Uncommanded thrust reverser deployment in flight

[Corrosion]

On Boeing 717, BR715 engines, there is mechanical connection from reverser scoop structure to LVDT via cable. Cannot recall indication as i am not fully certified engineer for that a/c type, might be transit/deployed light on EFIS. Another proxy sensors (unlock?) are located on lock actuators and senses if locking tongues moves out from locked position. This aircraft have full scoop reversers, like old DC/MD-series.
These reversers are locked mechanically by lock tongues, two/scoop, and overcentering links on pivot point. Canīt remember anymore is hydraulic rams pressurized to stowe all the time when commanded stowe.
Incidents happens for this a/c when reverser deploys during flight, manufacturer made some modifications to overcome this problem.

Like all a/c systems, these gadgets needs lots of maintenance.

[lomapaseo]

Like all a/c systems, these gadgets needs lots of maintenance.

The trouble is when you need lot's of maintenance it becomes a problem by itself to bite you.

Some fail-safe features have their own hidden faults and since they never get used in most of the life of an aircraft system, you don't know they are at fault or won't work when sorely needed.

I have seen too many assumptions in reliability of detection or prevention devices that were never validated during manufacture or installation.

Like a "jezzuz bolt we need balances in our attempts to minimize since we can't seem to prevent anything for sure.

[tdracer]

Originally Posted by paradoxbox

But.. When you are in the air, and the reverser has deployed, for example by aerodynamic forces, or mechanical failure (parts falling off etc) - will the indicators still appear in your aircraft? Is there some kind of mechanical or electrical sensor that physically or electronically detects that the reverser door or doors are not stowed completely? Does it detect when the door is only open slightly or do the doors need to open to the stops before the sensor detects it? If the cause of the deployment is due to faulty maintenance, will the sensor (i.e. in the Airbus) still be able to detect that the door is open?

I'm only speaking for Boeing here (although I'd be surprised if they other guys are meaningfully different). The T/R translating sleeve position is monitored electronically - prox switches, L(R)VDT, or a combination there of, and works exactly the same for an uncommanded deployment in-flight as it does on the ground (one of the last words on the Lauda flight recorder was "it's deployed" - not that the knowledge helped them much).

Sounds like some aircraft have a physical link to the throttles which slams the throttle lever down. That sounds like a nasty hospital bill but certainly better than crashing. What about in the Boeings and others i.e. CRJ?

Again, speaking strictly Boeing here but would be surprised if the others are meaningfully different. There is a mechanical linkage to something called the 'strut drum box' - the position feedback to this box serves a dual purpose by preventing advancement of the throttle above idle while the T/R is in transit and moving the throttle to idle if the T/R isn't in the commanded position. FADEC eliminated the need by doing everything electronically.
IIRC, the F100 crash occurred when the T/R deployed, the crew didn't realize why the throttle retarded and were able to force it back forward (basically they managed to override the safety device). In the aftermath we were tasked to determine if was physically possible for the crew to override the strut drum box on the Boeing aircraft that used that feedback (we determined they couldn't).

EMIT, again going by memory here, but I recall for Lauda is that there was a recurring T/R feedback fault to the FADEC, unrelated to actual cause of the deployment (although the troubleshooting of that fault might have contributed - one of the things we discovered during the investigation was the maintenance manual was junk). On the event flight, what we believed happened was the 'reverser stowed' prox sensor was slightly miss-rigged and would occasionally indicated the T/R wasn't stowed - this caused the auto-restow system to open the hydraulic isolation valve which would pull the T/R up tight, the prox sensor would indicated stowed, auto-restow would close the isolation valve. Normal vibration would then allow the T/R to move slightly, the prox sensor would indicate not-stowed and the cycle would repeat. On one of those cycles, when auto-restow opened the isolation valve - for reasons that were never firmly established - the directional control valve changed state to deploy... Whoops
The FDR was destroyed in Lauda and there was no usable data - so most of what we know is from the Voice Recorder and the non-volatile fault memory of the event engine FADEC. The FADEC was recording a T/R position feedback fault every time the T/R deployed - so there was a list of 'normal' deploy conditions - 500 ft., Mach .23, 800 ft. Mach .25, etc., then 24,000 ft. Mach .78.
The first time I looked at that fault dump was one of the darkest days of my life...

[galaxy flyer]

tdracer,

Was it that uncontrollable based on the aero analysis? Why I ask is the C-5 had a TR deploy in cruise and while certainly attention getting, it was controllable and didn't cause structural damage. Yes, at low altitude, not survivable.

[sheppey]

In your aircraft, are there definite indicators that the reverser has deployed,

I can only go by what happens when in- flight thrust reversal is actuated via the instructor panel under the main title of engine malfunctions. And that is the noise caused by the thrust reversal is quite unmistakeable just as you would expect when high reverse is used normally on the landing roll. In the 737 Classic simulator there is very little roll and yaw when the malfunction is initiated but that could be a fidelity issue. In fact it is easily controllable by the average pilot (if there is such an animal)

Basically it says to reduce thrust (Idle), Reduce Airspeed, Shut the engine down, control the aircraft with full rudder trim and sufficient bank to maintain heading.

I don't understand the "full rudder trim" point. What aircraft type are you referring to? Depending on the type of rudder trim but certainly in the 737 Classic it takes 29 seconds to operate the electrical rudder trim from neutral to against its stop in one direction and alone would be useless in terms of controlling yaw. Presumably considerable rudder would be required similar to engine failure in combination of aileron and flight spoilers if needed. Rudder trim would only be needed for subsequent prolonged cruise to reduce leg forces

[lomapaseo]

tdtracer

..... Normal vibration would then allow the T/R to move slightly, the prox sensor would indicate not-stowed and the cycle would repeat. On one of those cycles, when auto-restow opened the isolation valve - for reasons that were never firmly established - the directional control valve changed state to deploy... Whoops

Yes and the critical valve initially could not be found even among the parts being flogged off at the local flea markets. Later I heard that it was found after having been rebuilt in some native's hut.

Your comment about vibration reminded me of the reverser deployment that turned the aircraft over 360 degrees and was only discovered in the data days later with three of us reviewing the DFDR I still remember the remark by one of us as "holy-****" when the roll passed through 180 degrees Again flickering lights. When we tried to duplicate it it wouldn't deploy on the ground during runups (fail safes worked) until somebody gave a rap to the side of the engine with his hand and it deployed (specific vibration floated a fail-safe spring)

lessons learned which I will never forget !!! but now the regs call for three levels of safety .... but it's memories of what we all screwed up that I shall never forget

[Vessbot]

Originally Posted by galaxy flyer

tdracer,

Was it that uncontrollable based on the aero analysis? Why I ask is the C-5 had a TR deploy in cruise and while certainly attention getting, it was controllable and didn't cause structural damage. Yes, at low altitude, not survivable.

All else aside, I would expect a quarter of your thrust reversers deploying to have a less severe effect on controllability than half of them.

[tdracer]

Originally Posted by galaxy flyer

tdracer,

Was it that uncontrollable based on the aero analysis? Why I ask is the C-5 had a TR deploy in cruise and while certainly attention getting, it was controllable and didn't cause structural damage. Yes, at low altitude, not survivable.

Interesting that you brought that part up... There had been some partial in-flight deployments on 747s over the years, and while I suspect it was certainly exciting they landed safely. But aerodynamically the engine installation on a big twin is quite different from a quad.

Initially, the aero analysis said the 767 should be controllable - and in fact they did flight test it during the original aircraft cert - at a stable 200 knots, 10k, engine at idle, they initiated a reverser deployment and were able to control the aircraft.
So we went to the wind tunnel - this was during the development of the 777 and they had a half model of a 777 with blown reverser. The 767 aerodynamically looks very much like a 777, just smaller, so we changed the 777 model from the normal landing configuration to a 'clean wing' cruise with a deployed reverser.
Before we started the test, there was an aero S&C guy that kept proclaiming it was controllable and the flight crew had blown it - going so far as saying he go on a flight test where they duplicated the in-flight deployment. But as the testing progressed he started getting really quiet. By the time the testing wrapped up a week later he wouldn't talk about it...
When they updated the simulation with the updated aero characteristics from the wind tunnel testing, it quickly became apparent the flight crew never had a chance. When they tested the actual deployment scenario, a few pilots could save it when they knew it was coming and it happened during daylight with outside visual queues. When did it at night in the clouds (the actual Lauda conditions), no one could save it even when they knew it was coming
When we finished up with the instrumented wind tunnel testing, they did a flow visualization test - and the result was dramatic. Being a propulsion guy, I didn't really understand all those aero S&C coefficients and such, but I could readily understand the visualization of separated flow. Nearly the entire upper wing was separated, along with a good share of the tail surfaces.

Yes and the critical valve initially could not be found even among the parts being flogged off at the local flea markets. Later I heard that it was found after having been rebuilt in some native's hut.

IIRC, they finally got the Directional Control Valve (DCV) after offering a rather large reward, but it quickly became apparent the DCV had been disassembled then carelessly re-assembled to collect the award, destroying any possible evidence of what caused it to change state.
There are several theories as to what did cause the DCV to change state - one TV show I saw postulated it was a short circuit but most of us involved in the investigation think that unlikely (during testing the circuit breaker always popped before the valve moved). My personal theory is was a hydraulic hammer effect from the repeated cycling of the isolation valve.

[galaxy flyer]

That's very interesting, tdracer. IIRC, they were a former EAL crew on contract to Lauda. Knowing a bit about the efflux problem on the C-5 (engines very inboard compared to the 747) and the mid-ish span location on a twin, I understand what the testing showed.

[Goldenrivett]

Hi sheppey,

I don't understand the "full rudder trim" point. What aircraft type are you referring to?

A320. From FCOM below:
"If Buffet:
The warning alone, without buffet or vibration, may be a false warning.
MAX SPEED... 240 KT
ENG MASTER... (affected engine) OFF
If reverser is actually deployed:
RUD TRIM... FULL R (L)
CONTROL HDG WITH ROLL"

In Normal Law, it may be due to fact there is no feed back through side stick of the amount of aileron required and possibility of running out of aileron authority unless lots of rudder is applied.

[EMIT]

Lauda

Very interesting details, tdracer!

As far as training is concerned, when the accident happened, apparently the simulators still had a failure mode of REV deployed inflight. After the modification of the real 767 aircraft system, so that REV deployment inflight was really moved into the 10 to the minus umpteen regime, the failure mode was removed from the simulators. By the time I had a sim session and wanted to "try" whether an inflight deployment would be recoverable, it was no longer possible to have the failure simulated.
Reading the replies from tdracer, I assume the simulator fidelity of the REV deployed inflight scenario would not have been realistic anyway.

[Jwscud]

We used to train this in the sim on the Lear 45. The thrust lever automatically snapped to idle if the reverser deployed and we were trained immediately to shut the engine down.

The aircraft was marginally controllable and between stick shaker and V2 until you shut it down. You could just about attain a positive rate. One wily TRE also trained knowing the location of and instinctively pulling the deploy CB (4th back on the bottom row since you ask ) as a backup.

It was a good sim exercise, but it was seriously demanding even if you were in "sim" mode, and over beers we wondered whether it would be survivable if it happened for real.

[paradoxbox]

So do people think it's worth the risk of setting power to idle without knowing for certain if a reverser is the cause?

Some people have mentioned that the vibrations and noise would indicate a reverser problem, but in the aircraft I have flown, spoilers also cause a lot of vibration and noise and yaw/roll action. Without being able to see the back of the aircraft I think it would be hard to differentiate between an uncommanded reverser and an uncommanded assymetric spoiler deployment very quickly - unless you had some kind of experience, sim or otherwise, that let you judge otherwise.

I agree with the guy who mentions combat / aerobatic training as being extremely useful in this kind of situation. I think that a typical civilian crew that has "positive g, altitude" on their minds 24/7 may be in trouble. Unloading the aircraft (0g) would probably buy you a lot of time.

I wonder how unloading the aircraft (0g / 0 or low AoA) would affect things in a thrust reverser deployed situation. I do not trust sims to get this right but I also do not want to try it in a real aircraft, thankyouverymuch. Any test pilots with parachutes want to give this a shot?

[underfire]

Interesting. I seem to remember a few aircraft that would use deliberately use the thrust reverse in flight, the DC8? The Concorde to get from SS speeds.
It is commonly used on a few military aircraft, I remember the ride on a C17 diving in...without warning, it does feel as if you have been shot down. you could not do that with paying customers, at least ones you like.

[lomapaseo]

So do people think it's worth the risk of setting power to idle without knowing for certain if a reverser is the cause?

Risk ??? of what?

In an actual event, the crew reported flickering unlock lights during initial climb. They discussed the possibility of false warnings (loose connections, wiring etc.) but decided it would be prudent to follow the FCOM and pull the affected engine back to idle. The unlock light ceased so they once again advanced the throttle to maintain climb profile. The reverser then deployed and rolled the aircraft over to 180 deg in a matter of seconds. 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.

[underfire]

In an actual event, the crew reported flickering unlock lights during initial climb. They discussed the possibility of false warnings (loose connections, wiring etc.) but decided it would be prudent to follow the FCOM and pull the affected engine back to idle. The unlock light ceased so they once again advanced the throttle to maintain climb profile. The reverser then deployed and rolled the aircraft over to 180 deg in a matter of seconds. The airforced trained PF then completed the roll through 360 deg and requested clearance to return to the field.

WOW, what ac was this?

[Vessbot]

Originally Posted by underfire

Interesting. I seem to remember a few aircraft that would use deliberately use the thrust reverse in flight, the DC8? The Concorde to get from SS speeds.
It is commonly used on a few military aircraft, I remember the ride on a C17 diving in...without warning, it does feel as if you have been shot down. you could not do that with paying customers, at least ones you like.

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.

[TURIN]

First of all, thank you to all the contributors to this thread, I've not been this gripped since the Concorde thread started.

lessons learned which I will never forget !!! but now the regs call for three levels of safety .... but it's memories of what we all screwed up that I shall never forget

There for the grace of...etc.

I'm still amazed at how languid some pilots/maintenance organisations are about T/R faults. After the Lauda incident, which I remember well, I had occasion to lock out faulty T/Rs on Boeings and Airbus' and the attitude of some was quite disturbing.

[blind pew]

In flight reverse

On the trident we used in without any problems although it wasn't very pleasant down the back.
Think the limit was around 11,000 but we were allowed to use emergency reverse in the flare.
The dc8 was reverse on inners.
In the mid 70s Concorde had an engine go into reverse in cruise. It returned to lhr, got another crew and after trouble shooting was sent on its way only to have the same fault after rotation.
It crossed over my cottage, near to Newbury, so low that I rushed outside thinking there was a crash; about 40 track miles; it only got above three grand crossing the Bristol Channel. Iirc the engineeer got some sort of award for shutting the engine down velly quickly.

[Chu Chu]

It sure sounds like a good thing the DCV was carelessly reassembled -- if it had be done carefully but with one "small" oversight, it might have really ruined someone's day . . .