Lion Air crash

[DoubleWasp]

I'm starting to see the picture here. So some engineer(s) basically go, "Hey, remember Colgan 3407? We need to make the stick pull hard enough that another pilot can't do what he did.".

Any clue why the pilot would have pulled the nose up and moderately increased power to get out of a stall? Did he perhaps think he was too low to nose down out of a stall? From the report, it seems like he was REALLY determined to get that nose up and keep it up.

I see what you mean about gauging the results of this specific type of system being a murky process. It seems like this whole situation is like Whack-A-Mole. You try to slap one mole down, and another just pops up and some other disaster occurs.

 

[Astro14]

Originally Posted by DoubleWasp
I'm starting to see the picture here. So some engineer(s) basically go, "Hey, remember Colgan 3407? We need to make the stick pull hard enough that another pilot can't do what he did.".

Any clue why the pilot would have pulled the nose up and moderately increased power to get out of a stall? Did he perhaps think he was too low to nose down out of a stall? From the report, it seems like he was REALLY determined to get that nose up and keep it up.

I see what you mean about gauging the results of this specific type of system being a murky process. It seems like this whole situation is like Whack-A-Mole. You try to slap one mole down, and another just pops up and some other disaster occurs.


It's a bit like whack-a-mole - as you take authority away from pilots and put it in systems, then you're going to see systems failures as the root cause of crashes more often.

I am of the opinion that the pendulum (balancing between pilot authority and system authority) has swung a bit too far. The F-16 and the A-320 (as well as 330 and 340) limit the pilot. They simply won't respond to inputs that the flight control system believes exceeds limits. That's great. IF the flight control system is functioning correctly and actually knows what the limits are. But if the system is hampered by bad sensor inputs then the degradation may not (Air France 447) be graceful, and there are numerous F-16 crashes in which flight control limits are implicated.

The F/A-18 and the 777 warn the pilot of the limits but allow the pilot to exceed them.

Way better philosophy.

The pilot retains ultimate control, and if the flight control system is receiving bad inputs, the pilot can over-ride and fly the airplane.

I think (I don't have access to the flight manual for the MAX) that the Max is the same way, and what we see in this Lion Air crash is the intersection of system fault, greater (but not complete) system authority, and poorly trained crews. Remove any one of those factors, e.g. a better trained crew, a working AOA sensor, less authority from MACS to push the nose down, and this crash would've been avoided.

By the way, I've completely conflated my Airbus crashes and so I described a crash and gave you a link to another crash. The link was to an A-330 at Toulouse. Also a system/pilot interface study, but NOT the one I was thinking of.

No, the one that I was thinking of was an A-320, and it turns out that crash was in Habsheim, at an airshow. And most of the people on board survived as the airplane mushed into the trees. The pilot believed that the flight control logic limits would prevent the stall of the airplane, and he was demonstrating those limits by flying very slow.

But, he was supposed to be at 100 feet, and instead, he flew by at 30 feet. Impressive. But it put the airplane into different flight control logic. And it allowed him to get so slow, and at 30 feet, with the airplane nearly stalled, he relied on the flight control logic* to get him out of a very precarious situation.

Didn't work out that way, the airplane stayed near stall, then engines began to spool up as it hit the trees, and the ingestion of trees and leaves killed the thrust, so the airplane mushed forward in a staill into the trees.

The whack-a-mole approach isn't working.

That's why I am a big fan of the approach we are taking at United - analysis of flight parameters day in and day out to identify, and correct, threats and errors before they result in a crash. We capture and analyze the "near-misses" and even the "no big deal" events in a pro-active framework, partnered among FAA, pilots and Company. FOQA (Flight Operations Quality Assurance) data from the on board recorder is down loaded every night and sent to ALPA volunteers to be analyzed. The FOQA data capture is extensive (and I should point out, the equipment is not designed to survive a crash, that's the job of the FDR) and the vast majority of data shows nothing of concern. But sometimes, a flight crew had a deviation that they were unaware of, and that deviation can be analyzed, trends tracked, and pro-active solutions developed and trained before we bend metal in a crash.

We've made air travel incredibly safe. One crash every 10 million flight hours, or about that. Lower in places, higher in others, but overall, particularly in the US, incredibly safe. But engineers, and our historic approach to flying and training, have reached a limit. We've been sitting on this incredibly low mishap rate for decades despite improved airplanes, and systems. The only way to get the crash rate down even further is through better understanding of the flight environment, minor pilot mistakes and deviations, so that we can respond and correct things before the crash.



* The logic element was a protection mode known as "Alpha Floor". When the airplane reaches a critical AOA, much higher (i.e. at a much slower airspeed) than normal AOA, the engine thrust goes to full power, regardless of throttle lever position, and the light controls stop the AOA increase. Stall prevention. Get near a stall, and you get full engine thrust, and AOA limiting no matter what you've done with throttles or stick. A great idea.

When it works. But it is disabled below 50' to prevent momentary deviations in AOA from causing a landing mishap. At 30 feet, this airplane wouldn't go into Alpha Floor. I've had an Alpha Floor event. Moderate to severe turbulence at 1,500' AGL going into Reno, NV and the AOA sensors registered critical AOA momentarily. "Alpha Floor" displayed in the FMA, followed quickly by "TOGA Lock" which meant that the engine thrust was locked at maximum. I pulled the throttles back to mid-range and disconnected the autothrust to regain control of the airplane. We landed in Reno. Doubt the passengers or the Flight Attendants even knew. But the airplane responded as designed to a situation that existed only momentarily, and I had to respond very quickly to get the airplane back under control, into appropriate logic, for the actual, not momentary, flight regime.

 

[Astro14]

Originally Posted by DoubleWasp
I'm starting to see the picture here. So some engineer(s) basically go, "Hey, remember Colgan 3407? We need to make the stick pull hard enough that another pilot can't do what he did.".

Any clue why the pilot would have pulled the nose up and moderately increased power to get out of a stall? Did he perhaps think he was too low to nose down out of a stall? From the report, it seems like he was REALLY determined to get that nose up and keep it up.

I see what you mean about gauging the results of this specific type of system being a murky process. It seems like this whole situation is like Whack-A-Mole. You try to slap one mole down, and another just pops up and some other disaster occurs.


There are two separate questions here, and I wanted to keep them separate.

I answered the pilot/system balance in the previous post. The questions on the performance of this captain (crew, really) persist.

Several threats were present: bad weather. Icing. A First Officer who worked part-time as a barista, and had left her parents' house in Washington state the night before, flown to New York on FedEx, and had gone without sleep for over 24 hours. A Captain who had repeated problems/failures in training, particularly with stalls. A Captain with very little time in the left seat.

Several mistakes were made: 1. the decision to fly in icing. 2. failure to fly at the proper speed. 3. failure to respond to a stall warning. 4. raising the flaps in a stall recovery.

The mishap report from this flight goes through all these in detail.

I'll summarize by saying that, like any mishap, if you break one link in the chain of events, if you remove one factor, the mishap likely could've been avoided. If the weather was nice, or the FO hadn't been tired, or a different Captain was assigned, or, if the Captain hadn't let the airspeed decay, then you wouldn't have this crash and 50 people would be here today.

There is just so much about this crash that the industry needed to address: fatigue, commuting, inexperienced FOs, flight into icing, and yeah, pilots that don't do well in training.

We've addressed many of those, industry wide, in the US (fatigue played a big factor with the Air Canada near miss at SFO last year), at least. We've upped the experience requirements for FOs, we've changed fatigue rules, we've (at least some of us) started FOQA collection.

But what do we do about pilots who perform poorly in training?

First, the FAA has changed the test standards for stalls in training as a direct response to this crash. They used to require that you lose no more than 100 feet in a stall recovery in the simulator. That wasn't appropriate for a stall, it was appropriate for an approach to a stall, for slow flight, but the test standard itself used to shape pilot response to a stall into "must stay on altitude" where every fighter pilot, every military pilot, who has experienced REAL stalls, and maneuvered airplanes well below stall speed, knows that the correct response to a stall is "must control AOA".

The FAA has adopted that. Finally.

I was fortunate to be trained by some of the very best instructors in the industry. They knew the FAA PTS (practical test standards) were focused on the wrong thing: altitude loss and so I wa trained, in the 747, in the 757/767 and the A320, to control AOA.

Glad to see the FAA catch up to reality.

So, the FAA owns a tiny bit of this crash, by establishing standards that focused on the wrong thing. This Colgan Captain kept the nose up to minimize altitude loss. Absolutely the wrong thing to do. The airplane wasn't flying in an approach to stall, it was deeply stalled, and only lowering the nose, and going to full power, was going to correct that stall. He responded in part to the way he was trained.

But he also failed to go to full power.

That's harder to explain because he's not here to talk about it. It's my best guess that he didn't recognize how serious the situation was. He had problems in training, true. And poor stall recoveries were part of that documented problem set.

No one can explain why the FO moved the flaps up. That's just a complete mystery. It doesn't help with stall recovery, and in fact, would induce a stall at low speed. Perhaps she thought the control problem was asymmetric flaps (it wasn't) or that the flap drag was causing the delay in recovery (it wasn't). She's not here, either, to tell us what she was thinking.

But as we look at improper response to a situation in flight, I can't help but circle back to this: he had problems in training.

That's the crux of the issue for me. Problems in training. I've heard plenty of pilots say, "I can fly the airplane fine, I just can't fly the simulator" and ladies and gentlemen, I'm here to tell you that is complete hogwash. The guys who fly the airplane well fly the sim well. The guys who claim the above, are often unable to see their defects when those defects are pointed out to them. Dunning-Kruger in the cockpit.

I will tell you from personal experience that one of the hardest things I've ever had to do was tell a young man* that he did not meet standards. F-14 night carrier landings, 747 engine out profile, 757 landings, it doesn't matter. You're telling somebody that their performance in their life avocation isn't measuring up.

Depending on the training event, it might mean more training, or...

It might be the end of a career.

A few careers have ended with me. Not because I wanted them to, it was always in spite of my very best effort to help these guys.

But for their own futures, for the safety of the people in their airplanes and on the ground, the decision had to be made.

No matter the circumstances, it's a very difficult thing to do.

I believe that it should have been done in this Captain's case. I am aware of training programs that have difficulty in making the hard call. That's an area that will always need emphasis and real work. Not everyone is cut out to be a pilot. In the training programs with which I've been associated (USN, and UAL), we've always had integrity in maintaining those standards. It's hard, though, really hard, for everyone involved.

As we face an enormous pilot shortage, and tens of thousands of pilots will be hired and trained in the next decade in the US alone, that's a real question: do the training programs have enough rigor to wash out the pilots that should not be there?



*It's always been a man in my case, I've had lots of great female students, none of them performed poorly.

 

[DoubleWasp]

So much insight. Sorry to make you keep banging away at the keys, but it really helps to have a real world professional translation of what I was reading. Reading a lot of these reports tends to bring up a lot of questions. Reading the answers tends to bring up even more, but you've explained it perfectly.

One more question. I promise.

What's the exact story with AF 447? I read the report and the flight recorder transcript, and I kind of understand what happened here. But what's the possible rationale for the pilot literally just riding the plane with the stick pulled back all the way down into the ocean? I see it says that at one point, the plane slowed to nearly 60 MPH. Was it just sheer panic? Was the instrument failure really just so severe that he couldn't tell that he was stalling so bad he was darn near coming to a stop and plummeting? I mean, I'm playing Monday morning QB here of course, but wouldn't his SOG been suuuuuuuuper slow at that point, and altimeter giving away a full on plummet down the chute? It seems like when the third officer came into the cockpit, he immediately knew that the plane was completely doomed. So what was going on with the crew on duty (Other than unknowingly fighting a secret control war with eachother?) ?

The picture I'm getting here is that most plane crashes are basically Murphy's Law situations where it all goes wrong at the the same time. As you said, with one fault removed, no crash.

 

[Astro14]

Originally Posted by DoubleWasp
So much insight. Sorry to make you keep banging away at the keys, but it really helps to have a real world professional translation of what I was reading. Reading a lot of these reports tends to bring up a lot of questions. Reading the answers tends to bring up even more, but you've explained it perfectly.

One more question. I promise.

What's the exact story with AF 447? I read the report and the flight recorder transcript, and I kind of understand what happened here. But what's the possible rationale for the pilot literally just riding the plane with the stick pulled back all the way down into the ocean? I see it says that at one point, the plane slowed to nearly 60 MPH. Was it just sheer panic? Was the instrument failure really just so severe that he couldn't tell that he was stalling so bad he was darn near coming to a stop and plummeting? I mean, I'm playing Monday morning QB here of course, but wouldn't his SOG been suuuuuuuuper slow at that point, and altimeter giving away a full on plummet down the chute? It seems like when the third officer came into the cockpit, he immediately knew that the plane was completely doomed. So what was going on with the crew on duty (Other than unknowingly fighting a secret control war with eachother?) ?

The picture I'm getting here is that most plane crashes are basically Murphy's Law situations where it all goes wrong at the the same time. As you said, with one fault removed, no crash.


Ah, AF 447...fascinating story.

Some good discussion several years ago in this thread: https://www.bobistheoilguy.com/forums/ubbthreads.php/topics/2249217/1

I covered it again, a bit, in earlier posts in this thread.

Simply put: the system failed. It gave the pilots bad data, and then complete control of the airplane. They were startled, and didn't respond logically, or calmly as the airplane filled the cockpit with alarms and warning lights. They responded as they were trained.

And it didn't work.

But when the airplane started giving them good data, the results were so incredible (60 knots? That CANNOT BE right!) that the crew didn't believe their instruments.

It's fascinating because it's hard to understand if you're not a pilot, and you're here, looking at your keyboard, under 1G of acceleration, and at 0 knots, and no one's life is depending on your responses or on your decisions.

While they were flying on a dark night, through thunderstorms, when everything stopped making sense. They applied normal procedures:

1. Overspeed exists, the instruments say so, so, pull power and raise the nose, right?
2. Airplane is descending, so add power and raise the nose.
3. Airplane is stalled, stall recovery is full backstick and TOGA power.

And that airplane hit the water with full nose up (backstick) elevator trim, in a full/deep stall, at full engine power.

 

[DoubleWasp]

You're 100% right. Meant no disrespect. Just trying to grasp my brain around it. Gonna give it a good read.

 

[jstert]

im not a pilot, though my older son is one in the usaf. i worked in jakarta for 3 1/2 years and studiously only flew garuda. lion air is a wreck waiting to happen. that said, when i was there boeing had just won its hugely lucrative sales contracts. despite indonesian air carriers being boeing's biggest foreign market at one point and despite pleas from the u.s. government that aided boeing, boeing declined to put any full time staff in jakarta. i heard first hand how much boeing executives enjoyed their comfortable expatriate lifestyle in singapore, hated to visit jakarta and simply "dialed in" whatever they did for their indonesian customers.

 

[FowVay]

I see in the news today that Navy divers found the black box from this crash. Hopefully we learn a bit more about the final moments of the flight.

 

[Astro14]

Originally Posted by FowVay
I see in the news today that Navy divers found the black box from this crash. Hopefully we learn a bit more about the final moments of the flight.


I hope so, too.

The thought process of the pilots, and what they observed the airplane trying to do, should be captured in their conversation.

 

[DoubleWasp]

How long does it usually take for the release of the voice recordings?

 

[Astro14]

They are often never released to the public. IF they are, it can take years.

It really depends on who is doing the investigation.

No idea what the protocol is with the Indonesian investigators.

 

[Yah-Tah-Hey]

No matter how hard the manufacturers try to incorporate safety features into their products, someone in that part of the world will find a way to defeat those features and wind up at the bottom of an ocean.

 

[Cujet]

Originally Posted by Yah-Tah-Hey
No matter how hard the manufacturers try to incorporate safety features into their products, someone in that part of the world will find a way to defeat those features and wind up at the bottom of an ocean.


I don't see it that way at all.

What we regularly see in the high end corporate jet world, is the fact that systems engineers can't possibly foresee every possible situation, and therefore, the system fails. Often not due to defeating the system, but encountering an unplanned situation.

We had 2 FADEC engines roll back and throttles lock up at near idle (fuel metering sleeve locked in position) in cruise flight. It took the engineers years to determine why. The in-flight fix was to shutdown and restart one engine at a time.

The reason was that the FADEC system was unable to deal with outside air temperatures above freezing at 49,000 feet. Note: There is never a time where OAT is above freezing at FL490. However, when TAT (total air temp) probes ice up due to ejecta from a storm below, then the heaters melt the ice, the melted ice=water and is well ABOVE 32F as it floods the temp sensor. A parameter so far outside of normal was never envisioned. We were the lucky ones to find this out, while over the Pacific ocean, hours away from land.

As far as I know, the "fix" is still to cycle the engine switches off and back on. The FAA has not approved any new FADEC software/firmware.

 

[john_pifer]

Originally Posted by Cujet
Originally Posted by Yah-Tah-Hey
No matter how hard the manufacturers try to incorporate safety features into their products, someone in that part of the world will find a way to defeat those features and wind up at the bottom of an ocean.


I don't see it that way at all.

What we regularly see in the high end corporate jet world, is the fact that systems engineers can't possibly foresee every possible situation, and therefore, the system fails. Often not due to defeating the system, but encountering an unplanned situation.

We had 2 FADEC engines roll back and throttles lock up at near idle (fuel metering sleeve locked in position) in cruise flight. It took the engineers years to determine why. The in-flight fix was to shutdown and restart one engine at a time.

The reason was that the FADEC system was unable to deal with outside air temperatures above freezing at 49,000 feet. Note: There is never a time where OAT is above freezing at FL490. However, when TAT (total air temp) probes ice up due to ejecta from a storm below, then the heaters melt the ice, the melted ice=water and is well ABOVE 32F as it floods the temp sensor. A parameter so far outside of normal was never envisioned. We were the lucky ones to find this out, while over the Pacific ocean, hours away from land.

As far as I know, the "fix" is still to cycle the engine switches off and back on. The FAA has not approved any new FADEC software/firmware.




Geez, man. That's a pretty scary situation.

You have your shark repellent with you?

 

[DoubleWasp]

I hate to be macabre, but I'm going to guess that situation would not end with survivors.

 

[Kamele0N]

https://samchui.com/2019/02/16/lion-air-boeing-737-skids-off-runway/#.XGgA-lwzaUk

 

[Astro14]

Gee...wrecked another one?

Who would've thought...oh, wait...