View Full Version : Uncontrollable Engine Fires

24th Sep 2007, 08:45
G'day all,
Had a question, seeing as though I've got command training on the 767 (which I've never flown) coming up. What is the best strategy for managing an uncontrollable engine fire after take off? First, let me set the scenario.

Take off with 20 kts HW (ie return on reciprical runway not available)
ILS available
Wx is right on Landing minima
Radar is unavailable
Uncontrollable engine fire during rotateThe 2 options I've been thinking about are:

1. Turn back to the field and when within the circling area, descend to the circling minima. Remain within the circling area and basically carry out a circling approach on 1 engine.
Pros : gets you back on the ground the quickest
Cons : Not sure if the A/P will handle such a late intercept on 1 eng.
Late configuration change

2. Fly the full procedural ILS
Cons : Takes a lot longer to get on the ground :uhoh:

What do you guys think the best way to go is?

24th Sep 2007, 09:13
Off the top of my head..... which is how it would be in the real world.

1. Declare MAYDAY (obviously) - lets ATC know to get everyone else out of the way.

2. Are you familiar with field / local obstacles. Would there be a problem with just turning downwind and climbing to 2000'agl? (Other than potential traffic conflicts). With no radar available there are possible problems with following an EOSID and then trying to get yourself back into a position to make an approach - of course if you HAVE to follow an EOSID due to terrain you have no choice.

3. Use your nav display / navaids to position yourself for a base leg and fly the ILS. It has to be done urgently BUT you don't want to rush it so you end up fast/high etc.

4. Whilst all this is going on, a quick PA to the effect that we are making an immediate return to land and a quick ding dong to the No1 to get them to make a quick sweep for secure cabin.

5. With WX at minima and a burning engine/wing .... are you going to go around at minima? Could be argued that if you do it may be your last act of aviating....if you don't - likewise :\

OK. That took me more time to think about and type than you would probably have in the real event........ a fire is an unknown quantity and an unextinguishable fire is a nightmare...you just don't know what's going on out on that wing.

Bottom line, you have to get it on the ground ASAP but don't rush to the extent that you put yourself in a position from which a landing is impossible.
Remember, a Commander may deviate from standard practice if in his opinion it is necessary to protect the immediate safety of the aircraft and passengers. Perhaps this scenario is one of the few where it really applies?

Good Luck.


reduce to minimum
24th Sep 2007, 09:16
By the time you have followed your EFCOP to MSA or minimum radar vectoring altitude - depending on surrounding terrain - quite a few minutes will have passed anyway, especially at MTOM.
Also: what circling criteria are you talking abaout?
PANSOPS: 5,28km (Cat D), or TERPS (around 2,3km I believe).

If ATC is capable (read: trustworthy, depending on the part of the world you are in.....), I would always go for a tight radar vector pattern: gives you time to warn the cabin, get the A/C properly set up (F20, autobrakes, EGPWS inhibits etc.) and you will have a fairly stabilized approach. Better spending 30 extra seconds to get things under control than to screw up the approach and have to go around....
If ATC is no good, do your own short line-up according to terrain and keep ATC informed.

BTW: I found SE circling at or near MTOM on the 767 the only maneuver that can indeed be a bit tricky........never ever get below target speed!!

Just my 2 Euro-Cents....

Good luck!

24th Sep 2007, 11:22
Thanks for the replies :). Some very valid points raised which will make life easier when I have to fly this in the sim or, God forbid, in real life.

Just to try and be a little more specific and to answer some your queries, lets just say we're departing from 27L EGLL. Just bear in mind that the radar is out, so radar vectors aren't an option.

On the chart I have, the circling minima is 1080', based on PANSOPS 5.28 nm.
The full ILS procedure begins at 2500' and 7.5 DME LON, which is 9.6nm from the 27L threshold.
The applicable MSA is 2300'
There's no Eng Out procedure to follow.

The way I'm imagining things would go is :

Eng fire at rotate.
By the time you've accomplished the recall actions, fired both fire bottles and realised that the fire IS uncontrollable, I'm guessing you will be at 2000'-3000' and, depending on the SID you're following, ouside the circling area.

What do you do?

Do you maintain 2500' and track for the full ILS? ; or

Turn to be within the circling area, descend to 1080' and then circle for a 2-3nm final and intercept the ILS?

The second option is obviously the quickest and sounds fine on paper. But the real question is: on the 767 will the A/P in APP mode handle such a close in intercept of the ILS on 1 engine? The only 2 heavy types I've flown are the 747 Classic and the 744. The classic definitely would NOT cope with such a manoeuvre, whereas I think the 744 probably would, with a little bit of nursing around the corner in Hdg Sel. I've absolutely no idea what the 767 is like:eek:

There is always the option of disconnecting the A/P if it's not performing, but under the circumstances not my idea of a fun day out:rolleyes:


24th Sep 2007, 11:23
The scenerio you describe sounds like something you would see in a sim. The odds of all those conditions coming up in real life are remote to the extreme. Of course that doesn't mean it can't happen. In the sim you are trying to get though the scenerio, so as long as you don't waste too much time and get the airplane on the ground in one piece there should be no problem. In real life you don't know what is happening out on that wing so you should get the plane on the ground in the absolute minimum time. If you have time to talk to the F/A and pax you are taking too much time. Just tune the ILS and get to it quick. Do not use minimums, the autopilot will find the runway. Max braking to a stop and evacuate.


25th Sep 2007, 16:27
ILS - Chart

Whilst the ILS Chart says those figures, ATC can do better things. Dunno about LHR but the best I've had is descent to 1600' with a 5nm intercept at Glasgow. Certainly not what the chart had in mind...Tell ATC, if they can, I'm sure they will help!

25th Sep 2007, 16:43
and another little thing is your burning engine, in the first phase of the take off will be giving you 100% thrust (maybe even a bit more ...) so you'll be clearing a good deal of obstacles very fast before shut down. you'll much more than your single engine climb gradient.

so msa or radar safety altitude will be reached much earlier. and clearly as someone said before it's better to take a little more time and make it in one approach than going around or overshooting your loc or g/s.


26th Sep 2007, 02:11
As the B767 has wing mounted podded engines, the best scenario it would seem to me would be for the burning engine to separate from the wing.
If this were a B707 (instead of a 767) the normal drill would be to accelerate to the barber pole and hope that the fire at least is reduced (perhaps extinguished)...this has been tried in the past with this particular aeroplane, and it has worked.
Recall the BOAC B707 that departed LHR long ago, the engine separated, however, the tank valve was not closed, therefore the wing continued to burn.
Follow the fire drill exactly for best results.

26th Sep 2007, 17:21
I believe that Cathay had a similar incident a few years back on a 747-200 ( gear-box caught fire and the two shots failed to extinguish it). As you point out, by the time you have completed the fire drill correctly and determined that the fire has not extinguished, you will be almost through the flap retraction and may be at approx. 2000 feet. Descending to circling minima will not help if the weather is at landing minimums!.....rather having declared a Mayday and quickly sorted through the appropriate checks and c/c brief + short PA, aim to couple up with the ILS and carry out an auto-approach to a manual landing ( in accordance with your company's SOP's, of course). By the way, the 747 classic would cope with a quick return, as long as you carefully monitor the automatics and be prepared to go manual to capture the localiser/glideslope. For the 744, no problem -manual landing with Flaps 25 and the Vref close to the departure V2.
Good luck - don't forget the aim of the game is a safe conclusion!

26th Sep 2007, 22:32
Thanks for the replies everyone.
Skiesfull, now we're getting to the crux of the original question:ok: Firstly, I realise the example I've given at EGLL with no radar is a little unrealisitic, but there is at least one port we operate from with no radar, and a few others where ATC is a little ..cough.. unreliable:) and I would be very hesitant asking for vectors for the ILS with Wx at minima.

The idea of descending to circling minima, with the radar out and Wx at minima, is to then remain within the circling area (5.28nm), thereby maintaining obstacle clearance, and circle for a short final (2-3nm) and couple up to the ILS.

The second option is to maintain MSA and carry out the full ILS procedure.

FWIW, the TSB of Canada conducted a study into inflight fire/smoke events and found that the average time from first detection of a fire until the aircraft landed/ditched/crashed was 17 mins.

I had a bit of a play around with both the above options on MSFS (sad, but this has been bugging me) and the first option (2-3nm final) takes about 7-8 mins to get back on the ground. The second option (full ILS) takes about 14-15 mins, which when you compare to the average 17 mins quoted above, could be pushing your luck.

Obviously the first option is the best as far as time is concerned. However, what I am unsure about is if the A/P on the 767 will cope with such a close in intercept of the ILS on 1 engine, or would you be setting yourself up for an unstable approach?


27th Sep 2007, 01:30
How about climb straight ahead to MSA, carry out a 45-degree procedure turn and land on the reciprocal runway via the ILS for that runway? You indicated that a return on the reciprocal was not an option because of the 20-knot tailwind, but is that necessarily so? Sure it's above the limit, but with an uncontrollable fire that might be the safest option, assuming of course that there's enough runway length available.

The CX incident mentioned earlier involved a fully-loaded 744 departing off 13 at Kai-Tak with an engine fire at V2 or thereabouts. They immediately turned around and landed on the reciprocal runway with the No.2 (I think) engine on fire. I think they were airborne for approximately 12 minutes. Admittedly they didn't have to cope with a 20-knot tailwind, but they were damn close to MTOW when they landed and the Kai Tak runway wasn't overly long.

Just some food for thought - I guess the answer is to know your aircraft and what it can do in those types of situations.

27th Sep 2007, 02:06
Good point, BuzzBox.

I had considered that option also, and reckon it would shave about another 1-2 mins off the time to return. Even though the 20 kts TW is outside limits, I estimate you would need about 3000m of runway to land a 767, which is less than the majority of airports we operate from.

I was tending to shy away from this option, in the given circumstances, because:

1. At heavy weights, I'm not sure that brake energy limits wouldn't be exceeded.
2. You may be compounding the problem with a brake fire.

Happy to be convinced otherwise:)

27th Sep 2007, 02:10
Usual rules apply. Aviate, Navigate, Communicate, and then Manage. In this instance manage requires time to have high priority.
Management can be aided by knowledge; so find out what the capabilities of the autopilot are – I doubt that a short turn in will provide the reassuring success that you need in this instance. You have one approach, no GA (self set time limit), so get it right – expedite but don’t rush.
Knowledge of certification can be reassuring; some engine fires may not be extinguished, but they should be contained for quite some time. The structural and system design provides a margin – I won’t quote a limit, but note the Nimrod (Comet) that successfully ditched with an engine fire inside the wing after quite a long flight time. Also IIRC ’hloics quote (Canadian report) was for cabin related fires, so beware any preconceptions or bias.
The probability is that a standard emergency return to the airport (as briefed) will provide a higher success rate than attempting some unpractised, non standard, poorly planned alternative.

28th Sep 2007, 22:48
there is a danger in finding a shortcut and having the shortcut kill you instead of the fire.
fly the plane

follow the checklist

advise atc ,get their input BUT you make the decision.

advise f/a's/pax

and land.

now, if you are in a concorde situation in which the plane's wing is actually burning off...land anywhere and evacuate.

chances are this is more to see if you will CRACK and screw up something else...and yes the engine will probably fall off eventually...

29th Sep 2007, 01:40
The probability is that a standard emergency return to the airport (as briefed) will provide a higher success rate than attempting some unpractised, non standard, poorly planned alternative.

I agree, no need to panic. As someone pointed out earlier use V-2 for approach speed and perform the checklists as 411A pointed out. If the blasted engine falls off, it falls off. Hopefully not injuring anyone on the ground.

Jumbo Driver
29th Sep 2007, 09:34
This (http://aviation-safety.net/database/record.php?id=19680408-0) may provide some food for thought ...


29th Sep 2007, 11:23

If I could just add a couple of points to the already excellent advice given by others:

As you quote EGLL, you can still use the radar minmum charts, even if no radar service available, to get safely to a lower altitude as long as you are absolutely sure of your position. Its also feasible to utilise non-precision approach data (if available) to enable earlier descents i.e with 5 deg of approach course before intrecepting the localiser.

You quote 3,000 metres landing distance required for the B767 with 20kts tailwind. Are you sure you have that right? Could that figure possibly be the scheduled (planning required) distance as opposed to the unfactored advisory distance?

I don't have 767 data but B757 figures at max TOM (113T) with 20kt TW for landing give 4,745 FEET dry rw (6,525 Ft wet rw) unfactored and 8,700 FEET scheduled(wet by definition). I know the weights are not comparable and so would be interested to see 767 figures.

29th Sep 2007, 14:12
Hmmm, I note that many of the responders from the UK still apparently have the 'Comet syndrome', as regards turbojet engine fires.
An engine fire, altho serious, is much less do in a wing mounted podded engine, provided the correct engine fire drills are carried out within a reasonable time.
To go steaming around at low(er) altitudes, trying to get back on the ground, pronto, is many times less safe than actually using common sense, provided however that any remains on the FD in the first place.
I recall not long ago, with engine fires just after takeoff, where some UK pilots were in the habit of insisting that the fire drill be carried out at the lowest possible altitude, never mind the fact that in many cases (excepting severe apparent damage or separation) the engine may well still be producing useable thrust.
The FAA attitude (generally) is to leave the engine operating until clearly a safe altitude is reached, not start moving throttles/fuel levers/switches just off the ground...and certainly not rushing that landing, with poorly thought out off the cuff procedures.

And, before some wise guy says...'this would never be allowed with the UKCAA', they would be absolutely wrong, as I have done a few 1179 sim rides with two CAA inspectors, and they both have accepted my ideas, simply because....I have lots of command hours in the specific type.

To 'rush' an approach simply because an engine is on fire, especially in IMC conditions, is to likely result in bypassing the hospital, and proceeding directly to the cemetary.

Take your (reasonable) time and do it....right.

30th Sep 2007, 08:59
Whilst I absolutely agree with those comments guarding against "rushing around at low altitudes uneccessarily" there maybe times when it is actually necessary and so is worth thinking about from time to time on how this could be achieved safely.

Just change the uncontrollable engine fire scenario to one of dense uncontrollable smoke in the F/D or cabin and the required justification may be there and even a 20kt tail wind may be worth considering.

Also worth noting, from the FSF bulletin linked above:

In the case of -WE (BOAC 707 LHR) the No. 2 engine, together with part of its pylon, became detached about 1.5 minutes after the start of the fire BUT this was attributed to the fuel valve not being closed as required by the drill (QRH or recall equivalent). So again the comments about ensuring drills are correctly completed in addition to expediting are spot on.

30th Sep 2007, 11:04
It seems to me that since the engines are podded below the wings on most modern airliners, and they are there to keep them separate from the wing structure in a fire, wouldn't it make sense for there to be a mechanism (explosive bolts or whatever) that would actually jettison an engine pod in an extreme case like this one? Has this been done?

3rd Oct 2007, 23:47
G'day all,

Starburst, you are absolutely correct in that the landing distance of 3000m I quoted was factored. You are also correct that the main thrust of my original question was more to do with returning to land in a time critical situation when it is imperative that you get the aircraft on the ground ASAP.

Having said that, there's been some excellent information posted on engine fires. At the risk of starting a thread drift to my own question, I was wondering how long you could leave an engine fire without carrying out the recall actions before the fire spread and/or a catastrophic structural failure.

To try and put this in a practical context, say you are conducting a Low Vis approach and have a fire warning. Obviously, you would declare a Mayday, continue the approach and autoland. If the fire warning occurred at 2000' you would have enough time to carry out the recall actions. If it occurred at 50', you wouldn't. What height would you use as a cutoff between actioning recall items or not actioning them?

4th Oct 2007, 10:19
hello again 'holic

before I offer another tuppence worth, I would just caution against trying to have specific answers to all possible problems because that's just not possible. I don't think that is what you are doing here and are justt exploring some options but be careful nonetheless.

That said, as far as a cut-off height goes for recall items, many companies stipulate no actions below 500 ft AGL except for cancelling noises; raising gear; applying full thrust (if required) and commencing emergency turn if required. So in other words no drills (recall or otherwise) below this height, not even analysing the fault.

Most of that certainly implies a take-off scenario but in terms of drills can easily and often is applied to the approach situation as well BUT it will depend upon each situation because the most ipmortant consideration always is "Fly the aircraft". Easily said but often easily forgotten and especially by the PNF! The PNF's job in these situations is crucial and they must monitor PF's flying and ensure a/c is correctly configured on the correct flight path before even considering any other actions including ATC. Only when you consider it safe to action recall items without compromising flight path safety should you request them. Even 1,000 ft AGL (for example) could be considered too low in some circumstances whilst just above 500 ft may be acceptable in others. The danger comes from a breakdown in proper monitoring of those recall items and you just cannot afford to get it wrong close to the ground. You could consider briefing your colleague to action recall items once on ground at a safe speed BUT even that needs care. Whatever you decide, a Mayday call even on short finals will get help speeding towards you should it be required.

Also worth noting than the vast majority of Engine Fire Warnings are generated by hot gas leaks and often disappear when thrust is reduced towards idle and that could be considered without shutting down the engine at low heights and avoids reconfiguring a/c but will also inscrease workload somewhat with asymmetry and speed control. (much less of a problem cat2/3, just add some thrust)

Sorry, but that all sound a bit like preaching on re-read but I do believe its important and worthy of discussion.

4th Oct 2007, 16:11
With engine fires especially, it would seem that the 'Comet syndrome' is alive and well with some pilots.
'OMG, we have to do something now'...when in actual fact, engine fires, although sometimes serious, often to not require immediate action...such as on short final at low altitudes, or indeed at very low altitudes after takeoff.

Haste makes waste, is many times true.