Detrimental Effects of Re-working Previously Cold Expanded Holes (Fatigue) 1

[Cevers1988]

Firstly, this is my first time posting, so hello everyone! I am an F&DT Engineer working in Aerospace, so that sets the scene.

Ok to the point...

During a repair scenario (aerospace wing repairs), I have come accross repair manuals that depict the following:

Oversizing of Cold Worked Holes
The maximum recommended allowable oversizing of previously cold worked holes for repair purposes only are given below:
- From nominal diameter 4.83 mm (0.1902 in) thru 12.7 mm (0.500 in) the maximum allowable oversizing is up to second oversize (nominal + 0.794 mm (0.0313 in)),
[highlight #FCE94F]- For nominal diameter 14.29 mm (0.5626 in) the maximum allowable oversizing is up to next nominal,[/highlight]- From nominal diameter 15.88 mm (0.6252 in) thru 28.58 mm (1.1252 in) the maximum allowable oversizing is up to second oversize (nominal + 0.794 mm (0.0313 in)).

This apparent 'hyperbolic' behavior seems unusal; given that generally, CX residual stress fields increase in dia proportionately to bore dia. I have a few suggestions as to why:

1) Availability of fastener sizes in that category (repair procurment is less predictable than FAL for example).
2) Original justification for such limits may have arrived from a specific case, one of which has proven, through in-service experience, to remain valid.
3) Tolling optimisation at this bore size providing extra benefit (very tenuous )

Now to the questions...

An addendum was recently circulated stipulating 'all locations above 2nd oversize will require re-cold expansion'.

Naturally, which do we believe to be correct (both carry no formal justifaction, that i know of)?

Posing the follow on question, if we are to re-coldwork a location with 'existing compressive residual stress' (due to the conflicting nature of the higlighted repair manual and the addendum) would there be much cause for concern regarding overloading and causing some compressive corrosion cracking?

having looked around I can only find one reference that touches the subject, but I am reluctant to buy it as I am unsure about the angle the analysis is covering (potentially looking at the effects of relaxation etc..)

I look forward to picking the brains around here!

Evo

 

[rb1957]

i'm interested to see Will Taylor's response to this.

i take it that you mean "just" reeming out the CX'd holes to accept a larger fastener. it sounds very odd to me that this is acceptable, hopefully someone has some hard data on it and not just "i heard that someone read it in a manual". the point as i see it is that the material on the bore of the hole has had the most compression, so reeming it out seems like a bad thing to do ?

 

[Cevers1988]

rb1957...

The stress field occupies a region roughly 2D from the bore centre, which varys depending on material, level of CX (4% in most cases) etc. The reaming and oversizing eats into this region as we O/S.

To reiterate; I am unable to quantify the compounding effects of additional compressive stress due to re-cold expansion, on the fatigue life of the bore.

Evo

 

[rb1957]

yes, i know the compression field extends into the body of the part, but it's peak is at the bore of the hole.

reeming out the hole removes the most beneficial material.

if i had to reinstall a CX'd fastener, i'd repeat the CX process.

 

[Cevers1988]

We assume a minimum benefit at O/S levels, and incorporate that into DFM calcs. Allowing an inbuilt repairability, keeping maintainance cost down.

Your suggestion would imply that there is no knockon effect of re-expanding with existing Compressive Stress in the body and bore. Any literature on that?

Evo

 

[rb1957]

no literature, just a hunch.

but if you're adjusting the calcs to reduce the interference effect that sounds ok (mind you, where's your literature for that ? c/- tests ?)

 

[Cevers1988]

The SRM guidelines are were justified years ago and have not been associated with any failures (so can be assumed to be an acceptable baseline). I just haven’t been able to locate the specifics of the calcs.

The second addendum was issued to simplify the repair procedure for a specific repair retrofit task; One that is very complex and covers thousands of bolt locations. And as the rework of CX is carried out ahead of the SRM guidelines, it would seem to be a 'conservative approach' was created whilst simplifying. As a consequence, not many of the staff here are looking too closely at the effects of such compounding (if it even exists!).

As I cannot locate the original calcs with ease, I was hoping to get some second opinions. Not to challenge the original SRM per say, but rather to rationalise the latter addendum and it's innate assumptions or lack of.

I hope that was concurrent.

 

[rb1957]

am i right in reading that the later addendum changed only the 9/16" fasteners ? (down from next nominal size to 2nd OS)
(since i read the OP as saying all other sizes were limited to 2nd OS)

isn't much of a change, maybe done to standardise maximum size ?

 

[Cevers1988]

Yes, that is probably why it was changed - It was in response to confusion in the repair instructions being handed to the shop floor.

As you said before; your standard practice is to re-cold expand at any opportunity (if taking the peak benefit at the bore into the calcs), which makes total sense. However, you could be entering realms of 'over expansion' if the reaming doesn't remove all of the yielded material.

The same comparison could be made for the reduction in limits of re-work to 2[sup]nd[/sup] O/S from the 1[sup]st[/sup] nominal original.

I will park this for now and look for more detail. I think I may be barking up the wrong tree, 4% expansion - followed by 'early' 4% expansion may lead to >4% expansion of the hole, depending on penetration of the stress field. Which can lead to destructive effects as shown in the link...

Evo

 

[rb1957]

interesting link ! ... i'd've thought it was tough (in the real world) to CX a cracked hole.

 

[WKTaylor]

CEVERS1988...

The world experts on cold expanding holes for a huge variety of purposes are the folks at Fatigue Technology Inc [FTI

http://www.fatiguetech.com/

]. They are the ones that patented split-sleeve hole cold working [cold expand "CX"], CX to size, CX crack-stop, Forcemate bushing [CX-Bush in-place], Forcetec/Forcetec-Lite [CX nutplates in-place], etc... for both metallic and non-metallic structures. They are happy to help engineers, working with their various CX systems, understand the nature of each and every process.

The class FTI-CX split sleeve cold work process You are describing has been tested extensively and the residual stress mechanics are well understood by the FTI guys: this is why there are limits on potential rework that seem squirrely.

See

http://www.fatiguetech.com/coldEX_Split_sleeve.html

for the classic split-sleeve CX process.

When a hole is CX’ed within normal limits there is a residual compressive [plastic] strain field induced around the hole bore… and around this compression field is a wider [elastic] tension field. As I remember it, the FTI folks stated that the practical maximum ream-out for any hole before loosing life enhancement value is ~1/4R for high-CX holes in aluminum… and much smaller [~1/8R or smaller] limits for VHS/UHS [high yield] steel or titanium, etc. When a hole is enlarged, the important inner-layer of the residual compression yield field is eaten-away around the hole bore... to a certain limit where the residual stress is too low to be of an useful value for fatigue protection [a whole another discussion as to "what/why" that is]. When this limit is reached the hole MUST be re-expanded to regain the fatugue-protective residual compressive field for the intended replacement fastener diameter.

WARNING: the path for routine oversizing of CX’ed holes is NOT always a straight forward path to take. The new pre-CX starter hole has to be larger than the existing hole, but significantly smaller than the intended fastener to allow for the hole expansion [forced enlargement] and finish-reaming to the intended hole size. This is where practical aspects factor-in: Up-to 1/2-inch Dia, fasteners increment are in ~1/16's of an inch + 1/64 [1st OS] and 1/32 [2nd OS] oversizes. The next theoretical increment of 3/64" [3rd OS] is rarely available in lock-bolts [some HL pins, limited availability] and are never available in protruding or flush-head bolts of any kind. So the size increment is usually forced to go from 2OS to next larger nominal diameter. This situation gets really dicey for fasteners larger than 1/2-in [0.500 Dia]: the increments typically increase in 1/8-inch [0.125 Dia] + 1OS and 2OS…. Leaving a whopping ~3/32” size-jump for the next larger nominal diameter. Soooo… for large diameter fasteners the picture gets ugly fast… and really bizarre when replacing large diameter taper-Loks in thick structure [truly butt ugly what happens]. There are work-around procedures for “some” of these large/odd-Dia OS fasteners, such as force-mating repair bushings in each layer, to establish a small hole bore for reaming-for/installation of a nominal diameter fastener… or sleeving a nominal diameter bolt [old back-pocket liaison-last-resort trick of the trade that not all will agree with]. Where this whole process of CX’ing holes reaches an ultimate limit is when the edge margin X material thickness [e/d X t] falls below a certain critical value [FTI has tested for]… and the CX causes deformation to the edge or will not hold the residual surrounding tension field

CAUTION.
My discussion uses terms used by FTI. Suggest You absorb tech info available on the FTI site and contact an FTI tech rep to assist You understanding the full implications of the CX process/variables/etc… much better than this rambling tirade. This is NOT a rocket science technology… but does have some sophisticated elements worth knowing beforehand… especially when repairing critical structure.

Hope this make sense: I’ll probably wonder what bozo wrote this, tomorrow (I’m just getting back to life after a tough bout with the flu).


Regards, Wil Taylor

Trust - But Verify!

We believe to be true what we prefer to be true.

For those who believe, no proof is required; for those who cannot believe, no proof is possible.

 

[Cevers1988]

wktaylor...

A very succinct reply, covering most bases apart from the actual question at hand: 'Potential destructive compounding of compression with multiple applications of CX post O/S. For one, does it exist as a practical issue (theoretically it can happen if the expansion >5.5% roughly). Or are we confident that reapplication is always a beneficial process, in the context of reducing the reworking limits set by the SRM?

I feel we are deviating from the point 'tis all

The O/S of the fasteners and the process of CX is familiar, and we don't (not a standard practice) use 3rd O/S, next nominal post 2nd O/S is the norm here.

Evo

 

[WKTaylor]

OK... You proved my theory.. I wasn't all-there, last reply. I’ll take another shot…

Re-CX-expanding a hole that has already been expanded with the same mandrel and [new] sleeve has NO real/net effect on the hole and surrounding material.

NOTE.
In the real world this should not happen since each CX'ed hole has a residual lengthwise ridge formed when the mandrel is pulled thru the sleeve and the sleeve opens up large enough to form a tiny gap between the sleeve edges [length of hole bore] which forms the distinctive ridge: a dead giveaway that the hole has been split sleeve CX’ed [SS-CX’ed].

Exception: The only [trivial] effect on SS-CX’ing a hole is if the [new] sleeve is rotated so that edge-over-lap is rotated-away from the original sleeve orientation: the old sleeve ridge will be crushed-down considerably; and the old/new sleeve-ridges will be barely visible [if at all]. Per FTI testing, the sleeve-ridge and its orientation make little/no difference on fatigue performance.

OK You are determined that the old hole needs to be SS-CX'ed again. For this to happen, the hole will have to be reamed to a new starting diameter that is larger than the existing diameter... within the constraints of the existing FTI CX tooling. Guess what? You generally run into the FTI next repair size tooling that would be required. The new starter hole will effectively bore out enough of the strained material so that the next round of tooling would meet intent getting the hole ready for the next practical size fastener with relatively nominal residual expansion. These FTI folks are very clever indeed.

OK so You are really determined to get the hole supersized: somehow You identify a starter hole size, mandrel and sleeves barely small enough to fit the old CX'ed hole; or a mechanic drills the hole over size slightly, so you can stack a sleeve on a sleeve. You have now entered the twilight zone of SS-CX.
Holes already cold expanded will see greater mandrel/sleeve expansion/drag forces if re-expansion attempt is made under these circumstances. You have not lived until You’ve collapsed a sleeve, broken a mandrel or had a mandrel stall during the pull.

But suppose You some how get away with these higher pulling forces and are able to hyper-re-expand a hole to the same degree it was originally expanded. There is likely to be massive swelling/puckering/distortion around the hole lips. A single thick part [such as a lug] will have obvious volcano lips [both sides] that would have to be ground/machined down to restore contour. This machining requirement is also indicative of another problem: the residual compression-tension fields and machined-away distortion will likely grossly reduce static strength of the lug.
NOTE.
Where single expansion may not have remarkable distortion, lugs may also suffer terrible distortion [bowing/twisting/stretching/swelling] with hyper-expansion.

When this hyper expansion occurs in multiple layers of material that are not super-tightly compressed together [almost impossible] then the hole lips of each layer independently experience this swelling/puckering/distortion phenomena… leading to the entire stack-up swelling/gapping thru the thickness. This situation results in an entirely unacceptable structural condition due to the embedded stresses, gapping of surfaces between holes and the subsequent shanking of fasteners… if post SS-CX surface machining is not accomplished.

I am SURE the FTI folks could tell You a lot more about abusive SS-CX [etc] practices: but when it comes to grossly irrational actions they probably don’t test beyond conditions/limits that would never [should never**] exist.

** OK OK OK OK: nothing can be made fool proof because fools are so ingenious. Also, anything that can go wrong will eventually go wrong.


Interesting exercise.


Regards, Wil Taylor

Trust - But Verify!

We believe to be true what we prefer to be true.

For those who believe, no proof is required; for those who cannot believe, no proof is possible.