Sandwich Repair 3

[irq]

Hello,

Please see attachment - there is a sandwich panel. Let's imagine it is loaded by pressure acting on the top, so that the skins are loaded in tension/compression. The aim is to design the part joining two separate sandwich parts. In many repair instructions we can see the method (1), where the core in the middle is replaced by the new one. I want to propose method (2), glue both panels and add extra plies on the top/bottom. Do you see any disadvantages compared to the method (1)?

Is the re-qualification of the component or further analysis after repair according to common techniques (e.g. ply overlapping 60t or other techniques according to repair manuals etc.) required? Do you have any good reference for repair of sandwich constructions? Can the materials based on phenol and epoxy resin be mixed together?

Thanks for your advices.

 

[RPstress]

What is the difference between the 'adhesive splice' and the 'adhesive bonding'?

Structurally, as long as you've got sufficient core-shear and skin-endload connectivity it should be ok. Just foaming adhesive (possibly referred to as a 'core splice') should work for shear. Make sure you have adequate pressure applied. You don't want the core splice to be too thick, so try to keep the gap between the pieces of core controlled. A maximum gap of 1/2", and preferably less. Make sure enough foaming adhevive is put in to fill it (duh! Manufacturer's data sheet should guide as to amount of foaming, but experience probably more important). If the core is honeycomb the gap should be a rough average.

You show the original skins overlapping your 'adhesive splice.' Is the intent to 'core out' some of the original core material? Having a big gap will increase the chances of an undersirable outcome. There's very little support given to the skins by the adhesive as it foams. I don't actualy know what the internal foam pressure is but I'd be surprised it's a couple of psi.

Your shear overlaps should of course be adequate. While the overlaps shown aren't too bad, often 0.5" per ply would be used and in an overlap patch each ply should probably have an overlap of 20t or so. A simple pyramid patch (usually match the parent layup properties plus one overall 45° ply) would be probably be ok (I assume composite skin—if Al or similar no need for such elaboration). While just cure under vacuum might be enough, 30–50 psi (keep it as low as the original cure of the honeycomb panel) in an autoclave would help. NB: in diagram had to add the '-'s at line start as for some reason it made multiple spaces into one space in the preview window.
[tt]
------------ ________
------- ____/________\____
-- ____/__________________\____
===============....==================
-| | | | | | | |..| | | | | | |
-| | | | | | | |..| | | | | | |
[/tt]

Well, I *think* this monospaced 'diagram' make senses. The 'preview' area is a bit variable.

Some people would splice the core as a separate op from putting the patch on. That way there's no chance of a composite patch sinking into the gap between the original skins and the splice is easy to clean up and inspect.

I think you'd only need a separate block of core if the gap to be filled was a bit big. To sort of reiterate, if it's more than 1/4" to 1/2" I'd think about making it even bigger and putting in some core like (1). You run the risk of inadequate through-thickness support during cure otherwise.

Your second para concerning qualification depends a lot on what rules this is being done under. Is it a production aerospace concession? A field aircraft repair? A public walkway bridge handrail? A bit of your garden shed? Etc. I personally would always expect some sensible calcs to be done, but that's my background and there may well be no need. Even with aircraft, it used to be the case that a lot of repair was done based on common sense and replacement of area (I may well be out of date here; this is pretty remote memories from the '80s).

I don't have any documents explicitly covering repair of sandwich. The usual suspects such as the DIAB and Hexcel manuals will help, but you may already have seen these.

http://www.diabgroup.com/europe/literature/e_pdf_files/man_pdf/sandwich_hb.pdf

http://www.hexcel.com/Resources/Dat...eets/Honeycomb_Sandwich_Design_Technology.pdf

http://www.hexcel.com/Resources/Dat...heets/Honeycomb_Attributes_and_Properties.pdf

http://www.hexcel.com/Resources/DataSheets/Brochure-Data-Sheets/Composite_Repair.pdf

You might also check out Care and Repair of Advanced Composites by Armstrong. A bunch of $/£/€, but worth it.

With regard to 'mixing' phenolic and epoxy, certainly after they are cured they may be bonded together, although the bond to the phenolic won't be as good as you might think (very roughly halve values for bonding cured epoxy with epoxy adhesives). Co-bonding (with either the epoxy or the phenolic uncured and the other cured) should sort of work, with a similar reservation on the shear strength. People with more experience than I of things like aircraft interiors may well know more. Also of course beware that epoxy will be more dangerous in any environment where it might burn.

 

[Compositepro]

Here are some useful documents.

http://www.tc.faa.gov/its/worldpac/techrpt/ar00-46.pdf

http://www.tc.faa.gov/its/worldpac/techrpt/ar03-74.pdf

http://www.hexcel.com/Resources/DataSheets/Brochure-Data-Sheets/Composite_Repair.pdf

 

[blakmax]

If you use foaming adhesive it is essential that you do not apply high vacuum to pressurise the repair because that will result in over-expansion of the foam and result in large void cavities with low strength. Our guidelines are to use a maximum of 10 inches Hg.

With regard to the single shot approach where the repair is performed in one hit, you run the risk of having excessive foaming adhesive which may cause lifting of the external patch, or insufficient foaming adhesive which may result (as already stated by RPStress) in the composite patch distorting inwards into the cavity. I would strongly recommend a two stage process. Cure the insert, then sand off the excess adhesive from the surface and apply the external patch.

Another issue to be considered is weight. Foaming adhesive is much heavier than core, so if your repair is on a weight-critical structure, the additional weight may be unacceptable.

Regards

Blakmax

 

[irq]

Thanks to all of you for very informative suggestions regarding option (1).
All difficulties with expansion of foaming adhesive and appropriate support of the skin can be avoided by more simple cold repair method (2), which was very little commented (most likely due to lack of more detailed description). In first step both parts can be produced either with or without foaming adhesive (if the first option does not work). In case foaming adhesive is not used some of the core material at the edge to be bonded will be removed, the core cells at the edges filled with adhesive (see attachment, blue area) and sanded. Finally both parts are glued together, repair plies with overlap of 0.5" applied on the top/bottom and painted with resign (wet lay-up repair) as already shown in the first picture. One additional ply can be added to account for a loss in strength caused by differences in repair processing (curing in room temperature) and part fabrication (press/autoclave). The gap already mentioned by -RPstress- will be controlled since large areas of adhesives can make the repair stiffer and heavier than original part and the changes in load path should be avoided. Any comments regarding this technique?

 

[Compositepro]

In repair panels that I have done there is no gap. The repair plug is cut slightly large so that the cell walls of the separate pieces of core actually interdigitate.
In most cases, I think you want full vacuum during cure. It removes air from the repair which will expand during cure of the patch. This is particularly crucial for non-autoclave repairs where you have to be careful not to lift-off the patch.

 

[SWComposites]

1) the "glue" used to bond the two parts of filled core may not restore the required core shear strength
2) you cannot "guess" that one additional ply is sufficient to restore strength with a lower strength repair material; you must have material properties for both the original and repair material and do a suitable analysis
3) adding a "large area' of adhesive will not necessarily make the repair area stiffer
4) making a part heavier does NOT automatically make it stronger

 

[irq]

1) the "glue" used to bond the two parts of filled core may not restore the required core shear strength
--- the adhesive strength which bond both cores is higher then the shear strength of the core = ok
2) you cannot "guess" that one additional ply is sufficient to restore strength with a lower strength repair material; you must have material properties for both the original and repair material and do a suitable analysis
--- repair is usually made with equivalent material as original part, the only difference is the manufacturing method, which impact its strength. One repair ply is added to connect both parts (or more to rebuild / get the similar lay-up), one extra ply is included for safety (see my last post)

 

[SWComposites]

1) the "glue" used to bond the two parts of filled core may not restore the required core shear strength
--- the adhesive strength which bond both cores is higher then the shear strength of the core = ok
> do you have sandwich panel shear data with this type of joint to validate this claim? Using adhesive lap shear data vs core shear strength is not appropriate

2) you cannot "guess" that one additional ply is sufficient to restore strength with a lower strength repair material; you must have material properties for both the original and repair material and do a suitable analysis
--- repair is usually made with equivalent material as original part, the only difference is the manufacturing method, which impact its strength. One repair ply is added to connect both parts (or more to rebuild / get the similar lay-up), one extra ply is included for safety (see my last post)
> you are only using one repair ply to connect the two pieces? how is that sufficient?
> do you have test data or an analysis using material properties to justify that only one extra ply is sufficient? you cannot just say, well it looks good, should be enough ........