Whitley test "produced a failure at 98% of the design load" is that good?
 

I have acquired an old Armstrong Whitworth Aircraft Ltd book and in the part about fuselages it states regarding the Whitley:-

"A full scale test at R.A.E. produced a failure at 98% of the design load."

Are there any aircraft engineers out there, as I would have thought that an aircraft must meet, or exceed 100% of its design load to be airworthy especially when it would be required for AWO (all weather operations)?

Although I understand the design loading is higher than the design load for level flight, due to the necessity to be able to dive the aircraft, pull the aircraft out of a dive, make manoeuvres and turn, it is clear that the Whitley failed on test at "98% of the design load."

Also you have weather forces, icing and turbulence to contend with too.

Any comments?

Mark

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

In general, this is not a good thing. however, reports like this are quite common. We need more information on the test and the follow up before we declare the Whitley a bad aircraft. In my experience, test results like this can be refined by more detailed analysis of the real design load, and the test setup. I had a wing spar fail at apparently 98% load back in the 1970s. A detailed review and recalculation of the test setup, minus some time saving shortcuts used in the initial test design, convinced us the spar had actually failed at 101%.

Also, failures like this can lead to minor redesigns of the failed area, and then further testing or even just analysis to prove the structure airworthy. They can also lead to minor changes in the flight envelope (like speeds and g-loads) until a fix is completed on the whole fleet. Again we can't condemn this aircraft without knowing if this sort of subsequent activity took place.

One also needs to know the purpose of the test. Was this part of the original qualification testing? Was this an early development article, with the test results being used to refine the design? Was this a high time in-service airframe? There are a number of possibilities, all with different implications of this failure.

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Bill

Thanks, for those interesting comments you make, especially your comment, about when an aircraft weakness was discovered, it was then likely subsequently modified.

However, a file in AVIA 46 produced at our National Archives Kew, for the Air Ministry Whitley Aircraft Contracts was originally PA to Archives for permanent preservation, but susequently weeded and then PA to Archives twice again, with the numbered Minutes (Index Sheet) relating to correspondence of the early Whitley aircraft contract life struck out with "destroyed" written in, with the corresponding file parts removed.

A question regarding the Whitley was asked in 1940 to the makers Armstrong Whitworth about the dive speed restrictions on the Whitley.

I have seen a genuine copy of that 1940 reply from Armstrong Whitworth stating that the restrictions were due to the rear of the Whitley aircraft and applicable to all Whitley Marks with the Merlin engine. Regarding the restrictions it states:- "The limiting strength consideration is the rear fuselage in bend, probably just aft of the spar."

I have got some photographs of a Whitley V which has managed to force land, but the rear elevator and rudders have completely collapsed and the empennage damaged.

Mark

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Mark, any airplane will break if you push it hard enough. The description of the AW's concern (fuselage bending just aft of the spar) doesn't match your description of rudder, elevator and empenage damage. These were likely two unrelated cases.

It is possible that AW decided it was quicker and cheaper to limit diving speed than to reinforce the rear fuselage after the test failure. Given the pressures of the time, the RAF appears to have agreed. This goes on a lot, even today. The "design load" of any component is based on some worst case combination of speed, power, g-loading, etc. The combination is sort of arbitrary, and delivering an airplane with some design case at 98% of what it started out at may not be a bad thing. If everybody involved had taken the time and money to strengthen the rear fusleage to get back to the original design load, there is still no protection for anyone who goes over the design load conditions in actual service. Something else would have broken, somewhere else.

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

98% is obviously not great, but it's not a desaster either, and without knowing if any measures were taken afterwards to bring it to 100% or above (such as with the Airbus A380 that failed the wing load test at 147% instead of 150%, or 98% of test requirement), it's impossible to say whether this had any impact in the real world. Even without mitigating measures it might not have had any impact, since you would need to know first what design load actually was (and in the era of pre-computing, I believe engineers were far more generous with margins than they are today). For example, (making this up) design load could have been 200% of maximum expected load in a dive at maximum angle. In which case 98% is still more than fine.

Please note that none of the above is supposed to argue that the Whitley was a great plane.

All the best

Andreas

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Hello Bill and Andreas

A Whitley aircraft was twice as likely to be lost on ops than any other Bomber Command aircraft, simply by taking avoiding action.

Twice as many Whitleys were lost over the German searchlight belt according to a 1942 Op Research Section report.

I managed to get a copy from the AHB several years ago (still withheld in the UK), but the Canadians have put the report online along with a summary, click on the link for a pdf copy, which you can save.

http://www.canadianmilitaryhistory.c...operations.pdf

The RAF Whitley was twice as likely to crash due to simply taking avoiding action, diving, turning, or a combination of both, so there must have been something peculiar to the Whitley aircraft itself, in the design?

Mark

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Or simply because Whitley was less manoeuvrable than the others and clearly slower than Hampden, it was easier target to AA guns and to night fighters which operated inside those S/L zones. And thanks for the interesting link!

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Fascinating note, certainly, but in no way does it justify the phrase "twice as likely to crash simply taking avoiding action", as it says nothing about crashing nor about taking avoiding action! Nor, given the lack of individual detail about the losses, could it. It states that a Whitley is twice as likely as any other type to be shot down in these circumstances, but offers no explanation for this.

I think Juha has stated likelier reasons for these losses. Had there been any specific weakness in the handling or structure of the Whitley this would have shown itself in other ways, and evidence would be available from other sources - raised losses in training or general operations, for example.

It might be interesting to find when the corkscrew became a standard evasion manoeuvre, for the Whitley was something of a old lady for such a violent manoeuvre, but the result is likelier to have been a somewhat stately corkscrew rather than design load exceedance.

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Paragraph 3 (1st part). “It seems probable that the lack of manoeuvrability of the Whitley and the consequently greater difficulties of getting out of searchlights when once illuminated contribute very largely to the greater losses of Whitley bombers as compared with other types. Whitleys on the other hand appear quite suitable for attacking targets which do not involve crossing the searchlight belts.”

Paragraph 3 (2nd part). “When the Whitley aircraft do not cross a searchlight belt, their losses are about the same as the losses for other aircraft or less than the average for other types.”

So if it was that Whitleys were slow to get out of the way and more got shot down as you suggest, over these searchlight belt areas, then why did this higher statistic of Whitley losses (compared with other bomber types) not follow on other parts of the route to and from the target (where all bomber types were exposed to AA)?

I felt the report distinctly gave the impression that the greater losses of Whitleys than other bomber types was due to evasive action in getting out of searchlights.

1. I do not feel that the report is saying that more Whitleys were victims to A.A., but either way it is still flying limitations, due to design considerations.
or
2. That Pilots did make an attempt to turn quicker than the flying limitations permitted and the Whitley dived away and became a statistic.

A Test Pilot and several Whitley Pilots according to a 1941 report obviously did get back to report parts of the main plane skin torn away.

The Test Pilot turned the Whitley which stalled in the turn, then dived away. However the Test Pilot managed to regain control and land. The damaged Whitley main plane is pictured and you will notice that the structural bracing in the Whitley wing has begun to break apart.

Some of the RAF crews who found themselves in a severe dive and got back, managed to get the elevator to shift and regain control by applying trim tab.

However, when this obviously failed and the tabs / elevator were found damaged in a crashed aircraft, the Pilot was apparently sometimes blamed for having too much trim tab.

In November 1941 Bomber Command commissioned an RAE Engineering Report into diving speed limitations. The Whitley design only permitted a maximum dive of 240 ASI, but Pilots did not like to exceed 200 to 220 ASI.

She may have given the impression of a fine old lady, but Pilots who got back were reporting that the Whitley V was impossible to pull out of the dive, requiring strengthened Pilots cabin floor, strengthened control column floor bracket and strengthened control column pivot pin and other structural mods to the main plane, fuselage and tail. It was clear that huge pressures must have been exerted on the control column.

Other Whitleys were seen/heard to turn and dive, or dive to the ground at varying angles in the UK and that is all we know about the last moments.

Mark

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Point IV is advaisable in many planes. Using powerful trim tab to get out from high speed dive easily overloaded a plane.

For natural reasons AA guns and night fighters early in the war were concentrated to S/L zones, not much idea simply get the plane into S/L cone and note "there it goes". Especially for night fighters operating at S/L zones time was paramount. They must see the target, position themselves for the attack, attack and shot the bomber down before it departed from the zone to the dark night.

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

From everything I've read (including this latest information) there is no doubt that the Whitley was an old design, obsolete before the war began. This probably accounts for the large losses. I'm missing the link between this and the original question about failure of a static test.

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

A couple of observations. Implied but not stated is that the 'Failure' at 98% was during an Ultimate Load under static conditions. Assumption not fact (for me) is that it was a static wing load test with dead weight sand bags distributed like an assumed loading for a dive pullout or coupled with the fuselage in some way to predict symmetric bending loads on the fuselage for a pullout or even ditching/hard landing..

Nobody in the airframe business was attempting to analyze asymmetric load conditions except for possibly lateral rudder/vertical stabilizer loads during a yaw condition during dive. Offhand, such analysis could yield useful insight to the fuselage failure modes while taking out the empennage loads in bending and torsion at the carry through structure/longerons.

98% sounds pretty damn good to me but I'm sure the structures chief would have preferred 102.

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Don't ask me for a reference, but I seem to recall asymmetric loading being discussed at the university level in pre-war papers. (The classic rolling pull-up.) Whether it had made its way into RAF requirements is another question.

In today's terms, ultimate load can be 150% design load or 150% service load. As I said, without knowing more it is hard to draw firm conclusions about this test result.

And I have had more than one boss who would tell you that failure at 102% meant the @#$%& thing was too heavy!

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Bill - I don't doubt that there was a reference to bending analysis of asymmetric loads in text, but curious how the aero loads would have been simulated accurately.

It is relatively straight forward to calculate the sectional moments of inertia for the fuselage/tail station by station but quite a bit more to look at a moment by moment load/fuselage response with no finite element/relaxation methods applied via computer simulation in the 30's (or 40's or 50's or early 60's)...

having said that my practical experience didn't start until late 60's so my comments are to be taken with grain of salt about WWII methodology-

Ditto on the presumed reaction of airframe structures chief when confronted with 2% above design load structural integrity.

And ditto on Limit Load @ 100% Design Gross Weight for the configuration/condition analyzed with 150%. for Ultimate.. Simple rules for steel/aluminum with homogeneous metals - with departure when stress/strain curves were not linear in elastic range.

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

My exerience started in the late 1960s as well, I'm going by conversations with the old hands. Maybe I'm over simplifying, but dynamic load cases can be treated as quasi-static, with inertia forces added to the free body diagrams. Statically indeterminate loads were handled by hand iterative calculations way back before airplanes, I think early iron boat hulls were analyzed this way. When I started writing Fortran code in 1970 to do this the process was well understood. Computers just replaced rooms full of clerks with slide rules or mechanical calculators. Look for a book called "Slide Rule" by Neville Chamberlin for an interesting description of hand cranked finite element analysis of dirigible structures in the 1920s. Lots of people, using lots of paper and lots of pencils.

I think the limit here may have been in defining the aero loads to apply, as you suggest. If you look at old airworthiness codes, it seems to have been a fairly emprical process. "We know airplanes THIS strong don't break, so make yours just as strong". When that failed, the required loads were upped a little bit in the next amendment. Even today, the fine print in FAR 25 loads state that rolling pullouts may not be covered with great accuracy - so the manufacturers tell you their published g limit is for symetrical loading only. In other words, don't do rolling pull outs in your airliner.

Having said all that, the designer really has no special insight into what is required in an airplane. They have to rely on the customer (like the RAF for the Whitley) asking for the right things. Wrong assumptions can lead to perfectly designed wrong airplanes. It has happened over and over.

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Very well summarized..

Interesting on the Fortran. before I was immersed in Nastran in late sixties - I co-op'ed at NASA and Lockheed. At NASA I programmed the LoW Orbital Degradation model, first on an IBM 620, then 7090 (I couldn't imagine a Bigger Machine at the time) and was introduced to Bessel and Hankel Functions applied to spheres or oblate Spheroids for first time.

Even when building 'real' airframe models with NASTRAN, using RODS as beams, and shear panels (PLATES way too stiff), getting either static or dynamic loads wasn't in the application at the time..

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

"A full scale test at R.A.E. produced a failure at 98% of the design load."

The significance of 98% depends on the definition of "design load". If it meant the highest expected operational load, then they were in trouble. If it meant the test load including a safety factor on the highest expected operational load, then they were fine. 2% was well within the noise of their calculation accuracy. Pre-war, the standard safety factor for metal structures (i.e. steel) was 2x. This was reduced in the war by decree to 1.6x to save materials. I am sure that wooden structures would have had a higher safety factor to take into account the non-homogenuity of wood. This would be one of the advantages of using plywood which can be considered homogeneous, at least in 2 dimensions. So 98% of 2x is still 1.96x, an insignificant degradation. I doubt that there is any corrolation between this report and the combat losses.

Mark

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Thank you all for your comments.

Found this http://www.flightglobal.com/pdfarchi...0-%201470.html but it is 1949. Whitley spec was 1934. I did find an ARC or RAE article, but cannot find it now.

The Captain and crew of a Whitley V survived a forced landing when their tail beam (beam which also holds the rear wheel) failed at the end of July 1940 and within two weeks another occurred at Boscombe Down. By December 1940 a modified tail beam kit had been designed by AW and was being supplied to 4 Group and Contractors. Any weakness appeared to be in the tail.

Later in WW2 a second Whitley tail beam modification was supplied to supersede the first modified Dec 1940 tail beam.

Mark

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Hello Mark
you are surely aware that John Nesbitt-Dufort in his Scramble (London 1970) gives a short description (2 pages) what it was fly Whitley Mk V in ops.

Juha

ADDITION: ans not surprising a more positive analyze by Charles Turner-Hughes (former Chief Test Pilot, AW) in Air Enthusiast Number Nine.

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

Thanks Juha

I have another article about AW Test Pilot Turner-Hughes, but did not know about Air Enthusiast No.9.

The other reference was also unknown to me.

Another photograph in December 1940 Flight shows all of the wing skin virtually stripped clean off a Whitley wing (due it claims) to A.A. fire? But the skin looks too cleanly ripped off to be A.A. fire!

My Grandfather was in one of the tail failure Whitleys, which force landed!

Thanks again, Mark

Re: Whitley test "produced a failure at 98% of the design load" is that good?
 

The wing skins could have been a classic "zipper" failure. I'm assuming the Whitley wing skins were metal panels riveted on. If a few rivets fail (or are shot away) the panel load re-distributes over the remaing rivets. At some point this re-distributed load will exceed the breaking strength of a rivet, this one pops, the load re-distributes and the next one pops, and so on until the entire panel can fall free. Like undoing a zipper. If a corner or edge of the panel lifts and gets high speed air under the panel, the effect can be quite quick.

There have been cases where one panel tearing off damages rivets at an overlapping seam, and the next panel proceeds to unzip as well.