Transfer Displacement results from one mesh to another

[femasters]

Hello Experts,

I need to transfer displacement results from fine mesh to coarse mesh. Also the vice-versa is also needed - coarse to fine.

The transfer of results should be very much close to the actual simulation results.


Thanks,

 

[rb1957]

1) linear interpolation
2) rerun the model ? ... why would you want to copy a coarse mesh displacement onto a fine mesh ?

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

 

[Dave442]

if you are using a commercial FE package you might be able to use sub-modelling to do this.

 

[ThomasH]

Hi

If I understand it correct, in model 1 the displacement is a result while in model 2 it can be seen as a "load"?

If you just make sure that the node numbers in the models are consistent is some way, it should not be to difficult. What I mean by the "consistent" numbers is that node 1 should have the same location in both models so the displacement field becomes the same in both models.

In Femap you should be able to import the results from model 1 into model 2 and then create "Load from results". And I doubt that is a unique feature in Femap. Sounds like you may be playing with the concept of a multigrid solver.

What is the application (if you can share)?

Thomas

 

[femasters]

Thanks Everyone for your answers.

@rb1957, I need to do this displacement transfer for Fine Hex8 to Coarse Hex8. Will Linear Interpolation work fine ?

@ThomasH, Actually I need to write a program in C++ to do this transfer from one mesh to another. Is there any algorithm available ?

 

[ThomasH]

femasters

So you want to create new data as well. Take a look at how a mass matrix handles data, that might do the trick. A linear interpolation might work depending on how accurate you want to be and what the deformation looks like.

Thomas

 

[rb1957]

1) ok, you want to apply the coarse grid deflections as enforced displacements to the fine grid model. why will this be more efficient than rerunning the fine grid model with the coarse grid loads ? I think applying coarse grid deflections to the fine grid model doesn't validate the coarse grid model.

2) if HEX8 is a linear element then linear interpolation is reasonable. if 2nd order then parabolic interpolation would be better. maybe you need to create a deformed surface that matches the coarse grid deflections ?

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