VH-OQA’s number two engine was destroyed mid-flight on 4 November 2010. (Richard de Crespigny)

Slaying the black swan

When the engine of a Qantas A380 exploded shortly after take-off from Singapore in 2010, the world expected catastrophe – but saw a miracle, instead. Fifteen years on from flight QF32, Captain Richard de Crespigny shares the memories and lessons from Australia’s worst air disaster that never was.

When Qantas Captain Richard Champion de Crespigny and his flight crew – First Officer Matt Hicks and Second Officer Mark Johnson, as well as Senior Check Captain Dave Evans and Check Captain Harry Wubben – showed up at Singapore’s Changi Airport on the morning of 4 November 2010, it seemed like a routine flight to Sydney was all they needed to expect. The aircraft, VH-OQA “Nancy-Bird Walton”, was Qantas’ first A380, then just two years old and state-of-the-art, was powered by four Rolls-Royce Trent 900 engines, and was set to carry 440 passengers and 29 crew on the roughly eight-hour flight.Only 39 A380s were operating worldwide at the time, with six in the Qantas fleet, compared to almost 160 today, with 10 at Qantas (two others having been scrapped). Recalling those first years of the A380, Captain de Crespigny – who was in 2010 a seasoned pilot with 35 years of flying under his belt – tells Australian Aviation of the excitement of the modern flight deck with fly-by-wire controls, the quiet engines with soundproofed cabin and cockpit, and the many other improvements over older four-engine jets like the Boeing 747.

“It’s the most extraordinary machine. I was very good friends, and still am, with the person who designed all the aerodynamics,” he says. “There is nothing wrong with the A380, and if you take it hub to hub, it’s the only aircraft that will transfer passengers when you’re in a slot-limited airport or saturated airspace.

This content is available exclusively to Australian Aviation members.
Login
Become a Member
To continue reading the rest of this article, please login.

or

To unlock all Australian Aviation magazine content and again unlimited access to our daily news and features, become a member today!
A monthly membership is only $5.99 or save with our annual plans.
PRINT
$49.95 for 1 year Become a Member
See benefits
  • Australian Aviation quarterly print & digital magazines
  • Access to In Focus reports every month on our website
PRINT + DIGITAL
$99.95 for 1 year Become a Member
$179.95 for 2 years Become a Member
See benefits
  • Unlimited access to all Australian Aviation digital content
  • Access to the Australian Aviation app
  • Australian Aviation quarterly print & digital magazines
  • Access to In Focus reports every month on our website
  • Access to our Behind the Lens photo galleries and other exclusive content
  • Daily news updates via our email bulletin
DIGITAL
$5.99 Monthly Become a Member
$59.95 Annual Become a Member
See benefits
  • Unlimited access to all Australian Aviation digital content
  • Access to the Australian Aviation app
  • Australian Aviation quarterly print & digital magazines
  • Access to In Focus reports every month on our website
  • Access to our Behind the Lens photo galleries and other exclusive content
  • Daily news updates via our email bulletin

“Because of the engines and the wings, it’s about 10 per cent less efficient than the new jets, and fuel makes up 30 per cent of the cost of an airline, and that’s why some people question the A380. But apart from that 10 per cent, the A380 is the most extraordinary aircraft ever built. People will criticise the A380, perhaps, because if it’s not full, it’s an anchor, which is terrible – but if it’s full, there’s nothing that will touch it.”

QF32 and its aftermath would prove to be a trial by fire for the A380 that its designers – and its crew – never saw coming.

25 seconds of chaos

At 10:01am Singapore time, about four minutes after take-off on Qantas flight 32, the quiet of the A380 was shattered by two large explosions from the left wing, followed by a blare of alarms and warnings all through the cockpit. Speaking at an Airport Safety Week event in Bankstown last year, de Crespigny remembered the harrowing first 25 to 30 seconds after the explosions.

“I don’t know what they were. I don’t know what caused them, but it didn’t matter,” he said. “But the alarms are now going off. We press the Cancel button to stop the alert, but it comes back. Matt’s pressing the cancel button, but it keeps coming. Bloody noisy, bloody distracting. I press the altitude hold button, which would cause the nose to lower, and the thrust should come off. I’m pressing the cancel button, the noise won’t stop and the thrust should come off, but it doesn’t. Then I press the heading hold button to stop the aircraft following a curved departure, so now we’ve got a constant trajectory. Then I’m on to speed – the speed’s increasing. I don’t know why. I can’t think about it. No one’s saying a word. The top panel’s a sea of red lights. The ECAM treatments are flooding the screen. No one is saying a word. I grabbed the thrust levers and pulled them back, and controlled the speed. Then, I said the first words in about 25 seconds after the explosion – we’re at a constant altitude, we’re at a constant heading, our speed is constant, we’re above the mountains, we are safe – ‘ECAM actions!’”

Unbeknownst to the crew, a badly-manufactured oil stub pipe had caused a massive turbine disc in the number two engine – the one inboard on the port side – to disintegrate, shredding the engine and sending shrapnel flying through the wing, smashing into vital systems and igniting one of the fuel tanks.

“There was an oil pipe running from the outside of engine two into the internals to feed a bearing support for the shafts that run through the centre of the engine,” de Crespigny explains. “The oil pipe came in from the side and connected to a stub pipe, which was just a short pipe to connect into the bearing mount. That stub pipe where this thing plugs in had been badly manufactured. There was a crack. It was thin-walled. It cracked. Oil leaked into a part of the engine that should be dry, nothing in it, but it sits in the cooling air that keeps it cool, which is at 800 degrees Celsius, right? And that’s cooling air that feeds into the turbine blades. But at 800 degrees, the oil auto-ignites, catches fire, and heats up around the shaft and the turbine hub. The hub gets weak and soft. The 50,000 horsepower that the turbines take out of the air twists and torques it off. The shaft comes loose, overspeeds. There are meant to be cookie cutters that, as the turbine moves back, the blades get cut off. That didn’t work. The fuel system actually detected a lower thrust and so put more fuel into the engine. So the first there was a series of bangs, but the second bang was this turbine overspeeding. It’s meant to run at about 7,500 rpm. Who knows what it was at?”

At two-and-a-half times the speed of sound, the pieces of the 160-kilogram, one-metre-diameter turbine disc came flying out of the shattered engine, with three of them hitting the plane. As for what happened next, de Crespigny continues:

“So, this disc came out, energised all these other particles, a bit like energising a cluster bomb. Now, there are two torque tubes that hold the engine to the wing to prevent the 72,000 pounds of thrust from ripping it off. They were both broken. We’re lucky the engine didn’t rip itself off the wing and do what it did to that freighter aircraft [UPS Airlines Flight 2976]. There were 400 impacts on the aircraft: 100 under the wing, 100 on the leading edge, and 200 above the wing. There were impacts up to the top of the eight-metre-high tail plane. For passengers, it sounded like someone was throwing a bag of marbles over a corrugated roof.

“Those pieces came through the wing, through 15 millimetres of aluminium, through a fuel tank full of fuel, creating shock waves, fracturing the earthing leads, creating chaos in the fuel tanks. We had multiple holes in multiple fuel tanks, multiple leaks in multiple systems, outside limits. All the wires that were broken at the leading edge, there were backup wires that went down through the belly to the trailing edge – unfortunately, the third piece that came out of the engine went through the belly and took out two wiring looms with 250 wires each, took out a lot of the backup looms. That’s why we started to lose spoilers. We lost 65 per cent of our ailerons. We lost our leading-edge slats. We lost most of our brakes.”

This was, as de Crespigny details, an event that is never supposed to happen, one so unlikely that it can’t possibly be prepared for: the dreaded “black swan”.

“A black swan, curiously, is named after the black swans in Perth. The colonisers came to Perth thinking that all swans were white – so they were shocked,” he says. “They didn’t know it existed. A black swan event is similar to that: it’s an event that you never thought would happen, unthinkable. You just couldn’t imagine. It has a very low probability of happening, but it has a significant effect. So, for most people, 9/11 was a black swan event. For most people, the global financial crisis was a black swan event.

“QF32 was a black swan, because when that engine number two exploded, certification says the aircraft has to be able to take one piece of that disc that generally explodes into three pieces. You have to take only one piece and survive. Aircraft manufacturers such as Boeing and Airbus say the piece has infinite energy. And we have to stagger the engines forward and backwards so that the exploding engine doesn’t take out the others. And we have to have defences in the fuel systems, any critical system, we have to survive one piece, right?

“The A380 was designed to take one piece and survive – but we had three pieces that hit the aircraft, and it took out all those wires, half the networks. And so the damage was much in excess of what any certification standard has. Aircraft certification takes you up to one in a billion flying hours. We were a million times less probable than that, and so it was unlucky – it was a black swan event.”

Aviate, navigate, communicate

Faced with the loss of hundreds of wires, half the networks, 21 out of 22 systems, and even the sensors that could tell them exactly how many other things were broken, the QF32 crew had fallen back on the first principles of flying: aviate, navigate, communicate. In those 30 seconds, the first two steps were complete – the plane was still flying and on a heading away from danger, so now the proper work could begin of entering a holding pattern and figuring out what exactly had gone wrong, what they had left to work with, and how to get everyone safely back on the ground.

“We don’t look at controlling an engine fire – that can wait. We just have to make sure we’re alive in a crisis,” says de Crespigny. “In the case of some of the famous aircraft crashes, they get tunnel-visioned on a broken light globe that they can’t resolve, and they run out of fuel, and they crash. So, this is what aviation has learned through all the human factors of all the crashes that have killed people. Regulations and laws are written in blood, and this is what we do now to cover it. And again, the crew, not just me, the whole crew, I’m proud of what they did. We worked as a central, unified team, and I think we did a great job.”

Back to those first words de Crespigny spoke after getting the plane on a stable footing: “ECAM actions”. ECAM, or the Electronic Centralised Aircraft Monitoring system, listens to the A380’s 250,000 sensors and parameters for errors, finds the potential causes in its database, and presents solutions to try to mitigate or fix the problem. Unfortunately for the QF32 crew, so much had been damaged in the initial explosions that the ECAM system had gone into overdrive, queueing up page after page of checklists that – had they been the focus – could have caused a lethal distraction.

“If we’d started ECAM to try and fix the problem that we didn’t even know what it was, we might have been halfway through our checklist when we hit the ground inverted,” says de Crespigny. “ECAM is threat and error management laid out in a linear way. Matt’s running through it, though it would take him an hour to get to the end of the first pass of ECAM, and his voice fails because he’s talking through it. We had many ECAMs. We had so many sensors broken, many of them on the CAN bus, which is a bus designed by Bosch for cars but also used on aircraft to bring in third-party products. The CAN bus is a two-wire network, and if you break one of those wires, the gauge reads zero.

Now, does that mean zero oil pressure, or does it say there’s no data? And that was the problem with CAN bus. So, we had broken wires, and we saw many displays on the aircraft in the cruise where we saw the air pressure on the oil tank and the oil pressure level coming down, then low oil pressure, then the ECAM – ‘system failure, turn off these pumps’. We didn’t find out until four months after QF2 that every one of those indications was incorrect.

“Tom Enders, the CEO of Airbus, wrote to me apologizing for ECAM, saying ‘I’m sorry you had so many incorrect checklists’, and there are many more I could talk about – but there are many checklists that we actioned straight away, there are many checklists that we thought about and continued with, and there were many checklists that we absolutely ignored, because if we’d followed those ECAM checklists we might not be here today. That reinforces to every pilot that you must understand your system down to its foundations. You’ve got to understand that when automation fails, you need a secondary or tertiary way to see what’s working – and, of course, always just be driven by facts.”

From Apollo 13 to Qantas 32

Of course, their safe holding pattern couldn’t last forever: sooner or later, the A380’s three working engines would drain the last of its leaking fuel and turn the superjumbo into a glider. The crew had to get the aircraft back on the ground before that happened – but first, they had to deal with the damage not just to the physical structure of the plane, but to their mental models of it, as de Crespigny explains.

“You have a mental model, having arms and legs. You can close your eyes and touch your nose with your finger, because you have a mental model, and when you drive your car, you know where your wheel is to within two inches of hitting the curb. Well, pilots have that too,” he says. “When you know the aeroplane, you wear it like a glove. Feels like part of your body. When things degrade – there are 4 million pieces on an A380 – and the checklist comes out, and we’re losing systems, I’m adjusting my mental model, and eventually there are so many primary and secondary failures. We’ve lost 65 per cent of our roll control – that’s alright, but now we’ve got three fuel imbalances that are out of limits.

“ECAM didn’t say, ‘Hey, are you going to be able to control this aircraft when you come in to land?’ You know, we’ve lost the leading edge slats, so we have to speed up. And I said, ‘That’s okay if I’ve got brakes, but the aircraft didn’t know we’ve lost 60 per cent of our brakes’. The question is, will we be able to stop on the runway? So, my mental model was collapsing because I couldn’t keep track of the primary and secondary failures.”

To regain some control over the situation, de Crespigny turned to another near-disaster in aerospace: Apollo 13 in 1970, which famously saw three astronauts safely return to Earth despite the oxygen tanks on their spacecraft exploding and venting into space. It was the lessons from this return to Earth that would help the QF32 crew attempt their own successful touchdown four decades later.

“I reversed the logic, which is what happened with [NASA lead flight director] Gene Kranz during Apollo 13,” says de Crespigny. “Apollo 13 is going to the moon, and the oxygen tank explodes. The actual spacecraft had 10 million pieces on it, and the mission control is in Houston, with an average age of 26, melting down with all the error messages. They didn’t know what was going on. They were so confused. And Gene Kranz stood up, and he said, ‘Gentlemen, stop wondering about what’s broken, and focus on what’s working.’ Invert the logic. Look at what you’ve got. They did in Houston, and Apollo 13 is probably the most successful failure in human history.

“We inverted the logic on QF32, and then all we knew was forgetting what’s broken. We didn’t need all the transfer galleries, the fuel management computers. We didn’t need boost pumps. We just had fuel sitting over the engines, which could drip down into them. And so when you invert the logic and just say ‘We’ve got two and a half hours of fuel in that tank before that engine runs dry, and three and a half for those engines’, because we had a lot of fuel leaks here, time gives you options. So, you try to find fuel, because fuel gives you time. Time gives you options.

“When we inverted the logic, and I knew we had two-and-a-half hours before this guy goes quiet, the stress came down. Now it has become understandable. Now, you know, we’re building like a Cessna or a Piper aircraft out of a junkyard, and you don’t need much. And providing you understand aircraft and are confident hand-flying it, and you don’t overcomplicate it, and you’re not intimidated by electronics that might even not be working, if you invert the logic, it really changed my whole mindset.”

The ghosts of Amsterdam

As the crew prepared to bring QF32 back to land at Singapore Airport, yet another aerospace disaster was looming large in their minds: El Al Flight 1862. On 4 October 1992, the Israeli 747-200 freighter lost both engines on its right wing shortly after take-off from Amsterdam’s Schiphol Airport, and when it came in to land after circling Amsterdam twice, the pilots lost control and crashed into a residential neighbourhood, killing all four people on board and 43 on the ground.

“They could fly around in the holding pattern quite happily,” says de Crespigny. “When they came into land, when they put the flaps out, there were no slats here. The flaps didn’t work. They’d been damaged, and so those flaps came out, the aircraft rolled inverted into an apartment block, and everyone was killed. So, we learned from that. I’m not criticising these pilots – with aviation, we learn from people’s experiences.”

If QF32 was going to land safely at Changi Airport, the crew needed to prove it was safe to make the attempt with multiple holes in the wing. That meant a control check, as the wing’s coefficient of lift had been badly impacted by the explosion, and the flight controls were working to counteract the roll that resulted.

“We did a dress rehearsal at 4,000 feet – in other words, slowing down, putting flaps out one stage at a time, slowing down the next stage, and then slowing down, then finally, the gears down,” says de Crespigny. “And if the aircraft started to roll uncontrollably, I would aggressively manoeuvre the aircraft to see what it would do, and it was controllable. But if it did start to roll and what we call depart, then you would know that that was an unwise choice, and you should go back. I’m trying to prevent myself from getting to this El Al flight, right? We proved the aircraft safe to land at the approach speed in the landing configuration we had selected.

“That was the most important thing on the flight, because as we came down finals, we were getting speed warnings, which are low energy warnings, when we slowed down one knot, and when we slowed down about three knots. We got a stall warning. Now, I just have to reiterate this. You should never get speed or stall warnings. There should have been a 23 per cent margin, maybe 19 with no slats, between our approach speed and the stall speed. We should never have heard that. So we’re coming down, hearing the speed and stall warnings and alarm bells are going off, saying, ‘what in the hell is going on?’”

Unlike the other warnings and alarms that had plagued the flight crew over the past hour or so in the air, they knew this stall warning was real – it came directly from angle-of-attack sensors, with no other computers that could affect it, de Crespigny recalls.

“Something was really wrong. If we had gone around because we weren’t happy – obviously something was wrong with all our approach, though, we’re coming in far too slow – if I’d gone around, engines one and four that had no overthrust protection would have destroyed themselves, and then it’d be all over. But we made that. We did the control check, we proved the aircraft safe to land. We continued through the speed and stall warnings. I just pushed up one of the thrust levers a little bit. I mean, this is a binary logical aircraft, so plus one knot can stop the alarm every time. I just knew I had to be careful.

“The point I’m saying is, those control checks to prove the aircraft with an unknown wing – the aircraft we landed with was not an A380 wing. It was some other wing that had been damaged and produced less lift, but we proved it safe, and we landed despite all the warnings that we got.”

What made the landing even more remarkable was that the A380 computer’s landing distance performance application (LDPA) had initially told the crew it was impossible – that they couldn’t land at Changi 50 tonnes over maximum weight, that it would overshoot and end in disaster. As de Crespigny relates, the crew had to adjust the inputs to suit the circumstances they found themselves in: for example, removing data for a wet runway, as they knew the runway was dry.

“Dave Evans, the check captain, was in charge of all the calculation landing data. He initially said, ‘We can’t do it. Won’t calculate it. It just refuses’,” he says. “So, we took out a lot of things that we thought weren’t relevant, and eventually Dave came up with 130 metres. That means, at the four-kilometre runway, we would need 3,870 metres. And people say that’s terrible, but I say, no, invert the logic. We have 130 metres. We just have to understand how the test pilots certify the aircraft, which is seven seconds from 50 feet to touchdown. I mean, pilots should know this stuff, right, in case this comes up – seven seconds from 50 feet to touchdown.

If you can do that, you’re following the certification. That’s what I knew. In fact, we did five seconds from 50 feet to touchdown, so we touched down early, but I knew that this is how we’d have to land it. And when we did put on the brakes there, because there were so few brakes left, and only half the spoilers, and the ailerons wouldn’t act as speed brakes, it was a very slow deceleration. But eventually it did stop, and we stopped about 100 metres from the end of the runway, and then eight fire trucks surrounded us to protect the passengers if we had to evacuate.”

Out of the frying pan

After the harrowing effort to land QF32, the passengers and crew likely thought they were out of the woods – but they weren’t until they were safely away from the plane, and as the flight crew discovered, that too was easier said than done.

“After we landed, I was just about to pat Matt on the shoulder and say, ‘well done, it’s over’ – but it’s like a bush fire where the wind changes direction, and now we get 20 more ECAMs that start up,” says de Crespigny. “When we shut down the engines, the APU that’s meant to give us electrics and pneumatics, neither of those works because of multiple damages. So now we’ve got a hot cabin and a 30-degree Singapore day. We’ve got no electrics. We’ve lost six out of our seven radios. We’ve lost the radios we’re trying to talk to the fire controller with, we’ve lost nine out of our 10 screens, and eventually, when the fireman says no, he won’t approach the aircraft because we’ve got to shut down the engines. That’s after we finally managed to contact him using the last working radio. We said, ‘No, the engines are shut down.’ He said, ‘No, they’re not.’ And I stuck my head out, and engine number one is still turning. We’d lost the wires. We’d lost two systems, each using two independent pairs of wires, to the low-pressure and the high-pressure shut-off valves in two different parts of that strut and engine. So, in other words, the black swan got worse, because now we can’t shut down that engine.”

It took the fire crews dousing the engine with water until it eventually flamed out to make sure the plane was safe to disembark – another two hours or so on the ground – at which point the crew made the unusual decision to get the passengers off not via the slides, but via the air stairs. According to de Crespigny, the crew judged that, despite the potential risks of fire, the safest place for the passengers while the engine was dealt with was on board, and that the stairs were the safest way to get everyone from plane to tarmac– and were vindicated.

“When you understand the way that the fire service is set up, you understand what they’re going to do,” he says. “You understand the certification of getting 853 passengers out of half the doors in 90 seconds, and we only had 440 passengers. When you put all this into your risk calculator, we came up saying they’re safer on board right now. You get this sort of intuition after being a pilot in the air for many decades, and we had the most extraordinary pilots in the cockpit that day, and that decision to stay on board was correct. We got 440 people off that aircraft with zero injuries.”

In the aftermath, the QF32 crew – and other Qantas staff members – were quick to debrief the passengers and make sure they were ready to speak to the media, with de Crespigny himself even taking an extraordinary step to make sure any of their concerns were heard.

“I said to the passengers, ‘If you think that Qantas is not doing what you want, if you think that Qantas doesn’t care about you, then here’s my mobile number. Call me’,” he says. “So, I gave my mobile number to every passenger, and to this day, 15 years later, zero have called to complain. Zero have called to ask for help, but about 20 people used it to say, ‘Hey, thanks. Can we have a beer?’ Many of the passengers have become lifelong friends. Many send me pictures of their families and say, ‘Thank you, I have this family now because of you and your crew.’ The aftereffects of QF2 will last as long as all these passengers are alive.

“There were eight teams in this event, and I have to thank all of them, including the passengers, for basically turning a black swan event into an incident where the Qantas brand, which was under threat at the time because it couldn’t get a trick with the media, came out after QF32 stronger than when we went in. There’s not one picture anywhere on the internet of one single passenger crying, coming out of that event, out of that black swan. And that’s because of these eight teams, and that’s why I’m happy here to talk to you and to sing all their praises.”

And while A380 aircraft were grounded worldwide after the near-disaster, de Crespigny says that the plane itself, crippled as it was, still played a starring role in getting everyone safely back to earth.

“I just have to reinforce that the design of this aircraft was extraordinary. I gave a presentation to the 600 leaders at Airbus at their annual conference, and I stood up in front of them, and I said that 469 passengers and crew – and for every person that died, there’ll be 70 people that will go to their funeral – so, 32,000 people would like to thank Airbus for building a resilient aircraft, and that’s what I honestly think. I think all these Airbus fly-by-wire aircraft are amazing. They’re efficient, and I have the highest respect for them. And in the case of QF32, it absolutely proved itself resilient.”

Want to see more stories from trusted news sources?
Make Australian Aviation a preferred news source on Google.
Click here to add Australian Aviation as a preferred news source.

You need to be a member to post comments. Become a member today!
Momentum Media Logo
Most Innovative Company
Copyright © 2007-2026 MOMENTUMMEDIA