Uncommanded thrust reverser deployment in flight

[josephfeatherweight]

This is a brilliant thread - some very knowledgeable contributors, for which I am grateful.

[RAT 5]

The airforced trained PF then completed the roll through 360 deg and requested clearance to return to the field.
On ground examination found nothing wrong.

Ah, the famous "ground tested found satis" tech log reply. I wonder what the FDR showed. That would have been interesting. Indeed, following the a/c all the way through the manoeuvre, rather than reversing it, might sometimes be the best action. Not recommended with looping high nose attitude though.

I can't remember accurately the follow up to Lauda. They went in the sim and experimented and I think they all failed. I don't recall if they tried just closing both TL's immediately to gain control and descending. In fact, with the design of the day, would they have ended up with one at forward idea & one at reverse idle? I suspect the roll rate would have been eye watering and uncontrollable with rudder & aileron, but with both at idle, what then? There were a/c with too small rudders, (VMCA) when with an engine failure, and the other at full power, is was necessary to reduce thrust to gain control. Not the first instinct unless trained.

[paradoxbox]

Originally Posted by RAT 5

There were a/c with too small rudders, (VMCA) when with an engine failure, and the other at full power, is was necessary to reduce thrust to gain control. Not the first instinct unless trained.

This is golden info, thank you very much.

Someone asked me "Risk of what?" when I queried as to whether reducing thrust is safe to do without a definite indicator of thrust reverser deployment. The risk I imagine is making your energy state worse on takeoff or climb by reducing throttle on one or both engines before you know exactly what the problem is.

In a few incidents I can think of, the crew instinctively shut down either the wrong engine, or shut down an engine still providing forward thrust, just a little after takeoff.

There was a Taiwanese Dash 8 Q400 that crashed because of this. The aircraft had been experiencing engine trouble on previous flights and assumed that it was the same engine causing the problem this time. The pilot shut it down without confirming that it was indeed that engine experiencing the problem. In actuality it was the other engine which was not performing correctly this time and he shut down the only engine still operating normally.

Another incident I can think of is the Concorde crash, where the flight engineer shut down an engine (or both? Can't remember) because of the fire warnings going off. That engine was still providing thrust and if he had not done that, they may have had a chance of getting more speed and altitude. Of course they were in trouble due to the severity of the fire burning through the wings, but if the engine had not been shut down they may have had more time to deal with the problems, possibly saving the aircraft.

Those types of incidents make me very wary of cutting throttle in takeoff or climb before knowing what is causing the problem for certain.

Is it safe to assume that if the reversers do deploy for whatever reason, we can expect reliable indicators and should just reduce thrust? Again, would unloading the wings save you from uncontrollable roll at lower airspeeds like 190-230 knots? Perhaps lower wing loading would give you enough rudder/roll authority to at least look for other possible causes.

[lomapaseo]

Someone asked me "Risk of what?" when I queried as to whether reducing thrust is safe to do without a definite indicator of thrust reverser deployment. The risk I imagine is making your energy state worse on takeoff or climb by reducing throttle on one or both engines before you know exactly what the problem is.

The thrust reverser control problem has been at it's most difficult to control when some control surfaces were inhibited in later climb stages. If this was the case where you were asking about the possibility of inappropriate engine reduction then I would judge it not to be of significant risk since your aircraft normally has plenty of ground clearance and control speed to maintain safe flight.

I am however sensitive to quick reactions by the crew in pulling back wrong engines and as such far prefer automatic (via FADEC etc.) reductions and not hard shutdowns.

Of course I expect pilot training now covers all these concerns quite well.

[underfire]

In those cases the TR deployment is symmetrical, starts at idle, and power goes up to some predetermined limit that's flight tested and certified to not pose any issues wrt. efflux blanking out controls, etc. Not sudden, on one side, and at cruise or higher power.

Understood, but as the aircraft had procedures to deploy reverse thrusters inflight, it seems reasonable that procedures were in place if only one of the pair deployed? This may be a good learning experience on that adventure? (ie on the Concorde, when using rev thrust to slow from supersonic, and only one deployed, what was SOP, that seems a bit of an extreme case)

[galaxy flyer]

Again, would unloading the wings save you from uncontrollable roll at lower airspeeds like 190-230 knots?

I would think unloading wouldn't increase control because the problem is the reverse efflux is blanking out the flow over the wing and even zero G won't change that.

[paradoxbox]

Does this apply to aircraft with tail mounted engines as well, to the same extent as with wing mounted engines?

[tdracer]

I've never spent significant time working an aircraft with tail mounted engines (I was responsible for the 727 for a while, but it was decades out of production at the time and there wasn't much going on). However my engineering judgement says that with tail mounted engines, the problem is going to be dramatically reduced tail effectiveness due to the reverser efflux blanking the tail control surfaces. So it would be a fundamentally different problem than the one faced by Lauda where the aircraft almost instantly rolled into a dive due to the dramatic loss of lift on one side.
My understanding of the F100 crash (where the reverser deployed shortly after takeoff) was they would have been OK if they'd just left the engine alone after the feedback pulled the throttle to idle. But somehow they managed to brute force the throttle back up to high power and that's when they lost control.

BTW, the FARs (25.939) still don't mandate the 3rd level of protection, still spelling out that an uncommanded T/R deployment must be controllable. To certify, Boeing has had to petition for an "Equivalent Level of Safety" (ELOS) - basically showing that it's as safe as if the aircraft was controllable because it's shown it'll never happen. Providing an ELOS is a royal pain in the rear compared to showing direct compliance. EASA is better - they updated the corresponding CS to say that you either have to show it's controllable - OR - demonstrate that an uncommanded deployment is "extremely improbable" (which is defined as 10E-9 hr. or less).

[galaxy flyer]

No, it doesn't, the efflux is too far aft, depending on specific design. It's controllable in sim profiles on the Challenger and Global types bug more yaw that simple engine failure. I can't say as to the fidelity of the aero data, however.

The tail-mounted designs usually have a T-tail, so above the efflux. The Falcons with cruciform tail design only use a reverser on the center engine. Whether that is related to uncommanded deployment I cannot say.

In the C-5, it was inboards only for in-flight use. Effective, but I wouldn't use them below 10,000' so as not to get caught with a hung up reverser.

[t7a]

tdracer is absolutely correct when he says that the Tarom F100 crew would have been OK if they had not forced the thrust lever open (3 times) and then shut the engine down. After this accident our sim check rides included uncommanded reverser operations (normally just after Vr). Although it was an eye-watering experience the aircraft was controllable provided the engine was immediately shut down.

[RAT 5]

I think the idea of handling, successfully, uncommanded open TR's in flight by reducing thrust is going to depend on height. If climbing through FL 240 converting your a/c into a glider and sorting out any roll problems is a possibility if PF has some aerobatic experience. If there is enough height there is no need for high g manoeuvring, but controlled inputs, and as someone said, let the a/c roll all the way round and catch it on the way out. Doing this at idle thrust removes a few of the very severe problems. Meanwhile the airspeed is off the clock, but hey, it's still flying. Steady the wings, gently pull out of the dive ignoring the EGPWS, trade see for height and calmly decide which is the engine to use to stabilise the a/c under control.

Then the dream is over and I wake up to loud applause and take a bow, with my wife saying WTF.

[bafanguy]

Speaking of T/Rs deployed in flight, this promo has an interesting bit at 5:28:


https://www.youtube.com/watch?v=fMyHz33X5Ko

[EMIT]

Zero gee

Quote
Again, would unloading the wings save you from uncontrollable roll at lower airspeeds like 190-230 knots?
Unquote

Galaxy Flyer
Quote
I would think unloading wouldn't increase control because the problem is the reverse efflux is blanking out the flow over the wing and even zero G won't change that.
Unquote


If the efflux over the wing with uncommanded thrust reverse causes massive loss of lift over that wing, then the “normal” amount of lift over the non-effected wing will cause a roll, against which the normal roll controls are not powerful enough.
Unloading to zero g will not magically make the normal roll controls more powerful, but it will reduce the lift on the “normal” wing to zero as well. In that condition, the normal roll controls will be able to supply normal roll control.

Problem with zero g is, (one), passengers and cabin crew will float through cabin, if not strapped in and ,( two), zero g will cause a ballistic path towards the earth, so it cannot be maintained forever. However, a temporary use of zero g may allow roll control during the time that is needed to reduce thrust to idle. Once thrust on the bad engine is at idle (or the bad engine is shut down completely), normal flight with enough normal control effectiveness should be possible again.

For the Lauda accident: the fact that the event happened at night, in cloud, should not have made any difference: the big attitude display of the 767 should have been sufficient for recovery from any unusual attitude. (this remark is in no way meant to detract from the magnificent information given by tdracer, nor meant to criticize the Lauda crew).

[galaxy flyer]

EMIT,

Quite correct, I hadn't thought thru the implications of reducfing lift on the unaffected side. I'd guess night, zero G maneuvers and instructing the PM to shutdown the engine would earn one a early retirement.

[tdracer]

EMIT, I believe the key was they had to respond to the T/R deployment immediately - and lacking the outside visual queues and making sense of the attitude display took a second or two that they didn't have.
The FADEC fault dump showed faults consistent with the aircraft breaking up ~10 seconds after the T/R deployed (and with going supersonic before that). It happened fast.

Disclaimer - I wasn't involved in the simulator testing itself, I'm simple relaying what I was told by those who were...

[pattern_is_full]

@tdracer - yet another thanks for the insight into the loss-of-lift dynamics in the 767.

Did Boeing apply any "lessons learned" from the 767 lift-blanking to the 737 classics and NGs, as the larger engines required a location change from under-wing to "forward and upward?"

I realize the Classic design was already in production (1984) before the Lauda Air accident (1991). I just wondered if there were 737 retrofits of extra interlocks added post-Lauda, or if the 737NG already had those, or if the thrust dynamics are just different even with the more "767-like" engine position in newer 737s.

[paradoxbox]

so, TR deployment @ near max cruise seems to be an extremely dangerous situation with seconds separating recovery from breaking the sound barrier and your plane.

airbus and boeing seem to have taken care of it with electronic wizardry - not that i trust it farther than i can throw it. 10E9 probability with my luck it'll happen to me a dozen times on the same flight. At cruise if you experience sudden extreme roll is it a safe bet to slam the throttles to idle? Levitating PAX be damned if it means you won't die.

I would like to know if anyone has done any test flights to determine the roll rate at various airspeeds during thrust reverser deployment at different power settings.

Back when dinosaurs lived, I did aerobatics and was a huge fan of 0 or negative G (or low AoA) to get myself out of all kinds of bad situations. Of course jets are very different from props but I cannot help but think that there must be a way to buy time in this situation. With the aerobatic planes I flew, unloading the aircraft generally resulted in superhuman maneuverability, no doubt the propwash playing a part but I would imagine that even in larger jets, you might be able to get some roll rate back either through rudder or aileron if you unload. Depending on the g load necessary to restore roll the amount of time you could gain could be significant especially if you have a monitoring pilot to help. Keep in mind I am thinking primarily about lower airspeeds here, as it seems that at cruise speed your only choice is to cut the throttle immediately or end up inverted and passing mach 1.

I'm not entirely convinced that a TR indicator is going to show up on the engine screen and neither am I convinced that I am going to be able to hear or identify that it's a TR if I am in a large jet with the cockpit a mile away from the engines.

[lomapaseo]

I would like to know if anyone has done any test flights to determine the roll rate at various airspeeds during thrust reverser deployment at different power settings.

It's hard to get pilots to fly a test like this.

They much prefer some lesser tests and lots of extrapolation.

at any rate airspeed is your friend as it destroys the reverser efflux and sensitizes the wing control surfaces to work as planned.

Of course nothing is 100% so the FAA now requires a few more hoops to jump through before saying that it's good enough.

[tdracer]

I realize the Classic design was already in production (1984) before the Lauda Air accident (1991). I just wondered if there were 737 retrofits of extra interlocks added post-Lauda, or if the 737NG already had those, or if the thrust dynamics are just different even with the more "767-like" engine position in newer 737s.

My understanding was that in the aftermath of Lauda, ALL the then in production Boeing aircraft added the third lock (aka "Sync Lock" since they locked the synchronizing shaft that ties the actuators together to make sure they move at the same rate and don't jam the sleeve) - either in production or by AD mandated retrofit. That included the "classic" 737-3/4/500 although I don't think it was extended to the "Jurassic" 737-1/200.
A big part of what Boeing learned in the aftermath of Lauda was how dramatically the move to ever high bypass engines - hung out in front of the wing - changed the dynamics of a T/R deployment in flight.

airbus and boeing seem to have taken care of it with electronic wizardry

Paradox, what electronic wizardry are you talking about? At least on Boeing every new aircraft has a third physical lock - enabled only on the ground - that will prevent even a powered T/R deployment. On the newer Boeings, in part because the sync lock has some undesirable failure characteristics (it's difficult to tell if it's failed without actually testing it) they went to a track lock - a physical block in the T/R slider track that will prevent reverser deployment, the health of which is easily monitored with prox sensors. If you're really unlucky enough to have an uncommanded T/R deployment in-flight on a Boeing equipped with the third lock, I'd suggest you never go outside because if you do you're going to hit by lightning....twice....on the same day

[flash8]

There has always been to me some vagueness as to the actual events that caused the loss of this hull (Mozart), and I do recall (pretty young at the time, before my own flying career) a lot of people being very discomfited, including the Austrian equivalent of the CAA/FAA.


Thankyou trdracer for your most valued input, the most comprehensive explanation I have read about the actual events outside of the report.