Boeing amend 737 de-icing procedures & propose aircraft mods after near-stall inciden

[captplaystation]

Scary stuff


Norwegian near-stall to spur 737 modification (http://www.flightglobal.com/news/articles/norwegian-near-stall-to-spur-737-modification-393527/)

[Dufo]

Scary indeed, how after almost 50 years of operating this type, we are inventing new deadly procedures.

In my opinion, horizontal stabilizer should be deiced manually, without fluid touching the most important flying surface. It is inverted airfoil anyway. Anyone deicing the bottom part?

[carlstraub]

"But it notes that, during the initial ascent, there was “no attempt” to disengage the autopilot, autothrust or manually adjust the stabiliser trim to nose-down."

UNBELIEVABLE !

I've been retired a long time but cannot imagine current crews are so "auto dependent"

[lomapaseo]

but because PT2 or whatever is called tube if not properly heated and filled with ice crystals

This is not applicable to the problem that is being discussed. It might lead one to conclude that adding more heat would solve the problem. It's the residence time of the crystals against the probe. I liken it to light snow bouncing off a car windshield until you heat the windshield and a light layer of water-slush needs your wipers to become clear.

[SetStandard]

carlstraub

Have you flown the 737? Applying force to the control column (especially 920N) should disconnect the Autopilot.

Sounds to me like the crew averted an attempted Autopilot induced stall. I wouldn't call that "auto dependent"

[Captaintcas]

No Safety recommendation yet from Boeing, so all pilots should apply normal published procedures until adviced.

[Finn47]

SHT preliminary report here (.pdf):

http://www.aibn.no/Luftfart/Rapporter/13-18?iid=14939&pid=SHT-Report-Attachments.Native-InnerFile-File&attach=1

[Green Guard]

to compare PT2 heating with I liken it to light snow bouncing off a car windshield until you heat the windshield and a light layer of water-slush needs your wipers to become clear. is like to compare apples and oranges

[flarepilot]

very interesting incident and I'm glad to hear that a safe landing was accomplished.

set standard is right...enough force on controls should disengage the autopilot

often times in situations like this, the perfection of translation can come into question.

haven't we learned that in icing situations one should hand fly? of course that pre supposes that one knows how to hand fly on instruments well...

I always hated the deicing procedure for the tail of the 737...what a pain!

[safetypee]

There appears to be similarities with control freezing or control difficulties due to de-icing fluids experienced several years ago.
The de-icing fluid manufactures have implemented changes to the fluid composition, but there is still some residual risk of ‘evaporative dry-out’ and ‘rehydration’ during flight, resulting freezing/blocked control systems. Many manufactures require inspection and/or wash-off action to remove accumulations of ‘gel’ from previous fluid application.

De Ice Fluid residue (Gel) in a control system bearing – not B737.

http://i41.tinypic.com/2zz8cqc.jpg

[captplaystation]

Indeed, I had colleagues back in 2001 that experienced something not dissimilar on a Fokker 100 (no unusual attitudes, but loss of longitudinal control )

[flarepilot]

wondering out loud if new 737 icing thing has any connection to russian 737 crash.

I really doubt it, but

[KBPsen]

It's seven strokes to one:

[happybiker]

Fluid residue and rehydration can be a serious issue as highlighted by safetypee. A lot of work was done by the CAA 10 years ago following a number of elevator freezing incidents on the 146. It is worthwhile reviewing the CAA winter operations web site to refresh the memory on some of the hazards that are out there each winter. De-Icing and Anti-Icing Fluid Cautions | Air Operations | Operations and Safety (http://www.caa.co.uk/default.aspx?catid=2520&pagetype=90&pageid=14018)

[Jwscud]

It's a pretty concerning incident. The sound of the trim wheel continuously running has been drummed into my head from Day 1 in the sim as a major cue that "something is not right"

That photo from safetypee does not fill my heart with warm feelings!

[Basil]

SHT says the pilots’ efforts managed to bring the nose down. But it notes that, during the initial ascent, there was “no attempt” to disengage the autopilot, autothrust or manually adjust the stabiliser trim to nose-down.
<<manually adjust the stabiliser trim to nose-down>> or would that be to nose-up if they had a jammed elevator?
Don't know how the B737 system works.

[Dropp the Pilot]

The aircraft subsequently landed safely after a 30min hold for the crew to conduct a systems check.

That is a very polite way of describing the practice of checking one's underpants for evidence of incontinence. Was this written by a British chap, perhaps? A model for us all.

[Baja]

I confess ignorance on such matters but that doesn't stop me from asking a stupid question.

If this is "de-ice" fluid, then it seems counter intuitive that it would freeze. What causes this fluid to freeze?

[tightcircuit]

Basil,

The trim on the 73 (and I think all boeings) works by adjusting the angle of attack of the whole horizontal stabilizer. Trimming nose down with a stuck elevator will lower the nose. In fact in some circumstances, because of the large power-pitch couple the elevator is not powerful enough to lower the nose and as a result the stall recovery drill includes using nose down pitch trim if nose down elevator does not have the desired effect:

• Initiate the recovery:
• Hold the control column
firmly.
• Disconnect autopilot and
autothrottle.
• Smoothly apply nose down
elevator to reduce the angle of
attack until buffet or stick
shaker stops. Nose down
stabilizer trim may be
needed.

This is also the case for a nose high upset recovery:

• Disconnect autopilot and autothrottle
• Apply as much as full nose-down
elevator
• * Apply appropriate nose down
stabilizer trim
• Reduce thrust
• * Roll (adjust bank angle) to obtain a
nose down pitch rate
• Complete the recovery:
- When approaching the horizon, roll to
wings level
- Check airspeed and adjust thrust
- Establish pitch attitude

Exactly the same applies on the 757 and 767

In this incident the crew do not appear to have attempted either of these drills.

[safetypee]

Baja, in very simple terms, de-icing fluid is a mix of glycol and thickening agent/dye. If this mix is retained in a crevice or bay, and/or not removed by the airflow during flight – parked overnight, then the glycol can evaporate. The residual powder may absorb water and expand by a factor ~x10, forming a ‘gel’, which can freeze.
Thus there can be a period without incident due to the absence of liquid water (sub zero / cold dry operations), but later exposure to wet conditions re-hydrates the ‘gel’ which may freeze during flight.

[captplaystation]

Hey Dude, don't know which procedures your " independant preferred receiver" of you the "independant service provider" (I hesitate to use the word employer) is using but I have never seen any of yours messing with flaps when they have been de-icing around me, and certainly the only mention of flaps we have is to consider large accumulations of slush/crud when deciding to retract (or not) after landing, and to delay deploying them until after de-icing (or approaching departure runway if taxi-way severely contaminated.)

Cycling horizontal stab is another thing, but it seems with the new procedure (which we have adopted) that will now be a thing of the past.

[ImbracableCrunk]

Re: Cycling of the flaps

I don't think many US carriers follow this procedure.

[ManaAdaSystem]

Never heard about cycling the flaps after deice, and this procedure is not in any of our Boeing manuals.
Home cooked procedure for sure.

[Skyjob]

Cycling of flaps is only achieved in Freezing conditions (<+3C), not Icing conditions (<+10C), when deicing has occurred, etc... Only with a set of worst criteria.

It is to ensure freedom of movement of controls and high lift devices and so on, secondly this also allows a flight control check if a jet has been parked for several days in sub temperatures on first flight back into service (we all seen the jets parked up and rotated throughout winter use).

Now I would like to suggest that any such check should be done only immediately prior to departure, such as at a holding point, therefore not halving the holdover time by up to 50% as riven by independent research. Unfortunately this seems not the case in practise.

[shaftsburn]

Recycling of flaps is definitely an operator specific item.
My first 737 company never mentioned it.
Second 737 company said it should be done if temp less than -18C.
Third 737 company has it in adverse weather supp, but no mention of a temp.

[Aluminium shuffler]

It used to be 737 post de-icing procedure to cycle the ailerons and elevators to full deflection in both directions in the before take off checks to drain de-icing fluid out of the balance cavities until just a few years ago, when the procedure was removed because the drain holes were drilled larger. I wonder if that old practice also had an unknown benefit of moving any slush off the control mechanisms, so masking the issue and it only coming to light recently. I still cycle the controls three times anyway - it might be un-necessary, but it costs me nothing.

[LNIDA]

Snow and frost tends to form on the upper surfaces, underwing frost is normally down to cold soak fuel freezing the moisture in the air it is in contact with, so highly unlikely for any to form on underneath of the stab? just saying....

[Aluminium shuffler]

I thought we were discussing ice accumulation on the PCUs, not stab frost. The three cycles of the elevator was for clearing fluid that had got inside the elevator balance cavities, nothing to do with the stab.

As for flap cycling, it appeared on my company's Freezing Conditions before take off checklist a year or so ago. Flap selection is deferred until entering the runway, but I can't remember off hand when the cycle is required, before taxy or immediately before flap selection. I think it's the latter, which would be more prudent.

[NinER fIveR]

Boeing 737-800 FCOM 1 Supplementary Procedures SP 16.6

Before Taxi

Flaps ............................................................ .................Check F/O
Move the flaps from Flaps up to Flaps 40 back to Flaps up (i.e., full
travel) to ensure freedom of movement.

If taxi route is through ice, snow, slush or standing water in low
temperatures or if precipitation is falling with temperatures below
freezing, taxi out with the flaps up. Taxiing with the flaps extended
subjects the flaps and flap drives to contamination. Leading edge devices
are also susceptible to slush accumulations.

Call “FLAPS ___” as needed. C

Flap lever ................................................ Set flaps, as needed F/O

[Skyjob]

Adverse Weather Conditions

Before Taxi Procedure

Do the normal Before Taxi Procedure with the following modifications:

GENERATOR 1 and 2 switches ........................................................ ON
Normally the IDG’s will stabilize within one minute, although due to cold oil, up to five minutes may be needed to produce steady power.

Flight controls ............................................................ .................... Check
An increase in control forces can be expected at low temperatures.

In freezing conditions:
CAUTION: The flap position indicator and the leading edge devices annunciator panel should be closely observed for positive movement. If the flaps should stop, the flap lever should be placed immediately in the same position as indicated.
Flaps ............................................................ .............................. Check
Move the flaps from Flaps Up to Flaps 40 back to Flaps up (i.e., full travel) to ensure freedom of movement.
If remote de-icing is planned or if the taxi route is through slush or standing water in freezing conditions, or if precipitation is falling in freezing conditions, taxi out with the flaps up. Taxiing with the flaps extended subjects the flaps and flap drives to contamination. Leading edge devices are also susceptible to slush accumulations.

Flap lever ................................................. Set takeoff flaps, as needed

[ManaAdaSystem]

None of those procedures are in our manuals, but operators may get Boeing to include operator spesific procedures in the FCOM, but Boeing do not take any responsibilty for those procedures.

[NinER fIveR]

That was actually from the Generic Boeing Manual. It may be the other way around though - some companies have certain procedures scrapped or modified based on their operational needs.

[RAT 5]

Before taxi.

Flaps ............................................................ .............................. Check
Move the flaps from Flaps Up to Flaps 40 back to Flaps up (i.e., full travel) to ensure freedom of movement.
If remote de-icing is planned or if the taxi route is through slush or standing water in freezing conditions, or if precipitation is falling in freezing conditions, taxi out with the flaps up. Taxiing with the flaps extended subjects the flaps and flap drives to contamination. Leading edge devices are also susceptible to slush accumulations.

This always gave me trouble. My limited common sense wanted to adjust this; but had some resistance from F/O's & LTC's. The above statement is what was on my checklist. It was snowing at the gate, but de-icing was quite a distance from the apron; thus we needed to taxi a long way before being sprayed. I suggested we taxi out with flaps up and after de-icing do the F40 exercise thingy. Oh no. It was met with horror. I surmised that to open the flaps in snow, and then close then again, could trap snow in the gaps which would be compressed and not be influenced by the de-icing fluid. It was insisted that we exercise the flaps before taxi, de-ice and then taxi with flaps up to the holding point.
Opinions please. Am I mad, or what?

[Aluminium shuffler]

We are deiced flaps up (unless they were contaminated on previous ldg, in which case they're left at 15 until deiced). I just couldn't remember whether the C/L said to do the run check before taxi or at the hold point immediately before selecting TO flap, likewise the controls check. I'll be delaying the checks to the hold point so that the list can be done in order but such that I'm not shedding the diecing fluid by sloping the controls until the last safe moment...

If remote deicing, then I'd run the before taxi c/l after deicing, before the second taxi. Keeps everyone happy!

[captplaystation]

RAT 5

Nope, but common sense is a resource that is in limited supply in aviation. . . as in life in general.

[Aluminium shuffler]

Capt, you're sadly right. Too many trainers force line shags to strictly follow unrealistically literal (mis)interpretations of SOPs, which themselves are oft flawed or poorly written.

[ManaAdaSystem]

Not mad, but smart! I remember this procedure was in our books, but was only to be done in extremly cold wx, ie -30 celsius-ish. Same with the nose wheel steering left right, and idle for several minutes to let the IDGs stabilize. And having the ground crew heating the engines before starting.

I believe the full flap exercise is to make sure the flaps are not frozen in any way, but for that very reason I would deice before trying to move them.
We taxi to and from deice with flaps up, and only set take off flaps right before lining up. This way we don't expose the untreated areas to icing conditions before we take off.

[ManaAdaSystem]

Maybe we should as the Finns how they do it? :)

EFKT 082050Z AUTO 30001KT CAVOK M37/M41 Q1015

[captplaystation]

Not much ice/moisture there, dry as a Nuns - - - - I would wager. :eek:

[ManaAdaSystem]

It's not about the temp, it's all about the spread. And no, it's not that dry.

[BOAC]

"And no, it's not that dry" - actually it is pretty 'dry' - the same reason you do not necessarily require engine anti-ice at 40 and below!

[ManaAdaSystem]

Not correct. You get some of the worst icing conditions on ground at extremely low temps and foggy conditions.

EFRO 090850Z VRB03KT 040V120 5000 BR FEW001 M24/M26 Q1021=

Minus 24, and close to fog.

EFRO 090823Z 0909/1009 VRB03KT CAVOK TEMPO 0909/0912 1200 BCFG SCT001

[BOAC]

As Capt PS says, -24 is not -37! I would guess there would be no de-icing at EFKT during the day once any overnight condensate had been cleared.

[ManaAdaSystem]

If you say so, and that is why this is happening?

EFKT 091020Z 00000KT 9999 BKN003 M35/M39 Q1021 R16/72//75

EFKT 091000Z 0910/0917 VRB03KT CAVOK
TEMPO 0910/0912 4000 BR BKN002

[BOAC]

Yes, you get clouds in the arctic.

Let's rewind to your post #39. I take it you are seriously expecting to need to de-ice your nun here regardless?

EFKT 082050Z AUTO 30001KT CAVOK M37/M41 Q1015

[ManaAdaSystem]

Mist and fog does not form in bone dry conditions.
All i said was: No, those are not very dry conditions, and the weatherman agrees with me.

[RAT 5]

Is it not the case that you De-ice to get rid of contamination and you anti-ice to prevent contamination. The get rid of scenario is decided by sight & touch. The prevent scenario will depend on prevailing conditions. At -37c and CAVOK there might not yet be contamination. Will there become so is a judgement that thankfully I've not had to make. -20c was my worst.

[hawk37]

“Under certain circumstances it is possible that the input arms [to the power control units] may be exposed to fluid which in turn freezes solid and blocks [them],” says SHT, adding that Boeing was “not aware” of this potential problem before the investigation."

Can someone please tell me if this is the same as saying the aircraft experienced a jammed elevator, ie jammed yoke?

Or does the input arms exposed to freezing fluid somehow instead cause the elevator to deflect in one direction and not return?

[hawk37]

No one around that can shed any light on how a power control unit being blocked with ice can cause an aircraft to pitch up so aggressively in this manner?

Tightcircuit?

Safetypee?

[tightcircuit]

Its all to do with the power/pitch couple Hawk. The elevator was jammed but the pitch trim system (which adjusts the whole horizontal stabiliser) was still working. The report says that the autopilot adjusted the trim system nose up. I am not sure of the reason for that at that stage of the approach but it was probably just a response to reducing speed. This would normally have been countered by automatic nose down elevator but that was not available. As a result the nose rose and speed reduced further causing the auto thrust to increase power. The pitch up moment due to thrust increase is very strong in the 73 and can in extreme circumstances even overpower the elevator completely. In this case they had no elevator anyway so the nose would have kept on going up. The automatics would have been completely confused at that point and may even have added more nose up trim but the report doesn't tell us that.

Boeing altered the drills for stall recovery and nose high upset recovery to include the use of nose down pitch trim several years ago after a similar near disaster. In that case the elevator was not jammed but nevertheless the application of full power at very low speed caused the nose to pitch up uncontrollably.

[rbaiapinto]

the application of full power at very low speed cause the nose to pitch up uncontrollably

Reading this post, can't help to think about the Khazan accident!
Even more in a 23,5k engine in a 500.

[safetypee]

A speculative view assuming that the elevator was blocked with the autopilot engaged.
As speed reduced prior to GS capture the autopilot was unable to move the elevator to provide the necessary nose-up change in attitude, this inability would probably be interpreted as insufficient elevator response or a residual out-of-trim force by the autopilot, thus a trim follow-up input was commanded. The trim input continued because the 'blocked' force was not reduced.

As an analogy, the situation is like a pilot overpowering the autopilot nose down (hold the stick fixed when decelerating) resulting in an autopilot trim command; trim has far greater effect on pitch control than the elevator, thus the flight path deviation (most autopilots sense stick overpower and should disengage).

This doesn’t exclude a rare technical failure causing the trim to run, however there are multiple safe-guards against this, whilst for de-icing fluid freezing there are few if any.

When disengaging the autopilot the crew would have experienced the out-of-trim force requiring forward stick, which they were applying, but against a powerful trim input.
The key aspect of recovery is in recognising that a trim input is required (after disengaging the autopilot), but this is not always self-evident as indicated by similar low speed incidents and accidents.
Does modern training impart an expectation that aircraft will always be in trim, or not to appreciate stick-force cues (excluding certain FBW types)? Do pilots expect that centring the stick – by feel, to achieve zero force, will enable a low speed recovery – little awareness of the trimmed condition; is trim taught as a control system?

It would be interesting to know what aircraft types the incident pilots flew previously. I recall a similar incident in a conventional aircraft (not 737) where a nose-high overpower condition was induced due to poor control-feel awareness (previous / extensive FBW experience), yet the recovery was aided by the military FBW experience including controlling surprise and of low speed nose-high wing-over manoeuvres.

A de-icing fluid incident is described here: http://www.aaib.gov.uk/cms_resources.cfm?file=/dft_avsafety_pdf_022992.pdf but note that the BAe146 has a servo-tab control system and if this is blocked there is some ability to move the elevator directly – an advantage of ‘wire’ control ;)

[hawk37]

All very muddy to me, I guess we would need a detailed report to try to understand. Stuck elevator, ok that makes sense though the report didn't specifically refer to it as stuck/jammed. The power pitch couple I understand. However if the autopilot is in some kind of vertical mode, and the aircraft started to deviate from that and pitch up, I would expect the autopilot to gradually provide a nose down force to the elevator to command nose down. If that force is over a certain threshhold (which it would soon be if the elevator was jammed), then surely the trim system would start to trim nose DOWN, not up.

Must be more to this picture of course

Isn't there a simulator session for jammed elevator procedure that pilots would occasionally see?

thanks for your replies Safetypee

[tightcircuit]

The thing is Hawk that if the autopilot loses use of the elevator then who knows how it will react? I believe that the autopilot uses the elevator the way we were taught to, ie select an attitude and then trim. Autopilot trim is a bit coarse and at a slower rate than normal trim. I suspect it just couldn't keep up with a situation it had never been programmed to deal with. Add the actions of the auto thrust which knows nothing of pitch or trim and was only concentrating on speed, throw in a bit of turbulence and system lag perhaps. There you have it.

[framer]

So if I deice my 737 tomorrow, my manuals call for horizon stab full nose down. Is there a Boeing doc saying it should be otherwise. That seems to be what the report says. Have any of you had bulletins or manual changes?

[Thrush]

There is a Bulletin which came out back in October, I think. Your lot must be slow at publishing them to crews. Especially on such a critical subject...

Prior to de icing, set stab to calculated take off position.

[Lord Spandex Masher]

...or 5 units NU!

[framer]

Just searched and I can't find the bulletin.
Can somebody post a link ? Cheers

[Skyjob]

I can't find the BAOM reference, but here is the text from it:


IssueDate: October 16, 2013
Subject: Supplementary Procedure, Adverse Weather, De-icing/Anti-icing
Reason: To inform 737 flight crews that the stabilizer trim must be set to the calculated takeoff setting prior to de-icing the airplane.

Information in this bulletin is recommended by The Boeing Company, but may not be FAA approved at the time of writing. In the event of conflict with the FAA approved Airplane Flight Manual (AFM), the AFM shall supersede. The Boeing Company regards the information or procedures described herein as having a direct or indirect bearing on the safe operation of this model airplane.

Background Information
In the winter of 2012, a 737 customer reported an un-commanded pitch up on final approach followed by reduced elevator response rates during the upset correction. Root cause analysis determined that the reduced elevator rates were caused by ice accumulation on multiple elevator Power Control Unit (PCU) input levers. De-icing prior to takeoff caused fluid to enter the aft area that accumulated and froze on the PCU input levers during flight, which is believed to be the cause of this incident.

The current de-icing Supplementary Procedure calls for the stabilizer to be trimmed to full APL NOSE DOWN (AND), which positions the stabilizer leading edge to the full leading edge up position. Analysis and testing discovered that this position maximizes the amount of liquids that enter the tailcone near the elevator PCUs due to the opening in the tailcone sides left uncovered by stabilizer’s extreme leading edge up position. If enough fluid accumulates and freezes around the PCU input levers their movement may be restricted or jammed.

Follow-on analysis and testing of various stabilizer positions during de-icing demonstrated that setting the stabilizer trim to the calculated takeoff setting (in the green band) prior to de-icing significantly diminishes the amount of fluids entering the tailcone, reducing the likelihood of a similar event caused by ice accumulation on the PCU input levers.

Operating Instructions
Prior to de-icing the airplane, set the stabilizer trim to the calculated takeoff setting, rather than Full APL NOSE DOWN.

Administrative Information
Insert this bulletin behind the Operations Manual Bulletin Record page in Volume 1 of your Operations Manual. Amend the Operations Manual Bulletin Record to show bulletin "In Effect" (IE).
This bulletin will be cancelled in a future revision of the FCOM.
Please send all correspondence regarding Flight Crew Operations Manual Bulletin status, to the 737 Manager, Flight Technical Data, through the Service Requests Application (SR App) on the MyBoeingFleet home page.

[737ngpilot]

Our procedure call for Take Off trim setting...typically around 5 units

[ManaAdaSystem]

Why, stator vane, you don't set your stab before you take off?

The reason for trimming full nose down before deice was to prevent deicing fluid from entering the tail cone. Now we have to set it to take off position for the same reason?
Funny aircraft, this jurassic design still manages to surprise us after some 60 years of operation.

[heavy.airbourne]

(...) this jurassic design still manages to surprise us after some 60 years of operation.
...and it still kills people! :ugh:

[Lord Spandex Masher]

Why, stator vane, you don't set your stab before you take off?

The reason for trimming full nose down before deice was to prevent deicing fluid from entering the tail cone. Now we have to set it to take off position for the same reason?
Funny aircraft, this jurassic design still manages to surprise us after some 60 years of operation.

No, it was to allow the horizontal stabiliser to drain. Which is why you still set full nose down overnight but 5ish for deicing.

[putz]

Nope, it was to minimize the amount of fluid entering the elevator balance bays.

[hawk37]

Never flown the 737, nor anything that big. Does the certification procedure require these jets to demonstrate some controllability with a jammed elevator?

And if so, what could have been the difference here where the aircraft was almost not controllable at all?