Boeing Boeing 737 Cargo Plane Ditches off Honolulu

[russ_watters]

Jet engines will burn pretty much anything that can burn. I suspect that "fuel contamination" really means "fuel filter clogged by contaminants".

Water is a common contaminant.

 

[anorlunda]

They said the second engine was overheating. Does anyone know what makes a jet engine overheat?

 

[Vanadium 50]

Oil starvation.

 

[gmax137]

Do the engines share a common fuel supply? or does each have its own tank(s)?

 

[anorlunda]

Do the engines share a common fuel supply? or does each have its own tank(s)?

Typically, each engine has its own tank. There is a valve that can be opened to transfer from one tank to the other, but that valve is normally closed.

But both tanks may have been filled from a common source.

In searching for a common cause, I would skip all possibilities that do not lead to engine overheating. But the term overheating is not precise. EGT, exhaust gas temperature is my first guess, but there can be other temperature sensors on the engines.

Below is some discussion from another site. It suggests that internal damage can cause high EGT. That would make high EGT a secondary effect, not the primary one.

https://www.airliners.net/forum/viewtopic.php?t=762413
high EGT is due to high fuel flow. As an engine gets older the fuel flow and EGT will increase as the compressors and turbines get worn and have to work harder to do the job. I have seen over 50degC different EGT between two engines on the same aircraft, due to age. So you need to boroscope the engine to check the complete gas path for damage, not only blades and NGVs, but also the inside of the combuster for damage.. Take out the burners and check they are clean, they easily get clogged with carbon.

 

[cjl]

I’m about a month from my private pilot checkride. Some expansion/clarification: ABC is always overriding, but specific circumstances carry additional requirements, including checklists.

In a small/single engine plane engine failure, you have:

A - Airpeed for best glide ratio
B - [identify]Best place to land
C - Checkists; memory items first, more detailed ones as the situation permits.
D - Declare an emergency
E - Execute the landing

Obviously the specifics of the situation dictate the specific approach. I agree that at least it would be good to stay close to the airport.

https://studentpilotnews.com/2019/04/22/if-faced-with-an-engine-failure-remember-your-abcs/

Dual engine failures often have a common cause. It will be interesting to see what this one was.

I think "often" is underselling it a bit. Has there ever, in the history of commercial jet aviation, been a dual independent failure? I certainly can't think of one. There are certainly cases where all engines failed on multiengine jets, but every case I can think of is common cause (flying into an ash cloud, flying into a goose cloud, flaws in the design of fuel filters combined with extremely cold fuel causing ice buildup and fuel starvation, running out of fuel due to improper unit conversion, etc). In addition, single engine failure isn't all that uncommon, so it's extremely understandable that the pilots would've been perfectly comfortable exiting gliding range (but still staying relatively near to the airport) while troubleshooting, executing checklists, evaluating the aircraft behavior, etc.

At least on the surface level, this doesn't appear to be something that would require a major change in procedure or aircraft design philosophy to fix, aside from perhaps just advising pilots that it is best practice to remain within gliding distance (if possible) while performing evaluations and troubleshooting after an engine failure.

 

[cjl]

Do the engines share a common fuel supply? or does each have its own tank(s)?

Typically, tanks are in the wings, with each engine drawing from the tank on its own wing, but the 737 does have a center tank as well. In addition, there are pumps that can move fuel around between tanks, so it's at best only somewhat isolated. This is because a side to side weight imbalance would be a very serious problem, and the engines do not burn precisely identical amounts of fuel, so it's very common to move a little back and forth to keep the plane balanced during flight. Adding to this, once you have a single engine failure, you would only be burning fuel from one side, so if you did need to proceed for a substantial amount of time on one engine, you would definitely be transferring fuel from the shutdown side to the operating side to keep it in balance.

The center tank would only be used for long flights, as structurally it's better to carry fuel in the wings, and I do not know if it was used on this flight. That having been said, whether the center tank was used or not, it's fairly safe to consider both engines as effectively running off the same fuel, and it certainly would've been sourced from the same place and no care would've been taken to prevent mixing.

 

[cjl]

They said the second engine was overheating. Does anyone know what makes a jet engine overheat?

In addition to the oil starvation mentioned above (which I would honestly consider fairly unlikely), there are a number of other possible causes. A dirty or damaged compressor or turbine could cause this, as it would decrease efficiency and require more fuel to generate the same RPM and thrust. A bleed air leak would also lead to high EGTs, as it would reduce airflow to the combustor without reducing the power demand on the turbine. There could also be a problem with the fuel control unit, but on an aircraft this old, I'd probably suspect FOD, contamination, or wear on the compressor or a bleed leak before suspecting anything about the fuel control, since it's pretty simple on an engine like this (on a more modern engine, that could be a cause though).

EDIT: In addition, the fact that it only started overheating after the first engine failed supports the thought that it's related to compressor damage or wear (or turbine, but compressor seems more likely to me). Even with a worn, contaminated, or damaged compressor, the engine likely could perform OK at lower thrust settings or even for the brief high thrust at takeoff, but that would severely harm its ability to run continuously at higher thrust settings like would be required after the other engine failed. Granted, severe enough damage, contamination, or wear such that it couldn't even maintain level flight on a single engine is pretty extreme, but several accounts I'm reading about this airline and aircraft indicate that it was very near end of life, and the airline was suspected of poor maintenance practices, so that's my best guess right now.

 

[DaveE]

I think "often" is underselling it a bit. Has there ever, in the history of commercial jet aviation, been a dual independent failure? I certainly can't think of one. There are certainly cases where all engines failed on multiengine jets, but every case I can think of is common cause (flying into an ash cloud, flying into a goose cloud, flaws in the design of fuel filters combined with extremely cold fuel causing ice buildup and fuel starvation, running out of fuel due to improper unit conversion, etc). In addition, single engine failure isn't all that uncommon, so it's extremely understandable that the pilots would've been perfectly comfortable exiting gliding range (but still staying relatively near to the airport) while troubleshooting, executing checklists, evaluating the aircraft behavior, etc.

At least on the surface level, this doesn't appear to be something that would require a major change in procedure or aircraft design philosophy to fix, aside from perhaps just advising pilots that it is best practice to remain within gliding distance (if possible) while performing evaluations and troubleshooting after an engine failure.

There is the possibility that there was initially only a single engine problem, and pilot error in shutting down the good engine caused the subsequent failure. Shutting down and restarting a jet engine at low altitude isn't trivial, nor is it good for the engine. Yes, they ought to know not to make that mistake, but neither would they have been the first crew to do that.

 

[mfb]

This source says one engine failure every 375,000 flight hours. If an aircraft flies for an average of one hour after a single engine failure (?) and we neglect four-engine aircraft we should expect completely unrelated double-engine failures every 375,000 single engine failures. With ~15,000 aircraft in the air and assuming they all have two engines we expect such an incident every 500 years. Things are not completely unrelated of course - things like the age and airline lead to a correlated risk.

 

[russ_watters]

I think "often" is underselling it a bit.

Agreed.

In addition, single engine failure isn't all that uncommon, so it's extremely understandable that the pilots would've been perfectly comfortable exiting gliding range (but still staying relatively near to the airport) while troubleshooting, executing checklists, evaluating the aircraft behavior, etc.

At least on the surface level, this doesn't appear to be something that would require a major change in procedure or aircraft design philosophy to fix, aside from perhaps just advising pilots that it is best practice to remain within gliding distance (if possible) while performing evaluations and troubleshooting after an engine failure.

Isn't a single engine failure an instant/automatic emergency? I'd really like to know what the procedure/checklist says. I don't see what benefit you get from flying away from the airport vs orbiting off the end of the runway.

I'd also like to know if this was a true case of a rare dual engine failure with unrelated causes or if the pilots initially missed that the second engine was in the process of failing while they troubleshot the first engine. At least in some cases it is clear-cut whether you have an isolated single engine or dual engine failure.

If an aircraft flies for an average of one hour after a single engine failure (?) and we neglect four-engine aircraft we should expect completely unrelated double-engine failures every 375,000 single engine failures. With ~15,000 aircraft in the air and assuming they all have two engines we expect such an incident every 500 years.

Note that unlike the odds at a roulette table or of an individual radioactive atom decaying, those odds increase with time/age.

Things are not completely unrelated of course - things like the age and airline lead to a correlated risk.

Yes, what I was thinking was that after the first engine fails you are then forced to run the other engine at higher power/for longer, putting additional strain on it, further increasing the risk.

 

[anorlunda]

I don't see what benefit you get from flying away from the airport vs orbiting off the end of the runway.

The perceived benefit could be to put the plane on autopilot to allow both pilots to focus on the first engine restart. That would be a violation of cockpit resource management protocol.

It seems to echo the case of Eastern flight 401. The aftermath of that crash spawned the idea of cockpit resource management.

https://en.wikipedia.org/wiki/Eastern_Air_Lines_Flight_401#Cause

The final NTSB report cited the cause of the crash as pilot error, specifically: "the failure of the flight crew to monitor the flight instruments during the final four minutes of flight, and to detect an unexpected descent soon enough to prevent impact with the ground. Preoccupation with a malfunction of the nose landing gear position indicating system distracted the crew's attention from the instruments and allowed the descent to go unnoticed.

 

[Vanadium 50]

It is true that with one engine out, the odds of a second engine failure go way up - a factor of 1000 perhaps. However, it is also true that the vast majority of time (99.5%) you have a single engine failure, you have only a single engine failure. I'd be wary of changing the procedure to improve the outcome of the 1/200 case if it makes the 199/200 case potentially worse.

 

[berkeman]

And an interesting new development...

Air cargo company that ditched plane in ocean off Hawaii is grounded​

The Federal Aviation Administration said Friday that it will bar Rhoades Aviation of Honolulu from flying or doing maintenance inspections until it meets FAA regulations.

The agency did not detail Roades' alleged shortcomings. The company did not immediately respond to phone and email messages for comment.

The decision to ground the carrier, which operates as Transair, is separate from the investigation into the July 2 ditching of a Boeing 737, the FAA said. Two pilots were rescued by the Coast Guard after the nighttime crash.

The company had one plane still in operation this week, a Boeing 737-200 like the one that crashed.

The FAA said it began investigating Rhoades Aviation's maintenance and safety practices last fall and told the company about two weeks before the crash that it planned to revoke its authority to do maintenance inspections. The company did not appeal the FAA's decision within the 30 days as required if it wanted the case reconsidered, the FAA said.

 

[anorlunda]

To me, the surprising thing comes at 1:50 of the VASA radio transcript. Prior to that, flight 810 said that they had to return to the airport. The controller (after some confusion) cleared them to immediately return and to land. But 810 said, "Rhodes 810, we're going to have to run some checklists if we could get delay vectors."

That suggests that they changed their minds about an ASAP return to the airport and instead do something else. While doing the something else, they flew out to sea, away from the airport. I suspect the something else was the engine restart procedure, but the radio transcript does not say that explicitly.

I have listened to many such transcripts from VASA Aviation, and to every recreation produced on Air Crash Investigations. The "usual" decision in such cases is to return to the airport ASAP.

There is no need to restart the failed engine in order to make a safe landing. IMO, their choice should have been to begin the landing checklist, not to request a delay to do something else.

 

[russ_watters]

The only upside I can think of to their choice is if they were clawing for altitude then keeping the wings level helps. But that's a lose-lose proposition unless you can dump fuel to lighten the plane. Running checklists itself does not require flying away from the airport.

I did watch one of those VASA videos where a pilot orbited in an area abeam the runway, designated for fuel dumping.

I doubt they could have used autopilot on a compromised plane, but if so, autopilot can turn a plane.

 

[Vanadium 50]

For fun, I downloaded the 737 "Quick Reference Handbook, which at 822 pages stretches the definition of "handbook". And probably quick.

The single engine out checklist begins on page 7.16 (148) and you need to get to Step 13 to reach "Plan to land at the nearest airport". Step 12 allows for an attempt at an engine restart. A successful restart (11 steps) does not require an immediate landing, although laws, regulations, and common sense might.

The one engine inoperative landing checklist is 4 pages long, not counting the direction to check a table on page 378.

Page 7.6 is the six page gem "Loss of Thrust on Both Engines". Interestingly, the checklist does not give an action for what to do if neither engine restarts.

How much time were they running checklists before turning to HNL? As far as I can tell from posted audio, a few seconds under two minutes,

 

[anorlunda]

A big part of the pilot's responsibility is to decide which strategy and therefore, which checklist to use. Consider two cases.

1. An engine failure at 30000 feet, when it will take 20 minutes to descend with or without engines.
2. An engine failure during takeoff roll, or during final approach to landing, 10 seconds from touchdown.

Obviously, the engine restart checklist should be use in case 1, but not in case 2. But there will always be borderline cases where things are not obvious. That's when we most need a thoughtful and well trained pilot.

How much time were they running checklists before turning to HNL? As far as I can tell from posted audio, a few seconds under two minutes,

Right; it was 87 seconds (see #19). But they started at a place close to the runway to which they could glide with no engines, to (87 seconds later) a place miles out to sea, and out of visual contact with the airport. IMO, that was the primary error that caused the crash. If they had circled close to the airport, they could have survived the 2nd engine failure. If they had landed immediately, they would have been on the ground before the 2nd engine failed.

 

[gmax137]

Multiple engine failures has been discussed before, see

https://www.physicsforums.com/threa...n-free-fall-from-12-496km.991918/post-6374219

That post attached a report that says:

A DOUBLE ENGINE FLAMEOUT DUE TO WATER INGESTION WHICH OCCURRED AS A
RESULT OF AN INFLIGHT ENCOUNTER WITH AN AREA OF VERY HEAVY RAIN AND
HAIL. A CONTRIBUTING CAUSE OF THE INCIDENT WAS THE INADEQUATE DESIGN OF
THE ENGINES AND THE FAA WATER INGESTION CERTIFICATION STANDARDS WHICH
DID NOT REFLECT THE WATERFALL RATES THAT CAN BE EXPECTED IN MODERATE
OR HIGHER INTENSITY THUNDERSTORMS.

"Water ingestion" -- That's kind of surprising.

EDIT - I'm not suggesting that has any relevance to the incident in this thread.

 

[Vanadium 50]

I'm an outcomes guy, not a process guy. Clearly in this case it would have been better to have turned around. That doesn't mean that in general turning around is a better use of time than a restart checklist. We don't expect our pilots to be robots, but we don't expect them to be precognitive either: at the time of that message, they had only lost one engine.

Had the outcome been different - say a runway accident with 809 - because they turned back immediately after losing one engine, would we not criticize them for turning back too soon rather than attempting a restart? Would we not be saying "in the vast majority of cases, restart attempts are successful"?

 

[anorlunda]

That doesn't mean that in general turning around is a better use of time than a restart checklist.

No, you're missing the point of Cockpit (Crew) Resource Management (CRM). The idea is to split the responsibilities. One pilot looks after aviating, and the other can focus on checklists. It is not either or, it is "do both."

In the Eastern 401 crash linked above. All three (maybe 4) people in the cockpit focused on a malfunction, leaving nobody to fly the plane. Out of that accident came the doctrine of CRM.

Flying straight and level out to sea, allowed autopilot while both human pilots focused on something else. Flying circles near the airport would have forced them to split up, with one flying and the other doing something else.

 

[russ_watters]

Clearly in this case it would have been better to have turned around. That doesn't mean that in general turning around is a better use of time than a restart checklist. We don't expect our pilots to be robots, but we don't expect them to be precognitive either: at the time of that message, they had only lost one engine.

Had the outcome been different - say a runway accident with 809 - because they turned back immediately after losing one engine, would we not criticize them for turning back too soon rather than attempting a restart?

No criticism yet, just investigation. We have identified a potential better outcome if a different process (better process) were used (and more than one is on the table*). We don't yet have enough information to determine if either were foreseeable results.

0. Fly straight away from the airport while completing checklists, until...? (chosen option)
1. Orbit near the airport while completing checklists.
2. Return to land immediately without completing any more checklists than can be completed during an immediate return.

*I'm leaning toward Option 1. There's no guarantee that this would have helped. And like you said, Options 1 & 2 also carry risks (landing heavy, for example).

Also, ABCDE is a procedure guideline, not a checklist. As @anorlunda said, running a checklist and deciding whether you should are two different things.

 

[Vanadium 50]

No, you're missing the point of Cockpit (Crew) Resource Management (CRM). The idea is to split the responsibilities. One pilot looks after aviating, and the other can focus on checklists. It is not either or, it is "do both."

If you are saying "Do the checklist, but don't get too far from the airport", I sort of agree. I say "sort of" because of the geography of the area. Have you ever been there? North of the airport, you have highlands in about two miles. East of the airport you have downtown with several buildings over 400 feet. So I see where you might want to avoid that.

I don't know what a 737's turning circle is. Can it do a 360 in two minutes? That would mean it's a minute to turn around, or at 180 knots, 3 nautical miles. I wonder if they were thinking to line the plane up for 8R. (I had to look it up - I'm not that familiar with the airport) I can definitely understand them not wanting to be in the right spot but pointing in the wrong direction.

 

[russ_watters]

I don't know what a 737's turning circle is. Can it do a 360 in two minutes?

A "standard turn" for an airplane under 250 kts (which this one would have been) is 360 degrees in two minutes.
https://en.wikipedia.org/wiki/Standard_rate_turn

[Edit: fixed]

 

[Vanadium 50]

360 in one minute or two?

 

[mfb]

250 kts * 2 minutes / pi = 5 km diameter using the "standard turn" from Wikipedia.

I drew one 5 km and two 10 km diameter circles on the map (from Bloomberg), there should have been more than enough space to fly in circles. The turn towards the airport matches a 10 km diameter pretty well.

 

[russ_watters]

360 in one minute or two?

Oops...should have stopped posting right before bed. Fixed.

 

[anorlunda]

Let's wait for the release of the cockpit voice recorder (CVR) and the black box. That should conclusively tell us if both pilots were focused on the failed engine, or whether one pilot remained focused on flying the airplane.

The CVR may also reveal why at 10 seconds after takeoff clearance, they radioed an intention to return to airport, but 100 seconds later they asked for a delay in returning. That decision can then be compared with their training.

 

[Vanadium 50]

I think @mfb posted a pretty good map. Both the takeoff and the turnaround circles are about 6 miles in diameter. It also looks like they are about 6 miles away from what would be in hindsight the best position, which is not too far off from the famous 87 seconds.

I listened to the ATC again. Much oddness.

• While 810 had a declared emergency, ATC cleared T4 809 to land, and KH 56 to take off. I would have imagined they would have stopped all non-essential operations.
• Boy that poor air traffic controller had a lot to do!
• The plane should have been able to fly on one engine. They start descending right after the first turn, but it's not until after the second turn they say that the second engine is running hot. It looks like it was misbehaving well before then.

 

[DaveE]

While 810 had a declared emergency, ATC cleared T4 809 to land, and KH 56 to take off. I would have imagined they would have stopped all non-essential operations.

It is normal for ATC to continue some operations when handling an emergency aircraft. But, they have to have the bandwidth to do that and make sure not to interfere with the emergency operations. There's usually not a problem landing a plane or two while the other guys are running checklist or in the traffic pattern. They are experts at timing and sequencing operations, they do it all day long every day.

I heard some airline pilots (airlinepilotguy.com) commenting that the busy radio channel was partly the result of the emergency aircraft not clearly declaring an emergency with a "Pan Pan" or "Mayday" call. They said that while pilots are sort of monitoring other traffic comms, they are only really listening for specific stuff that effects them. But, they are hyper sensitive to hearing "Mayday" and they know to shut up and only transmit essential queries if they hear that. Aside from following ICAO standard comms, they said that it is important to clearly let everyone know that we aren't doing business as usual; get everyone's attention first, then you can explain later.

I also recall an ATC guy at JFK say that "Emergency", "Pan Pan", and "Mayday" all mean the same thing to them, they are processed the same way.