de Havilland Comet Design changes 1

[jpaero]

I am trying to learn the specific design related issues with the de Havilland Comet. I have read the investigation report which spoke of:
1. Underestimation of stresses at the corners
2. Overloading the fatigue specimen before the fatigue test itself
3. Testing only the fwd fuse and effect of end restraints
etc

What I am interested however is the specific changes that were made to the design to make it airworthy again. Possible amends I could think of (based on my limited knowledge) were:

1. Reduction of general stress levels by making the skins thicker
2. Better materials in terms of crack resistance/growth (fracture toughness?)
3. Changing some of the cutout profiles to reduce the Kts.
4. Adding stiffners around windows to allowing stress away from the cutouts

But I was hoping for some definitive answers rather than guessing. An ideal answer would be one which compares the old and new design almost side by side. While a lot seems to written about what led to the accident, I couldn't find much by the way of specific design changes that were incorporated to make it airworthy again. Any pointers in that direction would be great! I will also welcome less than ideal answers .

 

[WKTaylor]

The following Cranfield University website appears to have a large number of technical reports related to the De Havilland Comet [all series?]…

https://reports.aerade.cranfield.ac.uk/discover

I discovered an intriguing document up front...

Behaviour of Skin Fatigue Cracks at the Corners of Windows in a Comet I Fuselage
By R. J. ATKINSON, W. J. WINKWORTH and G. M. NORRIS
LONDON: HER MAJESTY'S STATIONERY OFFICE 1962 FOURTEEN SHILLINGS

Also stumbled-on …

https://www.sciencedirect.com/topics/engineering/de-havilland-comets

FATIGUE FAILURE OF THE DE HAVILLAND COMET I
P.A. WITHEY, in Failure Analysis Case Studies II, 2001

Abstract
The de Havilland Comet I entered service in 1952, and became the first commercial airliner to be powered by jet engines. It was introduced as the flagship aircraft on the routes of the British Overseas Airways Corporation, and was hailed as a triumph of British engineering. However there were a number of accidents involving this aircraft, culminating, in 1954, in the loss of two aircraft in similar circumstances. These were Comet G-ALYP near Elba, and Comet G-ALYY near Naples. A Court of Inquiry was convened, and the task of discovering the cause of these accidents was given to the Royal Aircraft Establishment at Farnborough. The investigation explored a number of avenues, and finally gave structural failure of the pressure cabin brought about by fatigue as the cause of the accidents. The use of fracture mechanics methods not used in 1954 has enabled the analysis of these fatigue cracks to be made, and the initial defect size has been estimated to be approximately 100 μm in the case of G-ALYP. This is not incompatible with the manufacturing techniques of the time, and information regarding cracks in the cabin identified during manufacture. © 1997 Elsevier Science Ltd.

and many more 'related' books for sale...

An intriguing thought from tiny brain. The Comet was De Haviland's first 'cut' at design/test/build/make of a large pressurized aircraft. Boeing already had hundreds of pressurized piston aircraft flying [B-29s and it's many derivatives for instance] and was developing the B-47... so, perhaps, many of the lessons of pressurization had been already experienced by Boeing when it eased into the 367-80 ['Dash 80'] prototype... then designed/tested/built the early 707s and the KC-135A. Maybe?

SWAG, my part... the placement of turbojet engines below mid fuselage, with their high sonic vibrations [and potential for catastrophic failure with fire possibilities]... may have been a 'wild-card' to pressurization and engine safety problems. Boeing eliminated these as issues on it's aircraft by placing the engines [isolated] on pylons hanging from the wing.

Regards, Wil Taylor
o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[rb1957]

"14 shillings" now that's an old report !

another day in paradise, or is paradise one day closer ?

 

[jpaero]

Thank you everyone for your inputs!! I guess I have found "the" details I was looking for:

1- A constructional method for minimising the hazard of catastrophic failure in a pressure-cabin
(

https://reports.aerade.cranfield.ac.uk/handle/1826.2/299)

2-Pressure-cabin design a discussion of some of the structural problems involved, with suggestions for their solution
(

https://reports.aerade.cranfield.ac.uk/handle/1826.2/239)

Both papers were published right after the enquiry.

@WK Taylor: thankyou for that goldmine!!

 

[rb1957]

these do seem to be what you asked for. Instead of the design changes incorporated they look to be design suggestions. And suggestions made immediately after an accident (a tragedy) so very conservative. Remember they're only now (then) learning what is a safe design space ... knowing what is unsafe.

Yes, 10in frame spacing would reduce the skin stress, but I'm pretty sure no one does that today (when we are more confident in our approach, as opposed to immediately after the accident), and I doubt the Comet was redesigned that way. Typical frame spacing in my experience is about 30". Less than that then the frames are broken at the window belt which introduces another design problem. Also they are using a high hoop stress ... 10' diameter, 0.028" skin thk, 9psi pressure >> hoop stress 19.5ksi ... this is about 50% higher than common practice today. Possibly they added sub frames and sills around the large cut-outs ... even though these were probably not the initiation site.

At this time, immediately after the accident, they didn't have time to appreciate the impact of using the static test article for the fatigue test.

Also, they had yet to drill down into the manufacturing processes used ... that punch drilling is sensible only if you're dimpling the skin (to save countersinking in the thin skin. Even then a two stage process (drilling then dimpling) would be used today. I don't think anyone punch drills rivet holes.

But still great references to add to the library ... it'd be interesting to compare their equivalent skin thickness approach for frames as compared with a FEA.

another day in paradise, or is paradise one day closer ?

 

[RoarkS]

Punch drilling...
Vans aircraft uses turret punch to make much of their aircraft sheet metal parts (starting to get into laser cut too, but that's another story). They pilot punch the rivet holes. Vans is quite adamant about match drilling, deburring, and edge prep... but I know a lot of builders who simply cleco the punched holes and start shooting rivets.

Looking at the anatomy of a hole punch process or even shear process at cross section on the sheetmetal it's obviously a 2 stage process where actual shearing occurs up to the point where compressive force exceeds that of the remaining materials shear strength then rips it out. That ripped out part is gross. I'd never leave that on an aircraft part... but I'm guessing they didn't know yet that back then???

 

[rb1957]

I figure they hadn't had any bad experience with it (yet). It would be efficient, to punch drill and dimple in one step ... and with no reason to change to a more expensive two step, why ?

"Vans aircraft" ... look to be kit planes (nothing wrong with that, I'm working on a kit project now), but definitely the shallow end of the pond. Very thin gauge sheet metal, very low cost manufacturing. If they take care to post-process and have done it for years ... who the heck are we to throw dispersions on their process.

another day in paradise, or is paradise one day closer ?

 

[Ng2020]

...kits... typically low working stresses.Unpressurized. Low utilisation too. So the fatigue qualities of airframe design details aren't as much of a concern as in transport aircraft. Typically other design/reliability issues (systems, propulsion) will dominate airframe fatigue.

 

[RoarkS]

I wasn't trying to throw shade at Vans... more so the builders who don't understand vans instructions to clean up the punched hole.

Fatigue qualities of airframe design details aren't as much of a concern as in transport aircraft... that was a hard line to read. But I suppose you're not wrong.

Though I am reminded of a 1967 cessna 182 I looked at a few months ago with 7400 hours total time. It was kinda rough... not to mention the reason I was looking at it was the lift strut AD... it was cracked.

 

[rb1957]

me neither, as Ng2020 points out, they live in a design space where simpler manufacturing processes which would be unacceptable to large (or small) OEMs today are effective ... 'cause they reduce cost and build expense.

and yes their fatigue expectations are different to OEMs ... their durability is lower 'cause their customers expect less. Cessna is not in this category, but yet a 50+ year airplane with <10000 hrs is very low utilisation.

another day in paradise, or is paradise one day closer ?

 

[WKTaylor]

A parallel/side note that i think is relevant to this discussion... about the importance of consistency/quality of general assembly workmanship.

MANY moons-ago a couple of high-time retired 727-? fuselages were disassembled piece-by-piece for a technical/research evaluation for damage such as corrosion, fatigue, fastener-failures, spot-weld failures, etc study.

The relatively short cockpit structure and tail cone fuselage structures [compound curvature] sections were very consistent in their damage profiles, throughout.

HOWEVER... The oddest thing was noticed by the tear-down crew and verified statistically: the upper-LH crown constant-section, upper-RH crown constant-section and the floor/belly constant-sections each had uniquely different damage profiles. The researchers were baffled. The belly section was relatively uniform and was dominated by corrosion and loose fasteners. However... One crown section had relatively minor cracking, corrosion and loose fasteners [etc]. The opposite-side crown section was dominated by widespread cracking, loose/smoking fasteners, corrosion out-breaks, sloppy fits/gaps, etc. YES, the researchers were baffled.

SO, they contacted Boeing for technical assistance. The big question 'WHY' also stumped the Boeing folks... There was absolutely NO technical reason [stress/strain/aero/materials/processes/etc] as-to 'why' the crown sections should appears to have such uniquely different damage levels. Then someone got the bright idea to go back to the production floor for answers. Since 727 production had just been terminated many of the 'build' crew members were still available. The resulting insight changed the company.

Each of the [5] sections [cockpit/entry door, tapered-aft fuselage, LH upper crown-constant, RH upper crown-constant and the belly-constant] were mostly made the 'old-fashioned-way', IE: built-up by hand in huge assembly fixtures. OK... Sooooo? Well, It was discovered that the crews and leads building these sections were kept on these same sections thru the production years... IE: there were dedicated section crews/teams that rarely changed personnel thru the years... and RARELY/NEVER worked on any other section. Hmmmm... That was the key, unlocking this mystery.

The production crewmembers for the [2] upper crown constant-sections were interviewed... and what emerged was clearly stunning.

One crew had 'older/experienced lead /workers' dating back to WWII, which exhibited work and quality protocols learned during those early high production 'git-er-done' rates of WWII 1940s. This older crew had been working together on jobs like this since B-17, B-29 years and felt most comfortable working together thru the 1960s.

The opposite crew had younger leads/workers with more recent quality and production training protocols that had a greater emphasis on consistent/standardized workmanship and quality practices, etc that were more methodical, and slower, than the 'older-gray-hair team'. Assembly-close-shimming-to-fit, deburring, added touch-up primer, recurring jig/fixture maintenance and alignment inspections, etc were emphasized on this team.

The cockpit and tapered-aft fuselage teams were more of a blend of old/new workers and leads.

Yep... the upper crown sections made by the older work-team were the ones exhibiting much higher levels of damage. The upper crown sections made by the made by the younger/newer-trained work-teams had far less detectable damage and issues. In real terms the upper crown-section made by the older crew was the primary maintenance cause for removing these jets from service, in the first place. IF both crown-section's quality had been uniform... consistent with the younger crew's workmanship... these jets could have flown safely and productively and with lower structural maintenance... for many more years/hours to come.

This one study starkly revealed the immense value of consistent high quality assembly practices on useful life [fatigue and structural maintenance]. The long term resolution for Boeing was to hasten the drive for major/critical structural assemblies to be machine-assembled [wherever possible], to eliminate the issues/inconsistencies found in common 'hand-assembly workmanship'. This also led to larger/fewer assemblies built-up all at one time with machine/computing-following for every aspect of quality... tight dimensional controls for individual parts for precision fit/clamp-up, drilling, fastening, etc... and for higher quality corrosion protective finishes on the parts before assembly, so that added/touch-up finishes... including fay sealant, etc... were minimized/eliminated. And projected useful service lives have climbed above 100,000-hours and 25-years.

I wish I could find that report... had a paper 'summary' copy for years.

I personally have experienced MIL-Jets where LH/RH parts or assemblies are dimensionally inconsistent and do cause a unique host of issues during replacements... and even 'aerodynamic oddities' that were known, but understood... but were livable-with… so they were simply ignored.

Regards, Wil Taylor
o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[3DDave]

Ha - a similar microcosm: I was on a job that was updating assembly drawings. They made left-hand and right-hand systems and I noticed that a shimming process on one was different than on the other. Went to the factory floor to see what the assemblers actually did - turns out they never noticed there was a difference, with each of two available assemblers expressing surprise at what the other one had been doing. This in an FDA regulated facility.

Having a steady and dependable workforce can avoid making the "stupid" mistakes but can also allow poor practices to continue for long periods of time resulting in dependably inferior products. Sometimes getting a new guy on the job that everyone has to explain "why" to can help understand when the "why" is flawed.

It would also be recommended that stress and design engineers be on the fabrication and assembly teams to see if what they thought was going to happen actually did happen. Too often they don't want to get near metal chips and, similarly, too often the factory guys "know the job and don't need anyone looking over their shoulders."

 

[RoarkS]

I legit had an intern call me out on an assembly last night.

He was match drilling some ribs to a trim tab spar. One of them was misaligned by about 1/32". I said absolutely not, throw that rib out and get a new one. He said where on the print does it call that out...

The moment he said it I realized he had me.

Took a new guy to find something out of control that's been in in production for 3 years.

 

[rb1957]

what sort of rivet usage) ? one in a line of rivets, or something "special" that Had to be on location ?

1/32" is only 0.03 ... typically within drawing dimension tolerances ?? no?

Unless this was misplaced with the match drilled partner ? In which case why not go up a size ? eD issues ?

are you Inspection or Liaison ? Was the part not to B/P ? Would it be flagged by Inspection ??
If so, then typical non conformance paperwork (but no "gotcha")
Possibly match drilling is a process with understandably tight tolerances ? (or the assumption of tight tolerances ?)

another day in paradise, or is paradise one day closer ?

 

[berkshire]

Here is something on punch riveting, I do not know if this is the system that De Havilland used but it sounds similar.

https://www.sae.org/publications/technical-papers/content/830600/although

the date is much later than the 1950s
B.E.

You are judged not by what you know, but by what you can do.

 

[WKTaylor]

I've encountered a weirdly similar sounding problem with a KC-135A, +25 years ago.

In this case a spar cap was corrosion damage and needed the segment replaced [cutout/inert]. A straight-forward repair [after I designed the splice and had it approved].

After drilling/punching-out the old 3/16 and 1/4-D rivets, eddy current NDI detected [2] perfect/parallel straight 'defect line-indications'... tightly spaced thru-the-hole... on a couple of widely spaced holes. Close inspection [10X] revealed the 'moon-shaped slivers' of smaller 1/8-Dia rivets that had been partially drilled-thru for the final rivets. Huhhh??? Getting help from Boeing, someone identified the fact that the sub-assembly drawing allowed used of 1/8-Dia tack-rivets at various locations for holding parts together... which should have been fully drilled-thru/enlarged during final Assy for the fastener installs... but never were. Unbelievable.

Regards, Wil Taylor
o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[RoarkS]

rb1957:

Rib was misplaced to the spar. would have had a crease in the skin it was so far off.

Nope. Just an engineer that was watching over in intern learning the shop floor by doing a few simple projects.

 

[rb1957]

well if it wouldn't fit that's a thing.

you were looking over the shoulder of an intern (who was doing liaison) ? it's way too common a trend ... having interns doing liaison. in my day only the oldest, narliest, crankiest guys did liaison.

or you were the intern and had ventured down to the floor (pit?) to see what was going on ?

another day in paradise, or is paradise one day closer ?

 

[Sparweb]

in my day only the oldest, narliest, crankiest guys did liaison.

Huh. Back when I was doing liaison, I didn't see you venture downstairs too often to join me, RB.

 

[RoarkS]

Them some fancy words. My initials are on just about every blueprint. I was just showing the new kid around. He'd never heard of match drilling, #40, #30, 470, 426... so on.

 

[rb1957]

let's get back to the Comet !

another day in paradise, or is paradise one day closer ?