Ethiopian crash raises safety questions
Sunday’s crash of an Ethiopian airliner is a terrible tragedy in which 157 people lost their lives. The scenes of anguish and destruction are horrific and heartbreaking.
Air travel’s safety record has improved tremendously, but any crash feels intolerable and everyone would like to completely eliminate such occurrences. This particular crash raises exceedingly disturbing safety questions because it appears to echo the circumstances of another recent crash that may have had something to do with the design and operation of the plane model involved, the Boeing 737 MAX, which is a relatively new design although substantially similar to an older one:
The accident drew immediate parallels to the Oct. 29 crash of a Lion Air plane that plunged from the skies in Indonesia and into the Java Sea, killing all 189 passengers and crew members.
A new MAX 8, an upgraded, more fuel-efficient aircraft from Boeing’s popular 737 line, also was involved in that calamity. In both instances, the pilots tried to return to the airport a few minutes after takeoff but were not able to make it back. And both flights experienced drastic speed fluctuations during ascent.
But experts warn that doesn’t mean the reasons they plummeted were the same.
But it’s hard to believe there wasn’t a connection, until proven otherwise. Air crashes have become less common than ever, but it’s my impression that crashes are particularly unusual with new planes and clear weather unless some sort of design flaw or error-encouraging design glitch is involved.
More:
The cause of the Lion Air disaster is still being examined 4½ months later, but investigators are looking into whether an incorrect reading from a sensor activated an automatic command to lower the plane’s nose. The pilots had tried to reverse the command.
Boeing did not take the costly step of retraining pilots on that new feature of the flight control system when it introduced the MAX 8 in 2017, based on the argument that the new models flew essentially the same way as the familiar 737s. Pilots complained that neither the company, the airlines nor the Federal Aviation Administration informed them of the change.
After the crash, Boeing sent out an advisory telling pilots how to override the software upgrade that created the problem.
American and Southwest currently fly the plane, but it’s a very small part of their fleets, 24 out of 1000 for American and 31 out of 750 for Southwest.
RIP all those who have perished, and condolences to all their their relatives and friends.
Okay, it’s the Daily Mail, and eyewitnesses notoriously see things that didn’t really happen. However, quoted is someone who saw flames and items falling from the rear of the plane before it crashed.
After the previous crash, I understood that Boeing had briefed all airlines having these planes. Ethiopian has been considered one of the better airlines.
I make that 55 tubes of potential disaster overhead. I’d be seeing if either Southwest and or American will take proactive measures in advance of any official report.
How to override means manipulating something down by the pilot’s shins!
DiFi wants the jet to be grounded in the US. Who the hell is she? The investigation hasn’t even begun!
A lot of these modifications made, I believe, to make the aircraft less “pilot error” prone. Ironic if this is one of them.
Sad, terrible. One Slovak (nationalist, I don’t like him) lost his wife, son, and daughter. Another Slovak Christian charity (I do like a lot) lost their altruistic worker.
Probably software error. Exactly the fear of auto-pilots in cars.
In the LionAir crash, the computer overrode the commands of the pilot, which is NEVER EVER a good thing. It’s likely that this happened here with the Ethiopian Air jet, as well, and we’ll know soon enough.
But computers aren’t perfect, and neither are pilots. Accidents will happen.
Harry, couldn’t agree more. The big issue with the Lion Air is that Boeing did not emphasize the modifications in manuals and training programs; and the system responded differently than to certain situations than previously.
Of course, ignorance of the modifications should not apply in this instance.
In my opinion, Airbus led the charge in the effort to have the engineers take the pilots out of the loop. Maybe Boeing is now following suit.
DiFi may not be senile; but, she fools me. Either way, like so many others she cannot resist weighing in on complicated situations despite gross ignorance.
However, there is another factor. The system response that was suspect in the Lion Air crash does not occur when the autopilot is engaged; only when the pilots are hand flying. Increasingly, pilots are taught not to hand fly the airplane–so, they do not develop the intuitive skills to react to situations. Before I retired I actually worked for a foreign manufacturer. When training pilots in the plane, there was a constant debate between American instructors, and the others. We wanted more emphasis on hand flying, while monitoring the system, so that they also were conversant with it. Since the airplanes were sold heavily in the Third World, they wanted to teach the pilots which buttons to push.
Nutz. I want a real human person in the hot seat, who knows how to fly the thing if the computer has a tantrum or the wings fall off or whatever.
Julie, my brother was a Marine pilot and then flew for decades for a major commercial airline. He tried to stay on 727s as long as he could because he wanted to actually fly the plane. He has commented to me on at least a couple of fatal accidents that he felt were caused by pilots who relied too much on autopilot devices and couldn’t react to an emergency situation.
Having said that, over the years flying has become much safer overall. There may be a catch-22: automated/computerized systems might have reduced the incidence of human error; even good pilots can be subject to fatigue and stress. But at the same time they have reduced the opportunity for pilots to practice their skills which are needed at rare but critical times.
FOAF – as the autopilots get better, there will be fewer human pilot error accidents, but a higher proportion of crashes will be because of special autopilot behavior / error. There will be increasing AI built into the autopilots.
I’d expect a near future software enhancement to have more “conversation/ feedback” between the human pilot and the AI-pilot.
Teaching pilots how to react in emergency situations, gaining skill at critical times, will be harder (worse results) and harder the better the AI-pilots get.
I thought I remembered something about a 747 that landed safely after about 1/3 of a wing was torn off. Didn’t find that story (my marbles long ago escaped an their former sites are now occupied by dust-bunnies or not at all), but Quora gives two other examples, and FWIW, notes:
https://www.quora.com/Can-a-plane-fly-with-only-one-wing
ANSWER:
In 1983, Israeli pilot Zivi Nedivi, found himself in a bit of a pickle after his F-15 collided mid-air with a Skyhawk.
The aircraft immediately fell into a tight spiral and started losing fuel at a high rate. Zivi was ordered to eject, but by then he had regained control and was confident he could land. He flew 10 miles to the nearest airfield, but again entered a spin upon decreasing speed. On the verge of ejecting, he decided to light the afterburners, which poured on speed, ended the spin, and again restored control.
So he lowered his tail hook and made an emergency high-speed landing at about 260 knots—twice the normal landing speed, and just managed to stop about 10 meters short of the arrester barrier.
Then he got out, turned around, and saw this: [photo at source]
There is no way any aircraft should fly with essentially all of one wing missing. However, part of the F15’s lift comes from the shape of the engine intakes and fuselage. In addition, its electronic stability augmentation system probably helped auto-compensate for the asymmetric lift.
Do not try this in a Buff.
…
There is, however, the case of Pan Am 843, a Boeing 707, which in 1965 suffered an explosive engine disintegration a few minutes after takoff, lost the wing beyond the outboard engine pylon, and remained flyable long enough to reach Travis AFB.
This image was supposedly captured in flight by a passenger camera: [photo]
Coincidentally, the 707 sent to Travis to carry the passengers on to Hawaii suffered a forward gear collapse and crashed, and a third plane had to be sent to retrieve the passengers. I’m guessing they were out of liquor by the time they reached the islands.
– – –
And of course there’s the famous “Miracle on the Hudson” landing by Chesley Sullenberger, 6/15/2009. Great photo at
https://abcnews.go.com/GMA/News/pieces-capt-chesley-sully-sullenberger-10th-anniversary-miraculous/story?id=60334892
Ethiopian Air has long had a very good safety record, so I would look at something going wrong in the aircraft instead of pilot error in this case. Maybe we’ll find out someday — I sure hope so, as there are a whole lot of 737s in service all over the world.
One of the criticisms I heard about Airbus from American pilots early in their history was that pilots could not easily override the computers. It would be sadly ironic if this turns out to be the case in the Lion Air and Ethiopian crashes.
As a glider instructor, I encourage people who want a power pilot rating to learn to fly a glider first, so they really understand what makes an aircraft fly. Understanding energy management in a glider makes for good instincts in a power plane.
“But computers aren’t perfect…”
https://www.dailymail.co.uk/news/article-6364901/More-500-village-Post-Office-staff-sue-company-accused-stealing-missing-money.html
As I understand the finding from the Lion Air investigations, there are two Angle of Attack sensors. If they detect too steep of an angle…which in this model could be caused by quickly adding power, causing the nose to rise, then the system will command nose-down trim. I’m pretty sure the flight crew always had the ability to disconnect and revert to manual trim.
Problem was, one of the AA sensors was bad, and apparently the system relied on a single-sensor input to command nose-down trim when it wasn’t really needed. Other problem was, the pilots were apparently unaware of this potential behavior and didn’t understand what was happening in time.
I don’t know whether the autotrim on this airplane would cause the manual trim wheel to spin, which could have been a visual cue indicating the system operation.
Way too early to say whether the same thing happened in this case or not. Almost certainly, the Ethiopian Air flight crew had been made aware of the potential problem.
I discussed some examples of automated systems that didn’t do what they were supposed to…in aviation, medicine, and maritime operations…here: When Humans and Robots Communicate, they do not always achieve Mutual Understanding
https://chicagoboyz.net/archives/39039.html
David Foster discusses information that I have heard from pilot friends who are familiar with the system. His point raises a couple of questions. One, was the Lion Air aircraft dispatched with one AOA system operating? Although the designers obviously felt that redundancy was necessary in such a critical system; it may be that governing authorities permit dispatch with only one. We don’t know if this is the case right now. Is there an implication that the other system failed? Haven’t seen that postulated. The supposition is that the crew got into a situation in which the system performed as intended; but not expected by the crew. Repeat supposition. What I have heard through knowledgeable, but unofficial, sources is that the system is designed to counteract the nose up pitching moment that occurs when large thrust increases are commanded to the under slung engines, at high angle of attack. If true, no one has explained why Lion Air was at such high angle of attack, and why a large thrust increase was needed.
The only known correlation is that both accidents involved the B737-Max8; and both involved third world airlines. I doubt anyone in the public, political, or media sectors knows whether there was any further correlation.
There are so many unknowns, that calls to ground the airplane are specious at best. One wonders if they originated at Airbus.
Having participated in a number of aircraft accident investigations, I can say with certainty that the actual cause is frequently not even considered in the early stages.
‘One wonders if they originated at Airbus’
That was the first thing that crossed my mind when Europe jumped on board. Knock Boeing’s stock down over 10% in a couple days and call into question one of their newer models in one fell swoop.
If in fact the aircraft was brought down by a technical or computer-related malfunction, it cannot be stressed enough that Murphy’s law is relentless and inexorable.
Which perhaps ought to make one wonder about some of the decisions that have been made by the aviation industry—particularly when a pilot, forced to take over in an emergency situation, is confronted with what appear to be counter-intuitive phenomena (when he/she has only a very short time to respond)…not to mention computer—and backup system—failure.
Air France #447, unfortunately, comes to mind.
(This should also be a red flag for self-driving vehicles.)
The question is, how many people is it “acceptable” to sacrifice on the altar of technological advancement; for technological “improvement”?
Regarding the current 737 model (via Instapundit):
https://www.dallasnews.com/business/airlines/2019/03/12/boeing-737-max-8-pilots-complained-feds-months-suspected-safety-flaw
The database that these news articles are referring to is the Aviation Safety Reporting System, a very useful system managed by NASA. Any pilot, mechanic, or controller can anonymously describe a safety issue that they are concerned about. (Including one where they themselves may have screwed up, which is one reason for the anonymity)
While the reporting of 5 safety-related issues with the Max 8 does sound bad, I wonder how it compares with reporting on other models? I doubt that there is any airplane ever made that *some* pilot or mechanic doesn’t see as having something that badly needs changing.
Trump grounds the 737 Max8.
I heard a retired commercial pilot who read those Aviation Safety Reporting System reports, and he said that he reversed his opinion towards a grounding, after reading them.
Then I heard some retired head or deputy head of the fed. Dept. of Transportation. He echoes Oldflyer’s comment a bit. Remember he’s a suit and may be talking out of a lack of aviation knowledge (or not).
He said Boeing “patched” an extra large (high bypass?) pair of engines on the old airframe to compete with the high fuel efficiency of Airbus’ new planes, and now the combo has more pitching moment under high thrust. He also vaguely suggested that the whole plane is just less stable, and that it really needs a redesign, rather than slapping new engines and mounting pylons onto the old structure.
All this mess with software and auto-pilot and anti-stall problems are really Boeing’s band-aid to fix these stability problems, according to him.
Further reports/speculation that appears to corroborate the above:
https://www.dailymail.co.uk/news/article-6803773/The-Boeing-safety-feature-linked-737-Max-8-jets-crashed-off.html
If these reports are correct, the crashes were caused by a “safety feature” introduced by Boeing because of the heavier “cleaner” engines used in the new 737 “MAX” version.
Irony.
Key graphs:
“Both planes were fitted with the Maneuvering Characteristics Augmentation System (MCAS) – an automated safety feature designed to prevent the plane from entering into a stall, or losing lift.
“Angle of attack sensors on the aircraft tell the MCAS to automatically point the nose of the plane down if it is in danger of going into a stall.
“This is done through horizontal stabilizers on the plane’s tail which are activated by the aircraft’s flight control computer.
“According to the flight data recorder, the pilots of Lion Air Flight 610 struggled to control the aircraft as the automated MCAS system repeatedly pushed the plane’s nose down following takeoff.
“The pilots of the Ethiopian Airlines plane reported similar difficulty before the aircraft plunged into the ground shortly after takeoff.
“A preliminary report on the Lion Air Flight 610 accident blamed it in part on a faulty angle of attack sensor that triggered the MCAS system and automatically forced the plane’s nose down.
From the BBC:
Boeing grounds entire crash aircraft fleet
1 hour ago
https://www.bbc.com/news/business-47562727
I had read that after the Lion crash, pilots were all told (supposedly) how to handle such a situation. But it may be that some pilots never got the extra training, or it may be that just being told wasn’t enough to allow them to successfully overcome the problem when it happens in real life, and there is so little time to react to it.
How tragic and terrible.
Neo and Barry,
The DOT guy I heard suggested that the plane or its loading may need to be redesigned so as to move the center of mass forward, or possibly re-positioning the engines or engine thrust. Apparently, these big MAX engines really change the characteristics of the plane.
Originally, I thought that there was a problem with the some “anti-stall” software feature. But Barry is correct in that it is the MCAS that is connected to the problem. “Maneuvering Characteristics Augmentation”? What the heck is that? My guess is that it is a software system that is intended to make a plane with shitty flying characteristics more “flyable” or at least moderately safe.
Factoid: Nearly all modern fighter jets will crash if they lose all computer systems, which is why they have several. The airframe is fundamentally unstable, and computer stabilizes it and makes it flyable. The combo of an unstable airframe with the computer creates a plane with superior performance. But for obvious reasons, all commercial and private planes are completely stable and flyable without computers. Except maybe the 737 Max8?
An example of an unstable aircraft and crash is Air Midwest 5481. It had three different problems but two were,
from Wikipedia