thank you very much thank you for giving up your evening and coming along thank you for the warm welcome yet it Thank You Ryan for inviting me and I've looked at the hit the Havilland comet for a number of years I got involved probably getting on for 25 years ago when I was asked to talk about fatigue as part of a documentary and it got me looking at the comets trying to understand a bit more about it and getting involved with some of the kind of the people that were involved at the time it's quite a few years now since it actually crashed since the investigation the benefit of doing the documentaries that I did was that I managed to meet a number of the protagonists from the original story and was able to actually talk to them about the the investigation about the aircraft so I'll be talking so some of the information and I'll be giving you tonight will actually be from the people that were there at the time I'm not old enough to have worked on the comet myself sadly I never flew on one so that that's unfortunate and but anyway what will work with that but it's it was a lovely aircraft first flew in 1949 absolutely beautiful aircraft that's in its Air France colors that 3-3 aircraft went to air for Air France I've got a particular affection for the Air France ones the the last remaining comets actually comet one in it in its full glory he's painted up in BOAC colours it cost for the aerospace museum and is actually one of the Air France comets so that's that's where that comes in but it was a lovely aircraft very clean lines and revolutionary for its day and I think that's one of the key things about it it really was an aircraft that put Britain at the forefront of aviation technology and what we'll do is we'll work our way through the story I mean everybody knows you're in an engineering department you mentioned de Havilland comet it usually gets a comment somewhere about between comments three and five is square windows so so we'll go through the square window story and we'll work our way on from there the general background at the time it was a bit of a kind of difficult time 1940s early 1940s middle of the war post-war air travel was dominated by large propeller driven aircraft though the air that the the aircraft that were kind of going to be flying at the time the US industry actually had developed some bespoke transport airframes as part of the war effort the UK tended to be trying to build bombers and so when it got to the end of the war the development of air transport was basically with modified bombers so the Avro yoke was a Lancaster bomber with a completely different airframe strapped underneath the wings and so they they tended to be looking at developing hybrids uncomfortable hybrids in some cases of aircraft to actually try and get some commercial interest from the airlines as all the world's Airlines started picking up and wanting to travel around the world after the war and airshow will becoming popular not as popular quite as today but people are expecting comfort and speed speed by today's standards not necessarily that much but they were expecting to get there in a level of comfort they were expecting that they would know get around the world um and to be honest although they were propeller on the propeller driven for a lot of the commercial products the military aircraft have been using gas turbine power plants since 1942 but that was only about two Air Forces de Havilland themselves had quite a good pedigree in this area and that's the de Havilland Goblin engine they designed the vampire which is there and it was a second RAF jet fighter to enter service in 1945 powered by one de Havilland ghost thirty three thousand five hundred pounds of static thrust which is peanuts that's in the loose change of today's civil aircraft but equally they've been involved in passenger aircraft they did Havilland for a completed the UK's first regular passenger flight in 1919 so they've been doing some passenger stuff and they develop this kind of parallel track of military and passenger aircraft military and civil so they've got the pedigree and anybody from rolls-royce here X Rolls Royce I'm gonna get in trouble now and the the Gloster Meteor was the first aircraft went to service with the RAF in terms of jet fighters the Gloster Meteor that the five two types that flew it was rolls-royce powered number one rolls-royce powered number two rolls-royce powered number three rolls-royce powered number four de Havilland powered number five Number five flew first because unlike every other engine rolls-royce was late with the first gas turbine yes anyway and moving swiftly on so so they haven't had a pedigree and actually their design of gas turbines was much more like the Whittle patent you actually go and pull the Whittle patent look at it the design that that Bishop and his team put into that looks much more like the Whittle pattern than Whittle ever stuck too which is interesting they basically pull the patent and copied it Whittle didn't but flight in 1942 actually kind of put some context around this it's a long sentence but the whole British Empire at present time has been up operational fleet of transport aircraft comprising conversions makeshift sand casts dose totally inadequate to represent the Empire in serving the air routes of the world in peace to come and so this was night this was December 1942 have we to rely upon other nations to do it for us the British aircraft industry is equal to the test the government should decide this vital question at once that's because they'd heard wind of the Brabazon Committee which was pulled together during the war after Winston Churchill got a trip in a liberator bomber a rather cold trip to Russia in winter I think but basically he how he got very cold in a liberator bomber and decided if we're going to do a fair travel after the war we needed something that was better than this and they set up the BRABUS in committee and the brothers in committee came back in 1945 and said we need all these different types of aircraft to be designed and built by the British industry to serve a number of different sectors and so type one very large transatlantic airliners serving high-volume routes London New York City its passengers in luxury for the 12-hour trip so there's all these kind of things cropping up and it split down into all sorts of types and look knocking in their type to be that was a bit of an argument with Vickers for an aircraft using new turboprop engine and courage by the Vickers that became a the by count lovely little aircraft one of the best-selling act British aircraft vault lovely lovely aircraft and so that big over windows that was that was actually put in by Vickers did Jeffrey de Havilland who was on the BRABUS in committee actually chipped in type four was a jet-powered hundred-seat high-speed transport and it ended up the personal merging of Jeffrey de Havilland funny that anyway that became the comet so the comet itself they did they started design work on it a type for high speed mail plane was where it really comes from design work began in September 46 the key features that they focused on aluminium construction they wanted it to be as lightweight as they could make it hydraulic actuation of the control surfaces now that was really radical this was the first aircraft to use hydraulics across all primary controls which meant no longer would the pilots yoke be attached the control services by wires and all the pilots of their day said will lose the feel this won't be any good anybody remember when their bus launched fly-by-wire and all the pilots said this won't be any good will 1940s that's what the pilot said wait till flyby light comes along and all the pilots say we it's not connected to anything we won't have anyway we'll just rehash the old arguments and the reason for that was that they felt it was the only way they're going to fly at high speed they needed the control they needed hydraulics to actually make the make it work it's just gonna be too high speed for control surfaces manually operated ones for de Havilland ghost engines slight up right from the goblins 36 passengers in the initial forum range one thousand seven hundred and fifty miles point-to-point so not a great range either cruising speed nearly five hundred miles an hour that's interesting that's slightly higher than today's aircraft but apart from Concorde which kind of sits out there on its own isn't too far off every other commercial aircraft that's ever come since you know 450 to 500 miles an hour and they're all in that bracket and everybody's sitting it's about the same speed foot the first gas turbine powered aircraft was the same cruising altitude 35,000 feet weight 50 tons some interesting data that I found put out by Frank Whittle he basically said that his gas turbine was going to be really really valuable pick a piston engine is there jets there and the comet I've added at the end but this is Frank whittles data looking at how the efficiency at high speed so you suddenly needed jets if you're going to go to high speed because you're capacity load which is the green bit at the top has completely disappeared for piston-engined aircraft and that was Frank's major argument in the 1930s and early 1940s and actually D Havilland bought this out and actually the only reason that their capacity load was there is that that's actually a slight this all of these he worked out on the London to Paris stage route so only an hour's flying this was actually for two hours so actually you could calculate back in there so it's not far of Frank's and regional calculations said they haven't weren't exceeding what was thought of as the time as being where they would actually start to play the data was very much in keeping with everybody else and some of the compromises they had to make gas turbine engines are more efficient at high altitude hence the height the passengers wouldn't need oxygen this was going to be the first time to an extent that they'd have to go and do some like pressurized cabins to any great extent because the passengers above 10,000 feet 12,000 feet you get quite poorly if you don't breathe so you had to go and have cab and they decided to pressurize the cabin they could have gone the other way and given everybody face masks they as pilots of high altitude Jets had face masks and separate oxygen resistant systems they decided they'll pressurize the cabin that that meant that the cabin had to withstand the pressure of you know eight and a quarter psi and I'm sorry for some of the unit's I'm going to dodge in and out of because of the age of the project I'm going to flip between metric and imperial that kind of random kind of intervals so that was they basically pressurized the cockpit through the equivalent of 8,000 feet and that's actually not being far off industry standard until the Boeing 787 which has just pushed it up or pushed it down to 6,000 feet and actually 8,000 feet was kind of the industry standard since 1952 so de havilland set a standard in terms of speed in cabin pressurization that hasn't really changed ghost engines weren't powerful by today's standards thrust-to-weight ratio of about two you know Trent's these days six plus airframe was made as light as possible to it to keep that matte that payload at its maximum and so they used something called Redux which is an aluminium gluing kind of technology and the aluminium alloy they used was DT D 546 which was a UK specification aluminium alloy which is very similar to a 2000 series alloy today so early generation aluminium alloy and some stuff that de Havilland had done the douve and to service in 1946 were to replace the dragon Rapide which is that one so they've gone from lots of bits of string wire and canvas and they went to this lovely aluminium framed of which also used redux so they got experiencing taking technology from and applying it so the redux seen was quite a good way to take the weight out and they also put in a whole load of stuff into this as a kind of tester so flaps these are variable speed constant speed propeller they got flat for retractable undercarriage and they've done some of this stuff which they'd learned and could apply to the comet Redux was actually developed out at Duxford it was a bank made of a liquid phenol resin at which they you applied and it had a load of use there was a lot of talk in the inquiry the subsequent committee inquiry about whether the redux had caused lots of the issues I think it's fairly safe to say from all the experience that they had with it it was not the redux although there's at least one expert witness who was absolutely convinced that it was and I just took that slide in to say it wasn't that it there it's been used by de Havilland it's been in service with the Nimrod for more years than many as have been flying so it's not actually a problem and it's also cropped up on a number of other aircraft but some of the information around the airframe itself 22 gauge skin 0.71 millimeters thick further built for the most of it around the windows it went up to point nine one mil thick which is about common Ferb's 77037 most things that are those kind of gauges certainly in early seventies all the same and the UTS of the alloy element ultimate tensile strength is about 450 mega Pascal's so reasonably strong and de havilland estimated that the stresses would reach about 190 mega Pascal's near the cutout so near escape hatches windows doors patent everything else where they taking a chunk out and had dealt with it they reckon the stress would get probably up to about 190 mega Pascal's and that was where they started to do their calculations but to be honest they didn't trust calculations they preferred to actually test things and we'll come on to that in a minute definition of fatigue at the time metals will break under a load which is repeatedly applied and then removed although they can support a much larger steady load without distress was about where it was in the mid-1940s when they started out on this journey it's been known about as a failure mechanism for about a hundred years since some railway axle failures in the middle of the 1800s military aircraft had suffered quite a bit of fatigue on some of the wing structures and there were some investigations during the war on that some REE work on fatigue of wing structures so there was it was known about it was a problem and people were trying to get their heads around it and see what they could do to deal with it and de Havilland understood it that was the typical kind of wartime construction so aluminium stringers ribs riveted construction and there was still fatigue in those kind of structures so they were kind of used to that the requirements of the day when they start in the design was that the proof pressure that they had to take it to was one and a third of the F P so that eight and a quarter psi 1 and a third of that pressurized the body if it didn't explode or didn't stress or Bend or whatever then that was good the design pressure had to be twice that so if you think that 118 mega Pascal's well was well within the 450 of the ultimate tensile strength so they felt they got those kind of safety safety margins but they actually used twice the proof pressure as a testing method method rather than one on third and two and a half times the pressure the design pressure was where they thought they could operate so they actually put in safety margins above the the legislation to see if they could cope with that now they also spend a lot of time talking to their registration board who would give them their certificate of airworthiness and they wanted to know that would that give them some kind of protection against fatigue and it was just okay fatigue was known they they kind of understood at the time the aluminium didn't particularly over threshold so it would continue to fatigue even at low stresses but they didn't know how to design against it and so both de Havilland's and the heir Registration Board and the Royal aircraft establishment at Farnborough all said well yeah you kind of over over design it you'll be fine um and that was the acceptable methodology there was no other way of doing it now there were some new draft requirements that came out in mid 1952 they were drafting 52 issued in 53 the Comets first commercial flight was in 52 so they knew they knew they wasn't coming and they said a static test to twice that pressure proof test two one on the third 15,000 applications of one and a quarter so they started to look at fatigue in their structural parts would need to take three times that so doors windows any cutouts would need three times that and they actually said that 15,000 cycles were seen as these the design life of the airframe if any aircraft ever saw 15,000 cycles it was going to get pensioned off now that's fine but we've seen boeing 737 with 91 thousand cycles and that failed in fatigue as well and the idea of the application one 1/4 p was to cover the scattering fatigue results and they still didn't understand fatigue but they were trying to do lots of stuff with stressing and loading on whatever to kind of fix the problem um and it was still not well enough understood to enable an accurate life in method methodology to be used they haven't actually got hold of to cabin section where they didn't need to get hold and they own them to caring sections the forward sections just under 8 meters long from the nose to just in front of the wing main spa midsection which covered the whole of the midsection at over the the wing main spa and they basically stuck those in a a chamber flooded it with water capped the ends and pressurized it and depressurize it pressure and they did all the tests on these two sections so it wasn't a whole airframe it was two sections both less than 8 meters long but they did that to it and by July 1953 the forward section has seen 16,000 cycles and the flying comics had not exceeded 2500 cycles and when one of these finally failed that just over that they thought they got enough margin against the the regulations they had fatigue tested these things to the point where they'd failed and they thought they they could then put the fleet up or continue taking have the fleeting service because they were ahead of the game in terms of a t-test the sections including cut out and prove tested to stress is higher than 2p and the de havilland there was lots of criticism of them about their design stressing they did not calculate the precise stresses at the windows they preferred to say it's about this but we'll test it and prove it they preferred to actually do a test than rely on calculations so first fly to the de Havilland comet 27th of July 1949 it had been a secret project they built it in as much secrecy as they could muster you had to have a pass to wander through although was one of the employees at the time said if you stick a file on your arm and walk like you meant it you could get through the hangar and and its first public displays in 1949 and the pilot on that first day was John Cunningham who was the chief test pilot for the project he was told roll it out for the press do a few runway kind of ups and downs everybody would be happy so he did he rolled it up and down the runway a couple of times the press took their photos and then he thought actually everybody went home about 6 o'clock in the evening who stood still doing room well try as he thought it feels quite good let's see what it can do and he took off at 6 o'clock in the evening the world's press had all gone home apart from one photographer they they it took them years to forgive John for that one the first ever flight of a commercial jet aircraft and it was after everybody gone home he said it flew beautifully it didn't take put up the undercarriage he kept the undercarriage down and flew a loop it was fine he had a great day John John loved the aircraft it is one of the people I was privileged enough to meet he was a test pilot he was also a night fighter pilot his idea of danger is not necessarily the same as other people's you tried to get out of a taxi it's in the in the middle lane outside some Kings Cross station because he needed to catch a train we tried to explain that the taxis moving you're in the middle lane it's a busy road his appreciation of danger was never quite what you anyway but he was a chief test pilot and broke many two passenger records it basically the point-to-point records Rome Cairo Copenhagen it just did point-to-point the reason for this was that they didn't know if air traffic control was going to get completely scrambled by somebody turning up twice as fast as everybody else all of a sudden they've got all these piston air engine aircraft turning up 200 miles now 250 miles an hour these things are going to turn up at 450 miles an hour into air traffic control holding patterns can air traffic control cope so they actually did lots of mail carrying to different places to just prove they could get everybody to cope and handle this new it new technology and it's entered service second of May 1952 yoke Peter there it is on the runway at Heathrow London to Joburg took off what waved off and it shaved hours off the flight I can show you some figures in a minute inaugurated villains in Tokyo service April 53 took one and a half days still to get there what's Tokyo 13 14 hours from Heathrow took one and a half days 89% load factors for the first year every really popular full aircraft everybody wanted to do it the reason it took one and a half days was that it had to fly as he says put UK industry at the forefront of aviation technology the reason that it had to it took so long was that he had a range of 1750 miles which meant when he left London the first place he landed was Rome you had to land at Rome because then you had to refuel and then depending on where you were going you could fly on to Beirut in laundry and you're hot you did this these kind of two-hour stages around the globe and it's still hard the journey times or you went from Rome to Cairo you went down to Joburg and so that was how it happened you basically flew in to our stages around the world and this these were the the routes that de havilland focused on BOAC were really pleased because it shaved hours days off the journey time he was having journey times it was taking 40% out of journey times so people wanted to do it they they could get a day and a half of their life back by just actually catching a faster plane so it was really popular in what in one year 370 hours per week hundred twenty two thousand miles twenty thousand seven hundred and eighty unduplicated miles but it also goes back to and the major incidents were both from Rome and you say well why Rome was here something to do with Rome Airport it was looked at in the investigation where there was somebody sabotaging planes at Rome Airport but the reason for Rome is every flight to and from London went through Rome so you have no choice yeah whichever way you were going you were going through a rope and it was just coincidence or unfortunate coincidence it was wrong both times there were a number of accidents 26 of October 1952 yoke zebra damaged beyond repair on takeoff the investigation found the nose was too high during takeoff caused the stall John Cunningham actually went in to try and replicate the situation with this and he found out that what happens is all the pilots are converted from piston engine aircraft and with a piston engine aircraft what happens is you can open up the throttles you get the engines working when you get to kind of rotate speed you can basically use the prop wash to give your wings lift and you haul yourself into the air using the propellers when you've got gas turbines you don't get the prop wash over the wings they stall you tip up too early the wings enter a stall and in this case it run off the end of the runway in an in a stalled condition and so John started to write the the rulebook that says you have v1v rotate v2 speeds and so what still used today by pilots was written by John to avoid situations like that unfortunately this plane was on a bit of a race Canadian Pacific airlines were taking delivery of CFC UN and they got as far as Karachi in what they were doing where they were attempting a record-breaking circumnavigation of the world it was it was dawn the pilots were tired they'd had sure rest periods long stage lengths they would they basically got into their it was dawn they couldn't see properly pulled back and John said it was unfortunate he could see all the marks of it there goes tail scrape down the runway they got the nose down but not in time they ran off the end of the runway hit a ditch and the plane burst into flames then second of May 1953 any conspiracy theorists out there like this one yoke Victor crashed in a tropical saw him climbing westbound from Calcutta a people have always said is this were the first of the examples of fatigue basically took off from Calcutta normal radio communications it was climbing up climbing out then just disappeared broke up in in midair crashed I people said was it fatigue the report says to storm to severe if the airframe tail broke away actually if you read the report you look at the pictures they're probably about right the storm was quite a major storm nobody else was flying through it nobody wanted to fly through it the pilot went through it unfortunately probably broke the tail off judging by the impact damage the tail probably went before the the the main airframe so it wasn't pressure carrying fatigue that we came to know and love later and then 10th of January 1954 yoke pizza that inaugural aircraft for the from the Joe Bergeron crashed into the sea after at 1290 pressurized flights this was a major event they it crashed into the sea second unexplained in facially unexplained crash at the time because the previous one had only been six months before basically they grounded all comets the BOAC grounded them and they started an University in an investigation but he crashed flying out of Rome II crashed into the sea they picked up a few bodies mail sacks floating wreckage seatbacks things like that which they brought back to the UK but they had nothing really to go on but Lord brothers and chaired a committee which looked at all the changes that could be made and they looked at all the different failure modes they talked about a load of stuff that could have happened they looked at engine failure whether it could have thrown what they called a turbine I'm not whether I think they mean a turbine disc but they basically then strengthen the portion of the airframe near near the engines to make sure there's no puncturing of the pressure capping they went round all sorts of stuff and there are no no definite reason for the accident has been established they still have no wreckage they know no way of finding it modifications are being embodied to cover every possibility that the imagination has suggested as a likely cause of the disaster when these modifications are completed and have been satisfactory flight tested the board sees no reason why passenger services should not be resumed flight services resumed 23rd of March 1954 so it took them a couple of months to actually get to that state yoyo took off less than two weeks later 8th of April 1954 - took off from Rome heading away it had a bit of an interesting flight it'd been held up in Rome until 9 o'clock in the evening it took off late they'd had a problem with I think it was one of the bits of electronic equipment and it was was held over I've actually got a photo of it taken that day by somebody who's actually who actually came along to it a talking said I've got a photo that his father took on that day at Rome Airport the Italian policeman his rather irate that he was taking phone photos of the plane but he's actually got a picture of yo-yo before it took off that night but he crashed into the sea at which point BOAC suspended all comic services and the certificate of airworthiness with was withdrawn major step 12th of April 1954 Ministry of supply instructed surrond Hall as the director of the Royal estatic a royal aircraft establishment to undertake a complete investigation of the whole problem presented by the accidents and to use all the resources at the disposal of the establishment so that was the story and sir Arnold Hall was called in and so they stopped a lot of other work and some people were rather cross because their projects got struck got stopped or delayed and they looked at all of these different things and the the final report from the royal aircraft establishment probably runs two three-ring binders this thick of all the things they looked at and reported out and so it covered just about everything they could think of can they rebuild the wreckage of the original crash comet yoyo crashed an even deeper water so was unlikely to be found but you know Peter reasonably shallow water can they find some bits fatigue tests of the pressure cabin that's kind of where we know with hindsight that state where the answer was but at the time they were looking at 40 tests on wings tail plane static strength damage the refueling refueling for the comet was high pressure refueling could that have caused the problem could they've got fuel in the place they shouldn't and it caught fire when yo Peter went down they could the fisherman said something some of the bits were on fire could it have been something to do with that also they had flight investigations head another comet and they they persuaded people to fly it and tell them what happened miscellaneous investigations and medical apps aspects of the accident which I will come back to miscellaneous investigations included making models firing firing bits of models off top of hangars filming exploding planes all sorts the two I'm going to focus on really today are the top two because we know that's roughly where the thread that came through but at the time and I'm talking to sort all about it he was very adamant that this wasn't the only strand that went on at that time every strand had equal importance and actually the tough bit was actually trying to get that one started 40 tests on the pressure cabin and he didn't get started straightaway and that's because he had to get hold of a comet and they decided that what they were going to do was stick it in a big water tank as a whole aircraft so they then had to build a water tank around it they had rubber kind of bits around the wings to stop it will eking out they built a steel structure around it and he had to get two water tanks built each holding nearly a million litres and he had to get permission from the water company to take that much water out of the system and that's why I had to build two tanks there was a holding tank in the main tank because he wasn't allowed to take a million million liters out of the system every time he wanted to fill the tank he had to pump it backwards and forwards between two tanks they let him do it once but not multiple times so the idea was that they would use simulated flight cycles on the wings using hydraulic rams and they basically had hydraulic rams on the wings and they pressurize the cabin to simulate a flight cycle and they knew that if they used water it was incompressible so any kind of damage any cracks as they said at the time if they used air it was like a 500-pound bomb going off in the cabin the evidence will be all over the the field said they were going to use water and they basically flew that simulated flight on the other plane measured the loads from the other flights and then did it simulated flight in five minutes and that's the simulated flight cycle so you've got this as a kind of what they what they actually saw and they impose that on there with the flight cycle and so in five to five minutes you could actually do a complete cycle of you know the normal kind of two-hour flight and that's what they did and it kind of kicked off in about April it with with the building they managed to get by May June time it was all up and running and there was some poor gentleman that was on a number of poor gentlemen who were asked to actually sleep on camp beds in the pump house how you slept in a pump house I'm not sure because the pump was going and everything was going off but they basically ran it day and night so they put yo Conklin the tank they removed all the internal cabin fixtures and replacing with lead weights they didn't want soggy seating so they kind of simulated that it had done huh what 1121 flights already passenger flights before they stripped it out and it another turn of the pressurized flights with de Havilland and it managed with 1826 simulated cycles before fatigue failure it's about three o'clock in the morning when the gentleman was awoken by the fact that he couldn't repressurize it and then because he'd been told to phoned up Arnold Hall as director of the REA and woke him up and basically he said it's not pressurizing with I think it's failed now this to an extent was a surprise total number of cycles completed 3,000 de Havilland had a piece in another tank that had done 16,000 cycles they had another piece the the bit that hadn't failed it done over 18,000 cycles at this point so they had these two bits that had carried on running they also went to put strain gauges on the comet and started to look at what the stresses were and as they got closer to the windows they suddenly realized that with 450 mega Pascal's as the ultimate tensile strength they started to see some really high stresses about 315 on a normal cycle just because of the cut out so they started to you know wouldn't about de havilland stressing calculations so they actually strain you know measured these these stresses and try to get a better handle on them and they turned out higher than they do have one reckon one hundred and ninety mega Pascal's and this is a put three hundred and fifty the other thing that they did was yoke Peter crashed in 600 feet 183 metres of water the Royal Navy which was very helpful were sent into the Mediterranean to find as much as it could and this was the first use of underwater TV to find an aircraft wreckage so they actually got a pods took a TV in it dropped it down and started searching the seabed for bits then you roughly worried to gone the re also made then that they understood the breakup sequence to an extent from what was said by some of the fishermen so they had models where they put plugs in these models and as they threw it off the hanger the plugs came out in a specific order and bits flew off and then they maps out where it fell on the floor and sent the Navy into that point on the Mediterranean and the Navy had quite a good job of it I mean they've seventy percent of the aircraft was located by September 1954 so it didn't work that badly they did a really good job Yoko was at rest that's well beyond anything they were going to do with the Navy's submersibles so they gradually be rebuilt yo yoke and they noted similarities between yoke Peter and yo yoke and that was that the flight profile was very similar and the forensic evidence was very similar and actually they did medical forensics on both sets of passengers and it turns out they both suffered from explosive decompression at altitude and so that and suffered from multiple impacts with bits of the cabin on the way through so they got evidence of a break up sequence for yoke Peter the tango crane plane separated fairly early in the whole sequence however damage to the tail plane there was a bit of carpet stuck in the leading edge so they knew that the carpet had come out before the tail plane broke off there was also a coin impact somewhere around here that actually they could work out what the coin was precisely from the impact of the coin on the tail so they knew what denomination and nationality of the coin said the knew the pressure capping must have failed first because it took tout all the in Syria and they started to rebuild it on a frame which kind of looks familiar to us now lots of investigations have done that but this was the first in terms of trying to build up a whole model on a frame to put it back together again and that's what they did they wanted to track down what had failed first and then it was a pressure cabin the other thing that they went back through is the heaviness manufacturing history and they started to pull together all sorts of bits of information about that manufacturing cracks were found in the airframe itself and in other airframes so both in yoke Peter which is this one and in other airframes riveting was likely to cause cracks in the skin around the rivet holes they punched the rivets through small cracks are put in and they they they cracked on inspection if they found them during manufacturing they were stopped with a 1.6 millimeter drill they were allowed to stop the crack just drill a hole through the end of it so they drill a hole through the end make it blunt not a problem except in this one in yolk Peter they drilled it stops it and it's still run on some cracks were seen to go beyond the hole some did not some stopped some didn't so it was a known issue but the secrets of the break-up that was this from the original report they knew that it had started to fail about here and it had opened up the the pressure cabin and unfortunately soon as the pressure coming peeled open the airframe was no longer a structural entity the wings snapped and the nose fell off and the tail broke in a downward motion they know that the wings failed in a downward motion because all three out of four of the engines failed that their shafts failed into gyroscopic loading as there as the engines just tipped forward and the shafts snapped because they were still under load at the time and this was the piece they found this was found by fishermen in about August 1954 and that was the roof of the pressure cabin and somewhere in there is the initiation and they could say the cracks started there and ran aft and it hit the window and run forward down and out and that's the crack the failure sequence as they got hold of it to show you some bit of fun or pass this round this is actually a piece of comet skin so you can see the thickness of it you can see the rivet holes we did find a manufacturing cracking one earlier it's a bit tatty this is actually comic one taken off the the the skin so you can see the thickness you can see the riveting that's the painted side and that was on there so if you pass that around um it's it's a real manufacturing piece he's probably got a couple of sharp corners on so no sticky in your neighbor or whatever so re went and looked at it they found the breakup sequence basically they got this problem origin of fatigue is actually it's a bolt hole here there were signs of fatigue they could map it they could look at it they got it on the microscope the opportunity cool microscope looked at it and said yet signs of fatigue this thing had spent a few months in the water from January all the way through to August it's a satyr the bottom of the Mediterranean saltwater aluminium doesn't cope well so it wasn't in great shape when it came up but they said there were signs of fatigue and so the re opinion at the end of that was that it was structural failure and the pressure Kelvin brought about by fatigue and they've reached that through a number of conclusions the evidence of the break up the fact that yoke uncle failed in the Tet in the tank test early and they went through and they listed these and it's they also said that same basic type was that producing the fatigue test so everything seemed to tie up and they said at the end I went to the absence of wreckage were unable to draw him a definite opinion about the the other name the yoke accident we draw the attention to the fact that the explanation offered for the above accident Adelbert appears to be applicable to that at Naples they put out that their accident report in September 1954 and that was sent straight to the court of inquiry which was set up and the court of inquiry accepted the royal aircraft establishment findings in terms of the cause and actually that was quite quickly dealt with in terms of the inquiry itself so the crash of yoke Peter was caused by fatigue fell of the cabin yo-yo twas probably the same cause and that was the first use of medical forensics to solve an air crash so number of firsts not necessarily all wanted and the course of inquiry also said that you haven't havilland will work he was working at or beyond the limits of naal but had taken all precautions to prevent failure so they actually went round they interviewed a lot of the design team that they interviewed a lot of other experts from the companies and basically said they had they were working beyond that that the knowledge of the day and that fatigue was one of those things that had just come out and caught them but they had done everything though they've gone beyond the air registration boards information and requirements to actually get this thing signed off but sadly it was the order they did the tests because if you remember going back there was 16,000 cycles on one of those pieces why hadn't that failed they did a 2p2 proof test on it before they started that is not a requirement for an airframe it was only a requirement that de havilland did a 2p2 proof test on a part and a fatigue test so they thought we'll do the proof test first then we'll fatigue it to failure two tests by doing a twice proof pressure test they probably blunted all the cracks and it then took on a further sixteen thousand cycles to get them sharp and failing and he was just the order they did the tests operating beyond the limits of knowledge today we'd look at that and go yeah I can see that then no idea it is very easy to look back with hindsight and be critical of de havilland with today's knowledge it is very difficult to do it with the knowledge of the day and to actually read the textbooks of the time and say from this information could I have made a different choice and professor Murphy came out with possibly one of the comments of the inquiry enough is now known about the fundamental physics of fatigue for engineers to be aware there is so much to be learnt well that's handy it's nearly as good as Donald Rumsfeld's unknown unknowns isn't it you know we're in that kind of League right so we get professor Murphy's quote but what what we know now I mean the general story the accident was caused by fatigue failure of the airframe square windows were the cause of high stresses there's a piece of information that the radius on those windows and the size of those windows was exactly the same to within about five percent of that of the Boeing 737 which is also a pressurized and see 9 and the number of other aircraft there were no square windows they were rectangular with rounded corners as every other plane is so the fact that they Havilland put oval windows into the later Marx was not because it was the squareness of the windows that caused failure and it was all reported in the media so what happens next well royal aircraft establishment has managed to get their hands on a number of airframes out of BOAC so they did all sorts of tank tests they did they need 40 test after fatigue tests they blew you Robert up filled it full of water until it went pop they actually welded up all the windows and the doors and then filled it full of water to see when the rest of the cabin went that was a really good pressure test and the Comets were no longer in service so they basically used as much as they could and they monitored stresses so the Royal aircraft establishment generated a lot of data on aircraft structures and they got charts like this which is number of cycles up there pressure pressure cycles crack length so they had maps of cycles against crack length and they basically sat on these reports they got lots of data that's just some of it and they end up putting more strain gauges on there and you've got the rivet rows next to the windows you got the edge of the window frame and you can see how the stress dropped down and we can start to say yeah they did lots of work they they basically pulled a load a lot of information off those airframes but actually the understanding fatigue developed over the coming years it was really a wake-up call to the the the kind of a scientific kind of groups if it could if fatigue could take planes out of the air then it needed to be something that we understood better and so people started to work a lot more in fatigue try to get some kind of understanding of it and methods for relating crack fatigue crack growth rates were instant to the instantaneous crack length apply and applies for somebody who published in 61 Paris an early gun put polish nail or in 1963 so by 1963 nearly nine years after the crash just get the laws today that we would so readily apply to this kind of situation and these empirical methods are now applied to metals and alloys DT D 546 the aluminium alloy was superseded and not particularly investigated further II kind of disappeared and people took on other alloys now if we were doing it today we'd use the Paris law first published 1963 by Paris and again the adn is a Delta K to the M so you've got your instantaneous crack growth rates so you can get that from this data and your stress intensity factor you can get your Paris exponent that tells you how good your alloy is and you can calculate the fracture toughness from this failure stress data so we could actually start to take get some of this stuff that we've generated the equations we've generated more recently and apply it back to the original data if we can get the data now fracture toughness we've got because re tended to blow you know see what what point the crack started to run themselves unfortunate for DT D there's little data it wasn't tested in if in any kind of manner that's mostly useful and so as it fell out of use so what do we have we've got these fatigue tests that the Royal aircraft establishment focused on so yoke uncle Robert would were tested and patched up and tested and patched up so every time they found a fatigue crack growing it got to a particular length they stopped it they carried on strain gauge from those tests and evidence of the rebuild from York Peter we got all this data and so we can start to say well if we get this data the re had of number of cycles against crack length which they had for a number of different cracks probably somewhere in the region of ten cracks that were well-documented we've also got the stressing from these strain gauges so we can work out the stress range that we've got because we now know exactly how long the cracks were from the previous ones and we know that the strain gauge data to give her stresses so we starting to build some kind of story about what we can get now on the comment the fuselage diameter was about 3.2 meters skin thickness varied and the coming between about point seven point nine millimeters carrying pressure of 59 mega Pascal's so we start to get if we as a tube we can start to work out the hoop stress of 128 mega Pascal's slightly higher than the major general stress of 69 mega Pascal founder and then re work but that's not really surprised because you got ribs and stringers in there so you've taken some of the stress out with the design of the aircraft five cracks were allowed to go to failure totally between 149 180 millimeters which gives us a fracture toughness of about 50 mega Pascal root meters and so we can start to say we can work out how what was the cable and see of these this particular alloy and if we stand start to say well using the data available we can calculate the delta K the stress intensity factor a is the crack length treated as a single crack at a hole in an infinite sheet because it's easier under bikes you're loading hoop and longitudinal stress the stress range can be calculated from the measurements in stress and the tank test that does vary with distance from the window so you have to factor that in and that sort varies with crack length and if you do all of that so you add those two bits of data together you can start to get a chart that looks a bit like that with a scattered data so if we then say we've got this tank test data yo-yo company yet Robert using that data we can take long for both sides that's the line through it by the way that I've seen worse data we their line plotted through it but it it's not great but that's the data we've got and I can't actually find any more data and so using the data for the 1950s we can calculate the Paris explain its aim for the material is about five which isn't too bad for kind of material of that it major fracture toughness of around 50 mega Pascal's so it's slightly worse than a material today but not cripplingly so that you'd start to worry that the material wasn't up to it however initial cracks were likely to be very small less than a millimeter in size so right start starting out at the start of the journey that's how big they'd have been therefore difficulty in to identify on inspection the initial manufacturing crack on you opt would like to be hidden but under the bolt head so you work your way back and you see that wasn't a great thing to look at that's actually the roof from yoke Peter it's actually in the Science Museum store it's the only bit of the comment that kept after they closed the investigation down and so that spent a number of months at the bottom of the Mediterranean they pull it back out this is actually not a great replica better ones about don't hand them around this is actually the replica of the bolt hole the bigger lump and the fatigue crack running off if you don't have a quick look at that so that's that's a replica of the actual crack from the yoke Peter so looking in that area that there's a doubler there's a plate that it was stuck to after the investigation but that's the crack in there that's the bolt hole and that's that's the in blue is the rip the inverse of that you get it on the SEM and you start to see well what has happened you've got this growth direction you've got possible striations possible striations I'd say it was yeah with yeah anyway with a bit of imagination but that's the backing plate but basically the fatigue crust crack grew down and went off a fast fracture so we know it failed after that point and that was the limit of the fatigue crack we actually get to the other end and it's not great but this is the chamfer that's the bolt hole so it's just a hole this is the Sham for that the bolt hole SAS that was made around the bolt hole and even though the skin thickness was only kind of 0.9 of a millimeter there was a chamfer on the on the bolt it was flush fitting for aerodynamic reasons and so you had this little crack that probably is about this big before the crack turned in the direction of major stress and that's probably it there so you got this kind of the Sham for the pre crack that was put in during manufacture before it just ran off in fatigue and so you've got this little area of fatigue that really was the kickoff of the the whole yoke Peter failure so the initial defect was around the bolt hole was about 1.5 millimeters long so not great at all and probably couldn't be seen because not only was it under the bolt head it was also painted over and it was also on a doubler plate that was underneath something else so apart from that it was really good too find and we knew that these cracks were there the rivets have been pushed through causing cracks the way they did them they they popped the rivets through they cracked them under the heads that piece of skins got at least one in there that's gone round um cracks could develop at the window rivets they grew towards the window and stopped usually you had two lines of rivets and you ended up with a really interesting kind of stopping mechanism if they were on this side they grew to the window and stops and they were they weren't long enough to fail the aircraft so if they were if they grow the other way they tended to go into the next hole the next rivet hole and stop anyway so actually a lot of the cracks that they found either grew to the window and stopped or grow to another rivet then stopped and so it was only if you had a crack that grew away from the second row of rivets that it didn't have anything to stop it otherwise it's self self stopped critical crack length needed to get up 265 millimeters and so that was reasonably easy to spot if it was in clear skin problem was the bolt hole on yoke Peter was 19 millimeters from the window it was a bolt hole up for other purposes the cracks groover both forward and after the bolt hole they grew to the window and probably after as well and at that point they grew beyond the critical limit it was also under a doubler plate so they wouldn't have found it was also on the ADF window which was their aerial Direction finding window which was on the roof so it was on the roof of the plane under a doubler plate on the coat of paint so then if you try to look at it from the inside there was all the cabin like there was a cabin insulation and the cabin lining so you couldn't there was no access easily to that part of the aircraft he's also and he'd done a few thousand you know less less than two thousand cycles so it hadn't done enough cycles to actually get to a place where it's going to be stripped and inspected so it ran forward now it was just that yoke Peter was had a crack that grew in just the wrong place at the wrong time so some modern analysis if we start with a calculated Paris equation we've got those from that scattergun data and if we substitute in for Delta K of that and we just assume that it's the single crack in infinite sheet and we kind of substitute back in we end up with that equation and so we end up you know feeding all the data in crack growth rates per cycle we've got this big factor here the stress range to the power 5 and then a crack growth rate we rearrange we can get that and then integrating we get that straightforward to do it just a bit messy and so we end up gathering all the terms together we get the crack length from that side we get the stress on this side and we can start to see exactly where we got to and we can use the hoop stress which can be calculated the initial crack length is 1.5 millimeters the final crack length probably about 165 millimeters total crack length from the investigation you feed the beast and you get it life for yolk Peter of 1272 cycles yo Peter crashed on the hood 1290 a flight didn't fiddle that and but that's how the numbers came out assuming that the crack length that I saw on the Sen was 1.5 mil was the number that I fed into there to get that data and so we're just feeding that into standard Paris law equations we end up with a life that isn't far off what yo Peter actually saw so the plane itself behaved exactly as we expected it to he did everything we expected and fell just at the wrong time so you kind of look at it with kind of modern eyes pressurized Calvin brought about you know fatigue I'm not disagreeing with in the 1954 findings the punch riveting rather than drilling drove rivet and glue is one of the causes and actually the upgraded windows the designer of the aircraft actually was asked whether he would rivet the windows or he'd redux them when he was first designing the plane he remembers many years later remember the guy coming into his office and saying do we read ox it or do we rivet it and he said the windows the rectangular windows the fixturing for redock Singh he's going to be really expensive we'll rivet them the Havilland went to oval windows on the subsequent marks because it was easier to Redux them in nothing to do with the stress concentration and it's purely to remove rivets nothing to do with the shape of the windows as I say the shape of the comic windows is very very similar to that on every other jet aircraft that's ever flown up to Boeing 787 so the bolt hole which failed when yo Peter had a defect in the sham fir which indicate the potential for manufacturing defects on all skin holes and the interaction between the stress and manufacturing defects was probably beyond any knowledge they had in 1950s so I agree that that was the way it worked out for them I don't disagree with the findings and it's interesting that modern analysis actually takes you to a place where had they been able to do it themselves they probably could have sorted out the stressing it was just nine years too early the comic for the comic two and three were actually in development at that point and was scrapped because they both they had this shape of Windows and they said they scrapped them off come with a fluid comic four different window designed to reduce riveting fewer manufacturing cracks that's always a good start and it was the first airliner to fly shed your service across the Atlantic on the 4th of October 1958 again the Havilland went and upset somebody by doing that Pan American were going to fly 707 and with big announcement they said passenger service will start across the Atlantic back end of October 1950 1958 do you have another quiet word with BOAC said could you slip slip a service in from early October and they went oh all right then and they did and and the Americans who had done this fanfare of the first jet service kind of came second and sadly it was with the 707 higher capacity and the rest is history but Boeing is dominated since but the aircraft itself Romania services in Nimrod 60 years after the first flight so the wings the tail and some of the fuselage sections actually room we retained and some of the design features like having the engines in the wing roots were really really useful for maritime reconnaissance because putting them close to the centerline of the plane meant that the the RAF could actually turn off three out of the four engines when they were just circling over the Atlantic at slow speeds they didn't need to burn the fuel it wasn't so far out they were sitting on a wing and would give them a massive turning moment and a handling problem so they just cycled which engine they turned off well turned on and and that extended the life of the aircraft so that was a quick jaunt through the comment [Applause]
