Avro Tudor 2 windows

Eddie
Ah no. The Comet, like the other early pressurised aircraft was designed to what’s called safe life philosophy which means by design there’s no cracks. If there’s no cracks you don’t need know about crack progression rates. It worked for Vickers, Douglas, Boeing and Lockheed and even DH when they finally got it right on Comet 4 /Nimrod.

The cracks progression work feed into the next generation of designs called damage tolerant philosophy which are intended to be safe even if cracked.

The basic understanding of metal fatigue started before WW1 and the very first aircraft fatigue test was conducted by the RAE in 1918. After WW1 the understanding of metal fatigue progressed in the UK, America and Germany, sometimes with quite independently discovering the same characteristics and the recommendation of similar approaches. Palmgrens basic fatigue linear progression description of metal fatigue failure from 1924 is all that was need to successfully design a pressure vessel cut out, and follow on work by Miner (e.g. ”Miners Rule”), etc was well known in the aircraft industry just before WW2

I would point out that the principles of fatigue crack growth are really more recent. Miner’s rule is more concerned with fatigue initiation, whereas fatigue crack growth (which was really the issue for the Comet) wasn’t really well understood until well into the late 1950’s/early 1960’s, where Paul Paris was able to show the relationship between stress intensity factor and fatigue crack growth rate. Some of the behaviours are still debated at length (e.g. effects of crack closure).

The ‘square windows’ is certainly an oversimplification. One thing that has occurred to me is that if you repeatedly pressurise a airframe of that type, so it fails through fatigue stress, it will always (eventually) fail at the windows (or other stress raising holes) no matter what shape they are.

On thing I’ve heard said about the Comet I is that because the engines were not as powerful as had been hoped the fuselage skin had to be reduced in thickness to save weight; if the fuselage had been thicker its failure would have occurred later, even with the ‘square’ windows.

One thing that puzzles me is that the investigation into the Comet I crash is always said to have discovered a completely unknown cause of failure, pressure stress fatigue failure, and that the loss of the Comets led to a new understanding of how to control this in subsequent (pressurised) airframes. What puzzles me is why this was only ever a problem for stress caused by pressurising airframes; the materials used are the same as non-pressurised airframes and they must have to resist stresses that are not caused by pressurisation (hence rounded corners on windows)?

Imagine a heavily laden Lancaster bouncing along at 18,000 feet in flak-filled sky; there must have been plenty of repeated loading and unloading (stressing) of parts of the stressed skin airframe but the Comet disasters were apparently caused by the single stressing and unstressing of the airframe due to pressurisation during every take-off and landing cycle? This is what the famous ‘water torture’ test was to investigate, and what it ‘discovered’.

My question is this: if the stresses caused in the airframe are no different if caused by pressurisation or normal loading, then why is the Comet disaster always assigned to something that was ‘not understood’?

Just like the myth of the square windows, the aero industry “discovered a completely unknown cause of failure, pressure stress fatigue failure, and that the loss of the Comets led to a new understanding of how to control this in subsequent (pressurised) airframes” is also a myth;- In reality DH learnt a lot about what the rest of the industry already knew. The Lockheed Constellation, Boeing B29/Statocruiser, Vickers Viscount where all pressurised designs from around the same time which didn’t suffer from the same problems. I believe Boeing where already designing crack stoppers into the B47/52 pressurised fuselages before the Comet failures.

The basic understanding of metal fatigue started before WW1 and the very first aircraft fatigue test was conducted by the RAE in 1918. After WW1 the understanding of metal fatigue progressed in the UK, America and Germany, sometimes with quite independently discovering the same characteristics and the recommendation of similar approaches. Palmgrens basic fatigue linear progression description of metal fatigue failure from 1924 is all that was need to successfully design a pressure vessel cut out, and follow on work by Miner (e.g. ”Miners Rule”), etc was well known in the aircraft industry just before WW2, that is except, DH……. who at the time were doing great things with wood which doesn’t fatigue. By the early 1940’s the aero industry was generally designing in metal and had a systematic stress sign off of component schemes prior to manufacture, whereas in DH, the component designer i.e. not a stress specialist, could decide if he wanted the stress engineer involved or not, perfectly acceptable for wood. Remember the Comet was not DH’s first disaster with metal construction, they got the basic static sizing of the first prototype DH110 wrong at about the same time they messed up Comet. In essence the majority of the aero transited from wood to metal in the 30’s and their early mistakes were largely hidden in WW2 losses. DH tried to catch up too quickly, on too big a project and failed very publically.

It’s not very well known but DH actually did a fatigue test of the Comet Fuselage prior to the crashes but their understanding of what this proved was so poor it merely led them into a false sense of security.

Unbelievably in these modern times, the then press at the time of the crash investigation, made excuses for DH mistakes and published myths that have endured to this day.

As for understanding the difference between a Lancaster and a Comet fuselage;- A non-pressured fuselage is in essences a beam in bending. Placing heavy items, such as gun turrets, at the end of the beam adds significantly to the sizing case for the beam, if not dominating it. A beam’s skin thickness may vary over its length and still support the end load. The operational limit load is big and the normal purtibations are small. But for a pressurised fuse it’s a whole different ball game, in that the vast majority of the fuse is now designed by the pressure load and its associated proof factor alone. The normal operational case is close to the limit load every flight;-Nasty. This envelops the beam bending cases and means the skin thickness must be constant over the whole pressured section. In general these older aircraft were not overdesigned, and alas tended more towards the under-design.

Hi Mike,

Not sure of the ultimate provenance but I found that quote here:

http://www.airpages.ru/eng/mn/lan_01.shtml

Claims to be extracts from the ‘Maintenance Manual’

Cherry Ripe
Is this quote referring to the Lancaster?

This framework is covered with alclad (a light aluminium alloy) sheet, the majority being 22 SWG but in some places 16 SWG. This is riveted to the formers and stringers with dome-head rivets.

If so, I wonder if you can give the source.
Many thanks
Mike

How do you work-out the stresses due to pressurisation?

It is a standard result, look up “thin-walled pressure vessel”.

For the Comet, pressure differential 8.25 psi:

The skin was generally 22 s.w.g. (0. 028 in.) except along the sides of the fuselage, where 20 s.w.g.
(0. 036 in.) skin contained the windows (Fig. 3). Skin material was D.T.D.546.

( I found this document which is very readable: http://naca.central.cranfield.ac.uk/reports/arc/rm/3248.pdf )

Alarmingly, that’s remarkably similar to the Lancaster, though using different alloy:

This framework is covered with alclad (a light aluminium alloy) sheet, the majority being 22 SWG but in some places 16 SWG. This is riveted to the formers and stringers with dome-head rivets.

From what I can find online, the revised Comet 4 had 18 swg fuselage skin thickening to 16 swg around cut-outs. Designed pressure differential was 8.5 psi.

The Britannia had fuselage skin gauge of 19 swg ( 0.040 inches ) for pressure differential of 8.2 psi.

Regardless of the reason for a crack starting , surely the result will likely be very different from a pressurised to a non pressurised aircraft ? A smallish crack in a non pressurised aircraft ( in a non critical part like around a window ) will likely cause no harm untill it can be repaired . The same crack in a pressurised situation could cause an explosive decompression ( and they dont call it ” expolsive ” for no reason ) .

Are the stresses that different? The fuselage skin must resist both in exactly the same way (and the only way that a thin flat sheet can resist); by being stretched end-to-end.

How do you work-out the stresses due to pressurisation? Good question!

Edit: Actually, if we look at a metre long section of fuselage, and assuming it fails fore-and-aft at the top along the centre-line, then the force that causes the failure is given by the pressure difference (above atmospheric) multiplied by the area that divides the left and right-hand sides of the fuselage…..maybe. (Static situation only, obviously.)

Question: how thick is the fuselage skin on a Comet I (and a Lancaster) and what is the diameter of the Comet I fuselage?

I bet there isn’t much difference in skin thickness; either the Comet is way too thin or the Lancaster is seriously over-engineered!

Agreed that the Lancaster was a semi-monocoque, although it might be interesting to compare the gauges of the fuselage skins and the wing skins, but the loads and hence stresses are very different between pressurisation and the bending of the fuselage (basically a large deep beam).

The incident in Hawaii was the Aloha flight 243 737

http://the.honoluluadvertiser.com/2001/Jan/18/118localnews1.html

The airframe had been poorly maintained and operated in a salty environment. There was a fatality, a Stewardess, not strapped in, was lost in the explosive decompression. I suppose the ‘fail-safe’ ‘rip-stop’ philosophy incorporated in the 707 and its descendants after the Comet I disasters saved this 737.

The redesign of the Comet could have been done with rectangular windows (with the correct corner radiuses) but I guess they needed a good PR story. It’s a bit curious that BOAC stayed loyal to the Comet but rejected the Vickers VC7 in the same timeframe as the Comet investigations

Boeing seem to life the more recent 737s at some 60,000 cycles but failures can happen earlier (thats statistics for you)
http://www.nytimes.com/2011/04/06/business/06air.html?_r=0

I would once have assumed that sheet metal, pressed out in a mechanical or hydraulic press, would produce a reliably standard form but working briefly in a factory feeding sheet steel blanks into a press opened my eyes….some came out perfect, some cracked…I think it partly depended on the amount of lubrication there was between the press tool face and the blank

I suspect that the Avro Lancaster is more a space-frame than a monocoque with the skins only lightly stressed and a much smaller hole up the middle. The difference with a pressurised aircraft is that when you blow it up like a balloon it inevitably stresses the whole skin circumferentially.

I’d say the Lancaster was a semi-monocoque rather than a space-frame and, even if lightly stressed the fuselage ‘windows’ have rounded corners, all of them.

As for pressurising an airframe, the stresses, although they may be greater in magnitude, still have to resisted by the skin in the same way.

I know we have stayed away from Cherry Ripe’s original post but it is all relevant I think. What I find incredible is that some of the skin cracking was there right from manufacture and had already been stop drilled. I spent my entire working life involved with metal airframes and the only occasions when cracks appeared in something you were making was if you bent it too tightly or bent some material that was too hard for the bend you were attempting, stop drilling was only ever used on secondery structure like fairings and cowlings and even then very rarely. So what were DH doing to a skin panel to cause cracking in the first place.
Pressurisation loads on a fuselage structure are huge and many airframes have a cycle life limit imposed when they are designed, sometimes extended when in service data is known or the manufacturer carries out structural testing. The Trident was an example of an airframe with a cycle limit of 22,000 (from memory) which BAe didnt budge from as too few were built to warrant any expenditure.
Richard

I get the impression that DH were quite daring with their structures to judge by the number and variety of disasters caused by total structural failure. I recall that the Comet windows were intended to be glued in place but this was changed due to difficulties with the process. The double row of rivets combined with the square-ish corners of the windows was about the worst way of doing it. Also worth remembering the Hawaiian? jumbo which suffered a similar failure in 1989 but somehow managed a landing held together by the seat rails.

The life of a Lancaster in flying hours was very short and some were lost due to structural failure, far more were lost due to accidents and enemy action.

I suspect that the Avro Lancaster is more a space-frame than a monocoque with the skins only lightly stressed and a much smaller hole up the middle. The difference with a pressurised aircraft is that when you blow it up like a balloon it inevitably stresses the whole skin circumferentially.

The ‘square windows’ is certainly an oversimplification. One thing that has occurred to me is that if you repeatedly pressurise a airframe of that type, so it fails through fatigue stress, it will always (eventually) fail at the windows (or other stress raising holes) no matter what shape they are.

On thing I’ve heard said about the Comet I is that because the engines were not as powerful as had been hoped the fuselage skin had to be reduced in thickness to save weight; if the fuselage had been thicker its failure would have occurred later, even with the ‘square’ windows.

One thing that puzzles me is that the investigation into the Comet I crash is always said to have discovered a completely unknown cause of failure, pressure stress fatigue failure, and that the loss of the Comets led to a new understanding of how to control this in subsequent (pressurised) airframes. What puzzles me is why this was only ever a problem for stress caused by pressurising airframes; the materials used are the same as non-pressurised airframes and they must have to resist stresses that are not caused by pressurisation (hence rounded corners on windows)?

Imagine a heavily laden Lancaster bouncing along at 18,000 feet in flak-filled sky; there must have been plenty of repeated loading and unloading (stressing) of parts of the stressed skin airframe but the Comet disasters were apparently caused by the single stressing and unstressing of the airframe due to pressurisation during every take-off and landing cycle? This is what the famous ‘water torture’ test was to investigate, and what it ‘discovered’.

My question is this: if the stresses caused in the airframe are no different if caused by pressurisation or normal loading, then why is the Comet disaster always assigned to something that was ‘not understood’?

The Comet 1 windows and its other square cutouts had rounded corners as this figure from the official report shows.

Wow, that completely changes my understanding of the cause!

Indeed here is the ADF window from G-ALYP at the Science Museum, in plain sight.

[ATTACH=CONFIG]235955[/ATTACH]

So the whole ‘square windows’ story is urban legend.

Looking back through the Flight reporting of the Inquiry, thickening of the skin around cut-outs and the revision and reduction of bolt-holes and manufacturing cracks were the actual remedies.

I consider myself re-educated!

Thanks.

The Comet 1 windows and its other square cutouts had rounded corners as this figure from the official report shows.
What shocked me when I read the report were the cracks found during manufacture that stop drilling hadn’t prevented from growing. These cracks never get a mention whenever the Comet is talked about but the ‘square’ windows do.
[ATTACH=CONFIG]235950[/ATTACH]

That’s an interesting question! I shall dig-out some DC- cutaways and inspect more closely..

Meanwhile I found this in Flight 1946, sort-of explaining the change to square-ular windows:

Each seat station is provided with a rectangualr window, set rather far
forward, the idea being that the traveller in the inside position can
get a view out—unfortunately this means that the “corner” seat occupant
has to crane forward to look out.

It also mentions that the original Tudor had a clever internal framing for its circular windows, with a large ‘spectacle’ frame around each pair of windows to provide the illusion of larger windows. The A380 uses something similar today, with a large internal pane masking a smaller exterior one.

http://www.flightglobal.com/pdfarchive/view/1946/1946%20-%200618.html

Aren’t windows corner-radiused on stressed-skin airframes, even non-pressurised ones, anyway?