Simulator Training for Engine Failure just below V1.

[Centaurus]

During simulator training, an engine failure is occasionally introduced 5-10 knots below V1 and generally the expectation is the captain will reject the take off.

Regardless of the status of the handling pilot (captain or first officer), the first definate indication of an engine failure (flame-out) is an increasing swing off the centreline (wing mounted engines). While the first instinct is to correct the swing by rudder, the PF is usually looking ahead while keeping straight down the centre line during the take off run, especially in marginal RVR. In other words, few PF can split their scans equally between outside cues and the engine instruments all at the same time. The PF's accent is outside the cockpit.

In turn, this means it falls to the PM to call at the first indication of a flame-out. For certification flight tests concerned with establishing rejected take off technique, the engine failure is deemed to be recognised by the PNF (or PF if test pilot) at a mere five percent drop of N1.

In real life, the PM, although responsible for observing engine parameters during take off, probably spends 50% of a typical 45 second take off run also monitoring the centre line tracking by the PF. Thus an engine failure (instant drop in N1 of 5% or more)) could momentarily go unnoticed by him if he just happens to be looking outside at the time. After all, it is surely not expected that for the whole duration of the take off run the PM has his eyes glued to the N1 gauges and in so doing disregard the what is going on outside his windscreen - especially in low RVR and/or heavy rain.

In fact, the first indication to the PM of an engine failure, could be a startled oath from the PF as he instinctively counters the sudden swing (which could be caused by a tyre burst?) and by the time the event is realised it is too late for the PM to quickly glance at the engine instruments - identify the failure as a definate failure - and advise the PF because by the time that happens the PF may have pulled back the thrust levers IF he intended to reject, and leaving the PM with no idea which engine (if any) has failed.

Further confusion will certainly occur if the PF is the first officer who then has to hand over control to the captain who conducts the abort.

And why am I writing all this? It is because I have observed much angst when simulator check pilots "arrange" an engine failure so close to V1 that the captain's decision to reject or continue the take off run could be equally correct - yet his decisions could be the subject of criticism because in theory the event (engine failure before V1) requires a reject decision.

To minimise the confusion, bad feelings and inevitable differences of opinion on the reject/go decision, perhaps it is better that the engine failure should be initiated by the instructor at (say) 15 - 20 knots below V1, rather than just below V1. Your educated and thoroughly professional comments would be greatly appreciated.:ok:

[flyboyike]

Hmm...interesting topic. At our airline, we are basically to assume an engine failure once the thing starts to yaw. The profile calls for the PF to immediately call "Set Max Thrust" as soon as the yaw is detected. Thence the rest of the single-engine profile is flown. At 1000' AGL, we accelerate to Vfs and at V2+20, the PF will call "Flaps Up, What is the problem?" If there, in fact was no engine failure, we revert back to normal ops and run whatever checklist is there for the actual problem.

As for that tight window of V1-5 to V1, I would think that if the "abort" call and decision are made before the aircraft actually makes it to V1, then the abort procedure is executed normally, but if the airplane makes it to V1 after the abort call, I would say it's too late for the actual abort procedure.

The pre-V1 failure in our sim training were done at V1-10 or so, probably not V1-5 that I can remember.

I suppose this is an issue of coordination more than anything. It's incumbent on the PM to notice and address the problem if it happens below V1, however close.

[bflyer]

HI....
If it can happen in real life..then we as pilots should train for it....so it is tough call???..so what...we are paid and trained to make that one tough call
You say it is unfair... so what......better to have it happen to me in the sim and see what i did,right or wrong and discuss it, than for it to happen for real and i only get one chance at performing correctly
Maybe our lords and masters will start thinking about an improved abort procedure

[Amateur Turbines]

Well Centaurus, I believe you make an interesting point however, in flight testing the critical engine failure occurs at Vef which I have not been able to find as a quanified speed. That being said the engine failure only needs to be recognized at V1 then the desicion made within I think it is two seconds(could be wrong).

The point of this rant is that in the sim I think it is an excellent exercise of CRM to see if the engine has been diagnosed timely and also if the problem has been communicated properly. I promise you however that every time the excersise did not go so well after a quick brief on roles during the T/O and another shot things were much improved.

Also you can not assume that the engine will always fail in real life just at the perfect moment. :ok:

[Back Seat Driver]

Boeing released an interesting paper dealing with go/no go decisions circa late 90's. Some additional info from Boeing to be found here (http://www.boeing.com/commercial/aeromagazine/aero_11/takeoff_story.html)

[Centaurus]

Back Seat Driver. Thanks for the Boeing link. It was invaluable.

[Piltdown Man]

I think that I can see where you are coming from Centaurus but you are damning your own organisation's training more than anything else. If you are criticised for making a "wrong" call this close to V1 then it the muppet who does the criticism needs training. In real life, if you make the this sort of wrong call, the "slack" in the performance will allow for this error not to result in an unpleasant outcome. This would also be demonstrable (?) in the sim. A writer who I think most of us respect often used the phrase "capable of being safely flown by the average pilot in the worst of circumstances" and this is one of the reasons. Those trainers who think that pilots should have the reactions of a cornered rattlesnake really are doing the wrong job. It's far better to make the correct call slowly than the wrong one quickly.

PM

[alf5071h]

Amateur Turbines has identified a source of the problem.
“Vef is the calibrated airspeed at which the critical engine is assumed to fail”. Certification Specifications For Large Aeroplanes (www.easa.eu.int/doc/Agency_Mesures/Certification_Spec/easa_cs25_amendment_1_12122005.pdf) part 25.107 (page 10).
“Vef must not be less than Vmcg”.
“V1 may not be less than Vef plus the speed gained with the critical engine inoperative during the time interval between the instant at which the critical engine is failed, and the instant at which the pilot recognises and reacts to the engine failure, as indicated by the pilot’s initiation of the first action (e.g. applying brakes, reducing thrust, deploying speed brakes) to stop the aeroplane during accelerate-stop tests”.

Another item from 25.109 is – “Allow the aeroplane to accelerate from Vef to the highest speed reached during the rejected take-off, assuming the critical engine fails at Vef and the pilot takes the first action to reject the take-off at the V1 for take-off from a dry runway… except that the runway is wet and the corresponding wet runway values of Vef and V1 are used”.

Vef is not a specific speed. Similarly the time delays are not specific, these are explained in AC 25-7A Flight test guide for certification of transport category airplanes (www.airweb.faa.gov/Regulatory_and_Guidance_Library/rgAdvisoryCircular.nsf/MainFrame?OpenFrameset) see Chap 2 (page 44). 2 sec is in the ball park. See the diagram on page 47 and the references about all-engine accel-stop distances; RTOs are not just for engine failure.

Perhaps the simulator training problem stems from the lack of knowledge of the certification process.
It would be better to refocus the training discussion on understanding the situation. This was addressed in the question. What are the cues of an engine failure in the chosen conditions – sight, sound, feel. Also, do not forget that very few simulations actually simulate side force and thus the important body sense cue may be absent or inadequately represented.

An RTO exercise is more about situation awareness than decision making or procedures/actions – if you understand the situation then the decision is easy. The technical procedures can be practiced independent of an engine failure; these are important, but not as critical as choosing the correct course of action – stop/go relative to the situation +/- V1, both driven by your understanding of the situation.

There are many more valuable simulator exercises than a ‘V1 cut’ and they all focus on awareness – the detection of cues that may determine a stop / go decision. Consider an engine surge (a series of 100 variants), tire failure (similar), and vibration alone or in combination with the previous. These can create far more interesting and valuable training scenarios than the standard ‘V1 cut’.

[Mad (Flt) Scientist]

Please be very careful when introducing the concept of Vef - which is purely associated with PERFORMANCE - into any discussion regarding HANDLING.

The delay times used for the Vef-V1 delay are not in any way representative of the assumed delay in reacting to the failure in the sense of handling. The delay is intentionally conservative.

[AirRabbit]

Hey alf5071h:
Thanks for the input. It’s nice to see that someone actually knows about the existence of some of the material available – the Advisory Circular you referenced being of particular importance when discussing issues like taking off or not taking off.

The material referenced by BackSeatDriver was largely (although not completely) developed during an FAA/Industry international working group effort, affectionately referred to as the “Propulsion Systems Malfunctions and Inappropriate Crew Responses International Working Group” (phew – what a title!), completed several years ago, that looked at accidents and incidents that occurred during takeoff when the flight crew might have made the wrong decision – or the right decision for the wrong reasons. The statistics developed during that study were interesting and some of the interviews and personal “war stories” were nothing less than eye-opening! Thanks for providing the link BSD!

I share Centaurus’ concerns about training and how much attention is given to tasks that are or are not going to be “on the check ride;” particularly as it pertains to the background behind those tasks. I think it would be good practice to have each pilot read through the kind of material referenced by alf5071h and BackSeatDriver at least once during his/her initial jet training and then probably once every 4 or 5 years thereafter, just to make sure that he/she keeps that information in the “operable zone” of the brain – AND to make sure they stay abreast of anything new or different.

I am continually surprised to learn what pilots know (or actually don’t know) about the real meaning of the numbers they use on a daily basis – particularly when the weather gets dicy and people start talking about “adjusting V1 speeds.” I think it also important to review just what it is expected that the PM actually be monitoring during those more intense (critical?) portions of the flight. And, while I’m at it, I also think it would be beneficial if, as an industry, we took another look at the pros and cons regarding changing control of the airplane at what is certainly one of the most critical times during the entire flight. By that, I mean if, for some reason, it is decided that the F/O (or the person in the right seat) cannot, or should not, make the decision and execute the rejected takeoff, perhaps we should start restricting takeoffs (because of the abort potentials) and landings (because after touchdown, the remainder is very much like an aborted takeoff) to the Captain (or the person in the left seat). I think we need to make sure that the tasks we authorize to be accomplished by the person at the controls (the PF) can be accomplished (including all of the logical deviations of those tasks) by that person, and the other person (the PM) knows what he or she is expected to monitor!

[411A]

<<The profile calls for the PF to immediately call "Set Max Thrust" as soon as the yaw is detected.>>

A bad idea, in my opinion, as it could certainly result, especially with a slippery runway, ending up in the weeds.
Derated takeoffs are predicated on the derated thrust set, and as such, meet all performance criteria.
It is poor practise to select MAX thrust at an inopportune time, especially just close to rotation.

[alf5071h]

MFS thanks for the reminder, but the critical training issue here is the identification of a problem that may require a RTO. Aircraft handling should be trained at the skill level, where a near automatic response follows the decision, be it Stop or Go. Thereafter if the procedures are followed the performance should not be an issue. Did you act as a ‘stoneman’ during engine cut trials?

411A, I agree with your concerns; do not interfere with the established situation or introduce any opportunity for confusion. More power may result in more yaw, roll, trim change, etc; anyone of which produces a new situation that requires assessment, i.e. getting the feel of the new aircraft configuration.

AirRabbit, thanks. The Boeing information is more likely based on the Runway Takeoff training Aid. ( www.airweb.faa.gov/Regulatory_and_Guidance_Library/rgAdvisoryCircular.nsf/0/fbffc4ad745e844b862569ea006960d0?OpenDocument)
The PSM+ICR report did present some interesting data; the report can be read on-line via the Flight Safety Foundation Digest Nov-Dec 1999 ( www.flightsafety.org/pubs/fsd_1999.html). The event types listed in Table 3 (Page 10) provide a range of realistic scenarios for simulator training. These events were either misdiagnosed and/or the choice of stop/go was incorrect – possibly a further lack of situation awareness (speed above V1), or a slow analysis process; all aspecs which can be practiced in the simulator.
The issue of negative training relating to V1 cuts is briefly discussed on page 24.

The complete PSM+ICR package, including follow-up reports is here, (www.faa.gov/aircraft/air_cert/design_approvals/engine_prop/engine_sp_topics/)
and
a reference for the 737 here. (www.b737.org.uk/enginemalfunctions.htm)
Another source including a link to the movie! (http://fromtheflightdeck.com/Stories/turbofan/)

[Linton Chilcott]

V1 is the speed by which the decision to reject should have been made.

This is not quite the same as: stop if an engine fails before V1, continue if an engine fails at or after V1.

If you haven't started saying "Stop, stop, stop..." before V1, then you should continue (in general, but follow company SOPs of course).

This does assume a fairly short decision making process. Normally a couple of seconds.

[FlapsOne]

UK CAA Pink 24/2006 - the Authority viewpoint!

............................

2 Decision Speed - (V1)

2.1 The decision speed (V1) is the highest speed by which the pilot should have made the mental decision to reject the take-off in the event of a relevant failure. Safety margins, in the form of delay times or extra distance on either side of V1, are built into the accelerated-stop distances published in performance data issued in compliance with JAR-OPS 1 and the Air Navigation (General) Regulations.

These margins address two considerations.

(a) A Recognition Delay Prior to V1. A pilot can only be expected to have made a rejected decision by V1 in response to a failure that has occurred a suitable time before V1 for him to have recognised it. The accelerated distance presumes an engine failure to occur at a speed, VEF, before V1 and the distance allows for the lower acceleration between that point and V1.

(b) Reaction Delays After V1. Allowance is made for taking the necessary actions to initiate the deceleration. In terms of time such delays are typically 1½ to 2 seconds. It is important to understand that this is not a margin to allow for a decision to stop beyond the V1 call out. It is a safety margin to allow for the delays that inevitably occur in taking the various control
actions (closing throttles, applying brakes, selecting ground spoilers, selecting reverse) even after the decision to stop has been made. This margin is in addition to the control activation times demonstrated for certification by a test crew responding to a premeditated test condition. However, the margin is no greater than might reasonably be anticipated for a well trained
and alert crew, taking into account the rare and unexpected nature of the event in operational service.

2.1.1 Thus, considerations of pilot recognition of a failure and reaction in response to that recognition are allowed for. However, V1 must be respected as the latest speed by which a decision to stop must have been taken to assure a stopping capability within the distance available on a limiting runway.

[callout]

Didn't have that problem when F/E's were on board. He or she made the call.

[togs737]

I have some input, I'm still in high-school, fly a Piper Cherokee, hope to be an airline pilot, yet do not know if this will have any validity what so ever!

Personally I think that engine failure on take-off a few knots below V1 should be practiced, I mean; don't airlines have to train for every scenario?
However, in saying that, by the time the PNF (or PF, or any other crew) notices an engine failure a couple of knots below V1, how quick can they make the decision? Surely you can gain those 3 odd knots in a second or so, beat the V1 speed, and fortunately/unfortunately (must) become airborne.
So...is there a point in practicing the 'few knots below' scenario?

As I said above hopefully there was something decent in my input. :sad:
Hopefully I can post as an airline pilot one day!!!

Kind regards,
Togs

[Miserlou]

Our SOP says high speed RTOs should be avoided. Effectively, this means you only stop if you suspect the the aircraft is structurely not airworthy.

An engine failure is a serious degradation of redundancy but the aircraft is still airworthy and, especially on a limited runway, it is safer to continue.

If you have the runway/obstacle clearance to spare there is no need to rotate at Vr. Accelerating to closer to V2 before rotating will also improve your situation; the lack of induced drag means you accelerate quicker and controlability is increased.

But it depends on the situation.

[keithl]

I would make two points.
One. that as a sim instructor I agree with Piltdown Man. When the decision is that close, the pros and cons should be discussed and analysed, not turned into right or wrong.
Two. That engine failures within 5kts of V1 do happen. One happened to me at max t/o weight for the type. The F/E called it (thanks, Thug!) and the sim training kicked in. We stopped safely.

[FlapsOne]

As an instructor I agree wholeheartedly with point one above by kethl.

Hope I never experience his point two! (except in the sim)

[mustafagander]

Miserlou, is it REALLY a good time to play test pilot when you just had an engine quit in the high speed (after V1) portion of the T/O roll? You MUST fly exactly to the numbers to guarantee your performance. Exceeding Vr is fraught with danger.

As an aside, every knot above V1 before the engine fails is a bonus for performance, but you cannot mess with the calculated speeds, and don't neglect to rotate a bit slower to be certain that you make V2.