Glue failure estimation

[Xprivate]

Hi all , i have a problem where my structure is a combination of glass material and aluminium. The glass is glued to the aluminum frame, also fastened to the frame . so basically the glass is fastened and glued to the aluminum frame.i am trying to simulate this on FEMAP. But when i run a glue contact the fasteners(CBUSH) will never see loads because the primary load path is always the glue connection.
1)can someone suggest me ideas on how i can post process the glue connection?
2)what loads can be used on femap to estimate the glue failure? .
3) in my case where glue and fasteners are working together, how to model to accurately estimate fastener loads ?

 

[SWComposites]

Fasteners in a bonded joint will not pick up load until the bond fails. Fasteners are less stiff than the bond.

Using stresses in bonded to predict strength does not work. You need to use fracture mechanics based methods.

 

[rb1957]

Practicalities aside ... why would the fasteners in the FEM have no load ? Because the adhesive is rigid ?

Gluing glass to Aluminium is probably not a good idea, depending on thermal effects, unless the glue is very low stiffness.

This is an aerospace application (I'm guessing) ?? Wow !? Can you explain more ?
Without knowing about this application, I'd rather see the glass constrained by a set of splice plates, with in-plane loads going around the glass.

I see a lot of issues with the model ... off-set loads, bending the glass.
How controlled is your glue application ? It may not be as good as you think ?

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[Xprivate]

thank you SW Composites thank you for your response. Also do you have any reference that could help ? THanks again


@rb1957 yes the adhesive is rigid in FEM, i dont know if we can control the rigidity of the bond in FEMAP.
yes it is aerospace application, and it is a cockpit door of an aircraft. i cannot give you more information as it is confidential.

the picture you see is just a test model i ran to see if the fastener will ever catch loads or moments, and this is also an image to give a context of my problem. my actual FEM is an aircraft door and frame. And coming to glass is again a different material that can be glued with special strong adhesives that is capable of holding the glass and frame together. Hope that helped . Thanks again @rb1957

 

[SWComposites]

there in an intro to fracture mechanics for bonded joints in a chapter in this book:

https://www.espcomposites.com/book/book.html

 

[rb1957]

"surely" you can control the stiffness of the glue modelled in the FEM ... rigid adhesive is making "nonsense" of your model.

is the door external (like on the fuselage shell) or internal (like on a bulkhead)?

this is not a part 25 aircraft, right ?

is this a side window or the windscreen ? If windscreen, how'll you treat birdstrike ?

Fastening glass to Aluminium is not a good idea. At a minimum you'll need a rubber bushing in the glass hole (to give it some compliance).

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[Xprivate]

@SW Composites , THanks for that.

@rb1957 i will research on how to control the glue stiffness. External on the fuseshell. Its Part23 aircraft , Side window and bird strike is accomodated. fastening is not principal, but bonding is. Fastening is just to show some safety compliance and are light fasteners that dont carry high loads

 

[rb1957]

ok, a very unusual design choice, I wish you luck with it.

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[Paul_M]

@Xprivate, If I'm correct, have you tried modeling the glue as a cohesive material/elements (MATCZ) and use Sol401? I suppose it should be the way to model the glue failure

 

[Xprivate]

Hi @Paul_M , i have been running everything under SOL101 for now. i have no idea about that as this is somewhat new area for me.

 

[rb1957]

can you model the adhesive with a CBUSH ?

Can you have an adhesive glue glass to Aluminium ?

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[SWComposites]

Any bond line (in aerospace we don't "glue" things) stresses from a linear FEM are not realistic or valid. If you are not capable of running a fracture analysis (cohesive or VCCT elements) or don't have the material properties for that analysis, then you just need to test the joint using representative test specimens.

 

[TG_eng]

Glue is going to be a 1:1 with displacements.

If you want some of sort of compliance with the "bond" you will have to use cohesive elements. But that requires the nonlinear solvers (401/402).

Perhaps some others have ways of modeling this type of thing without the advanced solvers. I am curious to hear.