > https://groups.google.com/d/msgid/projectchrono/c1db45a5-6d82-49bd-b5b1-d771300481d6n%40googlegroups.com > > <https://urldefense.com/v3/__https:/groups.google.com/d/msgid/projectchrono/c1db45a5-6d82-49bd-b5b1-d771300481d6n*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!Mak6IKo!MOF05q2wXkkfBDHMnzYQlr8F1XTyRnV9WN2zouxf4gzBc2kCjT162b9jv-2okTnEQyt42oYsuvxFJY84$> > . > -- You received this message because you are subscribed to the Google Groups "ProjectChrono" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/projectchrono/26a2f350-f711-4dc6-8e04-096006f1a040n%40googlegroups.com.
