FE modelling of a patch (surface to surface) 2

[johnnyc2]

Hi everybody,

I've been using Patran for a bit over a year and have a model of a carbon fibre aileron made out of surfaces. I have introduced a hole in to the top skin (as damage) and now I need to place a carbon fibre patch over the top to repair the damage.

To simplify the problem, imagine a flat plat with a 2inch diameter hole in it. I need to attach a 5inch diameter patch to go on top of this hole.

My problem is that I've never had to model surface-on-surface contact like this before and am unsure of which approach to take. I have received some suggestions from colleagues (using RBE2, modelling the adhesive as bricks with a surface patch on top) but nobody seems sure. I was wondering what people on here thought. Modelling of the adhesive is not necessary at this stage, but would be helpful to know how to do.

Any help would be much appreciated.

Regards,
John

 

[rb1957]

i suggest that you model the patch as a surface, which is probably blindingly obvious.

you have fundamental decision to make ...

1) you can offset the patch form the skin and fill the void with solid elements to represetn the adhesive; so you'd want to replicate the existing skin element mesh onto the repair patch (ok, you don't Have to, but it'll make the adhesive elements look nicer), modell the adhesive as shear effective only (real G, dummy E).

2) you can model the patch coincident with the skin (which unfortunately gives you coincident elements and nodes) and join node pairs with zero length springs (CELAS) ... i use this method for riveted patches.

for you application i think 1) is probably the better choice, but it will be "fussier" to work with ... all the skin elements will need to be plates. it'll catch to off-set bending (load transferred into the patch locally bends the skin), probably exagerate it.

 

[blakmax]

I am a specialist in composite and adhesive bonded repairs. My current approach uses closed form solutions, so I have not used FEA in many years, but I do know that modelling patches without modelling the adhesive will give you a gross overestimate of the effectiveness of your repair, and in my days if you modeled the adhesive as bricks, you got errors due to the aspect ratio of the elements. I remember that the FEA whiz kids used PAFEC and developed special elemets to represent the repair. I am sure if you did a literature search for the name Rhys Jones at ARL Melbourne you will find more useful data.

A further complication is that if you model the adhesive using elastic models, you will reject repairs because the adhesive elastic limit is exceeded. In practice, up to 80% of adhesive strength in properly designed joints is derived from plastic behaviour, so for a realistic model, you require elastic-plastic elements for the adhesive.

Regards

 

[kellnerp]

Here are a couple references

Composite Repairs of Cracked Metalic Structures

Damage Tolerance Assessment of
Bonded Composite Doubler Repairs for
Commercial Aircraft Applications *



TOP
CSWP, BSSE

http://www.engtran.com

http://www.niswug.org

"Node news is good news."

 

[blakmax]

Hi Kellnerp

I have to state that I have looked at this reference and I am not impressed.

"11 Sep 09 12:59
Here are a couple references
Composite Repairs of Cracked Metalic Structures"

Much of this data is "borrowed" from my work without my permission. Not the FEA stuff but a large proportion of the processes data is a direct lift of my work. As a cross-reference refer to DOT/FAA/AR – TN06/07, Apr 2007. Yes, the reference material pre-dates this publication, but that is hwhat happens when a much earlier draft is seconded by others.

blakmax