mesh gradual contact wheel-rotaia train

[Salvàtore_xy123]

hello to all and good year,
I am carrying out a static analysis with ansys workbench of a train on a rail; I would need to have a mesh with very small elements in the contact area wheel-rotary that gradually become larger moving away. I tried with the sphere of influence but I can not make more gradual the increase in the size of the elements. do you have any advice on how to do politely?. I attach the photo of the model. Thank you in advance.

 

[gerod]

You may use the contact as a mesh parameter. In the mesh reference criteria, use the contact criterion to finish the mesh close to the contact. decide how to do it.

 

[meccanicamg]

I propose a more crude solution that comes from less refined fem solutors.
If you detect a portion of geometry around the contact, let's say a footprint of 10 mm wide and separate the surface with a sketch you can make a small mesh. in the rest of the track make the biggest mesh. do not exist mesh control tools in ansys? I think so. But I guess you have to define an area.

 

[gerod]

in ansys there are mesh controls, you have several.
the idea of the area or footprint is solution that can go.
You could make a rough mesh and then leave the program to stick where it serves with a convergence criterion.

 

[Salvàtore_xy123]

in ansys there are mesh controls, you have several.
the idea of the area or footprint is solution that can go.
You could make a rough mesh and then leave the program to stick where it serves with a convergence criterion.

First of all thank you for the answers. Excuse me but could you kindly explain to me better how to check for a good mesh by inserting a convergence criterion in ansys workbench?

 

[Salvàtore_xy123]

I propose a more crude solution that comes from less refined fem solutors.
If you detect a portion of geometry around the contact, let's say a footprint of 10 mm wide and separate the surface with a sketch you can make a small mesh. in the rest of the track make the biggest mesh. do not exist mesh control tools in ansys? I think so. But I guess you have to define an area.

Thanks for the advice. about editing a cad file in ansys workbench, do you kindly know how I can update the geometry without having to re-enter all data in the preprocessor (such as connections and loads) every time? .

 

[gerod]

1. You have to study a little before you go to fem. There's a nice guide online. I can't assess if you've made a good mesh or not if I don't see the file. What version do you use?
2. you can edit the file if you have a licensed program with communication interface. If you don't have it, you have to put your hand on it every time!

 

[Salvàtore_xy123]

1. You have to study a little before you go to fem. There's a nice guide online. I can't assess if you've made a good mesh or not if I don't see the file. What version do you use?
2. you can edit the file if you have a licensed program with communication interface. If you don't have it, you have to put your hand on it every time!

I use version 19.2

 

[Salvàtore_xy123]

Hello everyone,
I managed to make this mesh (I attach photos and details). the problem is that for the moment the solutor after a while gives me this error: "an unknown error involved during solution. check the solver output on the solution information object for possible causes".
Can the problem be related to the fact that my PC fails to process such a simulation?. My goal is to bring the tensions into the contact zones wheel-rotary (I must find the maximum normal voltage). Moreover as soon as I change the parameters, ansys realizes mesh diversified between the tracks (on one side uses square elements, on the other triangular). Do you have any kind of advice? Thank you again.

 

[Onda]

Why not just models half, cutting it to the center and applying a symmetry bond? half the degree of freedom of the model. I have impression that you need too much time to turn
solid elements, must be square, linears are too rigid, but surely you have a command to choose the type of element to use.
in an image there are warnings, must be verified. .

 

[Stan9411]

you will have millions of knots... not enough to model a single wheel, reviewing the load conditions and taking into account the symmetry? warning is the one concerning processing time .. check case.

 

[gerod]

simplifies. consider only a binary conical. You have such useless mesh.
then if you put the convergence criterion on tensions be careful because it could be wrong. better moving. do not use the sphere of influence, use a contact surface or the inflating criterion on the bond

 

[Salvàtore_xy123]

I had thought of using the symmetry bond but in my case I do not think I can use it being the case of a curved train. I'm looking for a compromise. thanks again to all.

 

[Onda]

make a simplified model, such as beam, where you solve the balance. then cut the model and take the forces that this downloads on the part of interest, then study a single wheel. It's called submodelling, you find it explained in ansys' manual. In this way, you can analyze with contact only one wheel, the most loaded, keeping the correct loads. Otherwise I'm struggling to think you're out.

 

[Salvàtore_xy123]

Hello everyone, I am trying to resize the mesh so I have fewer elements but every time I change parameters ansys me realizes it diversified between the 2 tracks or gives me error. I would like to "copy" the mesh I want in the other track. I tried the mesh copy command but in my case I saw that I can't use it as the tracks are specular. do you kindly have any advice on what I can do? Thank you again.

 

[Stan9411]

a mesh is a discretization of a continuous domain (an object). change the domain, change the mesh. you can't copy a mesh. do it rework with the same criteria, it will be different but equally satisfying for your purposes.