Asymetric Thrust

[SLF3b]

The early part of the Spanair thread gives the impression an engine failure on the MD-80/82 is much harder to control than on twin engine passenger aircraft with the engines under the wings. I had always imagined that asymetric thrust was more manageable on the MD-80/82 because the engines are closer to the centre line of the plane. Also imagined a single thrust reverser deployment was less of an issue for the same reason.

Am I wrong?

I have no view on what caused the Spanair crash, or on the safety record of the MD-80/82, this question is simply to assuage my insatiable curiosity. And no, I am not a journalist.

[Chris Scott]

Your basic understanding is correct, SLF3b. Must admit that I have missed most of the many no-doubt valuable contributions to the Spanair thread, due to the sheer volume of rubbish through which they are fighting for attention.

So I don't know precisely what argument you are referring to, but my impression is that a popular theory involves uncommanded (or at least undesired) reverser deployment on one side, possibly at max reverse thrust, with the other engine at take-off thrust. That would represent a much greater asymmetry than the 'normal' engine-failure-on-take-off case, and would probably be beyond the capability of full rudder to counteract at take-off speeds.

Tail-mounted engines are, of course, further apart than in the days of the turbojet Sud-Aviation Caravelle (1957), because even the 1980s turbofans of the MD80 series are fatter.

[Mad (Flt) Scientist]

A couple of items.

While the engines on an aft-engined airliner are indeed closer to the centreline than an underwing type, and thus the yawing moment generated by engine asymmetry will be less, the plane's designers will take advantage of that fact to make the rudder smaller on the rear-engined aircraft. The goal, as in practically all aspects of aircraft design, is to make everything just the right size, and no more, because an over-large rudder means extra weight, more drag, and possibly other problems too.

So, as a result, for a 'simple' engine failure, one would expect an average rear-engined plane and an average under-wing engined plane to be broadly similar in terms of controllability.

Now, consider the specifics of the (theorized) TR deployment in the Spanair case. The direct yawing moment effect from the reverse thrust is proportional to the normal asymmetry, so again you'd expect the aircraft to be no more or less controllable than an underwing type with a similar failure. But there's something not yet accounted for - aerodynamic interference.

The effect of the reversers on the local flow is significant - and may well cause loss of quality flow around the tail - which means the rudder isn't getting as much bang for the buck. It might only be seeing an effective airspeed of 60 or 80 knots, rather than an actual 100 or 120 perhaps, due to the reverser interference. Obviously, underwing reversers don't affect the tail anything like as much - but might affect the wing, of course ...

[wileydog3]

The early part of the Spanair thread gives the impression an engine failure on the MD-80/82 is much harder to control than on twin engine passenger aircraft with the engines under the wings.

Just because the engines are mounted on the fuselage does not mean there is not significant yaw with an engine loss. The Learjet 23/24/25 and the MD-80 are examples where, when losing an engine, you go to the floor with the rudder and then back off a bit. Failure to put in FULL rudder may well complicate the problem. Any timidity in rudder input will complicate control.

The -80 has a very large vertical fin and rudder? An aero-engineer is not going to build a big vertical fin and rudder if it is not needed. Thus...

Also imagined a single thrust reverser deployment was less of an issue for the same reason.


An engine in reverse right after takeoff is going to be a challenge, no doubt. But again, one has to take the whole picture in view and consider flight test and certification. Do you think the FAA and JAA would certify an airplane and not consider the possibility of a reverser opening? Or that they would certify an airplane that called for extraordinary skill to fly? Certification is so that your average trained pilot can fly it.

Clue.. when you hear some stud talk about how difficult an airliner is to fly, it is probably more a statement of his flying skills than the handling of the airplane.

[SLF3b]

Thanks for the informative answers. The comment on the effect of the bucket deploying on the airflow around the tail is not something I had thought of. I guess the effect is greater for a rear engine because the tail is much smaller than the wing, but then as the other poster comments the surfaces will have been sized for the difference.

I repeat, no view on what brought Spanair down.

[stilton]

I spent 4 years on the MD 80 and can tell you most assuredly that you are using most, if not all of the available rudder deflection with an engine failure.

It does not have a particularly large fin and rudder, compare it to the 727 for example.

[Chris Scott]

Quote from wileydog3:
An engine in reverse right after takeoff is going to be a challenge, no doubt. But again, one has to take the whole picture in view and consider flight test and certification. Do you think the FAA and JAA would certify an airplane and not consider the possibility of a reverser opening?


Have never flown the MD-80 series, and have no Vmca or Vmcg figures. But Mad (Flt) Scientist has pointed out that manufacturers are unlikely to provide a rudder capability greater than is thought to be necessary, because of the possible penalties. stilton states that “most, if not all” the rudder will be required in what I’ve called a ‘normal’ engine failure, and wileydog3 seems to agree.

From the [U]take-off certification point of view, I’m not aware that allowance has to be made for an engine failure that involves it producing reverse (negative) thrust of any degree. It is common on many types of aircraft, when light, to nominate a Vr (rotation) speed which is at or just above Vmca (particularly on contaminated runways). For the benefit of SLF3b and others, it needs to be pointed out that Vmca is the minimum control speed in the event of the failure of any one engine when airborne.

Regarding the Spanair accident, one assumes that it was relatively heavy, with a much higher rotate speed, so there would have been a considerable margin over Vmca. Whether that would be even theoretically sufficient to cater for reverser deployment − including the rudder-airflow problem − I’m not going to speculate.

In the event of spontaneous (faulty) reverser deployment, there is supposed to be an automatic selection of idle thrust, but turbofans take a significant time to slow down, and that probably includes even the old JT8D? Bucket-type reversers are old technology, and may be more prone to undesired deployment than the more modern (but less effective) fan reversers. Nevertheless, this would be an extremely rare event, and to happen at such a critical moment of the flight is not something one would expect the certification to take into account.

[krujje]

Nevertheless, this would be an extremely rare event, and to happen at such a critical moment of the flight is not something one would expect the certification to take into account

TR deploy on take-off is considered during cert, however not from a controllability point of view, but within the context of the aircraft functional hazard analysis.

In aircraft I've seen, TR deploy on take-off is considered a catastrophic event, therefore effort is made through system reliability and redundancy to ensure the probability of it happening is better than 10-9.

[Green-dot]

krujje wrote,

In aircraft I've seen, TR deploy on take-off is considered a catastrophic event, therefore effort is made through system reliability and redundancy to ensure the probability of it happening is better than 10-9.


True regarding the probability calculation, hence the emphasis on system reliability. However, only catastrophic if the problem is not identified as such, the very remote event of an uncommanded TR deployment.

With hindsight, a Fokker 100 (TAM flight 402) accident on Oct. 31, 1996 could have been prevented if the crew had identified the problem to be an uncommanded #2 TR deployment. ATS disconnecting and the thrust lever retarding to idle (of the affected engine) are indications for such an unlikely event but apparently were not recognized as such by the crew. Whatever the perception of the crew was regarding the situation, human factors and situational awareness become vital elements. In an action to correct the problem, the thrust lever was advanced in order to "restore" thrust on the #2 engine, which consequently made the aircraft uncontrollable.

No alerts were presented to them as would have been anticipated to occur in case of a TR deployment during most phases of flight. But TR alerts are inhibited above 80kt until 1000ft AGL on this aircraft, the philosophy being to prevent high workload during the critical phase of a takeoff. Training and the AFM/FCOM cover this issue as part of the design and certification process.

Investigation later revealed there was a design flaw (a dormant failure) in the electrical control circuit which may have contributed to the uncommanded deployment. The electrical control circuit was later revised which now makes such an event even more remote.

Summarizing, if the symptoms of an ATS disconnect and a retarding thrust lever had been recognized for what this implied, corrective action would have been taken and the aircraft would have remained controllable with a deployed TR using the procedures in place to cater for such a remote event.

As is usually the case, hindsight has the benefit of 20/20 vision.

Regards,
Green-dot

[Chris Scott]

Forgive me for not researching the Fokker 100 accident, Green-dot (am on that dail-up connection again); but was the aeroplane airborne, and was its speed close to Vmca?

As I've argued above, even idle reverse would result in loss of control at Vmca. And how long would the Tay or JT8D turbofan take (automatically) to "spool down" after an uncommanded deployment of reverse at rated thrust? One or two seconds? Meanwhile, the aeroplane would be swinging rapidly into sideslip, even once full rudder had been applied. I'm arguing generally, of course, not in relation to any specific accident. And I've no idea what the Vmca is on either the Fokker100 or the MD82.

You have rightly alluded to the difficulty of swift crew recognition; particularly if flight-phase inhibition hides the conventional warning, and even if training has covered the likely reason for a snap-closure of one throttle lever. So, if it should happen close to Vmca, the crew would do very well to avoid a departure. I can only presume from your penultimate paragraph that this was not the case in the Fokker100 accident. As far as the recent MD82 accident is concerned, it seems unlikely that the aircraft would have been light enough to have been anywhere near Vmca if it was airborne at the scheduled speed.

A light take-off or go-around is not necessarily safer than a heavy one. Having said that, the wider adoption of optimised speeds on long runways (in order to use less thrust) means that aircraft are flown near Vmca less often than in the past.

Chris

[glawkshuter]

Chris, good point...and so does Flex/Reduced thrust reduce safety?

[Chris Scott]

Hi glawkshuter,

Personally, I never felt less safe using Flex/Reduced/Graduated (unless, as the far-end piano-keys loomed into view, I thought I might have screwed up the calculation!). TOGA, or whatever, is always available. As for Derated Thrust, I never flew an aeroplane that had the capability.

[krujje]

[...] only catastrophic if the problem is not identified as such, the very remote event of an uncommanded TR deployment.

I would say that point is debatable. Please correct me if I'm wrong, but there is no requirement to prove that uncommanded TR deployment at take-off is controllable, therefore airframers don't... preferring, as I said, to show 10-9 instead. Is there proof through test or analysis that the event is controllable in all normal take-off scenarios? If yes, does it hold for all modern aircraft, or only certain types?

With hindsight, a Fokker 100 (TAM flight 402) accident on Oct. 31, 1996 could have been prevented if the crew had identified the problem to be an uncommanded #2 TR deployment.

Can we really say that the accident could have been prevented had the crew known what the problem was? I think that that may venture too far into speculation. Although I say this not having read the report. I have only a cursory knowledge of the circumstances of the accident (I have not been able to find an english translation of the official report.)

[Green-dot]

Chris,


Forgive me for not researching the Fokker 100 accident, Green-dot (am on that dail-up connection again); but was the aeroplane airborne, and was its speed close to Vmca?


Yes the aircraft had just become airborne.
With reference to the Fokker 100 AFM performance section (an outdated copy I must add) I quote: "For the Fokker 100 Vmca is less than the stalling speed Vs for all take-off flap settings."

From the information available to me, the aircraft took off with flaps 8. Entering the Vs graph with a rough estimation of the aircraft gross weight and information of the actual airspeed recorded, the aircraft reached an airspeed well above Vmca. I estimate that Vs would have been somewhere between 108-112kt, hence Vmca is below this. The aircraft reached a speed of approx. 140kt which reduced to just below 120kt after the #2 reverser deployed. See links below.

krujje,

I would say that point is debatable. Please correct me if I'm wrong, but there is no requirement to prove that uncommanded TR deployment at take-off is controllable, therefore airframers don't... preferring, as I said, to show 10-9 instead. Is there proof through test or analysis that the event is controllable in all normal take-off scenarios? If yes, does it hold for all modern aircraft, or only certain types?

I don't know if there is or is no specific requirement at take-off. I do know that abnormal procedures exist for (uncommanded) TR deployment on the ground and in flight and the aircraft has demonstrated to be controllable with one TR deployed in flight as part of the certification process. I am also aware that with a failure chance of 10-9 (if I recall correctly it was even 10-11 for the Fo100), the emphasis should be on proper maintenance and operational procedures.

Can we really say that the accident could have been prevented had the crew known what the problem was? I think that that may venture too far into speculation. Although I say this not having read the report. I have only a cursory knowledge of the circumstances of the accident (I have not been able to find an english translation of the official report.)

As is usually the case with hind sight, it is easy to judge from a comfortable chair how things could have been handled differently. The surprise effect and the short time to analyse such a problem during a critical phase of flight obviously have contributed to the outcome. Especially when the remoteness of an uncommanded TR is taken into consideration. Would it, even today, be routinely in the back of the minds of pilots as one of the anticipated possibilities of this occurring during the take-off phase? Is such a scenario trained for on a regular interval?

The answers put emphasis on human factors and CRM.

To shed some light on what happened on that flight:
Here are 2 graphic presentations based on CVR/ATC/FDR data. The first without CVR/ATC text, the second including text. Please note #2 thrust lever position retarding to idle at moment of TR deployment and the subsequent thrust lever increase as explained in my previous post. Also note aileron/rudder inputs.

YouTube - TAM 402 - (http://www.youtube.com/watch?v=3lQOuHSodj0)

YouTube - TAM 402 de 1996 com legenda - transcrição caixa preta + FDR (http://www.youtube.com/watch?v=TH1L2HuE6o8&feature=related)


Regards,
Green-dot

[Spruit]

Hi,

After reading this discussion I have a few questions.

Would it be possible to reproduce this scenario of uncommanded TR deployment in a simulator environment for this type? And if so, would this produce results that are comparable with this incident? Or is that the question here anyway?

Also, are the older bucket type thrust reversers more effective than the more modern fan reversers on later aircraft?

Thanks in advance,

Spru!

[Chris Scott]

Quote from Green-dot:
Would it, even today, be routinely in the back of the minds of pilots as one of the anticipated possibilities of this occurring during the take-off phase? Is such a scenario trained for on a regular interval?
[Unquote]

Having retired over 6 years ago, admittedly from a type with fan-reversers, I'd answer: "probably not, unless there was a history of it on the type"; and "no, not as far as I can recall during my jet career (VC10, B707, BAC1-11, A310, DC10 and A320)." But none of these aircraft had bucket reversers.

[Furloughed]

Having flown the Md-80/Dc-9 series in the past I've experienced thrust reverser deployment just after rotation in the sim. While the aircraft yaw is controllable the reverse thrust and added drag of the deployed buckets over came the thrust of the 'good' engine causing an uncontrollable descent rate. Only way to survive was to immediately shut down the engine with the deployed reverser.
I found the old clam shell buckets to be much more effective than a sleeve type ( flew the 737 -200 and 3/400 series as well).
I believe an inflight reverser deployment caused the loss on an Air Lauda 767 some years ago.

[kijangnim]

Greetings,
Cascade reversers on by-pass engines reverse approximately 75% of the thrust, while the jet still produces 25% of forward thrust :}

[krujje]

I believe an inflight reverser deployment caused the loss on an Air Lauda 767 some years ago

26 May 1991, 223 souls on board. No.1 engine TR deployed about 15 minutes out from Bangkok. Engine thrust reduced to idle, however the reverser plume caused loss of lift over the wing leading to stall and uncontrolled descent. The aircraft broke up as the crew tried to arrest the descent.

See also:

Cessna Citation in 1983 in which both TR deployed at gear retract. Apparently circuit breakers had been pulled the day before and not reset...

Continental Air Lines Sabreliner in 1973; TR deploy after take-off, at about 1000 feet.

Air New Zealand DC-8 in 1966; TR deploy at take-off, insufficient time to correct before impact with ground.

[Green-dot]

Chris Scott,


Having retired over 6 years ago, admittedly from a type with fan-reversers, I'd answer: "probably not, unless there was a history of it on the type"; and "no, not as far as I can recall during my jet career (VC10, B707, BAC1-11, A310, DC10 and A320)." But none of these aircraft had bucket reversers.


With reference to the BAC1-11, I have witnessed one veering off the runway back in the '80s while it was landing at Schiphol. One of its (cascade) reversers failed to deploy and it came to rest in soggy clay to the right of rwy 01R (currently rwy 36R). Fortunately no one was injured but everyone had clay up to their knees after evacuating the aircraft. The aircraft only had minor damage and was fit to fly several days later.

Furloughed,


While the aircraft yaw is controllable the reverse thrust and added drag of the deployed buckets over came the thrust of the 'good' engine causing an uncontrollable descent rate. Only way to survive was to immediately shut down the engine with the deployed reverser.


From what I recall, the Fo100 with a deployed reverser is controllable/recoverable when the affected thrust lever retarding to idle is recognized as such (no triple chime ENG FAIL alert/relevant fuel lever light not on) and Vfto (final take-off speed) is established. All with a very small margin in rate of climb.

From memory (it has been several years) the logic was to leave it to the pilot to consider shutting down the affected engine. Conditions to do so depended on aircraft weight, field elevation and temperature (WAT conditions).

Because the aircraft should be controllable with the good engine fire-walled and gear retracted ASAP. Shutting down an engine would slightly improve aircraft performance, however, it also means losing system redundancy. Furthermore, human factors wise, even if with minimal rate of climb by leaving the affected engine at idle, would a pilot choose to shut down an otherwise functioning engine so close to the ground . . . . ?

If the aircraft remains controllable with minimal rate of climb, shutting down an engine would slightly increase performance but (apart from less system redundancy) would call for single engine procedures and (regarding the Fo100) a zero flap approach/landing in the case of a deployed reverser which could add to the complications already at hand. Therefore, as far as I can recall from years ago, shutting down the affected engine on the Fo100 was left to pilots discretion, depending on the actual (WAT) conditions he is confronted with.

I must emphasise that perhaps procedures and philosophy have changed over the years but this was the train of thought at the time. If anyone can update this to the current situation, please do so.

Regards,
Green-dot

[Chris Scott]

‘ Morning Green-dot,

Regret you may be − understandably − barking up the wrong tree. I think the BAC 1-11 incident you remember was a Dash200; landing on Rwy01R in 1981, in almost calm conditions. After a normal touchdown, the copilot, who was the PF, called for spoilers and full reverse iaw the SOP. The captain selected them and, as the Speys spooled up, and the nosewheels touched down, the aircraft swung alarmingly to the right. The aircraft was light, so the speed had already fallen to about 100kts. Also iaw SOPs, the captain immediately selected reverse idle, but the swing continued.

Meanwhile, the copilot had already selected full left rudder, which had no effect. [There is no rudder-fine (nosewheel) steering on Dash 200s.] Neither pilot attempted to arrest the swing using their steering handles; perhaps because the speed was above the 80-knot maximum in the SOP, and possibly because of the violence of the swing. The captain selected heavy left brake, but this seemed to make no difference to the continuing swing, and the aircraft left the right side of the runway; the captain shutting the engines down and operating the fire handles as it crossed a strip of grass and entered a ploughed (agricultural) field.

The aircraft had left the runway edge about 3 seconds after departing from the runway centreline. No one was hurt, and the 1-11 suffered little damage. The UK AIB investigated the incident, however. The FDR indicated that both engines and reversers had functioned correctly, and they were checked and found to be serviceable. Analysis of the tyre marks indicated that the nosewheels had been slightly offset as they touched down, but within the limits of system acceptability. However, they had apparently failed to caster straight in the normal way. Instead, they had remained fixed at an angle of about one or two degrees to the right. Some corrosion was found in the castering mechanism, but the proposed failure was never proven or reproduced, as far as I know.

Among other things, the incident did show the authority of undesired nosewheel-steering movements on a dry runway, particularly on short-wheelbase aircraft; the increase in its authority when mainwheel brakes are applied; and reminded us that rudder effectiveness is reduced when tail-mounted engines are in reverse thrust. It also taught us that, when your action seems to create a problem, the coincidence of the two events may be just that − a coincidence.

[Green-dot]

Morning Chris,


Regret you may be − understandably − barking up the wrong tree. I think the BAC 1-11 incident you remember was a Dash200; landing on Rwy01R in 1981, in almost calm conditions.


You are correct. Sorry for the mix-up between 01L and 01R :ugh:. Indeed it was 01R and the incident happened right in front of us while we were working on the appron at the Schiphol East hangar area.

Thanks for your detailed explanation regarding this incident. The only reference to the incident for me at the time was the press who in the papers I read and in the news on TV all referred to a malfunctioning reverser, I never read a copy of the final report. Having grown wiser over the years and having learned to take the press with a grain of salt regarding anything they write about aviation (and just about any subject for that matter) I stand corrected.


Regards,
Green-dot