Concorde Paris crash, questions, facts, opinions

[EXWOK]

The yaw induced by the reduced thrust on 4 was negligible, you really didn't notice it in handling terms.

[robertbartsch]

I have not seen much press on the progress of the French trial. At the start early in 2010, it was supposed to last 4 months.

I'm I correct in assuming that the French judges and not a jury will conclude who, if any parties are responsible for the crash and loss of life?

What are the possible sanctions under French law; monetary damages, jail sentances?

Thx..

[ChristiaanJ]

Robert,
There has been some reporting in the French press, and some discussions on French aviation forums.

Generally I tend to make a wide detour around canine manure (which I consider this court case to be), but if enough people want to know more, I can have a look and try and do a summary.

CJ

[lomapaseo]

What are the possible sanctions under French law; monetary damages, jail sentances?

Press releases both during and after the trial will probably suffice for the publics memory bank.

[robert f jones]

CristiaanJ

Reference "canine manure", we may have something in common after all !
My last flight in BA uniform was, at my request, a sny ride on Concorde to JFK and return. I knew a number of Concorde crew plus a few UK AAIB people and therefore am indoctrinated in the British attitude towards the Air France accident.
Where it appears we may have a mutual agreement is the judicial interference whereby the AAIB were barred from certain parts of the accident investigation, contrary to Annex 13, accepted by every other nation.
It would be appreciated if you could make serious interested parties know, by this forum, the outcome of the current court case when it finally concludes. Current thinking amongst my former colleagues (BA and CAA) is that it will be a foregone conclusion based on nine year old evidence.
The French people involved in the trial deserve better consideration after so long in the Industry.

[ChristiaanJ]

Originally Posted by robert f jones

...It would be appreciated if you could make serious interested parties know, by this forum, the outcome of the current court case when it finally concludes. Current thinking amongst my former colleagues (BA and CAA) is that it will be a foregone conclusion based on nine year old evidence.

Robert,
IIRC, the date of the final "verdict" is either December 2nd or 3rd. And yes, no problem about reporting it here once it happens.
It'll probably make the UK news too at the time, but I'll do my best to follow the French news and translate anything useful.

The French people involved in the trial deserve better consideration after so long in the Industry.

I totally agree there.

CJ

[atakacs]

One further question here... what was involved for an engine fire alarm to go off ? My understanding is that initially none of the engines were actually on fire ?!

[ROKNA]

@ Misterredmist

Engine 4 was always restricted until past 80 knots

Due to the vortex rolling off the side of the delta wing it caused the air flow entering engine 4 to rotate in the opposite direction to the engine turbine stages. Beyond 80 knots it didn't cause a problem.

This obviously only effected one side of the aircraft as on the other side, engine 1 the air flow coming off the side of the wing was rotating in the same direction as the engine.

So normal procedure of which the all concorde crew would be familiar

[ChristiaanJ]

Originally Posted by atakacs

One further question here... what was involved for an engine fire alarm to go off ? My understanding is that initially none of the engines were actually on fire ?!

atakacs,
If you're seriously interested, get yourself a copy of the Concorde flying manual (available on the internet). Until then, I think it's somewhat pointless trying to second-guess the BEA on such basic matters.

CJ

[M2dude]

I have deliberately not posted in this thread (although the totally one-sided view of things technical from France has tempted me on several occasions; no offence CJ). I'll continue to resist posting any of my own opinions here (and those of virtually the entire British Concorde family) as long as I can NO MATTER HOW MUCH ONE-SIDED the technical retorts are seen to be. (Can't promise though, and anyway if anyone wants my views on any of this stuff posted in this thread, feel free to pm me).
OK, just a couple of technical issues then:

atakacs

One further question here... what was involved for an engine fire alarm to go off ? My understanding is that initially none of the engines were actually on fire ?!

No need to worry about 'second guessing the BEA' in all its renowned abject impartiality, I'll give you a technical answer here no problem. The Graviner FFFD fire detection system used on Concorde used two dual capacitive detection loops; if a loop went short circuit it would generate a fault and NOT a fire warning, and for a FIRE warning to be generated BOTH loops would have to detect the fire. But, there is a big BUT here; if BOTH loops detect a fault, then after some modifications to the system carried out in the 1980's, this was deemed to mean that both loops had burned through, and a fire warning was generated. We can only postulate that maybe this happened here; the fire wire elements surrounded the engines, being mounted to the titanium centre wall and top heat shield, and around the jet pipes, as well as to the engines themselves.
ROKNA

Engine 4 was always restricted until past 80 knots Due to the vortex rolling off the side of the delta wing it caused the air flow entering engine 4 to rotate in the opposite direction to the engine turbine stages. Beyond 80 knots it didn't cause a problem. This obviously only effected one side of the aircraft as on the other side, engine 1 the air flow coming off the side of the wing was rotating in the same direction as the engine.

Good points here Rokna, but the figure was 60 KTS, not 80 KTS, above which a solenoid latched switch on the F/E's station was automatically released by an ADC speed signal. The engine control restriction was to N1 alone (By restricting the primary nozzle area)and not N2. the fold-over effect diminished to an acceptable level at 60 KTS, it took a little more IAS until it disappeared altogether. Although there obviously was a restriction to take-off thrust this was minor, and as EXWOK previously stated here, barely noticable during take off.

Dude

[henra]

The final nail in the coffin was the Flight Engineer shutting down an engine without instruction from the Captain.

Are we absolutely sure that it was a commanded Shutdown of Engine #1 ?

Looking at the FDR Data between Timestamp 97659 and 97661 the SFC and EGT still slightly increase whereas EPR starts to decline.
At 97661 also then SFC drops off.
In a comanded shutdown, I would expect rather the opposite.
On the other hand rudder right is immediately albeít only little increased. So the Pilot must have been expecting something or maybe he just noticed the engine spooling down

I agree that shutting down Engine #1 was the final blow.
You can come to that conclusion also from the FDR data.
It takes until 97661 until the aircraft starts yawing left (Lateral accel left) and until 665 until rolling left and pitching up, increasing altitude and at the same time loosing speed.
You also see the control column right input significantly increasing after 97665.
So the Sequence appears to be: Shutdown (voluntary or not ?) of Engine #1, followed by yaw to the left, followed then by roll left and pitch up, losing speed at the same time.

Seeing the data, I'm not really convinced that the wing/TE really folded up as some reports suggested. Pitch and AoA increase rather slowly and continuously (not aprubt) and also largely in sync. It appears the wing still produced significant lift.
But looking at the graphs I got the feeling that the Control surfaces on the left wing might have been lost/ineffective relatively early in the flight.
From 97647 on Control Wheel position largely correlated between Nose down and aileron right.
Looking at the graph between 647 and 651 control wheel right input + Nose Down was applied, vet the pitch increased 3- 4° Nose up at the same time
This to me implies that a roll right might have caused nose up pitch so the Pilot tried to compensate for that once he firgured out.
One could conclude from that that the Left elevon was ineffective/lost at that point in time.

Re initial cause i.e. metal strip.
That one should be rather doable to sort out.
On the vertical acceleration you see a bump at 97586. That should be the disintgration of the tire.
You would have to calculate the corresponding position based on speed and elapsed time and compare it to the place where they found the strip.
For BEA that should be feasible. So at least they should have a clue.

[atakacs]

the fire wire elements surrounded the engines, being mounted to the titanium centre wall and top heat shield, and around the jet pipes, as well as to the engines themselves

Thanks for this technical explanation which is all I was asking for, sparing me the need to dig into the Concorde engineering manuals

So from the above (sorry I'm sure I could find this info elsewhere too) they where multiple fire wires and shorting of two of them would yield to the fire alarm ?

[M2dude]

Yes that is quite correct atakacs. Originally at entry into service Concorde was fitted with an ultra violet detection system, designed to detect combustion chamber breakouts etc., but this was found to be totally ineffective (useless in fact) in service. When this system was replaced in the early 1980's with an enhanced version of the original fire detection firewire system, the fire detection logic was changed from it's original philosophy: Originally a single loop detecting a fire gave no warning at all; this being changed to a FAULT (not a FIRE) warning, and dual loop low insulation resistance now resulting in a FIRE warning, instead of just a FAULT. The basic Graviner FFFD philosophy was retained, where both elements would have to register a FIRE condition before giving the alarm. The fire detection network was arranged as two loops per engine in two long series independent loops running the entire length and breadth of the engine bay. If there was a break in the loop, then although the test system would not work, both halves of that loop would still give you adequate protection. A dual SHORT (which is what had occurred in the 1982 engine fire in G-BOAF) tends to occur when the loop wiring is burned through; hence the revised philosophy.

Dude

[PBL]

It is interesting (I use the word loosely ) to read here the speculation about the Concorde crash, much of it introduced and indeed resolved within a few years of the crash. I guess if unfounded speculation doesn't succeed the first time around, one can always wait 6 or so years and then reintroduce it.........

It prompted me to read again the analysis by Bernd Sieker, in his Diplom thesis at http://www.rvs.uni-bielefeld.de/publ...lom/sieker.pdf (the second half; the first half describes a causal-analysis SW toolset which has been superseded by the SERAS suite).

Contributing to this analysis was an exchange of information and views with the Chief Aerodynamicist of Concorde at the time of service introduction.

PBL

[Feathers McGraw]

Having read this thread after reading many of the previous Concorde crash threads on PPRUNE, I do note that no one has mentioned that a likely reason for the eventual failure of the No 1 engine was that it ingested pieces of runway edge lights damaged as the undercarriage ran off the left edge of the runway just before the aircraft became airborne.

Now the engine may well have surged and run down due to increasing hot gas ingestion but the surge margins were likely to have been compromised by FOD from that light.

I can't remember exactly what was said by the BEA about this and I don't want to go and read the report again (at 3am!) right now.

If I'm wildly wrong here I'll shut up. It's a small detail anyway, as others have said once the fire started after V1 (or even before I suspect) the result was inevitable. All extra thrust from the engine could have done was to alter the impact point.

What I am interested in, but don't think a definitive answer will ever emerge about it, is why the tank wall section was blown out.

I understand the hydraulic shock, but with an air space above the fuel I can't see how the internal tank pressure could rise enough for this to happen. Under this condition the air would have compressed, absorbing the energy. Now I'm not saying that the tank wasn't ruptured by a fuel shock wave inside it, but I thought the normal operating procedures were required to ensure that there was an air space in all the tanks.

The only scenario I can think of that leads to the ejection of that piece of tank wall is that the shock wave travelled backwards down the tank and the fuel surged up excluding the air at the back of the tank and rupturing it with a local over-pressure rather than one in the whole tank.

Does that make sense? It does worry me that what happened could not be recreated even with the knowledge of the size of the rubber piece, its probable velocity and the probable fuel state of that tank.

Does anyone know how these tests were done? I don't imagine they were able to repeat this on an actual Concorde airframe.

[ChristiaanJ]

Feathers McGraw,

To save you wading through the entire report:

- § 1.16.9 deals with the detailed inspection of the engines,
- § 1.16.7.2 deals with tank 5.

Originally Posted by Feathers McGraw

...a likely reason for the eventual failure of the No 1 engine was that it ingested pieces of runway edge lights damaged as the undercarriage ran off the left edge of the runway just before the aircraft became airborne.

See § 1.16.9.
In my view, the FOD damage to engine 1 sounds far more consequential than would be caused by hitting a frangible runway light.Also I would expect it to be crushed, with debris being thrown sideways and to the rear, not forward.

What I am interested in, but don't think a definitive answer will ever emerge about it, is why the tank wall section was blown out.

See § 1.16.7.2.
I would say that what happened is clear enough: a fragment of the tyre hit the underside of the tank, and an adjacent piece of the tank wall about a foot square was "punched out".
Exactly how is less clear, and yes, a definite answer is unlikely to emerge. But the various 'scenarios' mentioned in the report make sense, even if the various trials did not provide a clear-cut answer.

I understand the hydraulic shock, but with an air space above the fuel I can't see how the internal tank pressure could rise enough for this to happen. Under this condition the air would have compressed, absorbing the energy. Now I'm not saying that the tank wasn't ruptured by a fuel shock wave inside it, but I thought the normal operating procedures were required to ensure that there was an air space in all the tanks.

Not really... refuelling would normally be cut off automatically with the tank 95% full (94% in this case according to the BEA report). But it was also allowable to carefully "top up" the tanks even further, and by using the 'override' switch, some more fuel could also be moved during the taxying.
So the remaining air space may well have been minute.

The only scenario I can think of that leads to the ejection of that piece of tank wall is that the shock wave travelled backwards down the tank and the fuel surged up excluding the air at the back of the tank and rupturing it with a local over-pressure rather than one in the whole tank.

Read the report... the shock might well have reached the adjacent panel even before it had a chance to compress any remaining air... and, with the aircraft accelerating, that air would have been a small pocket in the front of the tank.

Does that make sense? It does worry me that what happened could not be recreated even with the knowledge of the size of the rubber piece, its probable velocity and the probable fuel state of that tank.

The report admits they could only do a limited amount of tests, and none probably fully reproduced the exact conditions of the accident.
It doesn't really 'worry' me... as I said we have a pretty good idea what happened, and plausible answers to how it happened.

Does anyone know how these tests were done? I don't imagine they were able to repeat this on an actual Concorde airframe.

It's described in § 1.16.7.2. At least one of the tests did use actual Concorde panels (I can't remember from which airframe), but indeed, none were done on a complete Concorde.

CJ

[Chu Chu]

This is probably far fetched, but I wonder if the impact and displacement of fuel could have caused a trapped pocket of vapor-laden air to ignite by "dieseling"? I guess it would require a trapped volume of air small enough to be greatly compressed (and heated by that compression) without absorbing all the excess pressure, but yet big enough to cause damage when it ignited. And, of course, the vapor concentration would have to be between the upper and lower explosive limits.

It may well be that all those conditions couldn't have occurred in the tank. But conceivably it's a way to explain tank damage beyond that attributable the excess pressure alone. I didn't see the possibility discussed in the BEA report -- though I guess the most likely explanation for the omission is that it wasn't a possibility . . .

[Feathers McGraw]

Thanks for that response Christiaan, I see several things there that I had not considered, particularly that rearwards movement of the tank 5 contents leading to no air space at the rear of the tank.

I know that M2Dude has said in the past that he was sure he'd seen a photo that was not included in the report showing something about the position of the tank 5 override switch. But there is nothing certain about that.

It does seem that, with some topping up, the conditions I was thinking about were more likely to be present in tank 5.

I'll withdraw my "worried" feeling, I think I can see more clearly why that piece of tank wall was punched out now.

As for engine no 1, I'm not quite sure I would regard a runway light as being very frangible. I see what you're saying about the likely path of such debris, but I was sure that I'd read somewhere that the investigators thought that it was likely to have entered that engine.

[ChristiaanJ]

Originally Posted by Chu Chu

This is probably far fetched, but I wonder if the impact and displacement of fuel could have caused a trapped pocket of vapor-laden air to ignite by "dieseling"?...I didn't see the possibility discussed in the BEA report -- though I guess the most likely explanation for the omission is that it wasn't a possibility . . .

Chu Chu, have a look at § 1.12.1.9 in the report. There is a large "splat" of unburnt kerosene at the point where the traces on the runway start, and only after that the soot (from the partially burnt fuel) starts. So the leak came first, the ignition only afterwards.

I like your reasoning... but what's the pressure in a diesel engine at the moment of ignition? And once the vapor-laden air had 'dieseled', that would be it; it still wouldn't have ignited the remaining fuel (no air left). Not to mention that a fuel tank is not a diesel engine cylinder... the 'diesel explosion' would have blown out the top of the tank above the air pocket.

We already have a hole and a big fuel leak, with enough ignition sources available outside, so IMO there's no real need to look for a more far-fatched scenario.

CJ

[M2dude]

FeathersMcGraw

I know that M2Dude has said in the past that he was sure he'd seen a photo that was not included in the report showing something about the position of the tank 5 override switch. But there is nothing certain about that.

I did some digging a few months back. It was not a photo as I first thought, but the verbal testimony of a former colleague who was assisting at the crash site in Gonez. He told me that that the Tank 5 inlet valve over-ride switch WAS broken, but the stub appeared to be locked in the OVER-RIDE OPEN position. (It was a DOWTY lever-locked type of switch; once engaged in the detent only a firm outwards motion could release it).
Either way I don't think that there is any real doubt about the final hydro-dynamics of the section of wing lower skin being 'blown out', the tests and investigations here do seem to have been extremely thorough.
ChristiaanJ

In my view, the FOD damage to engine 1 sounds far more consequential than would be caused by hitting a frangible runway light.Also I would expect it to be crushed, with debris being thrown sideways and to the rear, not forward.

Point taken my friend, but the damage could still be attributed in my view, at least partially, with the ingress of the runway light and it's fittings; we just cannot discount this totally. Either way, we are faced with an engine producing very little thrust and it's neighbour (still producing some thrust) being shut down some 20 KTS below V2 and below 100' AGL. The lack of serious damage to engine 2 compressor seems to indicate that the surging and lack of thrust was due to fuel ingestion and could in all probability have recovered, but I suppose we'll never know for sure.
One final point that I would like to re-iterate (although academic in terms of what finally happened to the aircraft. The BEA report states

The two loops must detect the fault simultaneously to set off the ENGINE FIRE warning. This results in a red flashing warning light lighting up on the fire handle of the engine in question, accompanied by an aural alarm (chime), then by a gong and the illumination of the corresponding red ENGINE warning light on the Main Warning System.

The BEA certainly missed yet another point here, in that a simultaneous loop fault (NOT JUST A DETECTED FIRE) would ALSO generate a fire warning; this was a deliberate design feature of modifications carried out to the original GRAVINER FFFD system in the mid-80's.
As far as the rest of this tragic event goes, well I'd rather not comment any further here; this is for once a generally well informed (and civilised) thread.

Dude