Best practices for FEA connector modeling.

[mfritze]

Hello,

I'm using SW simulation as analysis software for a piece aerospace hardware that will have a lot of oversight on our analysis.

What is the best practice for modeling fasteners in an assembly with up to 20 bolts? I am not interested in resolving bolt forces/stresses (these will be derived and analyzed by hand) but in the stresses in the components. In particular I would like to replicate the use of a CBUSH Nastran element as I know this is the practice used by the reviewer.

I do not wish to solve contact and instead would like to invest computational resources in mesh resolution.

Thanks.

 

[aerostress82]

If I got your question right, you are aiming at the component stresses being correct and not concerned about the loads that will come from your CBUSH elements.
You might want to have in plane & axial stiffness values calculated by Huth's Method and K=AE/L

And again for your CBUSH elements, you might want to put values around 10^8 rotational degrees of freedom that are in-plane of your sheet metal. The out of plane rotational degree of freedom should be around 100 to be conservative (at some level - you should really check this part depending on your assembly - but 100 is kind of standard for stress models).

Anything related to especially Solidworks, I don't know much. Just used Solidworks-Cosmosmotion-Cosmosworks combo in 2008 for a project, never modeled fasteners.

Hope above explanation helps at some level.

Spaceship!!
Aerospace Engineer, M.Sc. / Aircraft Stress Engineer

 

[mfritze]

Thanks!

I have not used Nastran but have viewed a lot of reports from it --hence why I'm attempting to replicate some of its best practices in sw.

One thing I never understood on the CBUSH use as a fastener is why it is ok not to consider the bearing stiffness? If we define axial stiffness as AE/L this works very well for when the bolt is in tension, but would be inaccurate for when the joint in compression. Can you explain why this acceptable practice?

On a side note, I think it would be very useful to have an element that had two stiffness for modeling fasteners -one for positive strains and one for negatives strains. While this would be technically nonlinear in reality would only require two iterations (first to solve for strain direction, second to solve with stiffness selected based on strain directions in first study). I suppose I could do this manually or write a script. Unfortunately solidworks doesn't let you scripts for things like this... time to ask for better software ...

 

[GregLocock]

Doesn't it rather depend on what effects you are trying to model? CBUSH for a bolt is rather primitive if you are looking at fatigue near the bolt, but will work fine for modal frequencies etc.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[aerostress82]

mfritze
When you have a bolt in compression, there will actually be no compressive load on your bolt as it will be transferred to the contacting faces of 2 parts in an assembly.
The general practice in aerospace is that the compressive loads are assumed as zero. They are not accounted for any fastener analysis. However, the resultant or in-plane shear loads on a fastener are used for fastener analysis for both compressive and tensile cases.

GregLocock
The CBUSH forces are again used for fatigue analysis - where as you say is a little more primitive than the automotive 3D hex fatigue models (with pre-tension, contacts &/or nonlinear geometry/material). But in the end, they are mostly on planar or very symmetrical geometries that are mostly as regular as possible (square, triangle etc).
However, in automotive, for example on a bumper assembly, you will have so much irregular shaped (trimmed composite shapes in Patran) that the aerospace methodologies don't come in handy for automotive parts.
In aerospace, they are pretty much used in almost every joint as CBUSH elements and all fatigue, damage tolerance calculations are made based on those models most of the time. Thus, it is a well accepted practice until we have supercomputers that will enable us model "everything" with hex and tria meshes....

Spaceship!!
Aerospace Engineer, M.Sc. / Aircraft Stress Engineer

 

[gravityandinertia]

mfritze,

SOLIDWORKS Simulation does allow for scripting through the API, it just isn't well-documented. However, if you know how to create macros and use the SOLIDWORKS API documentation there are ways for certain things.