Qantas A380 uncontained #2 engine failure

[rob_ginger]

Why are Qantas still only flying 2 A380's ?

Does anyone have any inside knowledge on why Qantas are still only flying 2 of their A380's ? There's an article in an Australian newspaper: Qantas receives its seventh A380 superjumbo from Airbus which says that Qantas now has 7 A380's. But only 2 are actually flying ? I saw in another thread that for the non-stop flights from LAX the A380 was limited by the maximum take-off thrust "safely" available - but the limitation appeared to be that the engines would require frequent inspection/replacement, rather than a hard take-off thrust limit. Why is is that Sinagapore and Lufthansa are flying all their A380s, and Qantas isn't ??

Even if the engines had to be changed/inspected frequently I would think that's an RR problem - Qantas could be rotating their fleet, flying a couple of aircraft to LAX while the others were sitting on the ground with RR inspecting/changing the engines.

Since this *is* a rumour site I will say that I was talking to an ex-QF employee today. He said that he thought that Qantas were worried about the failure of the A380 fuel transfer systems after the uncontained IPT failure. If that happened over the Pacific, where alternate airports are few and far between, the out-of-balance condition could become critical before an alternate was reached. Any comments on that ?

Also, does anyone know if Airbus are working on any modifications to the A380's redundant systems as a result of the QF IPT incident ? I know an uncontained IPT failure is a pretty catastrophic event, but even so I was surprised at the consequential system failures listed in the ATSB interim report. Apart from the complete loss of fuel transfer capability I saw that engines 1 & 4 were operating in "degraded" mode. I can understand an engine on the same wing being affected, but why the RH outer engine was affected by the failure of the LH inner engine seems strange to me ??

[NigelOnDraft]

Rob...

I would suggest the answers to (almost) all your queries / points are discussed / made clear in the thread

NoD

[hewlett]

Can someone explain "degraded mode" and / or the reasons #1 and #4 went into this mode?

[mm43]

Can someone explain "degraded mode" and / or the reasons #1 and #4 went into this mode?

Go back to Post #1738 and read on for the next 10 or more.

mm43

[mike-wsm]

Originally Posted by Old Engineer

. . . drawings of the individual 8 modules . . .

Does this help?

Silhouettes of engine modules
http://www.rolls-royce.com/Images/ga...tcm92-4977.pdf
Pages 14 and 15

[hewlett]

mm43 thankyou

[rob_ginger]

NoD & mm43:

I've been following this thread fairly closely, and I really don't think my points have been answered 100%. And I *did* re-read from post #1738.

If I understand the "degraded" mode correctly the engine controller falls back to a fail-safe mode because some sensor data is not available. I can understand how #1 engine degraded, because the wiring in the left wing was damaged. But #4 engine is on the other wing - so what went wrong to cause that to "degrade" ? If that's explained somewhere please let me know, 'cos I can't see it .

Apologies if I'm just illiterate/stupid !

[mm43]

rob ginger

I thought that 35YearPilot in Post #1751 had explained rather well the Alternate/Direct Law relationship that gave rise to the engine protections adopted. The Eng 2 first mentioned in the following, seems to be a typo, and he was referring to Eng 3, i.e. ATHR was not available and protections were applied.

Eng 2 in ALTN Mode (N1 with -4% thrust protection) (Due EEC having insufficient valid signals to calculate TPR)

1+4 EEC errors logged. Degraded mode is purely due to loss of sensor inputs/EEC errors. Think of it as the engine in Direct Law (no thrust protections!) (Due EEC not able to calc TPR demand. So N1 commanded as a function of TRA and altitude)

One engine in a degraded mode does not cause another symmetrical engine into degrade mode. Airbus never matches thrust - it's against their philosophy "manual thrust is manual thrust!" - not like pseudo Boeing "manual thrust".

So 1 & 4 in Degraded (think Direct Law) Mode (no protections)
3 in ALTN mode (think Alternate Law with protections)
2 was in Degraded mode (not surprising given there were so few sensors left)

NO engines were in a normal mode.

This is why 1,2,4 had N1 centric displays (Degraded) and 3 had an TPR centric display (ALTN) (though thrust set via N1 with 4% pad).

Thrust control on 1,3,4 all 100% correct (but differing protections)

Auto Thrust inop due to MORE THAN 2 eng in Degraded Mode

1+4 thrust levers held constant to: (not in any manuals)

permit very accurate (vernier) total thrust -> speed control on finals
minimise thrust missmatch->yaw->roll->flight control demand on finals with thrust changes (ESSENTIAL when little excess roll control available)


1+4 WAS NOT held at constant thrust due to them being in Degraded Mode (illogical).

Conclusion: The three level 380s thrust system has incredible fail operational capability.

From memory, I believe that Capt Evans said that a decision was made to use No.3 for speed control as less yaw correction would be required and Nos.1 & 4 were matched.

EDIT:: See Post #1931 for 35YearPilot's update.

mm43

[mike-wsm]

Originally Posted by Old Engineer

. . . spline wear . . . canary . . .

Our two gas turbine experts have explained to us that the pressures in the various chambers are such that the forces on the shafts tend to cancel and there is very little nett load on the bearings.

I wonder whether the problem in the Trent 972 might perhaps be deviation from the design pressures at high power, creating a force imbalance and leading to greater loads on the bearings and the bearing support structures? This in turn would lead to relative movement of the modules, with consequent spline wear and oil pipe fractures. Inspection of the splines would indicate the extent of any movement.

[Old Engineer]

mike-wsm:

Does this help?
Silhouettes of engine modules

Yes, thanks. Also liked the Vision of the Future on p13. Magnetic bearings... no more of that pesky oil; generator driven by windmilling fan... no more pesky shutdowns taking a bus off-line. Fuel, electricity, and thrust the only interfaces to the A/C... no control IF, or is it by radio? (or possibly by telephathy direct to pilot in degraded mode?) It's useful to see where the concerns lay, looking back 3 yrs. Prescient... and borne out by events.

OE

[mrdeux]

Does anyone have any inside knowledge on why Qantas are still only flying 2 of their A380's ? There's an article in an Australian newspaper: Qantas receives its seventh A380 superjumbo from Airbus which says that Qantas now has 7 A380's. But only 2 are actually flying ?

A is in Singapore
B is in Germany undergoing C check. Back soon.
C is in Sydney awaiting engine(s). C check due.
D, E, F all in service.
G goes into service tomorrow.

[Turbine D]

Old Engineer

I liked your analogy as to the spline wear being the "canary in the coal mine".

Here is why:

Spline coupling allows a shaft to be easily dismantled while providing a high-torque capacity for minimum size. In addition, it can allow relative axial and radial motions between the coupled shafts. Stability conditions can change should the alignment state between the driver and the driven machines (rotors) changes. In one manner, angular misalignment can greatly increase the slip, relative slip of small amplitude can occur between the contacting tooth surfaces. This can give rise to fretting damage which may limit the life of the coupling and has the potential to compromise integrity. This has been a particular concern in aircraft mechanical systems. In another manner, misalignment can cause high vibrations with different symptoms that sometimes cannot be explained. For a real system, there are two kinds of misalignments: static misalignment and dynamic misalignment (dynamic vibration displacement). In a real rotor-spline coupling system, all splines meshing tightly by transmitting large torque cause a deformation of each spline. At the same time, for coupling nodes vibrating with the system, there is relative displacement between splines in two half-couplings, which cause another deformation of each spline; both the deformations generate meshing force.
When the static misalignment appears and keeps constant, the following holds true:
(a) Meshing force changes linearly with dynamic vibration displacement, approximately, but there is an inflexion.
(b) The slope coefficient of meshing force curve reflects the stiffness of coupling, consequently, during the vibration of system, the stiffness of coupling is not a constant, it relates to dynamic vibration displacement and depends on static misalignment.
The dynamics of rotor-spline coupling system shows the following: 1X-rotating speed is the main response frequency when there is no misalignment; while 2X-rotating speed appears when misalignment is present; with the increase of misalignment, shaft orbit departs from original orbit, the magnitudes of all frequencies increase, and 2X-rotating speed increases rapidly. Thus, misalignment makes the vibration of a system more complicated.

This information is from:

Research Article
Meshing Force of Misaligned Spline Coupling and the Influence on Rotor System
Guang Zhao, Zhansheng Liu, and Feng Chen
School of Energy Science and Engineering, Harbin Institute of Technology, Harbin 150001, China

Could this be a viable explanation as to increased loading on the bearings and bearing structures at high thrust levels, e.g. takeoff?

Turbine D

[JFZ90]

Our two gas turbine experts have explained to us that the pressures in the various chambers are such that the forces on the shafts tend to cancel and there is very little nett load on the bearings.

If I understood correctly, they explained that the pressure forces are significant, and hence have a tangible effect on bearing axial loads. This is not the same as saying the bearing loads are very little.

The design aim will always be to minimise bearing axial loads as much as possible, as simpistically they then last longer - which is a good thing.

However, as explained in the CF6 link from the manufacturer, the axial thrust fan bearing load is still nearly 20,000lbs. This is not "very little", when compared to a total engine thrust of around 50,000lbs, though it is lower than the oft quoted thrust force that is created by the fan (i.e. 80% of the 50,000lbs). It would be interesting to know whether later gen engines have actually reduced these axial loads further through careful design of turbine section - as the CF6 data is quite old.

Is any of the bearing axial load discussion relevant to these failures - personally I suspect not, but it is quite an interesting aspect of the engine design.

[llagonne66]

QFA A380 fleet status

mrdeux,

We all know your background.
Thanks for this update that nobody will question.

Guess the one in FRA can be seen here :
Photo of Note: Superjumbo, Dreamjumbo, Superjumbo - FlightBlogger - Aviation News, Commentary and Analysis

[Old Engineer]

mike-wsm:

Our two gas turbine experts have explained to us that the pressures in the various chambers are such that the forces on the shafts tend to cancel and there is very little nett load on the bearings.

I'd agree with that for the forces involved with the HP and IP compressors and turbines, as the action of these is all internal to the casing. But with the LP system, there are forces external to the casing. First there is the hot exhaust, about 20% of the thrust, which reacts against the diverging walls of the exhaust gas nozzle, to include the diverging walls of the turbine stage, and possibly the preceding two stages.

The similar hot gas thrust in the preceding IP and HP turbines may also react against the casing with a forward thrust, but there will be an opposing thrust against the walls of the compressor stages. Presumably this will be less because the pressure is not yet augmented by the burning of fuel in the airstream.

And likewise, the LP circuit acts in (small?) part to compress air used internally in the engine, which associated forces will act through the casing. It is also true that the bypass air, accounting for 80% of the thrust, passes through a venturi-type nozzle. But no fuel is burned in this circuit, so the pressures opposing thrust and aiding thrust upon the walls of the venturi would seem to balance. I assume the relative low pressure at the venturi throat is used to drive the internal ventilation of the air buffers within the engine, and as such would reduce the net bypass thrust slightly below that created by the fan.

This is all a kind of "black-box" analysis, rather than blade ring by blade ring. Applying that concept to the assembled engine, the Y-yoke connection forward on the casing is shown with a vertical orientation terminating in a small attachment pad with 4 bolt holes at the corners-- this is an arrangement clearly designed to support the weight of the engine near its center of gravity, but not to impart any appreciable thrust to the pylon from this area of the casing.

So whether derived from tension in the casing or the tension in the LP shaft, which are on-axis forces when resolved into two single vector forces, the entire thrust is delivered to the pylon by the eccentricly mounted drag link, as I called it. Thrust link might be a better term. But certainly the diversion of forces away from the rear LP shaft bearing remains an interesting question.

NASA has published an interesting paper on how an aircraft turbine might be analyzed as to its behavior in a number of areas, in rather substantial detail. I haven't had time other than to glance at it. CFD, is it?..."Computerized fluid dynamics". Yes, real headache material that, at least in the days of the mechanical Frieden calculator. But today's desktops running XP should be able to handle it (maybe XP64 would be required), if only a free version running on C and Windows were available. A Fortran version, which would run on anything, is probably too much to hope for these days. Just to plug your ideas about the effects of overpressure/overtemperature into the example machine of this paper would give some insight.

For what it's worth (just turning the 20 pages is worth the effort of downloading it), the link to this paper is, with [] my file additions:

File Title: [NASA-TurbineAnalysis]TM-1998-208402.pdf
File Name: Tina's Work Drive-11234 Hall-11234 TM/P
Author: Tina Crawford, June 16, 1998, 4.63 MB, Adobe Reader 7.0
Link: http://gltrs.grc.nasa.gov/reports/19...998-208402.pdf

PS: I don't know where the "evil" faces came from; I lifted the File Name from the file properties, and pprune software converted them. Well, the original text at each is colon, cap E, no spaces. Perhaps a bit of black humor there. Be forewarned.

OE

[jcjeant]

Hi,

PS: I don't know where the "evil" faces came from; I lifted the File Name from the file properties, and PPRuNe software converted them. Well, the original text at each is colon, cap E, no spaces. Perhaps a bit of black humor there. Be forewarned.

That cause the code for "Evil" is

Code:

:E

[rob_ginger]

mm43

Thanks for taking the time to reply to my question. However something is still not clear to me:

1+4 EEC errors logged. Degraded mode is purely due to loss of sensor inputs/EEC errors. Think of it as the engine in Direct Law (no thrust protections!) (Due EEC not able to calc TPR demand. So N1 commanded as a function of TRA and altitude)

Now I thought that (from an aircraft systems point of view) each engine was completely independent from the other engines and any other aircraft systems. Each engine has its own generator which powers the fuel pumps and FADEC/EEC/whatever, so that with gravity fuel feed it will keep producing power through hell and high-water. (And indeed #1 couldn't be shut down after landing, and had to be drowned with foam after 4 or 5 hours.)

But why did #4 log EEC errors ? Here's an explanation of TPR from a Tech Log thread:

From the RR Trent training notes,

"TPR actual is derived from the P20, P30, T20 and EGT parameters."

From my understanding all these parameters are available from the engine itself. So to (finally !) get to my point - why did an undamaged engine on an undamaged wing generate EEC errors ? In my book that's an unintended consequential failure. Am I missing something ?

BTW, I'm not an anti-Airbus critic - after all it took many years for the 747 designers to realise that it was "a bad idea" to locate all four AC generator control units in the same bay directly under a galley.

[35YearPilot]

MM43 and rob_ginger

MM43>>>> I thought that 35YearPilot in Post #1751 had explained rather well the Alternate/Direct Law relationship that gave rise to the engine protections adopted. The Eng 2 first mentioned in the following, seems to be a typo, and he was referring to Eng 3.

Thanks MM43, you are correct. Post 1751 has been corrected.

rob_ginger

The FG (Flight Guidance) calculates the TPR! The FG is contained in the PRIM flight control computers. With > 40 wiring loom faults (breaks) logged - get the picture?

BTW: If you lose the PRIMs (so the FG), then the TLA and Altitude is used (within the EECs) to calculate an N1 demand. (GenFam 7-5). QF 32 was in Alternate Law (from the PRIMs.)

The A380 PRIMs are remarkable beasts. I think their architecture is unequaled in any other flying aircraft. If you think this incident suggests otherwise, then consider that the event was a "boundary case" of PR 10^-9 to 10^-12.

[Old Engineer]

Turbine D wrote:

Old Engineer

I liked your analogy as to the spline wear being the "canary in the coal mine".

Here is why:

Spline coupling allows a shaft to be easily dismantled while providing a high-torque capacity for minimum size. In addition, it can allow relative axial and radial motions between the coupled shafts. ...
.
.
.
... Thus, misalignment makes the vibration of a system more complicated.

This information is from:

Research Article
Meshing Force of Misaligned Spline Coupling and the Influence on Rotor System, [by] Guang Zhao, Zhansheng Liu, and Feng Chen
School of Energy Science and Engineering, Harbin Institute of Technology, Harbin 150001, China

Could this be a viable explanation as to increased loading on the bearings and bearing structures at high thrust levels, e.g. takeoff?

Turbine D

Yes, I believe it could. Thanks for posting that very graphic description of the view of the Chinese authors in this spline matter.

I had some background written up on the reasons involute gear section splines are used (a known and calculable tooth line bearing stress) and why they rotate in mesh (the clearances for free assembly preclude simultaneous contact of all the teeth, particularly when the shafts are misaligned. But the reply page timed out; I recovered only the quote. Maybe just as well, as it shortens the post.

I think that spline binding, loss of oil film therefrom, hence wear, hence loss of involute tooth form, hence loss of uniform motion, hence chipping of teeth from wear and increasing impact in mesh, will eventally lead to significant non-uniform rotational speed of the turbine and compressor disks. If severe enough, the gyroscopic effects may pound dents or grooves into the bearing races. Initially, it may just be loss of oil film in the bearings, roughening the races, leading to wear and heating.

Just some thoughts to pursue. Causes of gear teeth and bearing failure can be very hard to understand as to specific cause.

OE

[Jando]

The Qantas crew seems to think the airplane handled the incident quite well, all things considered:

From The Australian: QF32 crew still fans of Airbus A380

I'm sure Qantas, RR and Airbus don't mind a bit of positive news either.