Elastic - Plastic Hart-Smith Joint

[MNLiaison]

I have the shear stress-strain curves for a film adhesive and have constructed the Hart-Smith elastic-perfectly plastic idealization for the different temperatures. The process used to construct the lap shear test specimens to get the stress-strain curve was a PAA prep with an autoclave pressure of about 100 psi.

I would like to use the data above to analyze repairs where an alternate surface prep and cure pressure (vacuum bag - 20inch Hg) are used but recognize that these changes can affect the joint strength.

I have data on the max adhesive failure stresses of lap shear specimens for both the alternate surface prep method and cure pressure that I would like to use.

I am debating if I can use my Hart-Smith elastic-perfectly plastic graphs but knock down the plastic adhesive shear stress based upon the lap shear failure values of the alternate procedures. I would keep the adhesive plastic and elastic shear strain values from the original idealization.

I haven't been able to dig up any published data that I can reference to see if this correlates with test data.

Has anyone gone down this path before? Is there any publicly available data that correlates stress-strain curves for the scenario I have outlined? Or am I stuck and must construct stress-strain curves for each different process used?


Regards

 

[der8110]

Not only construct stress-strain curves for each process, but your own allowables test program for each process if you have any intention on using it for primary structure.

If it's not critical, then 'feel good' shop tests can get you there. What you propose is a very good start along narrowing down which variables to test.

 

[Compositepro]

I know John Hart-Smith and Ray Krieger who developed the KGR-1 shear stress-strain test method but I'm not sure I understand your question.

Cure pressure, per se, does not affect adhesive properties. Resin bleed, voids in the bond line, bondline thickness, and thermal history in the cure cycle can have dramatic effects on adhesive properties. These are very difficult to model and, in any case, the goal is to have a void-free bond line of consistant thickness.

Bagging proceedure, adhesive surface texture (rough and low tack versus smooth and tacky), and resin flow interact with pressure to affect bondline thickness and voids due to trapped air or excess bleed from the joint. You can spend several life times studying the affect of cure pressure with "all other variables held constant". The problem, actually is that many important parameters of a "good bond line" are not adequately controled.

For example air is easily trapped between a tacky adhesive film and an adherand. Many adhesive films are cast on embossed polyethylene film. This texture transfers to the adhesive. If the adhesive is cast on one side of the poly film there will be a cross-hatch pattern of grooves in the adhesive that allows air to be removed by vacuum. If the adhesive is cast on the other side of the poly film there will be a pattern of pits in the adhesive that trap an air bubble in the bondline at each pit. These can coalesce into larger bubbles in the bondline particularly if there is resin flow. Resin will bleed from the edges of bond lines. This will cause thin aluminum sheets to bow and the bondline can become thicker than the adhesive film in the center. This is where air bubbles can calesce.

I suggest closely examining what actually happens in adhesive bonding rather than developing simplified theoretical models. Adhesive films often seem so simple but getting good results with them can be more complicated than with paste adhesives where people can see what is happening and therfore technicians develop an intuitive understanding of how best to apply the adhesive.

I know this doesn't answer the question you asked. If you explain your question further perhaps we can help.

 

[MNLiaison]

DER8110: Thanks for the reply. I was 95% sure that I would have to perform a rigorous test program to get the data I wanted but I still had to ask. As a non OEM engineer it is very hard to come across bonded joint data that I can use for designing repairs to primary structure.

CompositePro: My intention was to use the methods and equations developed by Hart-Smith and as documented in the NASA CR papers to analyze bonded overlaps and not "develop simplified theoretical models" as you have suggested. My question was in regards to deriving the elastic-plastic idealization based upon actual adhesive shear stress-strain curves for a certain process and how it can change when the bond process changes.