Attachment Constraint Issues

[lancef]

Hi,

I'm modeling a knee and having a bit of difficulty with setting up my attachments. Using Simpleware, I've created the finite element model and duplicated the nodes at every contact surface in order to develop my constraints through ABAQUS. I'm not entirely sure as to which attachment method is the best (tie, MPC, etc). Or what might be the easiest way to go about it (time may be an issue). A few things worth noting about my model: 1) The attachment geometry is not perfectly smooth. For example, the ligament insertion points are not perfectly adjacent to the bone. 2) Some of the parts are finer than the others. For example, the meniscus has a finer resolution than the cartilage above and below it. 3) The model will be subjected to running-like loads and my primary analysis will be to look at the stress and strain inside of the femur, directly superior to the medial meniscus. Any bit of help on this will be sincerely appreciated!

Lance

 

[IceBreakerSours]

Lance:

Let me be honest with you because you seem to be lacking expert supervision:

You can not expect to make a model that gives reasonable predictions in a short time without any expert supervision. So, if time is a significant concern, there is a conversation to be had with the powers-that-be. Otherwise, writing is on the wall; failure is staring you in the eyes. Why? Because - at the very least - you have nonlinear materials, less than optimal meshing, and contact in a 3D dynamic problem.

With that ominous message out of the way, let's set some things straight:

What you have so far is not an FE model. You just have a geometry (from an image processing code) and a mesh (probably). That is it.

At the knee joint, you are going to have to specify kinematic constraints using ligaments. I hope they are just 1D springs or trusses with nonlinear stiffness/elasticity (appropriate for running) assigned to them. If they are in 3D, find yourself another project.

There is also the matter of contact! Let's leave it for another time because .. well, it is contact.

And, we haven't even talked about the solver: implicit vs explicit.

Honestly, under the aforementioned time constraints, this project should've been done using OpenSim/AnyBody + FEM. Using OpenSim/AnyBody etc., you could find out what the joint reaction forces and moments are during running and use those loads on a deformable femur to compute the stresses. Under the right set of circumstances, the whole thing could be finished in within 2 weeks - 6 months, depending upon the level of expertise of the individual.

Are you new to this forum? If so, please read these FAQ:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[lancef]

IceBreakerSours:

Thank you for the honest and perhaps a bit rude response. My question was to simply better understand what attachment constraints might be most appropriate here. The time issue is for the attachments and the attachments alone. I, in fact, have 2+ years to work on this. I admit I am new to ABAQUS, and a bit of help and insight was all I was looking for. Thanks for your time, anyways.

 

[rstupplebeen]

I agree with IBS that this is a very difficult problem especially if taken in 1 bite. Break the problem down and work on each chunk in parallel.

Create a loading scenario in a codes like OpenSim/AnyBody or from literature
Create very simplified geometry for initial quasi-static FEA with contact even starting as simple as a sphere indenting a block.
Slowly add complexity such as non-linear material properties, dynamics, geometric complexity

In 2+ years with help from here or with colleagues you can get it done you just need to me methodical. There is also a biomechanics forum here which may be appropriate for some questions down the line. I hope this helps, good luck and welcome to the forum.

Rob Stupplebeen

http://www.optimaldevice.com

https://sites.google.com/site/robertkstupplebeen/

 

[IceBreakerSours]

Ideally, having coincident nodes is what you want for attachment purposes. While it is problem-dependent, you should try hard to have coincident nodes because other kinematic constraints will result in a displacement discontinuity - which no solver is particularly fond of. But, if there is no way around it, *Tie is a good first step.

You can also project the nodes on a surface to a plane and subdivide the projected length into layers of elements, and then try to make the nodes coincident across the two opposing surfaces.

The hard part is to be keep the model as simple as possible without making it too simple or unrealistic. However, two years - with help, continued advice (as provided above; yes, some of it was bitter), and expert supervision - is sufficient, as long as you keep the larger picture front and center.

I do not think you can do the whole thing using FEM only in two years. But, if you still wish to, best of luck to you.

Are you new to this forum? If so, please read these FAQ:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[lancef]

Thanks for the responses. I hope to frequent this forum over the next few years!

 

[IceBreakerSours]

And we will be glad to help out. By the way, while this forum is an excellent place, there are other forums that you should consider visiting regularly: Abaqus Yahoo! Group, LinkedIn Abaqus FEA Group, and PolymerFEM. The first two happen to be more active than PolymerFEM.

Are you new to this forum? If so, please read these FAQ:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083