How to define RSURFU

[bues0022]

I'm having a bit of a hard time attempting to define an RSURFU. I don't actually know where to start on it. I've tried doing some internet searching, but I can't figure out how to define it. I basically have a quarter-symmetry model of a washer, and want to expand the ID to a larger diameter. I don't necessarily want to use an axisymmetric model because in some simulations the center of my washer is non-circular.

Furthermore, in previous simulations I defined a center radius change with a displacement constraint, but this didn't work with my X- and Y- symmetry constraints (different coordinate systems). I can't figure out how to get around this. Since I don't know how the RSURFU works, I'm telling this so I don't end up in the same place again.

The project is very urgent and I need some quick tips. Thanks again for your help with this.

 

[bues0022]

Update: I think I have some things figure out, but my model won't run. I'm getting an error: ErrElemMissingSection. I'm fairly certain that my model is defined correctly - but I'm obviously wrong though. I have my input file attached here. Can someone please run it and let me know what I have defined incorrectly? I've never created/used a rigid surface for an analysis, so I'm a bit puzzled how it all fits together.

 

[rstupplebeen]

You have a surface section defined for your balloon. I assume this is just to provide a pressure load to the artery. I would just skip the balloon and apply a pressure to the inside wall of the artery. Why are you running this in Explicit? Please see the attached picture for this example run in Standard. I hope this helps

Rob Stupplebeen

 

[bues0022]

Rob- Thanks for the help and thoughts. I think perhaps I should have explained some things a little better...

I have a large series of simulations where I applied the load directly to the center ID. It's interesting you were able to get that to run so quickly, I had problems running it in Standard and Explicit actually solved quicker for me. What was your pressure?

The reason I am using the surface to deform the ID of the quarter-symmetry is because I want to deform the part out to a known diameter. Prescribing a distance rather than a pressure load should also have an easier time computing. Furthermore, the next step is to try a non-linear center ID, prescribe a final ID, and see what happens.

Long explanation for saying that I really need the rigid surface to work. Thanks for the help!!

 

[rstupplebeen]

I applied a unit pressure. The images were scaled to show the deformation.

I would check out using plane strain for your model. In the bottom of the image I sent before there was hour glassing occurring. I assume that the artery is relatively long so this assumption would be reasonable.

If all you care about is the final diameter I would push a rigid cylinder of the desired diameter into the artery.

Is a fixed outer wall condition appropriate?

I hope this helps.


Rob Stupplebeen

 

[bues0022]

Well, first off I think I need to do a slight correction. The material is tissue, but not a vessel - it's a thin cross-section of tissue. Think of it like poking a whole in a shirt, and balloon-expanding that small hole. There likely will be some hourglassing.

The last thing you said: "If all you care about is the final diameter I would push a rigid cylinder of the desired diameter into the artery." is exactly what I'm trying to do - but can't get the rigid cylinder defined correctly.

A fixed outer wall should be far enough away from the center distortion so minimal effects are seen out that far. It eliminates the need to model a giant portion of tissue.

 

[rstupplebeen]

Thanks for the further explanation. Hour glassing is a modeling phenomenon not a physical one. I would increase your mesh density and possibly switch to second order elements.

In your model you were stretching the balloon not displacing it.

I would create an rigid part and translate it's reference point towards the tissue.

If you post your CAE file I will happily look at it.

I hope this helps.

Rob Stupplebeen

 

[bues0022]

Thanks for the offer Rob. I understand what you are saying by creating a rigid part and translating it towards the tissue rather than stretching the balloon. In your opinion, how much different will the solutions be? I am trying to mimic my experimental testing as closely as possible in which I use a dilation balloon. I think by translating a rigid surface towards the tissue I would see stresses and strains spread non-uniformly across the tissue because the ID near the symmetry planes would see deformation first, and the ID at 45 degrees see it last. If my expansion is too large for my material model (quite likely) then my results could be questionable. I guess the key is in the interpretation, right? If the simulation completes correctly, and I get concentric stress rings, then I'm set. If the solution errors out due to excessively distorted elements, and the solution is non-concentric rings, then I know the simulation did not complete correctly.

I am aware that my mesh is quite coarse. I have a free student copy of ABAQUS running here on my laptop that I am using to test assumptions and setups. Once everything is working I redefine the model on the supercomputers - but that's a remote login so the connection can be slow making this debugging process painfully slow at times.

I *think* the rigid surface finally works, but my contact doesn't. In any case, here's my CAE file (just over 2Mb). There's still the possibility that my contact doesn't work because of the surface. Thanks again for taking a look at this with me today!

Ryan

 

[rstupplebeen]

Sorry but I can't open the student version. Could you send your latest INP?

Rob Stupplebeen

 

[MechIrl]

Have you looked into any of the literature for modelling angioplasty balloon expansion? There's quite a bit out there, for example De Beule's work in University of Ghent.

Generally the application of an internal pressure is peferable for this sort of work as the application of a displacement is not as physically realistic.

Getting convergence with large deformaion of cylindrical tubes can be difficult. Have you looked at folded balloon configurations? They are more realistic and dont involve the excessive distortion your getting in your cylindrical tube model.

Also element choice is important, membrane elements seem to be most common.