Rejected Take-Off due Engine Fire

[All Spooled Up]

Remember the British Airtours B737-200 accident at Manchester in 1985? For those that don't, the aircraft had an uncontained engine failure during take-off which punctured a wing fuel tank access panel. The resulting fuel leak ignited, and a high speed RTO was initiated, the aircraft being turned off the runway onto a link taxiway at low speed. Unfortunately this resulted in the fuel fire being upwind of the fuselage, and being blown towards the cabin, with disastrous results. It should be pointed out that the wind at the time was only 7 knots.

The AAIB report (which is available at www.aaib.detr.gov.uk/formal/gbgjl/gbgjl.htm)suggests (http://www.aaib.detr.gov.uk/formal/gbgjl/gbgjl.htm)suggests) that "procedures should be developed to enable the crew to position an aircraft, when a ground fire emergency exists, with the fire downwind of the fuselage".

Does anybody have any experience of operating in accordance with this recommendation? Has anybody actually tried, in the sim for example, a high speed RTO with an engine fire, turning into wind at the last minute. A read of the above report is enough to convince anybody of the need to avoid spending any time at all with a fire upwind of the fuselage - but there are clear handling issues, particularly with RTO braking from a high speed abort on a short runway.

I've heard valid arguments in favour of stopping on runway heading regardless, or turning into wind at the last minute. Any further comments or advice would be much appreciated.

[fireflybob]

I was on the B737-200 both when this accident occurred and also when the accident report was published.

The company I was working for incorporated advice/recommedations in their Ops Manual that the take off briefing should give consideration to the prevailing wind and what heading the aircraft would be placed on in the event of an RTO for engine fire.

Obviously you need to do the RTO first but then, depending on time available etc., the aircraft should be turned accordingly.

That's the theory but when you are really on fire I suspect you are not going to hang around for long!

[411A]

At what speed was the RTO done? Many times it is better to take the aeroplane into the air, if the speed permits.
This will surely invite opinions from pilots of the Comet era and their slightly obtuse thinking.

[mustafagander]

411A,
Are you serious???
You have an engine fire.
You know about it.
You want to get into the air??? :confused:
As the wing structure rapidly softens I believe it's a choice you could seriously regret.
Personally, I'd prefer to go off the end of the r/w decelerating and being chased by the fire services.
Actually, my first choice is not to find myself in that situation.

[Zeroflap]

At KLMuk it is standard practice to attempt to turn such that a fire is downwind. This is no problem at all, taught right from the beginning and normally turning 30 degrees through the wind was considered sufficient.
The problem (confusion??) arose with a quartering tailwind when it was more expiditious to turn in what appeared to be the opposite sense.
The other thing to remember is that these were relatively small aircraft, ie FK100, BAe146, FK50, ATR72 and FK27. It might be a different story with a widebody. :p

[A and C]

Zeroflap it sounds like a good plan to me but i suspect that with a bigger aircraft you may not have the room on the hard surface.
I would think that the second best option would be to point the nose into wind.

[All Spooled Up]

411A, the uncontained engine failure occurred at 125kt, and the max IAS during the whole incident was 126kt. No engine fire warning occurred until passing 85kt during deceleration, owing to the fire emanating from the wing fuel tank access panel vs. the engine. Incidentally, the report states that the fuel access panel had an impact strength approx. one quarter that of the surrounding lower wing skin - and that if the ejected engine debris had struck the skin vs. the panel then penetration of the fuel tank would not have occurred. Very unfortunate.

Fireflybob, it sounds like you are in a similar situation to me in that my companies' ops manual had advice incorporated to the same effect - but it's not really practised e.g. in the sim. Zeroflap sounds like he's had a chance to actually give it a go which would seem like the way forward - given the intense workload / stress during an RTO from high speed, I think you really need this kind of response hardwired into your system. A high speed reject is not the time to be pondering the quartering tailwind case for the first time! I imagine in most companies (certainly my own) you can always ask to try it in the sim, but I personally feel it should be covered as an important and integral part of a conversion course. Who else actually includes this in their formal training syllabus?

There are parallels between this incident and the Air France Concorde crash, and one wonders whether the accepted go / no-go decision crieria are overly simplistic. Undoubtedly with a "straightforward" engine failure (if there is such a thing!), the concept of V1 would seem eminently sensible. But if the Air France Concorde had stopped, even if the RTO had been initiated above V1 (assuming they'd had appropriate cockpit information to indicate they had a serious problem, which I don't think they did) and they'd overrrun, surely that would be better than the tragedy which ensued? Yet the poor chaps in the front were acting entirely in accordance with accepted training. Similarly for the MAN 737 accident, at least some people escaped alive - had they continued the take-off, perhaps the wing spar would have failed with the heat, as mustafagander suggests. Certainly they had no means of extinguishing an airborne fuel tank fire other than airflow. One wonders whether there is a tendency for V1 to be considered as some kind of magic wand which guarantees safety - when all it really does is tell you if you'll stop within the ADSA or meet the appropriate performance criteria if you continue (including a mere 15' obstacle clearance from an obstacle-limited wet runway, incidentally!). Surely it could be acknowledged that there are circumstances where an overrun is the best of a bad set of options, again as mustafagander suggests?

[411A]

Thanks for the info, AllSpooledUp. Yes, indeed, in this instance, the only option was to stop.
Mustafagander--
Running off the end, at high speed (or any speed for that matter)is certainly not an option, at least for most. Most unfortunate that the fire was in the wing, not the engine. The question is, how is the crew to know? Podded engine designs offer a choice (speed permitting) for the crew to fly if the fire is in the engine and the engine is still producing usable thrust. Training is designed around just such a scenario.
Some will disagree of course.

[olivasnooze]

As has been mentioned, with a wide body the problem could well be going off the hard surface, resulting in fire/rescue vechiles bogging down etc.
The extra time turning into wind could be critical in an evacuation.
I believe that boeings reccomendation is if below V1 stop ASAP and evauate ASAP.

[Centaurus]

If I recall correctly, the abort was done relatively gently because the crew thought that the thump they felt was in fact, a tyre burst and not an engine problem. This, and the fact that the crew decided to clear the runway at the next taxyway before stopping to evacuate, took valuable seconds which sadly proved critical for cabin survival.

I believe that Boeing, (in commentating later about the accident) made the point that once a decision is made to abort following a serious problem, it should be made with maximum braking until the aircraft has stopped - and straight ahead. This is because the true seriousness of the problem may not be immediately apparent to the crew, and every second may be vital for evacuation.
Being very wise after the event, the Manchester accident could have resulted in less lives lost if this advice had been documented earlier. Like many manufacturer's, Boeing publish recommended procedures but rarely explain the rationale behind each procedure. For the more critical emergency procedures, I have always felt that a concise explanation would be of significant operational value.

[ 30 September 2001: Message edited by: Centaurus ]

[411A]

Many airlines choose to ignore the manufacturers recommendations and operating proceedures and instead develop their own, oftentimes derived from other equipment or at the Chief Pilots' misguided insistance.
The manufacturer designed, built, test flew, designed operating proceedures for, and sold their aircraft and expects it to be operated accordingly. But no, some dorks in airlines think that they "know better". It has been proven time and again that the most efficient and logical way to operate an
aircraft is to do so as designated by the builder. :rolleyes:

[Propellerhead]

411a, I agree that maybe some smaller companies may operate as you say. However, don't forget that when Boeing wrote their operations manual, they had NO experience of operating it in a commercial environment.
An airline may have 15+ years experience flying a particular type. In my company, the ops. manual and SOPs are still changing. It's called evolution. An airline will have the benefit of hundreds of thousands of hours of operations, including any accidents and lots of incidents they may have had. If, as a result of this experience, there are lessons to be learnt, then they will of course change their SOPs.

Interestingly, some of our SOPs changed recently as a result of a Boeing safety audit, a change that Boeing had brought in as a result of lessons learnt etc.

I think the V1 situation with an engine fire is quite simple. IF you can identify, decide, react, and INITIATE a stop at or below V1, then you should.

There does seem to be a lot of apathy regarding turning to get the wind and fire away form the fuselage. I don't see why it should be so difficult for a 737 / A320 etc. As you decelerate through the last couple of seconds it means turning the tiller left or right a bit. It's never going to be exact, but doesn't need to be. I think the report is quite clear that had the fire not been blown onto the fuselage many more would have escaped.

[Propellerhead]

The following is extracts from the report. Essential reading.
Check out the picture of the fire pattern: http://www.aaib.detr.gov.uk/formal/gbgjl/bgjl8f.jpg
http://www.aaib.detr.gov.uk/formal/gbgjl/bgjl8i.jpg http://www.aaib.detr.gov.uk/formal/gbgjl/bgjl8j.jpg http://www.aaib.detr.gov.uk/formal/gbgjl/bgjl8kl.jpg

Although the wind was only some 5-7 kt - a strength so slight that it would have been a relatively insignificant factor in terms of aircraft handling - there is a powerful body of evidence which clearly shows that the influence of the wind onthis accident was paramount. Not only did it drive the static fire plume against and beneath the hull, making a more rapid penetration of the aluminium alloy fuselage skins inevitable, it created an aerodynamic pressure field around the fuselage which, once doors and exits had been opened on the side opposite to the fire, induced the products of the external fire into and down the length of the cabin interior. In turn, some interior materials ignited leading to the development of a fire inside the cabin.


The effect of a crosswind blowing over a fuselage is principally to create a region of high aerodynamic pressure on the upwind side of the hull, and a low pressure region on the downwind side, relative to the ambient pressure (Appendix 8 Fig i). Consequently, once the fuselage is opened to the outside atmosphere, whether as a result of penetration by the fire or because of doors and escape hatches being opened, there will be flows set up through the cabin interior dependent upon the pressure differential between the various apertures in the hull. (It is of extreme importance to appreciate that the wind strength necessary for this pressure-field mechanism to operate in practice has been shown to be very low17 - as little as 1 or 2 kt is sufficient.(Appendix 8 Fig j) These flows are crucial, because they have the capability to draw fire and toxic combustion products

62 The wind was the principal factor controlling the fire's behaviour. It carried the external pooled fuel fire against and beneath the rear fuselage, giving rise to rapid fire penetration. Subsequently the wind induced aerodynamic pressure field around the fuselage drew fire products into the hull, through the cabin interior and out through open exists on the right side of the fuselage.


63 The initial fire penetration of the fuselage occurred within 20 seconds of the aircraft stopping, when the lower skin panels on the left side adjacent to the aft cargo hold were burnt through, followed shortly afterwards by penetration of the fibreglass accoustic insulation blanket. This gave the fire access to a cavity surrounding the cargo hold, from which it entered the aft cabin via floor-level air-conditioning grills located on each side of the aircraft.


64 It is estimated that within 1 minute of the aircraft stopping, the fire penetrated the cabin sidewalls just above floor level adjacent to seats 17A to 19A, giving the fire direct access to the cabin interior

4.1 Procedures should be developed to enable the crew to position an aircraft, when a ground fire emergency exists, with the fire downwind of the fuselage. Visual indicators of local wind direction located within the manoeuvre areas would be valuable aids to the implementation of such a procedure. (letter to CAA 14 March 1986)

[ 30 September 2001: Message edited by: Propellerhead ]

[ 30 September 2001: Message edited by: Propellerhead ]

[m&v]

In the airline I was formerly with for 30years.there was a definite reference to'turning' the aircraft,once slowed down,to face the wind to blow the smoke/fire from the fusalage.In the 'Sim' sessions it was very rarely remembered,but commented on in debrief..In both the Calgary/Manchester cases,where the fuel tank was ruptured by the engine debris,there was 'no'fire warnings in the cockpit..Man' it was Atc.YYC it was the F/A's..What did come out of the accidentswas-stay on the runway.In YYC the a/craft was on the taxiway and the trucks had difficulty positioning.
The question remains-who briefs(prior takeoff) for the final heading,on the runway,to be 'downwind' of any fire.Sim'exercises have shown quite a few guys 'run off' the runway(into the mud)precluding the 'trucks' from circling the fire!!! :eek:

[Checkboard]

During the sim sessions at Ansett, all crews turned the aircraft into wind at the end of the abort as a metter of course - it is a simple and easy thing to do as the aircraft is slowing to a stop.

This was instigated purley due to the Manchester accident. The original idea was to place the engine on the downwind side, however discussions withthe professional airport fire crews adjusted the prodecure to the "nose into wind" scenario.

Firecrews always like to tackle the fire from upwind. Here they can array their engines free of the smoke, get a clear view of the fire area, and their foam/water streams have up to twice the range flowing downwind as opposed to upwind. Placing the fire on the downwind side of the aircraft makes their job (and the survivor's chances) much, much more difficult.

The procedure was also incorporated into the cabin crew's training; they would expect the aircraft to be nose into wind, and they would direct evacuating passengers to run forward during an evacuation in this situation.

[Danish Pilot]

Interessant discusion, many valid points. But there is one issuse missing if you ask me. The 737-200 have underwing mounted engines, with no pylon. When that enigne failed (the P&W JT8) the engine was running at very high rpm, causing some very hot turbine pices to go straight through the engine houseing, and through the wing, like a bulit. That leaves a nice passage for fuel to go straight into the hot section, and ignite. To day, basicly every aircraft, that have underwing mounted engines, have the engines on pylons. No rotateing part of an engine (if an uncontained failure happens) will be able to pass through the fueltank, because the line from the last rotateing part, is in front of the fueltank, rather than straight below, like on the 737-200.

[ 01 October 2001: Message edited by: Danish Pilot ]

[tired]

In my mob, operating A340s, SOPs call for us to put nose into wind as much as possible at the end of an abort. It's practiced as such in the sim. too - generally seem to be able to get the nose 15 - 20 degrees nose away from the centreline before a danger of running off the runway.

[Max Angle]

Checkerboard,

That's a very interesting point and one I had not thought about. Our SOP is to try and put the engine downwind and a mention of any x-wind is included in the briefing. According to someone who did an RTO on one of our Bus's the thing stopped so fast that before he thought about turning the a/c had come to a halt. Perhaps I ought to arrange a visit to the fire station on my next airport standby!.

[Al Weaver]

Danish Pilot

Some of your points are valid. However the bigger issue/cause is the uncontained part ricochet off the runway and the fuel streaming onto the runway. with a failing engine near.

This issue is the same for pylon mounted engines. The issue is not there inflight.

[HOMER SIMPSONS LOVECHILD]

Just when I was about to give up ppruning due to bollox overload ,along comes Checkboard with that little gem and reminds me that we can learn new tricks as old dogs.Its so easy to get into the "lets go" mentality and I constantly try to run through the abort scenario pre t/o.I will now include this stuff in my mental drill along with the usual stop/handbrake on/fire bottles/get the flock out of there stuff.As to going with a fire.I'll take my chances on the ground thanks.

[ 05 October 2001: Message edited by: HOMER SIMPSONS LOVECHILD ]

[Ignition Override]

Mustafagander-411A's use of pilot judgement could help prevent some major problems.

I did not read each comment in detail, but with tail-mounted engines, which are always behind the wing and fuselage fuel tanks, a high-speed abort, i.e. within 10 knots of V1 on a short runway, might not be the best option. The fire bell and red lights can indicate only a very hot bleed air leak, and if the engine fire (i.e. fluids) is real, then it might still be producing normal engine power at V1 to quickly help get you into the air for a V2 minimum climbout. Once at Engine Failure Clean-up Altitude (at least 800' AGL or maybe higher) , the Engine Fire Checklist will be done. When the affected throttle is pulled to idle and IF the fire indications go out, the engine is allowed (on this old plane at this airline) to operate at idle power, at least in the air. There must be a reason for allowing the problem engine to operate, which means more generators (less chance of instrument failures?), hydraulic pumps and air cond packs available, without the very high workload of single-engine operation. The APU might have been inop per the MEL for days. In the last two weeks, we have flown two different planes, each with only one operative cockpit wing tank fuel gauge: this means one operative gauge among the two wing tanks needed, for each of those two flights.

If the actual aircraft weight is much higher than the "advertised" aircraft weight given to us in the cockpit, which happens to various crews more than we realize (how about 14,000# on a plane with a MTOW limit of 114,000?), I might rather be climbing out away from the ground than screeching at max braking and go off the end of the runway, whether on a dry runway or not. Also, at many US airports used in the winter only for takeoffs, if you ask the tower for the braking action, they will say "We don't know. Runway 21C (now 21R) etc is only for takeoff". !!

If this line of thought is always a bad idea, just say so, and why.

[ 07 October 2001: Message edited by: Ignition Override ]

[ 07 October 2001: Message edited by: Ignition Override ]