This is exactly the technique I had some objections with.
If you want to get anywhere, don't fly close to stall. Gliding faster (at best L/D) will get you further and leave you more Options where to put it down.

There is absolutely no benefit I could conceive in gliding down close to stall speed when having lost all engine power.
Ask the sail plane guys. They are flying with engine failure all the time...
And they would never, ever try to slow close to stall speed when trying to glide as far as possible.

It's quite possible that they didn't want those options. It seems that there was only one optimal touchdown point; beyond the field lay more buildings. What's the use in gliding at best L/D to end up overflying the landing field and smacking into a building? They had what appears to be the only possible field in front of them (and a short field at that), with power lines in the way. Over, or under? If over, how to lose the excess height & energy?

It seems to me that looking through their windshield the crew might have thought it quite possible that their flight path would have landed them where they wanted to be, after passing under the power lines, and perhaps with a final pull on the stick (having read Think Like A Bird I assume an ATR's flaps don't travel as quickly as a Beaver's).

Then #2 decided to intervene.

Then they were really bad, bad flyers. Perhaps as bad as some of the armchair pilots on this forum.
River was everywhere along their path and that's a perfect landing opportunity. And the bridges were sparse enough. Had they executed a controlled landing on the river everyone could have walked away.

You got this one right.

Actually, it appears that they were both feathered after about 02:53:31 ...

I await a correlation of the CVR with the FDR. Hopefully someone will post a link when it appears in the wild.

to shut down or not to
 

Clearly the #2 engine was what started this whole fendengo. That was the culprit engin...whether by analogous indications, unwanted autofeather, or just plane flameout.So that engine should have been shut down.reducing ITT is the easiest indication of an engine flameout.

If in their muddle they shut down engine one and left the shutting down of eng 2 incomplete..that was what gave them all that unwanted drag and lost speed control in the process whilst panicking where to terminate the flight.The river and its easuaries gave them plenty of ditching options and that rever bed being shallow would have been advantageous in terms of evacuation.

Perhaps having shut the wrong engine(had they realised it) better to have moved CL2 to FSO.Then dry motor eng #1. Before starting it to purge any fuel and insure against a hung start.With engine one started and eng 2 shut down..they woukld have been able to climb away to their single engine ceiling whilst sorting the problem with 2 or simply return and land in controllable condition.

I also wonder if PL2 was retarded out of the TO notch..because that hard roll over would have been caused by sudden excess power from eng2 if it it unfeathered itself and the PL was in the notch...especially if climb sequence wasn't yet performed and hence TO torque kicked in with the PMS selector on TO.
It was the drag mess they got themselves in that betrayed helpless state..

A few questions if you folks dont mind...
1. Are the torque spikes recorded due to the feathered props still slowly spinning in the air?
2. Am I seeing the correct indication of a gradual yaw to the right from 1100 feet to 1350 feet, and does that prove #2 feathered as normal flight path is 114-116 degrees through 2250 feet?
3. Could this be masked by an automatic rudder input after the feather of #2?
4. Does anybody else think they were headed for Sun Yat-sen Freeway till they got near it and realized it was unsuitable for landing as it is boardered on either side by elevated freeway and its too narrow to accommodate the wingspan?
(I believe that may have been where the real panic would set in)


Thanks in advance. I am also curious if there is more on the CVR than what has been released so far.

xmh53wrench
 

To your 3rd point.

Could this be masked by an automatic rudder input after the feather of #2?
The yaw damper(auto rudder as you put it) might have been at it.It is normal practce to have it on even with the AP disengaged till short finals to countermand strong crosswinds.In this case I don't know if it remained on if the AP was on at that stage or if disconnected by any assymetric yaw due to Eng 1 shut down.


We are taught early not to attempt forced landing on a busy freeway..especially with a wide wing span and electric poles and other obstacles.The rivers away from bridges offered a more sensible landing site option.I think they were planning that r they wouldn't have been that close to it.

Not entirely true, in fact even worse...
If operating experience shows 1% too high fuel burn or .5% too low dispatch reliability, something is changed.
If operating feedback from pilots identifies weak points of the aircraft or built in traps, maybe the pilot training is improved, but nobody would improve the aircraft until some serious incident results in according official safety recommendations.
Same applies for feedback from maintenance.
Aviation has always relied on in-service feedback, but this element is quickly fading away as nobody listens to pilots or mechanics any more, and management only worries about economic aspects.

Doesn't anyone find this strange...

The authorities were really quick to release some DFDR information (day after the crash) which at the moment is incomplete and lead us to think the pilots shut down the wrong engine, yet no English version/transcript of the CVR is available.

Before making a judgement, I would like to see more and explained DFDR information and certainly an English version from the CVR of what was said in the cockpit.

i smell a rat!
 

Yes we need an explained CVR full transcript corelating to the DFDR

It is also a wonder that even the world media have shut up on this news.What has ATR themselves got to account for in all this?

From what we have been told, that they shut down the wrong engine, got to the edge of stalling five times then spun it in, it doesn’t look like there was much spare capacity left to navigate around obstacles and/or pick somewhere for a forced landing. At the speeds they were flying at, there will have been very little manoeuvre margin, so any determined attempt at turning would have ended in a rapid departure, like the one in the final few seconds. The crash site was likely selected by the aircraft rather than the pilots, although the CVR should provide some enlightenment.

The crew makeup is another classic: two captains and a F/O. Loads of potential CRM, authority gradient and simple unfamiliarity issues. In my (Western) airline, we don’t allow this kind of crew composition unless the capt. in the RHS is a trainer and checked/current for the RHS. It’s also only rostered for training duties.

A few questions to those who are qualified on the 72-600:

I understand it has auto feather in TO mode, as well as uptrim for power and rudder auto-trim for any asymmetry. Is the auto-trim always engaged? Also, would you consider use of the autopilot in a non-normal situation like this or is that non-SOP?

If operating experience shows 1% too high fuel burn or .5% too low dispatch reliability, something is changed.
If operating feedback from pilots identifies weak points of the aircraft or built in traps, maybe the pilot training is improved, but nobody would improve the aircraft until some serious incident results in according official safety recommendations.
Same applies for feedback from maintenance.
Aviation has always relied on in-service feedback, but this element is quickly fading away as nobody listens to pilots or mechanics any more, and management only worries about economic aspects.

True words nowadays - I 100% agree! Thank you, Volume!

...I doubt there is much difference in distance with this configuration
between minsink and max glide speed.
it was all about getting some power back into the equation.

Wing drops on high aspect ratio aircraft can be as much about
drag as loss of lift. that is.. the wing is dragged back slowing it
even more.

Missing the point?
 

This weeks Flight has an article on TransAsia (page 7) which points out that since 1995 this airline has suffered 7 significant safety incidents including the loss of 3 other ATRs and 1 Airbus A321.

The most severe prior incident was the crash of an ATR 72-500 in Magong on 23 July 2014 which resulted in 48 deaths.

By all means discuss the technical aspects of this incident but the real question for me is why was the Airline permitted to keep flying when it has such a poor safety record?

Kaboy:

Asia: 4.4 billion people in 49 countries. Generalisation much? I also work in Asia, and it's my opinion that anyone who talks about "Asian culture" talks about human culture but hasn't worked out that not looking caucasian does not automatically equate to incompetence, lack of proficiency and foresight. Tell me how it is that Asia is the fastest growing economic region in the world. Is it through culturally embedded incompetence?

Maybe Air Asia has a safety problem. I don't specifically know. I certainly would want more than raw incident data to draw a conclusion. The United States has the largest prison population in the world, and the second-highest per-capita incarceration rate. Going by your mode of assessment, Americans are a bunch of criminals.

I post this because I think all the predominantly western criticism of Asian aviation safety on this forum, often based on a bunch of premature conclusions, needs to be rebutted.

I live in the Philippines, here many of the regional airports are little more than landing strips, taking my local airport for example, something like a 1,400m runway, no navigational aids, no runway or airfield lighting, no fuel, lots of high ground around, and the local airline, weather permitting, put ATR72's in and out up to 7 days a week and, fingers crossed, they do it safely.

And isn't Cathay Pacific an Asian airline? ... How many of you lot would like to fly the approach in to the old Kai Tak airport each and every time you were returning to home base?

Possible answer to your post is that Eastern WORK culture is similer in some ways to Europe was is the Industrial Revolution. The attitude is to make money quick and hang the consequences.Health & Safety naaa, Slave Labour pay yeeeess! Industrial Relations and Unionisation naaaa! Safety Culture was that then? Command Gradient, Alpine Black Run!
The East will inevitably change and modernise, but in the meantime we can expect more industrial accidents, which sadly includes aviation. Just my opinion, of course.

This might not sit very comfortably with the many professional aviators in the East, but the accident stats published speak volumes.

@bud leon

Thanks for your post - if there were a "like-button" I would definitely push it. Operating for decades into Asia I agree with you!

It's culture not intelligence that seems to be the problem.

Asian Culture
 

Now that this subject has been opened for discussion can we begin by defining what this "Asian culture" being referred to in here means? There is still a high proportion of non Asian western and other pilots serving the carriers in Near East and Far eastern Asian countries so let us be absolutely sure what we are projecting in bashing this Asian Culture.Let us be specific so we can arhue it out and be more learned shall we?

Yes a full analysis of the CVR and DFDR is needed. That takes time so don't expect it for a while.

Analysis of this detail ids done behind the scenes with expertise and joint party cooperation. It typically doesn't get published until it is understood and agreed by the majority.

The world media has already satisfied the majority of their readers with news releases and have moved on to other news of "todays" interest elsewhere.

ATR has no need to jump ahead of the investigation by announcing opinions until/unless they need to publish a recommendation specifically concerning their product.

lomapaseo
 

I agree with all you have quoted. I understand Accident investigation enough to know the lead times involved and dissemination of "filtered" information..but that doesn't stop the public demanding it.

Little was revealed to the public concerning the possible crew error..after they were hailed as heroes..it was left at that.Could it be the Taiwanese authorities have muscled their objection against any further media reporting?

Am sure if it was a systemic problem on the type under investigation EADS/ATR would have dissemninated an AD or SB by by now. It is upto ATR726 operating crew who are members of this forum to tell whatever they know that could benefit this forum with our curious analyses.It would seem this story has already lost momentum until something new develops.
Perhaps it is time to move on.

Airbus would do what?

ATR is an independent company from Airbus, even if the latter by name is a 50% shareholder and a subcontractor for the assembly line. The other 50% are with Italy...

If anyone publishes anything concerning ATR it will be ATR itself.

The has been a lot of discussion about stick and rudder skills - or perceived lack thereof.

This series currently showing in Australia on SBS TV has been very interesting viewed against this background of this discussion. OK, there's a bit journalistic drama added, but I would sit behind these pilots anywhere ....

Here is a youtube URL for the first in the series:

https://www.youtube.com/watch?v=xBu-2EP_eMU

It's an English production; maybe most PPruNers have seen it before. (It sometimes takes while for these things to reach the southern colonies).

FullOppositeRudder, thanks for the link.

Perhaps mistakenly, I'm trying to compare PNG flying with bush piloting in other parts of the world. The guys interviewed and filmed in your linked clip seem to be mostly Europeans looking to build hours and an entry to regionals and then majors in Europe. Pretty much the same in the Americas although for some reason I get the impression that North American bush pilots like what they do and stick with it.

I wonder, are there no Chinese, Indians, other Asians in Susi's PNG ops and those of their contemporaries?

Culture
 

I've lived and worked in Asia for 15 years (nothing to do with aviation) and I think I understand what is meant by such comments—deference to authority, rote learning, superstition.

I agree that much of this is not unique to Asia and is typically a question of economic and social development. But that doesn’t make it any less real.

Which countries in Asia are not top-down, hierarchical societies? Which have a strong rule of law? Which have vibrant democracies that give voice to ordinary people? Which have a free press that encourages accountability for those at the top? Which have a tradition of independent academic enquiry? Which have a welfare system that means people aren’t terrified of losing their jobs? Which are free from corruption?

The answer is that no country in Asia has all of these things, and most have none of them at all. It’s not about race or culture—it’s about the basic institutions of advanced democracies.

It’s not my place to say how much any of this matters when it comes to CRM or aviation safety, but my lay view would be that I notice less of this deferential conservatism in free but poor India, and more of it in rich but restrictive Japan/Korea. (Plenty in SE Asia and China too.)

Even so... I suspect that such things only have an effect at the margins, which is not to say that they are unimportant, but a far bigger challenge is how Asia will find the 500k+ new pilots (plus another 500k+ mechanics) that Boeing predicts the region will need during the next two decades. It seems unlikely that these new pilots will be getting extra hours in aerobatic planes (or even extra hours in a sim)...

Oblivia - very well put. Of course personality plays a role as well. There are plenty of narcissistic captains of all nationalities who would crucify you for correcting them if you turned out to be wrong.

Bashing "Asian culture" does seem to be a default position for some. However...

I think it fair to say that it is based on statistics...

http://www.fearofflyinganxiety.com/w...t-airlines.jpg

Go to page 20 of this report...
http://www.icao.int/safety/Documents...port_FINAL.pdf

Lies, damn lies and statistics?

Not a useful statistic for comparing safety by region
 

That chart shows only the percentage of accidents that occurred in each region. Without knowing how much air traffic there was in each region I'm unable to compare how many crashes there were per flight or per million miles....or any other sensible definition of safety.

Asian culture is.120 hrs By month...... on ATR It Looks like 120 legs a month.....Tired crew never do good job...EVEN in asia......thailand DCA have been downgraded to level B ...next Taiwan
My personal records on ATR
India ........125 hrs
Philippines .105hrs
Indonesia ...113 hrs (28 days)
Thailand .....129 hrs (28 days)
my colleagues records
Laos ........+120hrs ( 13 days duty w/o day off)
Indonesia ...140 hrs
Taiwan........120 hrs
In Asia ..they believe FDTL is minimum duty.... ;))

Asian Culture
 

Two points made in a recent article in the New Zealand Morning Herald:

1. The FO was born in Hong Kong, and spent 15 years in New Zealand, including secondary school and initial flying training, thus, highly formative years. Excessive deference due to cultural factors is less likely, then, I think (especially as Hong Kong, on casual observation, does not seem exactly hide-bound).

2. There was some discussion as to why someone with so many hours was FO, and the phrase "command failure" was bandied around. The article claims he was, in fact, a captain, called in to fill an FO's shift.

Alas, the article refers to him as "hero pilot," but we don't really know yet, do we?

http://http://www.nzherald.co.nz/nz/...ectid=11397934

If you read back to the relevant posts you'll see why it's postulated that his being a captain might have been part of the CRM problem (if there was a CRM problem - it looks that way right now, but we don't have CVR yet).

Culture
 

Any objective look at different cultures will reveal their different propensities to deferentiality. Japan remains quite deferential; it's a culture of which I have long direct experience. On the other hand, deferentiality can crop up anywhere. The cockpit of the KLM 747 at Tenerife turned out to have been a highly deferential society.

The CVR will eventually reveal whether undue deferentiality was a significant factor here.

pag 10 have a table showing the accident ratio per million departures

Apart from Xiamen Airlines, a minor player in China, none of the Chinese (China Airlines is Taiwan) airlines feature. Chinese airlines are the major people movers in Asia.

Just a thought for the culture club here:
How many years or decades did it take the "Western" airlines and cultures (and for that matter military cultures) to address cockpit gradient and develop other CRM concepts and tools that are now industry standards?

How much effort went into that, and still goes into that? It didn't just happen.

One of the challenges inherent in this progress is a bit of a paradox: the Captain bears ultimate accountability for the conduct of his flight. This won't change any time soon, and so the balancing act has to be dealt with as an ongoing challenge for any captain/aircraft commander. It takes effort and development of some of those "people skills" that don't fit nicely into boxes and checklists.

As we used to say about a lot of thing: "If this stuff was easy, anybody could do it."

Back to the accident at hand: if there were confusing indications of what was wrong with the #2 engine, wherein the interpretation of what the engine and its instrumentation were telling the crew, what were they? If it was enough to induce the captain to command the good engine to shut down, the community who fly this model of aircraft would likely benefit from understanding that anomaly.
Possible CRM issue: when you both can't agree on what's wrong with the engine ...
When available, a CVR transcript might be very educational.

I've been looking at the FDR traces that have been released, and would like to share my analysis.
My credentials: I'm an amateur at this, but I have been debugging computer systems for over 40 years, so I'm used to looking at what evidence there is, and trying to work out what happened.

So: "Correct me if I'm wrong" applies to all of the below! When I say "the system" I mean some part of the automation - I don't know enough to say which particular box is involved, and it doesn't really matter at this point. All times are approximate, as it's a bit tricky to interpolate between the marked times.

It looks like everything was OK until 52:30. Then (at about the same time):
Objective TQ1 rose to 100%
No.1 Bleed Valve was closed
Prop.2 started to feather

I believe this is "Autofeather and Uptrim" by the system, because it believed No.2 engine had a problem.

Engine 2 Flame Out Warning was given
Fuel Flow 2 started to reduce, followed by No.2's ITT and RPMs (NL, NH, NP) and torque.

This seems to be the system bringing No.2 back to Idle, but without moving the Power Lever (can it do that on this aircraft?)

I'm going to call the small amount of time over which the above happened: "Point 1".

52:40 As a result of the above, No.2 spooled down, and by now it was at idle, feathered. (Its torque trace starts a strange triangle-wave, with about a 4-second period, spiking from zero up to about 150%.)

Power Lever 1 is retarded a bit, I'm assuming by the crew. CVR would be useful at this point to find out who/why.
They are still climbing, albeit slowly, until:
53:04 PL1 is retarded a lot further, then further still, then PL2 is firewalled, then PL1 pulled back further still, with consequent reduction of TQ1 to practically zero. Prop.1 is still in normal pitch and NP1 is about 80%, so for the next five seconds or so it must be windmilling. I'm calling this "Point 2".

53:20 Prop.1 starts to feather and Fuel Shutoff 1 is closed, shutting down that engine, (and with its torque trace taking on the same triangular wave pattern as No.2, and then its oil pressure trace doing the same thing).

53:28 they have both props feathered, No.1 shut down, No.2 idling but with its PL still firewalled, and they're at about 1200', descending.

Without the CVR we can only imagine what happens in the flight deck for the next three-quarters of a minute, but then:
54:14 Fuel Shutoff 1 is reopened, PL1 advanced slightly, so this is a relight attempt. They are at about 700'. "Point 3"

54:19 No.1 is spooling up, but Prop.1 is still feathered and not really turning.

54:27 Power Lever 2 is retarded most of the way, the Flame Out 2 warning stops, and Prop.2 starts to unfeather, so the system seems to think No.2 is OK now, and starts to spool it up, but it never contributes any thrust - unfeather never completes and the RPMs and torque barely get off the stops before the end of the trace, so I don't see the "sudden thrust from No.2 raised the right wing" happening, as some have said.

As I see it these are the questions:

1. What caused the system to "think" there was a No.2 flameout at Point 1? Before that the engine indications should still have looked normal to the crew, but it seems they had no chance to override the spool-down that resulted, and perhaps no reason to think they needed to.
2. Why was PL1 retarded 10 seconds after Point 1? It wasn't to shut down the engine as it was only a 10degree change.
3. At Point 2 they seem to have ignored the normal readings of No.1 and the "Idle" readings of No.2, and shut down No.1 Is this a "confirmation" problem because they were expecting No.1 to be a problem before they set off? Also, would the windmilling No.1 "confirm" they were shutting down the bad engine?
4. Why did retarding PL2 clear the erroneous Flame Out condition (not just cancelling the warning, but also bringing up the fuel flow and unfeathering)?
5. If they had retarded PL2 at Point 3 instead of trying to restart No.1, would they have got out of the situation in time?
6. (Unimportant) What's with the three triangle-wave traces?

Finally, why are the FDR traces all engine related, except the Main Gear, VHF1 (what is this?) and height readings? Where is the airspeed, vertical speed, control positions, etc?

Cheers,
Howard

I would be very surprised to find out there is some kind of and "anomaly" which contributed to them shutting down the wrong engine.

It shouldn't be the Captain ordering which engine to shut down. It is the Flying Pilot(FP) who instructs the the Non-Flying Pilot(NFP) to do the engine shutdown drill (memory items) or engine shutdown checklist.

Shutting down an engine is not a one step process. First the NFP placed his hand on the power lever and asks the FP to confirm it is the correct one. Then the NFP will slowly bring the power lever to idle. After those first two steps you should know with 100% certainty that you have the correct engine.

If all looks normal to both pilots(three in this case) then the NFP places his hand on the Condition Lever and again asks the FP to confirm he has the correct Condition Lever. THEN you bring the Condition Lever to the Cut Off or Shut Off position.

I have done hundreds of shutdowns in the Simulator and occasionally someone will start by placing their hand on the wrong power lever. Usually the flying pilot will catch it. In a very few cases they will miss identify the power lever and end up pulling back the power on the working engine. There is instant feedback that you have the wrong engine and a simple push forward on the lever fixes everything.

It is very difficult to imagine how this experienced 3 man crew accidentally shut down the WRONG engine if they followed the correct engine shutdown procedure.

Yes, it will be VERY interesting to read the CVR transcripts.

Very interesting analysis HDRW.

1. Could the ATPCS system have erroneously activated and feathered prop #2? It looks like a distinct possibility.

2. We will have to wait for the CVR for that one.

3. We will have to wait for the CVR for that one.

4. I believe that as part of the start sequence for eng #1 the crew selected the PWR MGT switch to MCT, thereby cancelling the autofeather signal from the ATPCS system, this i think caused eng #2 to begin to revert to normal.

5. I don't think so, as I believe this to be an ATPCS issue.

6. I have seen DFDR spikes before and they seem to appear when the recorded value is 0, some sort of "noise" on the recording, I am certain that they are not relevant to the accident.

On the ATR you even have an intermediate position before fuel shut off, called feather, and you really don't want to go past that point unless you are 100% certain. I doesn't take that long to unfeather the propeller again.