Aerodynamic smoothness of rivets

[Sparweb]

Working on a turboprop aircraft whose SRM defines aerodynamic smoothness in terms of maximum protrusion beyond the skin. It's tight enough that everyone says you can't have protruding head rivets. However, the purpose of the rivets is a doubler, and the doubler is a reinforcement to an antenna, and the antenna is a blade 16" long with a 10" chord. A big shark-fin blade that has plenty of drag. I am looking at the overall situation and considering whether the rivets really have to be flush for "aero smoothness" when the antenna will add many pounds of drag in the free air stream.
Other considerations:
-Fracture mechanics of countersunk rivets is poor compared to protruding heads
-The skin is very thin, chem-milled. Even low-profile rivets (NAS1097) are knife-edge in 0.032 thick skin.
-Blade antenna is stabilized by rigid internal structure but that means there are out of plane loads on this doubler. Small, but not zero.
-Boundary layer thickness in the area is at least 1/4" (this is not a guess, I have experience with this)

I think I'm gearing up for fight but is there really one to be had?
If the purpose of the smoothness requirement in the SRM has more to do with blending edges of repairs, and overall fuel economy of the aircraft in service, and "nice to have" then could an argument for structural integrity beat it?

 

[Sparweb]

Like I said: I certainly have more reading to do.

and... Don't call me "surely"

 

[rb1957]

don't call me "shirley" ...

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

 

[Sparweb]

Circling back to something else you said:

I'd be surprised if we were looking for cracks under the head. I'd be looking for a 0.25" surface breaking crack (so the analytical crack length would be slightly longer) ... if I was using HFEC in the first place. Most of the time I try for visually detectable cracks, particularly if my plane is "a pick-up truck".

Something that I've noted among the 3rd party "modders" of the aviation industry is that there are certain individuals with out-sized influence on how the specialists in a given area behave. Since you and I are a continent apart, we are in different regions with different regulatory offices to report to. There are different individuals setting the standard at each regulator's office. Even if everyone is using Safarian's methods (and we aren't) many engineers who write these reports can't self-approve them, and fewer (if any) can approve the associated airworthiness limitations. So there's always someone looking over our shoulders. These engineers doing the government oversight get the last word, every time. That's how I can end up with a different methodology than you have to use, even if we took the same courses and read the same books.

When you suggest that nobody is "looking for cracks under the rivet head..." Oh golly, you aren't talking about Safarian or Eastin or any of that clan, are you? The folks looking over your shoulder may have steered you to use substantially different methods than my benefactors. If I ventured to omit such a detail...

Your other point about visually inspecting for cracks is well taken, and the crux of my decision today. Today's work on the design has led me to choose not to pursue an external doubler after all. The antenna is large enough to require reinforcing members (intercostals) on the internal structure. The combination of an external doubler and internal intercostals plus some other reinforcement bits will make it difficult to complete an inspection visually or by eddy current. The structure would cover both sides of the skin, preventing complete coverage of the inspection without complicating the inspection task to see 2 layers down. Not worth it.

The internal doubler concentrates all of the additional structure on the inside of the skin, leaving the outer skin surface completely exposed for DVI and HFEC (after removing the antenna of course).

 

[rb1957]

yeah, well understand different regions having different opinions. I will grow the crack from 0.05" but my typical detectable crack is a 1 pitch crack, whether I'm using detailed visual or LFEC. My idea is to make it simple for the operator ... more frequent visual inspections vs less frequent HFEC inspections. That was the way I was taught.

design is always compromise. Sure, leave the OML clear, and if you're not using CSK rivets then you've avoided one compromise. But you've still got three layers ? (skin, internal dblr, internal longeron ?). In my mind, a big advantage of an internal dblr is that you don't affect the OEM inspections (if you inspect the cabin shell); of course you're adding your own inspections.

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

 

[LiftDivergence]

... more frequent visual inspections vs less frequent HFEC inspections.

RB, just curious... how are you able to typically support ICAW based on visual inspection as the sole inspection option?

I understand that if you look at the SSID or MPD for commercial aircraft they are usually rife with GVI and surveillance inspections. And even service bulletins and DTR check forms will include visual inspections.

But in my experience they are generally used as a supplementary source in combination with other inspections to improve the DTR of a detail.

In the past when I have specified visual inspections in an ICAW it has been in situations where I was supplementing instrumented inspection, or I was in an SSI affected area or something where I wanted to layer the existing inspection with an additional one.

For something like a skin reinforcing doubler with interfastener crack details, etc. I would not generally expect that analysis would support a visual-only inspection program. I guess it depends on the net section width you assume if you are critical for residual strength. But typical POD requirements would put the detectable flaw length for DVI at no less than 1.00 inches. For a lot of details I would think once the crack reaches 1 inch long, you've not got enough life left to support 3 intervals between detectable and critical.


Keep em' Flying
//Fight Corrosion!

 

[rb1957]

took me a while to figure out "ICAW" is what I call "ICA" ... Instructions for Continued Airworthiness

visual inspections are perfectly legitimate limitations. Many OEMs use them. Of course with an external doubler you need LFEC which isn't that big of a problem (to our customers) ... no more than a typical OEM repair. Most of our mods (installing antennas) have little structural impact, and we kill it with Aluminium. The main reason for the dblr is to avoid questions if we didn't use one and to hide the CSKs of the rivets; it really does little to reinforce the hole (1" dia).

"critical crack length" ? "residual strength" ?? yeah, I know what you mean but I'm never close to those limits. Typically I'll have 1 pitch crack growth from 0.05" as the detectable (threshold) time and another pitch (making 3 in total, as the uncracked side of the original hole starts up) for the repeat interval. Max stress intensity less than 50 ksi.in^0.5 ... "nah bother" ... not worth continuing the analysis for a slightly longer interval.

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

 

[LiftDivergence]

Maybe chalk this up to differences between the FAA and Transport Canada, but:

1 pitch crack growth from 0.05" as the detectable (threshold) time

I have honestly never heard of that being used as a threshold setting criteria. In my experience with FAA ACO guidelines, if you are using LEFM to set a threshold it would be the time from initial to critical length divided by 2. There are situations where your criteria might be conservative, but very likely not if you are using a visual only inspection. With a visually detectable length of 1 inch and an initial flaw of 0.05", the overwhelming majority of the life will be spent before the crack is detectable and you could be well advanced into Region II before you ever start inspecting. My point is if you are using DVI only, it's not going to leave you much time to get multiple inspection opportunities in and your repeat would have to be very short.

The main reason for the dblr is to avoid questions if we didn't use one and to hide the CSKs of the rivets; it really does little to reinforce the hole (1" dia).

Maybe that's a bit tongue in cheek, but the primary purpose of the doubler is to provide a structural load path for the skin load in absence of the removed material, no? Even if the cutout is "small" at 1 inch.

another pitch (making 3 in total, as the uncracked side of the original hole starts up) for the repeat interval. Max stress intensity less than 50 ksi.in^0.5

So I understand that you are cutting the analysis off before failure due to residual strength or critical stress intensity, stopping at 3 pitches based on some kind of patching technique I assume. But,
1. I don't get what you mean by "another pitch (...) for the repeat interval. Are you taking that time and dividing it up? Or are you just saying the time from detectable length to 3 pitch length is your repeat? Because I don't see the relevance of that. Regulations would dictate that we need multiple inspection opportunities, right?
2. "as the uncracked side of the original hole starts up"... Again, maybe a difference between FAA and Transport Canada, but isn't the requirement to examine the simultaneous action of primary and secondary flaws?

visual inspections are perfectly legitimate limitations. Many OEMs use them.

Yeah, but not as the primary inspection basis for a reinforcing repair. I'm quite familiar with SRM standard skin repair inspection requirements. They might include visual inspections, but they always have some form of instrumented inspection. It's really about layering the inspection options to achieve the desired DTR.

Not trying to overly scrutinize / criticize. It's just I've worked in the FAA system a lot and with a lot of DERs, a lot of STCs and major repairs, and there are pretty standard ways of doing things. So I'm always curious when I see different techniques being used.

Keep em' Flying
//Fight Corrosion!

 

[rb1957]

threshold "time from initial to critical length" yes, that is the current typical (and what I do, I was trying to be economical with my words). Mind you in the very eary days Tom Swift didn't want threshold, and the first definition was time to detectable crack (which makes sense) and isn't that different to time to critical.

my 1 pitch detectable gives me plenty (or sufficient) time for repeats. We have low cycle planes, typically I get 5 year repeat intervals.

"Maybe that's a bit tongue in cheek" ... nope, a riveted on doubler does little to reduce the stress concentration at the hole; ok, it typically reduces the skin stress to 2/3rds. The most it does it to produce a higher fatigue issue at the load transfer rivets, which is what we're inspecting.

I start from 0.05" a hole loaded hole. grow to the next fastener. The next fastener has a "continuing damage" flaw, and the original hole has a 0.005" continuing damage. grow this to the next hole (3 pitches in total). inspection intervals are enough, residual strength is plenty; extra analysis not worth it.

"the requirement to examine the simultaneous action of primary and secondary flaws" ... yes, as explained above. The initial hole has a 0.05" and the others have 0.005". The initial hole is going to initial a crack at some time, but I don't "need" a 0.05" and a 0.005" at the initial hole.

"Yeah, but not as the primary inspection basis for a reinforcing repair." .. well that depends on the reinforcement. In the case being discussed (yes, let's get back to that !) with an internal doubler, I'd be happy with an external visual of the skin (being FCBS). Depending on design details I might not need inspection of the dblr (at all).

I've worked with OEMs and now with a "mod shop". What we do isn't the same as an OEM (mostly because we don't have the detailed data the OEM has). So we make simple conservative assumptions and analysis. Most of our work is on fuselage mods, which greatly simplifies the fatigue spectrum. (Btw, I hate the Chicago Office methodology, even though I know many people use it and it is of course "blessed" by the FAA.)

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

 

[LiftDivergence]

Wasn't referring to Chicago, I haven't really worked with them or adopted any of their methodology. I was speaking from experience mostly with Los Angeles, Seattle, Wichita, and Atlanta.

Keep em' Flying
//Fight Corrosion!

 

[rb1957]

ok ... I thought it was the Chicago Office that put out the methodology that most of the FAA used (as applied to the mod world) .... but no doubt your offices are like our regions ... they may be in the same family, but they do their own thing (when it suits).

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