PPRuNe Forums

PPRuNe Forums (https://www.pprune.org/)
-   Spectators Balcony (Spotters Corner) (https://www.pprune.org/spectators-balcony-spotters-corner-52/)
-   -   AF66 CDG-LAX diverts - uncontained engine failure over Atlantic (https://www.pprune.org/spectators-balcony-spotters-corner/600170-af66-cdg-lax-diverts-uncontained-engine-failure-over-atlantic.html)

EEngr 4th Oct 2017 03:05


Originally Posted by RatherBeFlying (Post 9913512)
Possibly a fan blade, but can't say for sure from the available photos. Cowl pieces are the other possibility, some fragments in the vanes look like composite.

It does look like a blade, with the spline that engages in the fan hub slot near the top (#176). I'd guess that if the fan did come loose and move forward, getting tangled in the cowling's Kevlar containment band made short work of the integrity of the fan. If they go looking for it, it won't be a fan stuck in the ice of Greenland. More likely a 20 mile trail of blades and pieces of cowling.

Nonbiased 4th Oct 2017 04:03

I have a video on my youtube account which has all the details from my perspective as a passenger in row 12 of LAST Class, the unlucky ones who went to ATL on the second flight. I can't post links though as I'm a noob. Does anyone want to DM me and put the link up to share the experience with you guys?

ManInJapan 4th Oct 2017 05:24


Originally Posted by lomapaseo (Post 9913467)
I take it that some have not noted what looks to be an intact fan blade and its attachments lodged in the vanes behind the fan?

That doesn't look like a fan blade to me, and if the fan blades (even just one) hit the guide vanes behind them, I think you'd see a lot more damage to the guide vanes. They look almost untouched. The debris in the photo could be anything really.

ManInJapan 4th Oct 2017 05:39


Originally Posted by EEngr (Post 9913451)
Just guessing here: The fan hub to shaft coupling is designed to withstand quite a bit of thrust (produced by the fan). But the fan hub to LP compressor barrel coupling not nearly so much. It's more for the torque of spinning the compressor. So an initial failure of the shaft to fan (where a jagged piece remains) would allow the fan to pull forward. The 24 (?) bolts holding the compressor stage failed under tension (photo in #194). That may be by design. To let the fan go forwardvwithout trying to pull the compressor rotor through the stator.

Like 'lomapaseo' I'm inclined to think that purposeful design of the fan to separate is unlikely, but to me your sequence of failure sounds right. I do find it surprising however that the bolts at the compressor flange all failed.

My armchair opinion is that the bolts in the flange were not tightened / sized / put in correctly.

A failure of the driveshaft would pull the whole LP compressor forward which would either pull it through the guide vanes and out the front of the engine with the fan. Or the guide vanes (stators) would hold (as they have done) and prevent the LP compressor from moving out of the engine. In this case the deceleration of the LP compressor and the fan would be high, but I doubt high enough to severe the bolts at the LP compressor flange as cleanly as what has occurred.

pax2908 4th Oct 2017 05:52

Still, perhaps it is the latter ... if the core comes to an "instant" stop, the choice might have been to let the fan "snap" off, rather than overstressing the engine attachment to the wing. But what do I know :)

procede 4th Oct 2017 06:08


Originally Posted by Barnsbury (Post 9913090)
The suggested failure modes still don't fully add up. The fan would have been producing a leisurely five or six tons or so of thrust, which wouldn't have troubled anything compared to the previous take-off loads, so this isn't an overload failure.

It probably happened during a step climb, thus at higher thrust levels.

The total load case is not just thrust, but also torque and centrifugal loads.

I would think that with the lower air density, the rotational speed of the fan will be considerably higher at altitude than at ground level for the same amount of thrust.

wrighar 4th Oct 2017 06:12

1 Attachment(s)

Originally Posted by ManInJapan (Post 9913573)
That doesn't look like a fan blade to me, and if the fan blades (even just one) hit the guide vanes behind them, I think you'd see a lot more damage to the guide vanes. They look almost untouched. The debris in the photo could be anything really.

Difficult to say with these pictures.

DaveReidUK 4th Oct 2017 06:35


Originally Posted by procede (Post 9913592)
I would think that with the lower air density, the rotational speed of the fan will be considerably higher at altitude than at ground level for the same amount of thrust.

No engine is capable of producing the same amount of thrust at cruise level compared to SL.

But of course it doesn't need to.

For a large turbofan, N1 in the cruise will typically be in the region of 80%.

procede 4th Oct 2017 07:29

I agree. Maximum thrust levels decrease with altitude.
During a step climb N1 will probably get close to 100% (max climb).
One more thing to mention is thermal stress, which could be higher at altitude. And then there are acoustic stresses...

Does the GP7000 have heating in the cone? Could well be that that failed...

Stumpy Grinder 4th Oct 2017 07:31

1 Attachment(s)
If it waddles like a duck and goes Quack then it's probably a duck......


That there is a fan blade, of that there can be no doubt.

Onceapilot 4th Oct 2017 08:27


Originally Posted by DaveReidUK (Post 9913605)
No engine is capable of producing the same amount of thrust at cruise level compared to SL.

This is, of course, incorrect. It would be correct to say that modern civil turbofans produce their highest or limiting thrust at low altitude.

scifi 4th Oct 2017 09:27

What is left of the engine should be able to spin without causing much vibration.
However, the passengers felt a lot of vibration for several seconds.
This seems to imply that the fan broke up in sequence... First one blade detached, then the vibration caused all the bolts to fail, and what was left of the fan disk spun off forwards.
.

DaveReidUK 4th Oct 2017 09:47


Originally Posted by Onceapilot (Post 9913696)
This is, of course, incorrect. It would be correct to say that modern civil turbofans produce their highest or limiting thrust at low altitude.

Hmmm. It sounds like we're saying the same thing in two different ways.

That aside, since maximum thrust on any jet engine is achieved at zero TAS, I think we can safely say that's not available at altitude. :O

Musician 4th Oct 2017 10:00

looks like that's a fan blade
 
4 Attachment(s)
Comparing these details (taken from the big twitter images that tricityb linked in post #175) with the picture of posted by Stumpy Grinder in #185 that shows a worker assembling the fan, I can identify the curved line separating the two types of surface on the blade and the rounded ridge that locks the blade to the hub.

CAAAD 4th Oct 2017 10:09

I think it may well be a fan blade.

The dovetail root length from the very useful GA is about half the span of the fan OGV, and the leading edge of the fan blade is about one OGV span from root to change of curvature in profile. Also , the root is the correct colour.

And a bit of containment on the blade tip.

Fan failures are very chaotic affairs and we cannot possibly predict the sequence of events from the limited material available.

I'm sure East Hartford will have a very good idea by now.

slip and turn 4th Oct 2017 10:17


Originally Posted by scifi (Post 9913762)
What is left of the engine should be able to spin without causing much vibration.

Might you mean pie in the sky? :}

However, the passengers felt a lot of vibration for several seconds.
Yes.

This seems to imply that the fan broke up in sequence... First one blade detached, then the vibration caused all the bolts to fail, and what was left of the fan disk spun off forwards.
.
Yes a sequence as long as your arm and more! For my money, and admittedly it is a statement of the bleedin' obvious, I reckon the bright sparks will eventually compute from the available remains that there was a lot more going on after the first thing broke than had previously been computer-modelled. And personally from a little reading around the subject, I have little doubt that at some stage - milliseconds or possibly whole seconds after the first thing broke - a somewhat massive lump came off in a dizzying non axi-symmetric whirl, having long since (well a few seconds anyway!) given up the momentum of its original purpose to something more of a sling shot effect, and in passing gave the wing leading edge (some distance behind the fan case not forward of it) quite a ding!

I really would love to see that tail fin video - but will we ever? Anyone know if footage is retained informally other than by passengers who happen to be filming the screens on the backs of seats?

Stumpy Grinder 4th Oct 2017 10:30

The post fan blade off (FBO) vibration is 2 phased, firstly a High Level Short Duration (HLSD) followed by (normally) Low Level Long Duration as the unbalanced engine windmills. All gas turbine and airframe parts are subjected to either test or analysis to demonstrate that they can remain attached and if necessary functional during and after a FBO event.

DType 4th Oct 2017 11:57

It could have been even more exciting if the fan had detached when still on the ground during take off. When they drop to the ground, they zip v v rapidly sideways, bouncing as they go. I've seen a few holes in test bed roofs, and liberated (turbine) discs in distant fields.

G-CPTN 4th Oct 2017 12:05

If, as is suggested, that the fan 'broke up', what chance of it being due to FOD damage?

Birdstrike or drone?

slip and turn 4th Oct 2017 12:28


Originally Posted by DaveReidUK (Post 9913905)
Notwithstanding the fan blade(s) embedded in the guide vanes, the engine did still satisfy that particular FAR, in that it neither caught fire nor detached from its mounting attachments.

Well now we're venturing into the bounds of semantics - when I studied the subject of how power is delivered, we sat down and opened learned reference works titled "propulsion" as opposed to "engines".

As I understand it, an aircraft engine these days comprises several modules each and every one of which is mounted in a obligatory fixed sense in relation to each adjacent module and at sundry well chosen points directly to the airframe. Plenty of suitable elasticity is built in, of course.

So the plan is that the collection of major engine bits on landing should still ideally be reasonably complete even after an FBO. That's as opposed to a scenario where the fan and fan containment module are merely expected to self-adhere to their master's heel through the worst of the thick and thin, but having given up their angular antics, are allowed to pop off in some as yet undeciphered direction when the final urging to leave gets overwhelming!

Bearing in mind that the few hundred kilo module delivering 80% of the propulsion is indeed our Elvis in this case, one wonders if the purport of § 33.94 seriously allows for an exit like that ... methinks not.

For the pernickety and legal-minded, the question might be 'What is meant by "mounting attachments" in § 33.94?', but I am sure more serious-minded engineers are somewhat beyond that.

musicrab 4th Oct 2017 13:10


Originally Posted by G-CPTN (Post 9913919)
If, as is suggested, that the fan 'broke up',

I don't think the fan broke up e.g. there's no evidence that any part of the fan was ingested; in my view the hub/shaft/bearing gave up resulting in the fan cutting up the engine cover/cowel before the complete fan assembly separated under or over the wing (thank goodness not through the wing).

EEngr 4th Oct 2017 14:43


Originally Posted by Stumpy Grinder (Post 9913817)
during and after a FBO event.

One fan blade or the entire fan?

Turbine D 4th Oct 2017 14:55

I agree with musicrab, I don't think the fan broke up. Hopefully the bearing remnants should still be there. The bearing is a high speed thrust bearing with a split inner ring race. The inner race is attached to the very rear end of the fan stub shaft and rotates with fan. The outer race is attached to the cylindrical cone and doesn't rotate. The split inner ring allows for some axial movement of the fan stub shaft forward or aft which occurs during acceleration or deceleration and at the same time, avoids putting undue pressure on the ballbearings as the overall bearing distorts slightly. If the bearing fails, the precise centering of the shaft connected to the fan is lost and IMHO, from there on all hell can break loose which appears to be what happened. Examination of the LPT module will give additional failure information.

Stumpy Grinder 4th Oct 2017 22:47


Originally Posted by EEngr (Post 9914069)
One fan blade or the entire fan?


1 blade, if the whole fan comes off then the imbalance (and hence vibration) is far less so 1 FBO is worst case in terms of dynamic vibration.


I'm seeing reports from passengers of high vibration, therefore it was possibly a FBO event that was contained (blade in the bypass duct) but some secondary and perhaps further failures leading to fan loss but it would have had to have been a very rapid failure sequence as the fan slows and hence has reduced loading.


That blade in the duct also appears to still have it's root; It could be a red herring, defo not a black Griffon........


This is indeed a curious one.

Musician 5th Oct 2017 02:34


Originally Posted by Stumpy Grinder (Post 9914555)
That blade in the duct also appears to still have it's root; It could be a red herring, defo not a black Griffon........

Well, once the front end departs, the blade would be free to slide out backward, especially if it got a bit jammed in the lining (or what was left of it)? and then the wind could have lodged it where it got stuck.

Another option: the fan ring did split, and it would split at the weakest points, aka where slots for the blades are, freeing the blades in those slots with their roots intact. This feels more plausible than a blade sliding out, but would the ring have split?

Onceapilot 5th Oct 2017 08:03

Any "official" comments seen by anyone on this event yet?

Squawk_ident 5th Oct 2017 08:23

ATC transcript (Gander Center/ AFR066)
 
There are various archives files on LiveATC where you can hear the R/T between AF66 and GANDER Center or Goose.
LATC is archiving its audio feeds every 30 minutes for each recorded station. I believe that each uploader has his own receptor hence the good or poor reception is depending of their position.

The N°4 engine shutdown happened over Greenland at 1351Z. At that time one can observe that AF66 starts its security descent.

On LiveATC I've found the following feeds starting at 1430z :
-CYYT GND/TWR/Gander Center
-CYYR Gander Center

At that time AF66 is at FL270

The R/T between AFR66 and Gander is starting at + 18'18" (14.48'18"z) on both feeds.
A word of caution:
Both feeds are very likely scanners that grab pre-programmed frequencies. Scanner(s) stops on an active frequency( R/T) and, when transmission is ended on this frequency then jumps to another preselected freq if it becomes active etc... therefore transmissions with AFR66 are not entirely covered if another frequency become active in between. Frustrating sometimes...
Also : it seems that exists some kind of overlapping between each 30 minutes feeds therefore times indicated may be wrong.

This is my transcript from both feeds. CYYt appeared to be less cut by other transmissions. I could get almost the entire transmission using both CYYR and CYYT feeds.
+18'18"
AFR66 : GANDER Control this is AIRFRANCE 066 MAYDAY did you copy?
G : AIR FRANCE 066 HEAVY GANDER CENTER good day I read you loud and clear go ahead

AF : AIR FRANCE 066 MAYDAY FL 270 on course YYR request priority landing engine number four out and we are estimating the airfield at 1536.
G: OK I check the engine number 4 and you can expect priority landing and I check your estimate of 1536.
AF: [garbled]... and if possible to have the best runway for arrival?
G : AFR066 you can expect the RNAV runway 26 that will be straight in you can proceed direct BIVMI that's Bravo India Victor Mike India
AF : repeat the direct AIR FRANCE 066 MAYDAY?
G : that's : bravo india victor mike india it will be the initial fix for the RNAV 26
AF: OK from position direct to BIVMI for RNAV 26 at Goose and request priority when arrival
G: AFR066 even expect the fire services on the ground and can you confirm the fire is out on engine number 4?
AF : yes sir fire is out on engine number 4 but we have small/some? Damage on slat and the right wing
G : AFR066 say again your last?
AF :we have little damage on the slat on engine number 4 and we have big damage engine number 4
G : AFR066 I check your remark.

At +22'40" Gander request AFR66 to squawk 6050 and confirms "radar identified at FL270" at +23'48"

-CYYT1500Z

At +0252 AFR066 request descent and is cleared 14000 Alt. setting 30.04 at Goose.
AFR66 is requested to contact Goose on 119.5 at +03'34"

The other feeds are CYYR DEL/GND/TWR/APP from 1500z to 1600z.
You can check further CYYR feeds until landing and after. Sorry but the zip file won't upload...

romiglups 5th Oct 2017 09:46

parts of engine found in greenland
 
In a statement in french on their website , BEA announces that some parts of engine have been detected in a desert area in Greenland 150 kms southeast of Paamiut by an helicopter sent by danish authorities. BEA and danish inquirers will organise parts recovery.

(sorry server denies to post URLs, this is on frontpage).

wiggy 5th Oct 2017 10:00

Romiglups

Thanks for that, hopefully this link to the BEA statement will work:

https://www.bea.aero/index.php?id=25...76709dddbb6b8f

SStreeter 5th Oct 2017 10:12

That is, according to translate.google.co.uk

<
The use of the data contained in the parameter recorder (FDR) was used to determine the place where the damage occurred during the cruise to the Airbus A380 engine number 4 and to delimit a search area for the detached.

It is a desert area, covered with ice, located about 150 kilometers south-east of the town of Paamiut, located on the west coast.

At the request of the Danish Accident Investigation Office, a helicopter from the Danish airline Air Greenland flew yesterday and spotted parts from the engine.

The BEA is in contact with its Danish counterpart to organize the recovery of these elements.

Work on the analysis of the data contained in the flight recorders will continue at the BEA laboratory. The engine computers will also be analyzed at the manufacturer of these equipment in the United States.

The BEA investigators who traveled to Goose Bay and Ottawa will return to France on Thursday, October 5th.
>

wiggy 5th Oct 2017 10:20

For the avoidance of any :ooh: s it might be worth pointing out the probably obvious fact that the best translation for

désertique
in the context of this BEA report is "wilderness", "deserted", or similar......

DaveReidUK 5th Oct 2017 11:24

Though dated 3rd October, I haven't seen this posted here yet:

"Following the accident to the Airbus A380, registered F-HPJE, during cruise over Greenland, the Danish Accident Investigation Board (AIB DK) have delegated to the BEA the opening and conduction of the safety investigation. The BEA represents France, State of the Operator, State of Registry and State of Design. Investigators from AIB DK, representing Greenland and Denmark, State of Occurrence, from NTSB, representing the United States of America, State of Engine Manufacturers, and TSB, representing Canada, State where the flight crew diverted, are taking part in this safety investigation.

Four BEA investigators travelled to Goose Bay on Sunday 1st October, accompanied by technical advisors from Airbus and Air France. NTSB investigators, accompanied by technical advisors from engine manufacturers General Electric and Pratt & Whitney, as well as TSB investigators, also travelled to the site.

A fifth BEA investigator travelled to Ottawa to attend the first read-out of the Flight Recorder data. This read-out enabled the investigators to determine at which point of the flight the failure occured.

The Airbus A380-800, operated by Air France, was performing the flight from Paris (France) to Los Angeles (United States of America) under the callsign AF066. It had taken off from Paris Charles de Gaulle Airport on Saturday 30th September 2017 around 09 h30 (UTC), with 497 passengers and 24 crew members on board. Following a failure on the number 4 engine while the plane was overflying Greenland, the flight crew diverted to Goose Bay Airport (Canada), where they landed at 15 h 42 (UTC) without any further incident.

Damage to the aircraft seems to be confined to the number 4 engine and its immediate surroundings. A visual check of the engine has shown that the fan, first rotating part at the front of the engine, has separated in flight, taking with it the whole air inlet cowl."

Information du 3 octobre 2017

VNAV PATH 5th Oct 2017 13:12

Debris have been located by an Air Greenland helicopter, 150 km south east of Paamiut on the west coast.

Do not have the british link of french BEA bulletin, but here is the essential.

https://www.bea.aero/index.php?id=25...76709dddbb6b8f



Information du 5 octobre 2017
L’exploitation des données contenues dans l’enregistreur de paramètres (FDR) a permis de déterminer le lieu où l’avarie est survenue en croisière au moteur numéro 4 de l’Airbus A380 et de délimiter une zone de recherche des éléments qui se sont détachés.
Il s’agit d’une zone désertique, couverte de glace, localisée à environ 150 kilomètres au sud-est de la ville de Paamiut, située sur la côte ouest.
À la demande du bureau danois d’enquêtes sur les accidents, un hélicoptère de la compagnie danoise Air Greenland a survolé hier cette zone et a repéré des pièces provenant du moteur.
Le BEA est en relation avec son homologue danois pour organiser la récupération de ces éléments.

radken 5th Oct 2017 13:30

I am definitely not an engineer, but I am a great admirer and respecter of the magnificent human achievement the GP7200 engine represents. That it, and it’s various brothers and sisters, can safely operate and perform as routinely and reliably as they do has wrought change and improvement to the world like no other invention... except the wheel and perhaps the taming of electricity itself.
At heroic effort and cost, the blending of concept and human knowledge, then actually producing a device which, with about 39 percent efficiency, can produce some 70,000 useable hp, — then coupling it up with a few others like it to go merrily flying across the oceans of the world (along with 500 or so other people) has to rank as one of the major wonders.
I believe the dynamic of huge, pure power, such as presented to the N1 fan assembly, itself shrunk to the minimum size possible to be complementary to the rest of the 6 or 7 ton engine, transports engineering to the very edge of what is possible.
That there has now been a mysterious and violent event occur at the dynamic point where, so to speak, the “rubber meets the road” it’s fair warning by the grace of God that something very major may be wrong. In aviation we often don’t get a warning something is amiss...but here we have, and all survived to tell about it. Recall the first (AA) DC-10 aft cargo door related decompression, at what, FL 14? The collapsing aft cabin floor tensioning control cables? The successful emer landing? The FAA caving then on an emer AD to ground the type? Just issue a service bulletin? Then later the Turkish AL decompression and loss (Orly?) with 346 on-board?!! We’ve had our warning. Methinks we need to bite the bullet. Stop the A380 for a bit and take a look.

Bergerie1 5th Oct 2017 15:54

radken,

No reason to ground the A380, it's not the aircraft it's the engine - and - different engine, different problem. The QANTAS incident was with a Rolls Royce engine, this one is a GP7200. The RR failure was a turbine disk, this one looks like an LP shaft failure.

jolihokistix 5th Oct 2017 16:50

Agreed re désertique. It seems to imply reassurance and relief that the route had been a wise choice, and the bits thankfully fell over an area devoid of any habitation.

RatherBeFlying 5th Oct 2017 19:24

What failed first?
 
We know one blade detached from the hub - perhaps more. The rotational forces may have overloaded the flange.

Or the flange failure resulted in blades rubbing the containment ring and initiating one or more hub attachment failures.

The possibilities range from just one blade coming out of the hub to all of them.

BEA has retrieved the engine computers. Hopefully the event remains on memory.

The cutaways we have seen show adjacent roller and thrust bearings just behind the flange. We don't yet know their status.

sardak 5th Oct 2017 21:09

Here's a map of the flight path from FlightAware and a 150 kilometer arc from the Paamuit airport. A few minutes later and engine pieces and parts might not have been so quickly found.
https://i.imgur.com/gPNrUL9.png
A larger size here: https://i.imgur.com/VlGN4rb.png

VNAV PATH 6th Oct 2017 18:34

http://nsa39.casimages.com/img/2017/...0158324047.jpg

wrighar 6th Oct 2017 18:54

1 Attachment(s)
And another


All times are GMT. The time now is 23:54.


Copyright © 2024 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.