Crosswind take-off. Upwind engine failure.

[Blip]

Can someone describe to me what would happen if during the take-off of a B737NG at maximum certified crosswind, the upwind engine failed just after reaching V1. That is to say a crosswind from the right and the right engine fails?

On a 45 metre runway, how far off centreline can we expect the aircraft to deviate even with maximum rudder input, or can we expect to happily track the centreline?

On the other hand think about what happens when the downwind side engine fails, ie crosswind from the right, left engine fails. Before the failure you would have a heap of right rudder, and possibly left control wheel ( the subject of another thread!). And when airbourne and climbing away, you would have a heap of left rudder, and control wheel basically centred. There would not be nearly the concern of tracking the centreline as the other engine failing, however there must be an interesting change in rudder input somewhere between the two!

Regarding the upwind engine failure, what does the certificaion process require with regards to directional control on the runway. I have looked on the internet and have found no reference of use. I found a PPruNe link to an FAA Avisory Circular but couldn't see any relevent information. Perhaps I missed it.

http://www2.airweb.faa.gov/Regulatory_and_Guidance_Library/rgAdvisoryCircular.nsf/0/BF33F5768F62B0E1862569B2006A63D7?OpenDocument&Highlight=23-8

[john_tullamarine]

(a) how far off centreline can we expect the aircraft to deviate even with maximum rudder input, or can we expect to happily track the centreline?
There will always be a deviation .. however, it only becomes significant/difficult as the V1 gets down near Vmcg. For V1 well above Vmcg, the centreline deviation should be minimal if the pilot is alert and competently trained. (excluding some of the older aircraft for which it is always interesting ...)

(b) Crosswind from the same side as the failure is destabilising, from the other side, stabilising .. ie the problems mount up in the former case.

(c) And when airborne and climbing away, you would have a heap of left rudder

.. due to "lights on but nobody home" syndrome, the next comment was written with AEO in mind .. which is not relevant to the OEI discussion .. left so that the following posts make sense ..

Why would you want to fly with rudder in during the initial climb ? .. at least after the head patting, tummy rubbing antics appropriate to the rotation flare ..

(d) I found a PPruNe link to an FAA Avisory Circular but couldn't see any relevent information
.. or perhaps it's not all that relevant to FAR 23 .. have a look at FAR 25.149(e) (http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=ecfr&sid=b581e83ade8bc750e047522ac466fcdb&rgn=div8&view=text&node=14:1.0.1.3.11.2.156.24&idno=14).

[Blip]

Thank you for the link jt.

I learned that following an engine failure at Vmcg the aircraft is permitted to deviate 30 ft from the centreline. But there is no mention of a failure at max crosswind limits. I will take your word that there will not necessarily be excessive deviation to centreline tracking above V1.

Why would you want to fly with rudder in during the initial climb ? .. at least after the head patting, tummy rubbing antics appropriate to the rotation flare ..

Because in my experience unless you have it configured with zero aileron input (which requires correct rudder input ("dead leg dead engine") I find it impossible to maintain the desired heading (and hence track) with wings level. Wings level with insufficient rudder equals aileron and constant heading change. No I am not in the habit of using 5 degrees bank towards the live engine.

[BOAC]

Why would you want to fly with rudder in during the initial climb - it is SOP, JT. Rudder applied to stop yaw when it is seen, especially on the runway. Blip is asking about the bit between V1 and Vr while the a/c is trying to stay on the runway. Rudder most definitely needed - and also in the intial climb asap to bring the c/column back to neutral and avoid large spoiler deployment.

Blip - taking off in 'max crosswind' is 'exciting' in itself - a little more excitement will probably go un-noticed:)

[john_tullamarine]

aaahhhh !! mea maxima culpa .. I misunderstood the question to relate to AEO (brain infarct when one considers the rest was clearly OEI) .. if we are talking OEI for this as well .. of course, one has a rudder input until time permits to trim it out. Other things being equal aim to climb out with around 2-3 deg bank into the operating engine(s) and rudder as required ... ask legions of turboprop pilots .. the DTs have nothing on the shaking OEI leg takeoff ..

... maybe I need a good night's sleep ...

Re deviation .. nil wind is the general rule .. if you want guarantees, let me offer you one .. a min V1 departure with a maximum crosswind .. will just about guarantee you a trip in the weeds before you know where you are .. the case is not considered at certification and the pilot is expected to exercise a bit of discretion in whether he should go in these circumstances ..

.. retires muttering to himself in rambling terms ....

[captainpaddy]

Re deviation .. nil wind is the general rule .. if you want guarantees, let me offer you one .. a min V1 departure with a maximum crosswind .. will just about guarantee you a trip in the weeds before you know where you are .. the case is not considered at certification and the pilot is expected to exercise a bit of discretion in whether he should go in these circumstances ..

Ah now, ya can't really mean that!
In still wind, a engine failure at V1 will require a certain amount of rudder deflection to control.
In a crosswind, the rudder will already be deflected to keep the aircraft tracking the centreline. If an engine then fails at V1 the rudder will need to be deflected further by a similar amount to the still air case. e.g.:
Still air:
Engine failure - 50% deflection required to control.
Crosswind:
Requires 40% deflection to maintain centreline.
Engine failure - 50% deflection to control.
Total 90%
During a crosswind or in still air the rudder needs to be moved the same amount following an engine failure. You can think of it as the initial reference position of the rudder being different in the crosswind case. If you feel that a trip to the weeds is inevitable then you are implying that the rudder will hit the stop and control will be impossible. You can't honestly think that possibility has been overlooked by authorities or manufacturers. At my company on the 321 there are adjustments made to V1 for crosswinds of 20 kts or greater, although this is primarily due to engine consideration which I don't understand. (Incidentally, the change is 1 knot!)
the case is not considered at certification and the pilot is expected to exercise a bit of discretion in whether he should go in these circumstances ..

Ha ha! I don't really think that you will find many manufactureres who would agree that their aircraft probably shouldn't take off at their published crosswind limit, just in case something goes wrong. Nor do I think that pilots will say "The limit is 30 knots, it's at 25 now, why don't we wait a while cos I'm worried about not having enough rudder if an engine fails". Tea and biscuits coming eventually if you keep that up! My question is - What crosswind component would you sugeest we wait for? 30% of the limit? 50%? 75%?
The only thing I would concede in the above discussion is certain operators poilcy of not taking gusts into account. If you're engine goes bye-bye at V1 which is limited by VMCG and at the same instant a gust that is 10 knots above your crosswind limit occurs, you may have trouble. You may however just simply be the most unlucky individual on the planet!

[john_tullamarine]

(a) Ah now, ya can't really mean that!

I mean just that - we are talking about the wind from the same side as the failure, of course .. from the other side the situation is improved in the crosswind condition. Keep in mind that the (small) deviation we routinely see during local asy training or in the sim does not increase gradually as speed decreases .. it tends to come in a bit of a rush as the cut speed approaches Vmcg. Now add a significant additional yawing moment due to the crosswind and it all turns pear shaped.

(b) In a crosswind, the rudder will already be deflected to keep the aircraft tracking the centreline. If an engine then fails at V1 the rudder will need to be deflected further by a similar amount to the still air case. e.g.:
Still air:
Engine failure - 50% deflection required to control.
Crosswind:
Requires 40% deflection to maintain centreline.
Engine failure - 50% deflection to control.
Total 90%

Doesn't work that way ... your scenario might be the case for a "normal" cut speed well above Vmcg .. for the min V1 case, you can expect full or near full rudder deflection before you start considering the effect of wind... ergo add wind and it's all excitement

(c) If you feel that a trip to the weeds is inevitable then you are implying that the rudder will hit the stop and control will be impossible.

That's the story ...

(d) You can't honestly think that possibility has been overlooked by authorities or manufacturers.

I know so. The certification requirement looks at nil wind (or 7 kts in some older cases). There is nothing stopping you going with a min speed schedule in a strong crosswind .. you just need to understand that there is an elevated risk of control loss if you have a critical failure.

(e) At my company on the 321 there are adjustments made to V1 for crosswinds of 20 kts or greater, although this is primarily due to engine consideration which I don't understand. (Incidentally, the change is 1 knot!)

I can't speak to your company's protocol other than to suggest that the correction isn't looking at crosswind.

(f) I don't really think that you will find many manufacturers who would agree that their aircraft probably shouldn't take off at their published crosswind limit, just in case something goes wrong.

Of course not .. the OEM will refer you to the certification design standard ..

(g) Nor do I think that pilots will say "The limit is 30 knots, it's at 25 now, why don't we wait a while cos I'm worried about not having enough rudder if an engine fails"

This pilot considers the following

- if I have a longer than necessary runway, can I increase the speed schedule to make the problem go away ?

- can I delay ?

- else, is it reasonable to go and accept the (small) risk of a critical failure ?

The reason that some of us bring this topic up every now and again in PPRuNe is that the very great majority of pilots have never had the certification problem explained to them .. ignorance is bliss.

(h) What crosswind component would you sugeest we wait for?

Not a question of crosswind .. a question of speed schedule ... First consider asking your flight standards/ops eng folk to find out if the OEM has the data .. most will. Then apply that to the min V1 (ie Vmcg-limited takeoff schedule) to come up with a corrected schedule. Otherwise my thought would be to go for something not less than 0.5kt/kt for twins and 1.0kt/kt for quads

(i) ..not taking gusts into account.

Difficult to do that .. but you can make a considered assessment

(j) If your engine goes bye-bye at V1 which is limited by VMCG and at the same instant a gust that is 10 knots above your crosswind limit occurs, you may have trouble.

You don't need a gust .. you already do have a problem.

(k) You may however just simply be the most unlucky individual on the planet!

We all accept that there are days when one should not have got out of bed.

[Bellerophon]

captainpaddy


...Ah now, ya can't really mean that!..

Oh, I can assure you he does!


...If you feel that a trip to the weeds is inevitable then you are implying that the rudder will hit the stop and control will be impossible....

Yes he is, under the circumstances he has outlined.


...You can't honestly think that possibility has been overlooked by authorities or manufacturers. ...

No, not overlooked. Just discounted, on the basis of probability of occurrence.


In a nutshell, one of the limiting factors in deciding how low V1 may be permitted to go is Vmcg, the minimum speed necessary to keep the aircraft (straight) on the runway between V1 and Vr, following an engine failure at V1.

If Vmcg is established in a zero cross wind situation, as it is on my current aircraft type, then it will be far below the correct value of Vmcg that should be used in a (maximum) 40kt X/W take-off.

Whilst I can't speak to the performance calculations on the A321 which you fly, on my current type the difference in the two Vmcg values is in excess of 53 kts. The nominal correction to be applied to Vmcg is +1.33kts/1kt X/W.

So, if by ill fortune you were using a Min V1, on a limiting cross wind take-off, and suffered an instantaneous upwind engine failure at/after V1, you may well be committed to the take-off some 50 kts below the actual Vmcg pertaining on the day.

Whether Vr, or the edge of the runway, arrives first will then be in the lap of the Gods, but a trip to the weeds seems to me to be a reasonable description of what may happen next!

What john_tullamarine is correctly pointing out is that many pilots misunderstand the parameters surrounding certification, and therefore have an unfounded confidence in their ability to keep the aircraft straight following an upwind engine failure, at light weight, using a Min V1, in a limiting cross wind.

Demonstrations in the simulator, during command courses or on Base Training Captain's courses, quickly induces a sobering sense of reality.


Regards

Bellerophon

[Mad (Flt) Scientist]

I don't really think that you will find many manufacturers who would agree that their aircraft probably shouldn't take off at their published crosswind limit, just in case something goes wrong.[/i]
Of course not .. the OEM will refer you to the certification design standard ..

*nods*

But if your OEM, off-the -record, still insists that everything will be fine in the conditions JT describes, I'd seriously question their competence.

[john_tullamarine]

Whether Vr, or the edge of the runway, arrives first will then be in the lap of the Gods, but a trip to the weeds seems to me to be a reasonable description of what may happen next!

The problem is associated with the nature of the departure. Based on the few aircraft with which I have had a Vmcg involvement, the deviation is fine down to a speed a few knots above Vmcg.. in the last few knots the deviation escalates markedly and rapidly becomes uncontrollable. If you are a knot or two below the real world Vmcg you probably will be OK due to CG and a bit of NWS capability as the rudder is deflected. However,

(a) if the delta is significant, and

(b) given that the average pilot, not having had any training exposure to the environment, will be caught a bit unawares

(c) the typical airline pilot is programmed to "go" at V1 ..

the expectation is that the weeds are the likely outcome.

I started to take a training interest in this while doing some contract work for an operator which routinely made use of high overspeed takeoff schedules .. while permitting one routine ferry sector to be done with min speed schedule. I used several guinea pigs in the sim to see what sort of comfort the guys had .. and it wasn't much. However .. after a few practice runs, the message was driven home.

Is this something over which you should go to bed at night in a lather of sweat ? .. of course not.

The point is, this usually is an easy risk to mitigate ... if you are aware of the problem in the first place.

[galaxy flyer]

Bellerphon:

Would your type be the Boeing Whale? That seems like a lot of correction for x-wind. The C-5 was almost exactly 1kt increase in Vmcg for each knot of x-wind, but the 744's engines are considerably further away from the centerline.

This is an interesting topic that, I believe, few line troops understand. I brought this issue up in a training session at my new job and everyone quoted the book answer for Vmcg and passed over the idea that crosswind had any effect on Vmcg. Even on a twin, tail-mounted at that, the effect can be significant at light weights. We can run into a problem far below max. demonstrated crosswind, only about 22 knots of x-wind will do it, based on 0.5 kt/kt x-wind.

GF

[Bellerophon]

galaxy flyer


...Would your type be the Boeing Whale?...

Yes, well guessed (if you didn’t look at my profile ;) )


...That seems like a lot of correction for x-wind...

It does, but then 40 kts of crosswind hitting the fin does create quite a turning moment. If we assume that the up-wind outboard engine fails at V1, obviously the lower the V1 speed, the smaller the maximum opposite turning moment the rudder can generate.


...and everyone quoted the book answer for Vmcg...

And they probably genuinely believe that, at V1, during a take-off in a strong crosswind, they are/must be above Vmcg.

It can take some explaining that, whilst they will be above the book Vmcg, they may very well not be above the actual Vmcg pertaining in those conditions.


...and passed over the idea that crosswind had any effect on Vmcg....

A demonstration in a simulator normally has a salutary educational effect.

I wouldn’t want to leave anyone with the impression that this is a major problem on every take-off, because it isn’t. Experiencing a limiting crosswind on take-off, in a lightly loaded B747, is a relatively rare event in the life of a longhaul pilot, and having the up-wind outboard engine fail catastrophically at V1 during such a take-off is extremely rare.

However, it could happen, and, like many problems in aviation, it will likely be handled much better if the handling pilot has a reasonable understanding of the potential problem, and an awareness of the conditions under which it could occur.

Regards

Bellerophon

[formulaben]

You guys have motivated me. When I go back for recurrent, I'm going to ask for max crosswind and a V1 cut. Should be fun to see what happens...

[john_tullamarine]

formulaben,

When I go back for recurrent, I'm going to ask for max crosswind and a V1 cut. Should be fun to see what happens...

.. but at min weight and speed schedule .. presumes that you are Vmcg-limited.

Slightly different suggestion ... presuming the sim panel organ grinder is amenable, time available, and your aircraft is Vmcg limited at low weight ...

(a) min weight, min speed schedule, max forward CG. Engine failure just before V1 .. it may take a couple of goes to get the hang of it as the whole thing goes in a flash at light weights.

(b) same but for mid CG ..

(c) same but for max aft CG. As the CG moves back, be prepared to use very positive aileron input (could be to max deflection) to catch the roll .. aileron needs to go in rapidly and then be backed off rapidly to suit the initial climbout .. will probably take a couple of goes to get the hang of it.

(d) same but with, say, engine seizure

(e) if you a lucky enough to have a decent bird strike programmed into the box, then try that

(f) aim is to handle the failure and then climb out back tracking the opposite end localiser.

.. and, of course, all this should be done raw data .. ie no FD.

If you thought you had a fast I/F scan .. then watch it really improve with several of these ... or else you will have to scrape the box off the red screen (or whatever passes for crash in your scenario). A lot depends on box fidelity of course .. some are pretty average, some very good.

If you think that is a big ask, it's not .. I've trained numerous endorsement crews to handle just this .. and then they NEVER have any problems with "normal" critical failures .. walk in the park.

Main thing is that, if you haven't done some practice here .. you DON'T want to see it first time for real .. my thought is that the majority of folk will have little chance of catching the bird before the ground does ...

[Captain Airclues]

Crosswind is not taken into account in the calculation of Vmcg because the certifying authorities consider that the chances of having an engine failure exactly at V1 are less than ten to the power of seven. When we used to fail engines during base training on the 747 we used to apply Bellerophons figure of +1.33kts/1kt crosswind to the Vmcg.
After the last discussion on this subject I decided to do an experiment on the sim. We did a heavy weight take-off on a wet, limiting runway with a 25kt crosswind from the right. The V1/Vmcg was 133kts and the Vr was 169kts. The very experienced (and willing) crew were briefed that the #4 engine was going to fail at exactly V1 and that they should apply, and hold, full left rudder. After the engine failure the aircraft yawed to the right, leaving the paved surface passing 140kts. it continued to yaw to the right although at a reducing rate until full directional control was regained at 165kts. At Vr (169kts) the SIM FRZ button was pressed and we then estimated that the aircraft had deviated about 300-400ft from the runway edge. At some airports you would be in rather more than weeds!

Airclues

[captainpaddy]

You guys have motivated me. When I go back for recurrent, I'm going to ask for max crosswind and a V1 cut. Should be fun to see what happens...


Jaysus lads! I think I'll do the same. You have me scared now! How could all those wonderful trainers I've met overlook this!!!!!!!!

(d) You can't honestly think that possibility has been overlooked by authorities or manufacturers.

I know so. The certification requirement looks at nil wind (or 7 kts in some older cases). There is nothing stopping you going with a min speed schedule in a strong crosswind .. you just need to understand that there is an elevated risk of control loss if you have a critical failure.

Could you tell me which regulation you are referring to? I think I need to look it up!!!

PPRUNE is scary. All it does is show you how ignorant you are and how much you thought you knew............

[formulaben]

Jaysus lads! I think I'll do the same. You have me scared now! How could all those wonderful trainers I've met overlook this!!!!!!!!

Yeah, with all the sadistic sim instructors I've seen, I'm surprised one of them hasn't used this one!

[john_tullamarine]

(a) You have me scared now!

.. no need to be scared, per se .. we are talking a low frequency/probability consideration here .. however, far better to have an understanding so that you have the background to figure a risk mitigation process where that might be feasible.

(b) .. which regulation ..

25.149(e) (http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=ecfr&sid=f843d8e5eee070d245ba39cd21c739d0&rgn=div8&view=text&node=14:1.0.1.3.11.2.156.24&idno=14) doesn't require wind .. ergo .. none. See, also, the FTG at AC 25-7A (http://www.airweb.faa.gov/Regulatory_and_Guidance_Library/rgAdvisoryCircular.nsf/0/4144773715d01e9a86256b9f0072f02a/$FILE/AC25-7A.pdf). This is a pretty useful book to read in any case as it gives you heaps of background on how things are done in the certification process ..

(c) all the sadistic sim instructors I've seen

unfortunately, they exist .. however, if the attitude of instructor and pilot is positive and sensible (and the sim fidelity half reasonable), the sim is a great training tool for seeing how things might go when things are a little out of the ordinary.

[Ignition Override]

Interesting.

Few pilots (guys/gals who have flown about 15-20 years) have had any real engine failure. I've flown with about two or three. None on the runway after or even near V1.

A pilot who had a more interesting series of compressor stalls in a 737 (300?)was departing Chicago Midway (MDW) and the power kept falling off then coming back several times. He had already finished a full 20-year career in the Air Force C-130. Ironically, the Air Force C-130 had no autofeather, as with the P-3, the E-2 Hawkeye or the Convair-580 (Allison).

How about in simulators? Has any trained or been checked for a sudden failure near V1 with a fairly strong crosswind?

[TyroPicard]

On the other hand think about what happens when the downwind side engine fails, ie crosswind from the right, left engine fails. Before the failure you would have a heap of right rudder, and possibly left control wheel...


Your crosswind take-offs sound even more interesting than my crosswind landings!
Presumably everyone else noticed but were too polite...
TP

[Blip]

[cough] :O

Oh dear. Yes that would look rather spectacular wouldn't it?

Thank you for pointing out the gross mistake TyroPicard.