How to use OpenRadioss

Victor

This is quite exciting. OpenRadioss is an open source explicit dynamics solver. Mecway's support for it is currently quite limited and some details might change in future because I'm not very familiar with it yet, so you have to turn it on in the Labs menu. In future versions, I intend to greatly improve the integration. Mecway 26's OpenRadioss features are:
- Export .rad input file.
- Run the solver when you click the solve button.
- Import solution with displacement, velocity, acceleration, stress (solid elements) and some others.
- Solid, shell, beam, truss, spring, damper elements.
- Linear elastic material
- Custom material
- Gravity load, which may be time dependent.
- Named selections are exported as groups (node, surface, line, brick, shell, 3-node shell, beam, truss, spring) with predictable IDs in the order they appear in the outline tree (alphabetical). So if you name them beginning with numbers starting from 1, you can easily refer to them with custom boundary conditions.
- Custom model definition

How to use it:

1. Download OpenRadioss from https://github.com/OpenRadioss/OpenRadioss/releases and extract it somewhere.
In Mecway:
2. Turn on Tools -> Options -> Labs.
3. Turn on Tools -> Labs -> OpenRadioss.
4. In Tools -> Options -> OpenRadioss, click Find OpenRadioss root directory ... and choose where you extracted it to.
5. Set Analysis type to Nonlinear Dynamic Response 3D.
6. Set solver to OpenRadioss.
7. Set the time step options that are enabled.
8. Make a model
9. Define boundary conditions in OpenRadioss -> custom model definition and most materials in Material properties -> OpenRadioss.
10. Refer to the Altair Radioss reference guide for how do write these. https://2022.help.altair.com/2022/simulation/pdfs/radopen/AltairRadioss_2022_ReferenceGuide.pdf
11. Click Solve
12. Solution has displacement and velocity by default and optionally acceleration and solid element stress.

The file for the video above is attached here. It's small enough to run in the free version of Mecway and solves in only a few seconds.

Sergio

❤❤❤❤❤ can't wait to start crashing things!

Thanks!

FERNANDO

JohnM

Nice

Sergio

@Victor, in the impact model, in which part of the Openradioss instructions you tell that the lower part is "fixed"? I'm tring to make a new similar model but my lower part flies away when it gets in contact with the impactor as it would be free :-(

Victor

@Sergio The void material (/MAT/LAW0) stays put on its own. Another way is /BCS for fixed displacement boundary conditions.

Sergio

Yes, I took the /BCS from the tubes examples and is working, even with symetries. Thanks a lot for bringing OpenRadius to us! Now that there are three solvers available, would be very interesting show the name of the solver in the post view, maybe with the model name

FERNANDO

Next, Openfoam.

Sergio

Uf, that would be incredible, having Openfoam running natively from Mecway interface. Do you know @FERNANDO that you can prepare Openfoam models in Mecway? But for solving you need the command line, and postprocess in Paraview.

FERNANDO

I didn't know. It seems interesting. But a pre and post process in Mecway would be fantastic.

FERNANDO

A very important result in explicit impact calculations is the energy balance.

Sergio

@FERNANDO, I did a few test yesterday, and even some more variables can be requested as results adding some cards in the custom engine file, for some reason I cannot postprocess as the native or CCX ones, units are fixed and when I choose one node the result is not showed in the time plot. I´m sure that in next versions the support will be improved. Probably there is a way to tell to the solver to write a compatible VTK file for post in Paraview for the moment.

JohnM

We have used Mecway to prepare OpenFoam models for many years, similar to how it's shown in this video. You basically need a program to convert your Mecway model to individual STL files for each patch, then a script to pick an approriate template. Not too hard!

Fatmac

Thanks for another great update. I'm particularly enjoying the addition of the openradioss solver. I'm struggling to get to grips with even the simplest boundary conditions or materials but I'm getting there and I feel it will be a very useful tool. While looking for help on the internet I stumbled across this document. Introduction to Explicit Analysis
using RADIOSS
A Study Guide https://designgyan.com/srp457/assets/images/papers/DG_17541091_1_Crash_Book_October_24_2017.pdf . Its a huge paper but section 4 really gave me some pointers -I hope it hepls some one else as much!

Sergio

@Fatmac we thought CCX had a complicated syntax...OpenRadioss is perverse, that fixed format is breaking my head!!!

Fatmac

its certainly challenging me - i found more infor on the web but not sure I'm making much progress at the moment - I can't even displace a node or make a time function yet!
i think this is my best source of info so far https://help.altair.com/hwsolvers/rad/topics/solvers/rad

Victor

@Fatmac Thanks for that Altair Study Guide. It's easier to read than the user guide and more helpful about recommended element options.

The fixed format does take getting used to. But there's also a lot of consistency so it gets easier.

Here's an example of imposed linear ramped displacement on node group 2 from time 0 to 1:

/FUNCT/13
13
0                      0
1                      0.13
/IMPDISP/1
1
13                 X                       2

#

Sergio

@FERNANDO , I was able to postrocess in Paraview and request/post energies!

FERNANDO

Brilliant. Paraview is a very powerful tool. The ideal would be to not have to exit the program.

Sergio

@FERNANDO, of course, working outisde Mecway is 10x more complex and time consuming. Still I don´t understand why also in Paraview as in Mecway cannot pick a node and get the field variable as with CCX results. In this example the 2d plot is from a csv value, but with CCX converted results I can pick a node and get similar plot along the time, and this plot is linked to the time of the 3d view/results as in MW (here is disconnected).

Victor

Not picking nodes in Mecway is because the OpenRadioss stresses/etc. are only on elements, not nodes. I intend to apply node averaging in Mecway so it has both.

Vinicius_Petrucci

how incredible!

polymathic

I've had a decent muck around with this now and it is really, really cool.

JohnM

I posted this on LinkedIn, got over 30k views. I think you are onto something Victor...

https://www.linkedin.com/posts/johnmannisto_openradioss-mecway-activity-7238656023182413824-O-6e?utm_source=share&utm_medium=member_desktop

Victor

@JohnM Amazing. Thanks. It was @Sergio who prompted me to do this. At first I thought it might be a less buggy alternative to CCX but this explicit dynamics thing is incredible. More integration coming in Mecway 27.

disla

Hi,

¿Do open radios also have eigen solver.? In that case, do you think it could be worth to use it know the actual limitations in Arpack.

Victor

I think it kind of does but the build from OpenRadioss don't have it enabled. In 2022, one of the maintainers said "It is not possible to provide the direct usage of /EIG functionality in OpenRadioss, yet." The only ARPACK alternative I know of is FEAST which is in MKL.

MikeMcMullen

OOH! this looks just right for me!! Smashing bridge rails here I come.

Sergio

I'm sharpening my arguments to propose OpenFOAM support as well :-)

MikeMcMullen

That might work well for the vehicle. They deform 6"-12" during crash tests. The bridge rails, not so much. Deformations may be as much as 6", but strains are usually well under 1%.