Check single mesh size in abaqus and comsol 1

[FEAppp]

Hello,

I am doing the same analysis in both abaqus and comsol, but they give me different result.
So I want to know whether the single mesh size is same in two softwares.
For example, I have total 35000 elements, how to check the length of the single element?

Thanks.

 

[FEA way]

In COMSOL you can control the mesh size using Mesh --> Size --> Custom (there it's possible to set minimum and maximum element size). There are no options to check the size of individual elements in this software.

In Abaqus you set approximate element size but there's a Query --> Distance tool that allows you to measure individual elements (using the distance between selected nodes). More can be done with Python scripting.

 

[rb1957]

can you export a NASTRAN bdf (or whatever code) from one and import it into the other ?

another day in paradise, or is paradise one day closer ?

 

[FEAppp]

Thanks, FEA way!
Do you have experience in modelling stress dependent elasticity in comsol?
My Young's modulus changes with stress. I defined a function for Young's modulus versus stress.
I tried to use weak contribution and auxiliary dependent variable to solve the problem with the circular variable dependency and introduce a new variable onto which the stress value is mapped.
But the question is that I don't know how to write the weak expression.
Thanks!

 

[FEA way]

Have you seen the example problem "Modeling Stress Dependent Elasticity" on COMSOL's website (Application Gallery) ? It should help you define this dependency.

 

[FEAppp]

Yes, I have checked that it. They gave the expression test(p)*(p-solid.pm). I am not quite sure what the expression means.
Thanks.

 

[FEA way]

This expression of weak contribution is used to map calculated pressure (solid.pm) to new dependent variable (p). First term is a test function of p. So basically what we have here is:

Built-in PDE + ∫_Ω δp*(p-press) dΩ = 0

Pressure is calculated as -(1/3)*trace(σ)

 

[FEAppp]

Hello FEA way,

In the stress-dependent-elasticity, the material model given is where Young's modulus is a function of the pressure. As for the map solid.pm to p, why it is not just test(p)*solid.pm (weighting function times force)? I don't know why it needs to be test(p)*(p-solid.pm)? To put it in the simple way, I am still confused what the expression inside the integral "δp*(p-press)" comes from.

In my application, my material model given is where Young's modulus is a function of the maximum stress. So I want to map calculated solid.sp1 to new dependent variable (p). I just guess the expression is test(p)*(p-solid.sp1). But it's not correct, right?

Thanks so much for your help.

 

[FEA way]

From what I understand, this expression simply equates new variable p and pressure:

δp*(p-press)=0 —-—> δp*p = δp*press

So in your case, if you want the elastic modulus to be a function of first principal stress, you should use test(p)*(p-solid.sp1), like you said.

 

[FEAppp]

Thanks FEA way, it makes more sense now!!

 

[FEAppp]

Hello FEA way,

When I define my material model that is Young's modulus is a function of the maximum stress.
The result says it's not converging.
However, it converges if I set a constant Young's modulus.
Do you have ideas where I should change in order for it to converge? It doesn't work if I increase iteration number.
Thanks a lot~

 

[FEA way]

You may have to adjust the function that you use to vary Young's modulus based on stress. Make sure that values are realistic and, for example, you don't start with very low stiffness.