Norwegian B737-800 almost stalls at EFKT?

[norpilot1]

Finnish aviation news site lentoposti.fi has posted a news item on a near stalling incident at Kittilä (EFKT), in Northern Finland:

Suomen onnettomuustutkijat avustamaan norjalaisia Kittilässä tapahtuneen vaaratilanteen tutkinnassa | lentoposti.fi (http://www.lentoposti.fi/uutiset/suomen_onnettomuustutkijat_avustamaan_norjalaisia_kittilassa _tapahtuneen_vaaratilanteen_tutk)

Short summary in English:

The B737 registration LN-BYM (NAX5630) was flying the ILS 34 approach to Kittilä (EFKT), in Northern Finland, when at 3250 ft on the glideslope with autopilot engaged and flaps partially extended the aircraft started a sudden climb with full thrust, at the same time letting the airspeed decrease near stalling speed. The aircraft climbed 1500 ft during the incident before the pilots regained control.

The next approach was successful and no technical faults have been found so far in the aircraft.

Norway's Accident Investigation board (SHT) is investigating this incident together with it's Finnish counterpart, NTSB and Boeing. SHT has the FDR and some aircraft parts in its possession.

Some facts which may be irrelevant in this case: EFKT34 ILS G/S is 3.4 degrees steep, with typically tailwind conditions (no precision approach to 16 available), 2.5 km runway and often slippery runway conditions. Depending on time and day the airport may be manned either with an ATCO (TWR) or an AFIS person.


Any ideas from B737 pilots?

[Hotel Charlie]

Maybe sombody pushed TOGA inadvertently ...

[Hotel Tango]

Yes...it shouldn't take 1500 feet to "regain control".

Hmm, it was a B737 not a DC-6. I would imagine that with the "wtf" moment combined with possibly a high rate of climb at high thrust it could easily take 1500 feet to recover.

[rudolf]

No TOGA below 2000ft RA on the 737 and Kittila has an elevation of about 600ft. Also, I would be surprised to see full thrust during a normal go-around; false glide-path?

[Craggenmore]

The aircraft climbed 1500 ft during the incident before the pilots regained control.

Was the journalist in the flight deck..?

[misd-agin]

"A/P off" and unexplained climbs are over in seconds.

[sky jet]

I have seen this once years before in a -900. I wrote it up after the landing and followed up with maintenance. No fault was ever found and we also came to believe it was due to a false glide slope. In level flight approaching the marker the G/S needle had just wiggled alive and the aircraft went off to the races in a climb. It can be easily controled by selecting AP off, but the flying pilot was a brand new ab initio FO. We climbed about 500 feet before I got him to release the yoke and let me get it back to altitude. Stuff happens. That is why they still put us in the pointy end.

[JanetFlight]

BYM???...DYM for sure.

[RAT 5]

Were they in Glide Slope capture? What was the MCP Alt set to? Did they engage VNAV or LVL CHG before G.S. capture with MCP set to MAA?

[BOAC]

None of which should cause a steep climb and an 'airspeed decrease near stalling speed.' It remains to be seen whether this was a system or human failure and with the FDR etc it should not be long.

My first look would be at manual throttle, poor speed control and autothrottle 'armed' kicking at 1.3Vs with TOGA and autopilot disconnect followed by big wide Scandi eyes, but as I say, it will all be in the data.

[Checkboard]

I had an incident, capturing the glide slope from above while in open descent, where the altitude window switched to the default upper limit (50,000'). The aircraft pitched up to climb, but as the power was at idle for the previous descent the aircraft washed off about 30 knots within a few seconds...

[M609]

Was the journalist in the flight deck..?

No, but the AIBN has some details......

From the Norwgian AIBN:

ENGLISH:

On the 26th of December 2012, Norwegian Air Shuttle with flight number NAX5630 serviced a regular passenger flight from Helsinki airport Vantaa (EFHK) to Kittilä airport (EFKT) in Finland. The aircraft in use was a Boeing 737-800 delivered new to Norwegian Air Shuttle in 2011.

Enroute the flight was uneventful and the aircraft was established on the approach to runway 34. When the aircraft passed approximately 3,250 ft on the instrument approach (ILS) with flaps 5 and autopilot engaged, the aircraft started an unintentional steep climb with full engine power. The aircraft climbed approximately 1,500 ft and the airspeed dropped so that the aircraft came close to a stall. However the pilots managed to regain control of the aircraft. Some later a new approach was executed and an uneventful landing was made in Kittilä. After the incident, the aircraft has been flown on a test without any remarks.

The Accident Investigation Board Norway has access to the aircraft digital flight data recorder and some aircraft components. In corporation with the US National Transportation Safety Board (NTSB), Boeing and the airline, it is initiated an ongoing investigation as a result of the incident.

......and it´s just as likely to be wide Baltic or Polish eyes these days. :E

[ant1]

I had it twice on an -800, GS -roller coaster type- captured far beyond what the good practices dictate. The thing started to climb like it would never end and indeed the A/T were doing a very poor job at maintaining speed. The outcome: A/P disengaged and reestablished ourselves on the right path.

[GlueBall]

This wouldn't have happened if the flying pilot had kept his paws on the wheel and thrust levers during final approach. Any Boeing A/P and A/T can INSTANTLY be switched OFF, or overridden with muscle power. :ooh:

[ant1]

Well, you know, it's the WTF is it doing sindrome vs instant click-click, click-click.

[HURZ]

Many moons ago I learned that there is a limit like max 25 nm out for the LOC and 17 nm for the GS. This is imho the earliest where you can receive reliable signals. I experienced something similar on the 747 as well but if you keep your hand on yoke and thrustlevers you just switch the damn A/P of instead of watching it trying to kill you.
If you don t understand the automatica at any moment revert to manual flying :eek:

[latetonite]

Regardless of the aircraft doing things, there were two pilots on board, no?

[Teg Bahadur]

Checkboard
*
Join Date: Aug 1998
Location: Ex-pat Aussie in the UK
Posts: 2,679
I had an incident, capturing the glide slope from above while in open descent, where the altitude window switched to the default upper limit (50,000'). The aircraft pitched up to climb, but as the power was at idle for the previous descent the aircraft washed off about 30 knots within a few seconds...


Had it happened to me once too! Investigation after my incident report was inconclusive eventhough cabin crew testified that a pax was messing around with his cell-phone. The only conclusion the techies drew was " electronic transient " or something like that.

[Matey]

Rudolf,

TOGA is indeed available above 2000RA on later B737NG with flaps not up or G/S captured. If the AFDS remained engaged then it would have commanded only sufficient thrust for 1-2000fpm climb, and a speed of 182 kts with flap 5. The cause lies elsewhere I think.

[misd-agin]

If it's related to GS capture, MCP altitude setting, etc, how does it even change 100' before the CVR hears 'click' and 'click'(sounds of A/P and A/T disconnecting)?

One of the worst things to watch is to see the a/c displacing from the intending flight path, especially fairly abruptly due to false GS capture, TOGA engagement, etc, and see the flying pilot try to figure out what mode the FMC/AP is in. Who cares? Turn it off, hand fly it along the desired flight path, and reestablish the appropriate level of automation.

'Click' 'click' is the fastest way to reestablish direct control of the airplane. Is it needed for most situations? No. Knowing how to work the FMC/AP is important. But equally important is understanding when manual control is required. Disengaging the AP when simple FMC/AP entries would fix the problem is not the goal either. Know the level of automation appropriate (hand flown, AP-V/S ALT HOLD, speed intervention, FMC-LNAV, VNAV, speed intervention, etc) and achieve the mode needed for the situation. IF the a/c isn't performing as desired/required YOU fix the problem.

Using the wrong level of automation seperates the men from the boys. Inappropriate mode use can be painful to watch.

[CancelIFR]

BOAC (http://www.pprune.org/members/14135-boac), with todays Norwegian Im guessing it wasnt big scandi eyes, probably english or southern european eyes...

[Petercwelch]

As a non professional, I wonder whether a manually flown approach every so often wouldn't encourage PF to click/click more quickly. As an occasional passenger on transport aircraft, my expectation is that reversion to a
manual approach shouldn't be a big deal. Sounds from PPrune like SOP forbids (discourages?) this in many companies or am I wrong? Shouldn't this be a routine part of airline flying? If it isn't, not just in simulators but in the aircraft, how certain can we be of good performance when the chips are down? 447?

[misd-agin]

Peter - U.S. carriers have fewer restrictions and the overwhelming majority of landings, and approaches, are hand flown as long as the weather is decent(300' and a mile visibility?). No data to back that up except for 30+ yrs of observation and experience.

Departures typically are hand flown until above 10,000' and the A/P is typically turned off below 10,000', and before 3,000', on arrival.

Different work loads and weather, or procedure driven automation requirements, are deciding factors in using a higher, or lower, level of automation.

The restrictive SOP's are typically non U.S. carriers.

[GlueBall]

misd-agin: "...Disengaging the AP when simple FMC/AP entries would fix the problem is not the goal either."

...that's OK only as long as attitude, altitude and flight path are not affected.
Making heads-down entries/activations and waiting for A/P response can quickly exasperate flight path deviation.

Too many of today's trained automatons have a complete aversion of disconnecting the automatics; even when exigent circumstances require immediate manual control, they momentarily hesitate in clicking off the A/P & A/T.


Boeing says: "When the automatic systems do not perform as expected, the PF should reduce the level of automation to ensure that proper control of the airplane is maintained. The PF should not attempt to restore higher levels of automation until after aircraft control is assured."

[misd-agin]

Glueball - we're in agreement.

Simple FMC fixes, done promptly, should be done while maintaining a higher level of automation. Knowing when to change levels of automation is important. Inappropriate use of levels of automation is painful to watch and is often a factor in incidents/accidents.

[5 APUs captain]

...'Click' 'click' is the fastest way to reestablish direct control of the airplane...

Yes, but it is not easy to explain that to "iPAD generation".

For M609: 90% of air Baltic FOs are western europeans

[M609]

And? When did Air Baltic get mentioned in my post? NAX has quite a few pilots (not just FOs) from the east.

[RAT 5]

It would seem "children of the magenta line" should be compulsory viewing on any TQ course for any type. I would even add, to all command courses as refresher.

[Bigmouth]

If a plane does a low altitude go-around at LHR, I know I will be reading about the poor, terrified passenger´s near death experience and how they were seconds away from certaindisaster.

Where are the big fat shock horror headlines in this case?

[Ancient Mariner]

From AIBN:
During approach to Kittilä (EFKT) in Finland on 26 December 2012, LN-DYM, a Boeing 737-800 NG on Norwegian Air Shuttle's (NAS') air service NAX5630 from Helsinki airport (EFHK), came close to stalling. The outcome of a stall would most likely have been catastrophic, primarily because the elevator system at that time did not function normally. The elevator system worked only at a ratio of 1:250.

AIBN's investigation has uncovered that de-icing fluid had ingressed the tail section and frozen on three or four of the input cranks for the aircraft's two elevator Power Control Units (PCUs) and thus prevented them from functioning as intended.

More, Report on the serious aircraft incident during approach to Kittilä airport in Finland (EFKT) on 26. December 2012 with a Boeing 737-800, LN-DYM, operated by Norwegian Air Shuttle ASA | aibn (http://www.aibn.no/Aviation/Published-reports/2015-01-eng)

[ImbracableCrunk]

http://avherald.com/h?article=4641efec

[de facto]

An oscillating pitch and speed calls for an immediate hands ON(if not already ON) and if there is obvious unstability,the auto flight must be disconnected.
Such issue may be avoided by capturing the glide at the correct altitude....(glide starts at 2500 feet,dont capture it at 6000)..
Such oscillations should be reported by pirep and deviations reported by SRS.
Airlines aware of such deviations should plan extra sim to ascertain such pilots manual skills.