Even as Flydubai Flight 981 took off from Dubai on March 18, the pilots knew they’d be in for a difficult flight. Bad weather lay ahead at their destination, the Russian city of Rostov-on-Don. As the plane skirted the Caspian Sea and crossed over the Balkans, the situation stayed iffy. By the time the plane approached Rostov airport, a landing looked challenging, but manageable, with rain and winds gusting to 40 mph.
Setting up for an approach from the northeast, the Boeing 737 broke through the cloud base at 1,800 feet and had the airport in sight directly ahead. But gusty conditions meant a risk of windshear—a sudden tailwind could cause the plane to drop out of the sky. Playing it safe, the flight crew did a “go-around,” increasing engine power and climbing away from the runway. For the next hour and a half the plane flew holding patterns, waiting for a break in the storm, but none came. Finally the pilots decided to bring it around for a second try. Once again they descended through the clouds, got the runway in sight, and set up to land. Once again, wild winds forced them to abort. The plane accelerated and nosed back up into the sky.
Later, security cameras on the ground would show the plane disappearing into the overcast sky—and then, mere seconds later, zooming back out of the clouds at a steep angle and impacting the runway in a fireball, instantly killing all 62 people aboard.
The reason for this tragedy, we now know, was not wind nor rain nor simple pilot error. It was an illusion.
For obvious reasons, initial speculation about what went wrong centered on the weather. Perhaps the plane had been hit by lightning or suffered particularly severe turbulence. Mechanical failure might have played a role, too. In several recent accidents, autopilot malfunction has caused planes to dive unexpectedly. And then there were potential psychological factors. Having already flown nearly two hours longer than they expected, with much of that time spent in turbulence, amid the stressful uncertainty of not knowing how and when they would get their passengers on the ground, the flight crew must have been tired. Pilot fatigue and challenging weather make a dangerous combination.
The picture became clearer this past Wednesday with the release of the official preliminary report (pdf) on the accident by Russian aviation officials. Data recovered from the plane’s black boxes ruled out mechanical failure or a violent weather event. The problem, most likely, was that the pilots fell victim to a pernicious form of disorientation called “somatogravic illusion.”
During a go-around after an aborted landing, a plane tends to be lighter than normal since it’s at the end of its flight and has burned up most of its fuel. That means its thrust-to-weight ratio is relatively high, so when the pilot pushes the throttle forward from idle to full thrust the plane accelerates with unusual alacrity. This acceleration pushes pilots back in their seats, which to the inner ear feels exactly the same as tilting upward.
In this case, the plane really is tilting upwards as it climbs away from the runway. But this weird sensation can throw off even seasoned pilots. As long as they can see the ground below them, the true orientation is clear. “When you initiate the go-around and still have some visual reference, you’re fine,” says aviation analyst Gerry Soejatman, “but once you get into the clouds, your senses start to play on you.”
Black-box data show that as the plane started to enter the cloud after the second go-around, the flight crew briefly pushed the controls forward so that its rate of climb decreased, as if the pilots were momentarily disoriented. Then the plane returned to its previous rate of climb. For a few seconds, all was normal. The flight crew members were almost certainly following their instruments, as years of experience had taught them to do. Then, as if suddenly disoriented and unable to believe their instruments were correct, the flight crew pushed the stick far forward. “It takes time for someone to go from ‘Oh, the instruments are saying this,’ to ‘No, no, no, this is all wrong!’ and start pushing,” Soejatman says.
The pilots probably believed they were preventing the plane from getting too nose-high, which could cause the plane to stall and crash. But in reality they were taking a safe situation and turning it deadly. The lurch downward would have caused them to rise up in their seats as though on a roller-coaster zooming over the top of a hill. By the time they rocketed out of the bottom of the cloud and gained a visual sense of their orientation, they were in a 50 degrees vertical dive at more than 370 mph and just a few seconds from impact. There was no time to pull out.
The violence of the resultant impact can be gauged by the by the condition of the remains recovered. From the 62 people aboard the plane, 4295 “samples of biological matter” were collected.
Somatogravic illusions don’t cause plane crashes often, but a 2013 study by the French transportation safety agency identified 16 similar incidents. One crash that happened just two and a half years prior to the FlyDubai crash was eerily similar. Coming into Kazan, Russia, Tatarstan Flight 363 aborted a landing amid low clouds and gusty winds, started to climb out, then suddenly pitched down and plunged into the ground at a steep angle and high speed. All 50 people aboard that 737 were killed.
Wednesday’s report was only a preliminary finding, meaning that investigators’ findings may change. For the time being, however, they’re recommending that pilots undergo fresh training in how to conduct go-arounds under different conditions and study how somatogravic illusions can occur.
This piece was originally published on the Popular Mechanics website.
In other words robots should be flying and driving for us.
@Jeff Wise:
Your link to the preliminary report fails. Here it is:
http://mak-iac.org/upload/iblock/be4/Interim%20Report%20A6-FDN%20%28en%29.pdf
@Gysbreght, Thanks! Updated.
@Jeff Wise.
This is quite a shift.
Does it mean you ended the discussions on MH370?
And if so. Why now?
@Jeff thanks, very interesting and good article; at least the victims had it quick here
@Brock
what a coincidence as my Monty Python piece was randomly picked and randomly turned over, while mentioning Madagascar too; not intent, really, I swear…
BTW, its wonderfull how you can apply the mathematics and statistics science together with some east end boys to west end potentionally crashed plane debris; real top science investigators were probably learning itself mostly, but teaching mathematics in our country is very old and bad style – too early digging into much of details of too complex and unaplicable things with synthetic uninteresting examples, contrary to teaching abstraction and separating real useful signals from the other noise – and nobody for example mentions interdisciplinar usability of things like linguistic analysis of language sentences to words and their semantics, despite language teaching goes deeper into the sentence building too, but not noticing about real usability of this at all… I personally realized somehow all this only at high school, that mathematics/statistics can be used together with linguistic sentence de/composition to write algorithms which can be used to detect some texts meaning etc… for sure heavily used today by NSA, GHCQ etc, and… for good reasons, to catch all the bad guys everywhere… so, no place to hide;
… unfortunatelly, now I cant do the good use of mathematics because of such bad experience in school, but its really wonderfull to see how it can be used and how much important, or most important it is; thanks
@Falken
Yes, math is wonderful. We may someday be able to figure out (with advances in text parsers) what Donald Rumsfeld was trying to say.
It seems quite clear the reason for this crash and many of the other ones after 9M-MRO. But still there is so much issues finding 9M-MRO. It seems to be much cover ups going on.
@Brock
Thanks for a very transparent analysis. What would happen if you used the known locations of the finds and plugged them into your analysis to reverse engineer it. Could they lead to a range of probable impact points?
Re: drift probability study (linked to in Jeff’s “Sidestep” post):
@Ge Rijn, @falken: thanks for your kindness and encouragement.
(Techie addendum #2: I used two stat functions whose syntax will confuse old versions of Excel, Easy fix: search & replace ‘_xlfn.GAMMA.INV’ with ‘GAMMAINV’, and ‘_xlfn.BINOM.INV’ with ‘CRITBINOM’.)
?????? What is this all??
The often critisized narcissistic, opportunistic ways of American journalism?
And then folks follow without any question asked?
Without any explanation of the writer?
What character you have?
My God.. First I need to hear a good explanation, for I regard it is quite unrespectfull to all the people who were and are so intensly involved with MH370.
@Brock
Yes, you’ve done about as well as possible with the model data available, and it does confirm what I have inferred from just eyeballing other models. I am always suspicious of the accuracy and stability of the drift forcing functions themselves not to mention the variability of the wind. Nothing you or anyone else can do about that.
Thanks.
@Jeff, @falken
“From the 62 people aboard the plane, 4295 “samples of biological matter” were collected.”
“very interesting and good article; at least the victims had it quick here”
On average, each people was divided into 69.27 pieces of about 1 Kg each (for an average 75 Kg person). I would not like to be a family member to collect my part of these “samples of biological matter”…
@Trip: you’re very welcome. The IPRC probabilities are in the forward direction only. Several studies attempted to back-trace the Réunion flaperon – linked below is what Jeff called my “meta-analysis” of drift studies, which compiled and analyzed as many as I could find.
Bottom line: particularly given WHERE it was found, the lone flaperon could not do much to narrow down the search zone (other than to strongly suggest the southern end of the search box was surprisingly UNlikely).
Essentially, drift experts threw up their hands, and said they needed more pieces before they could begin to triangulate it down. Yet I have not seen or heard anything from any drift expert in response to the March debris flurry.
Pure speculation on my part: I’m guessing they’ve found their models generally don’t jive with the debris record to date, but risk:reward analyses inhibit public comment.
https://drive.google.com/file/d/0B-r3yuaF2p72RnNtRVZuSVdHaG8/view?usp=sharing
@Ge Rijn (exclusively):
Since you like dutch proverbs: “spreken is zilver , zwijgen is goud”
Re: 981
IAC interim report linked below.
http://mak-iac.org/upload/iblock/19b/Interim%20Report%20A6-FDN%20(en).pdf
The PIC had only 1000 hours under his belt as a PIC on that aircraft type. Not a great deal of experience.
Aviation Week and other sources report a serious pilot shortage, particularly in the Asis/Pacific region. Carriers are considering lowering some barriers to entry that would make it easier for retired pilots to return to active status. Frankly, I would prefer that relative to accelerating the training of new pilots.
@Gysbreght
I spend many interesting, and mostly happy, days in Eindhoven driving BMW’s around testing navigation systems with my customer at the time, Siemens VDO. Good people, great town, incredible food,…
VDO was headquartered in Regensburg at the time, and I would use any excuse (even a butt chewing) to get there to experience the Regensburg Sausage Kitchen on the Danube. I could not get enough of that stuff.
@Ge Rijn, Don’t panic. If you had been reading this blog longer, you would know that while I am certainly heavily invested in MH370, I write about a wide range of aviation topics. You would also know that many readers who are equally invested in MH370 also are interested in these other topics. Finally, you would also know that the ongoing conversation in the comments section largely proceeds regardless of whatever the topic happens to be about.
@Brock. Thanks for your probablistic analysis of shoreline debris and ‘was the Curtin Boom mh370’. Very interesting. Also glad that you don’t discard the Maldives finds lightly.
But Madagascar’s lack of debris remains puzzling, in this scenario too.
Anyone going to Madagascar for the holidays?
@carla – I agree that more debris should have been found by now but if you read statements of some of those who found debris, you’ll see something in common: the local villagers and fishermen did not know about MH370 and/or did not know that debris was expected on their beaches.
@DennisW: out of curiosity – BMW and Siemens VDO are both very much German, so what brought you to Eindhoven?
@Gysbreght
That is where VDO had their car navigation development center at the time headed up by Toine Van Kessel – brilliant guy relative to routing algorithms, and the blending of GPS, inertial, and map matching.
https://www.linkedin.com/in/toine-van-kessel-708a212?authType=name&authToken=DOOk
The German corporate guys helped by keeping us informed about how much we were behind schedule, over budget, and that we all sucked.
@DennisW
I probably dont know what you exactly mean about Rumsfeld, maybe something about Iraqi WMDs? Fact is(?), that president Bush himself probably realized then that it is enough of stupid stuff and will be better to try something totally different, what is being implemented right now.
So what about Singapore and MH370?
Something went wrong.
Everything here is displayed in italics.
http://www.bbc.com/news/magazine-35706048
In depth article about MH17.
@Falken
Rumsfeld had a very difficult time with the English language. It was difficult to understand anything he said without significantly reorganizing the words.
Having been a pilot and a flight instructor; vertigo is a concern even with experienced pilots. Continued training includes vertigo situations. Looks like even more training is required. In some societies (Korean comes to mind) they have trouble flying by the seat of their pants. No private airplanes in Korea. This might be a benefit for they due trust the instruments. When flying into San Francisco by the seat of their pants; Korean air had a problem with their 777.
@ Gysbreght
A word on aerodynamic tolerances. Aerodynamic tolerances very based on; location on an airplane, the purpose of the airplane and the need to achieve laminar flow. For Aero smoothness; The leading edge of the wing/flight surface is the most critical with the aft section of the fuselage being the least critical. Other areas are between the two. Examples: Production smoothness tolerances on LE laminar flow parts used on a NASA test section were within .001 over a 2 inch distance for a part 22 feet long. Achieved on a 757 LE of a wing section used for an active laminar flow test. Open class and 15 meter sailplanes have similar tolerances on LE. The aft section of fuselages has lower tolerances; for some Boeing airplanes, where button head protruding fasteners are used, because the boundary layer thickness is measured in inches. e.g. Boundary layer is where there is turbulent flow next to the airplane. Thus no drag impact. Weight is the consideration.
So; fasteners on the TE of the wing, past the 2/3 point, will in general have a very slight protrusion + .001 inches allowed above the surface with a greater tolerance below the surface of maybe -.003 inches. Below surface tolerance is governed by strength requirements not aero tolerances. These tolerances are as tight as can be achieved, to reduce variation in design (improve strength calculations), within the production environment, without undue costs. Note: Some fasteners are shaved smooth on metal wing box aero surfaces.
So, No filling of fasteners is required or even allowed unless for other purposes ( i.e. electrical/static requirements; like the stuff that comes from clouds). Paint fills slight variations for appearance purposes.
I have a question born of idle curiosity: Why is the font for this post and the ensuing comments in all italics?
Link below to a Monte Carlo simulation by Connor Myhrvold which appeared in Fast Company about a year or so ago. It is interesting because Connor includes all his Python Code. It is also interesting for the early link to a Harvard lecture series, AM207, on the Monte Carlo Method which is actually quite good IMO for those who want to brush up or dive in. You cannot beat free.
http://nbviewer.jupyter.org/github/myhrvold/MH370_MCMC/blob/master/MH370_MC_ConorMyhrvold-V3.ipynb?create=1
@Nate:
“I have a question born of idle curiosity: Why is the font for this post and the ensuing comments in all italics?”
This is the first part of the hack. The second part includes redirecting all the messages to an IP address somewhere in the middle of the South Indian Ocean. Printed messages will then appear two years later on a beach somewhere in East Africa…
@Falken @Dennis Not meaning to speak for Dennis W, but here is popular example of Rumsfeldspeak:
“Reports that say that something hasn’t happened are always interesting to me, because as we know, there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns – the ones we don’t know we don’t know. And if one looks throughout the history of our country and other free countries, it is the latter category that tend to be the difficult ones.”
I think I know that he means…
The fastener we’re talking about is not at the trailing edge. It attaches the #3 upper forward torque box panel to the front spar of the main torque box at between 25% and 30% chord of the horizontal tail.
Boeing’s aerodynamic smoothness specifications are not made by NASA. If you know what they are, please let me know. Please explain also what the reason is for the aerodynamic PKC that I referred to.
@ DennisW – re: Connor Myhrvold….what i would give to understand/comprehend that link you suggested….boy o boy, to an interested layperson like myself that is some kind of possible….well im at a loss for words…only spent about 20 minutes scanning it until laymans vertigo got the best of me…thanks…G.C.
Some possible factors influencing the FlyDubai Crash. Just my oppinion restricted to the flying ops at the last go around, not to be misunderstood as fact
1. Low weight high power go around, leading to quick acceleration and steep climb angle without correction
2. Improper CRM. An instrument go around starts with all eyes to the instrument panel, one guy flying, the other guy monitoring. Awareness in the instruments is the only way to stay within the defined operating limits.
3. Due to 1. and 2. the PF and maybe the PNF as well were falling for somatogravic illusion, feeling being in a much more steep climb than they really were. Reducing the climb rate without adjusting the power was increasing the acceleration, being sensed as a steepening of the climb. Cycling of the flaps due to load relief was contributing to the confusion and sensoric overload.
4. Late recognition of this illusion led to overcorrection by bunting the nose down, resulting into -1g according to the report. Even fighter crews try to avoid negative g, and this in a B737 close to ground was a totally unknown feeling for the crew.
5. The funny thing is the nose down trim running for 12 seconds. I can see no situation were a sound pilot would do that for whatever reason. Could be that something or even one of the pilots got misplaced due to this -1g which in the following blocked the trim and the steering wheel in the nose down position.
6. Overall main points: Automation dependency, unsatisfactory power and pitch control, lack of appropriate CRM procedures and maybe unsafe cockpit for landing (loose items, not fastend seat belts)
Oleksandr – I haven’t done any recent beach surveying at all but since we have been discussing the towelette thousands of these things will have arrived in Perth. And I reckon the sun would have made a much bigger impression on the towelette than the water. Plastics not designed for UV tend not to hold up?
I also wonder what recovery actions were undertaken and when they started. Over 50° nose-down pitch at impact …? We’ll have to wait for more data in the finbal report.
@Jeff, @Greg Long.
Are the italics everywhere intentional? I’m finding it hard to read.
@Brock. Thanks for your great new analysis; on first reading it does suggest that the current search area is way wrong. I’ll study it carefully as soon as I have time.
@Gysbreght
Early speculation was lack of adequate crew rest might have been a factor. Certainly it was a long and tedious flight by the time the second approach was attempted after 1.5 hours of holding, and getting bounced around. Not a lot of combined hours of experience in the cockpit either. Seems like the classic accumulation of contributing factors.
@Donald Rumsfeld
“Stuff Happens” – commenting on Abu Ghraib
@Brock – great job on the analysis.
@George C
Yes, it can be overwhelming if you are not “into” that particular thing. I have a similar feeling to yours when I look at a Picasso sketch of a butterfly or sit near Segovia while he is playing a classic guitar. Not to say Connor is in the same zip code as Picasso or Segovia relative to his chosen interest.
I once roomed with a mechanical engineer, and a very good classic guitarist on a consulting assignment. After I finished a piece he would label it as a “variation on theme by XXXX”. He was very PC.
@all, Sorry for the italics. I neglected to close the tag in the title. FWIW, I had just finished 36 hours in Disneyworld with two young boys and could not be held responsible for my actions.
@Retired F4, Wonderful insights, thank you.
@All Pamela Mayer’s YouTube video on liespotting makes some great points on deception. She says that innocent people will always want to help in every way possible regardless of where it might lead and guilty ones will always try to obscure the facts and point elsewhere. Think about the nations involved and ask which ones immediately stepped forward and which ones held back.
@Jeff
We hold you responsible for everything whether it stems from your actions or not.
@Dennis, 🙂
@Jeff,
Great article here and great you are increasing awareness of this somatogravic illusion disorientation problem. My heart goes out to the FlyDubai victim’s families too.
@DennisW
Interesting about an Asia/Pacific pilot shortage. That would have made Zaharie more valuable had MH370 not happened. There is something to be said about vast experience vs. fresh out of training younger pilots I agree. Wouldn’t the salaries for the younger guys be less as well? What is more cost effective bringing the old guys back or retraining the younger ones?
@Cheryl
I cannot speak to the economics – old guys returning vs young guys coming on.
@Jeff, Thanks for fixing the italics. The fairly new list “The Discussion” on the top RHS is very useful BTW.
@Shadynuk @DennisW
thanks, and – well, so we are back again here
“I dont know prime minister; I dont know, what you dont know”
https://www.youtube.com/watch?v=IKQlQlQ6_pk