Wheeled vehicle simulation on deformable terrain?
hj
Radu Serban
Yes, of course you can use wheeled vehicles on deformable terrain.
There are demos in the Chrono distributions that illustrate that. Which particular deformable terrain formulation are you interested in?
Finally, would you please let us know where in the tutorials is mentioned that only tracked vehicles work with deformable terrain?!
--Radu
--
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
projectchron...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/projectchrono/67c948f8-4d62-43cb-8b2d-bf4e59d0347fn%40googlegroups.com.
Radu Serban
HJ,
First, please keep these conversations on the Chrono user mailing list.
Well, the statement you quoted is pretty clear. It doesn’t say that a wheeled vehicle cannot be used with deformable terrain, but just that certain tire models are not compatible with a deformable terrain. These tire models (Pacejka, TMeasy, Fiala) were developed and meant for rigid terrain; and in fact, pretty smooth rigid terrain at that. Note that we didn’t come up with these tire models, but just implemented them in Chrono. They are standard tire models that are widely used for *on-road* vehicle simulations.
As for which Chrono deformable terrain formulation to use, I would not recommend FEA. That is not what we commonly use, it’s a pretty inefficient way of modeling soil deformation, and is quite limiting. I would certainly start with SCM because it’s simpler and fast. If you want something higher-fidelity, I suggest looking at the continuum representation of deformable terrain provided through Chrono::FSI (although that will require an NVIDIA GPU).
As I indicated on the tutorial page you linked to, for now you will have to stick with a rigid tire or an FEA tire. Again, for simplicity and efficiency, I recommend using a rigid tire for now.
--Radu
From: hj <mhadi...@gmail.com>
Sent: Friday, 27 January 2023 14:57
To: Radu Serban <ser...@wisc.edu>
Subject: Private message regarding: [chrono] Wheeled vehicle simulation on deformable terrain?
Thank you vey much for clarifying Radu. Yes I actually saw the tutorials but the part that confused me was from https://api.projectchrono.org/vehicle_terrain.html that is says :
"Note however that these quantities are relevant only for the interaction with the so-called semi-empirical tire models. As such, they are not used for the case of deformable terrain (SCM, granular, or FEA-based) which can only work in conjunction with rigid or FEA tire models and with tracked vehicles (as they rely on the underlying Chrono collision and contact system)."
So I just wanted t omake sure. I think I'll go with FEA or SCM terrain.
Thank you
hj
hj
Dan Negrut
Hi hj,
You want to use SCM as much as possible. We used it in the past for mud and almost rigid terrain (small sinkage). It’s pretty much an empirical approach. In general, SCM can run at Real Time Factor close to 1 (that is, real time).
We do not have a FEA model – the Chrono continuum terrain model is solved with SPH and it’s called Continuous Representation Model (CRM). CRM is your next fastest choice, RTF is anywhere from 50 to 200.
If you go to DEM, the RTF probably is going to reach into 1000.
So, if you want to generate training data and need to get tons of data, try, in this order, SCM, CRM, DEM.
See here, and here, for recent papers on this topic. For DEM, we’ll have a paper out soon, but for an older version, see here.
Good luck.
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
---------------------------------------------
To view this discussion on the web visit https://groups.google.com/d/msgid/projectchrono/48ed9365-641b-4a3a-918d-7e54721ae2can%40googlegroups.com.
hj
Radu Serban
HJ,
Keep in mind that SCM is just a framework that generalizes the semi-empirical Bekker theory to arbitrary terrain topography and impactor geometry. As such, in terms of modeling a particular type of soil, you will need to look for appropriate parameter values which typically come from bevameter experiments. Look for example in the book by Wong, or any other source for such parameters.
The FEA deformable soil model in Chrono::Vehicle is not really maintained (it builds and should work, but we have not used it for any relevant simulations). That modeling approach is computationally quite expensive, relies on parameters that are difficult to come by, and (unlike granular dynamics or the CRM approach that Dan mentioned) cannot capture larger soil deformations. If you are interested in an FEA-based approach, look also at work by Yamashita and Sugyiama on hybrid FEA-DEM formulation (by the way, their terramechanics code can be coupled to Chrono vehicle simulations through the Chrono::Vehicle co-simulation framework).
But I second Dan’s suggestion to look at the deformable soil formulations available in Chrono, in the order SCM, CRM, DEM.
To view this discussion on the web visit https://groups.google.com/d/msgid/projectchrono/00d9de13-79a5-4001-9943-e14c234f28d0n%40googlegroups.com.