Wideroe DH8A at Svolvaer on Dec 2nd 2010, aircraft rapidly descended on base turn, recovered at 25 meters AGL

Last Update: November 23, 2016 / 17:47:50 GMT/Zulu time

Bookmark this article
Incident Facts

Date of incident
Dec 2, 2010

Classification
Incident

Airline
Wideroe Airlines

Aircraft Registration
LN-WIU

Aircraft Type
De Havilland Dash 8 (100)

ICAO Type Designator
DH8A

Airport
Svolvær Helle Airport, Svolvær

Airport ICAO Code
ENSH

On Nov 23rd 2016 the AIBN released their final report in Norwegian only. Although an English version had been indicated previously (when the AIBN had told the Aviation Herald, the English version of the preliminary report had been released a few days after the Norwegian report and was identical in content with the Norwegian version), the English version of the website just provides a brief summary and the link to the Norwegian report, too. Based on the announcement of the English version becoming available, The Aviation Herald decided to skip the translation of the Norwegian Report, however, focus on some issues of the occurrence that were object of an e-mail exchange between AVH and AIBN following the discovery of the English version of the interim report, which may - or may not - have bearing into the final report.

The English version of the website provides following summary for the final report:

During approach in darkness, LN-WIU suddenly had a significant loss of both speed and altitude. The flight crew regained control, but the airplane came very low (83 ft (25 m) above ground). During the recovery the airplane was exposed to high g-force and the engine torque limits were exceeded. The crew aborted the approach and continued to Leknes Airport where a normal landing was performed.

The AIBN investigation indicates that the airplane at low altitude was exposed to significant wind shear (microburst), probably from a Cumulonimbus cloud (CB) in the area.

The Accident Investigation Board has, as part of this investigation, in particular focused on risk management and safety margins in connection with circling approaches in darkness. No obvious systemic failures or other deficiencies that could have influenced on the chain of events or causal factors was identified.

At one point during the chain of events, the First Officer took over the flight controls. The Commander did not oppose this. Apart from agreeing that this happened, the crew has given, in part, different descriptions of the chain of events. Based on the available facts, the Accident Investigation Board has not been able to determine neither which pilot did what, exactly when, and in which order, nor the effects of each action, seen in isolation. It has therefore not been possible to draw any solid conclusions about the significance of the actions of the first officer. The collective actions of the crew did, however, most likely lead to a recovery that was initiated in time to prevent the airplane from colliding with the ground.

The maneuvering to regain control was executed with sparse visual references and without a visible horizon. This investigation has revealed that the Commander was exposed to conditions that could provoke a somatogravic illusion. The Accident Investigation Board has, however, not found evidence to conclude that a spatial disorientation, if any, did have any effect on the Commanders handling of the wind shear.

This incident is an important reminder of the vulnerability connected to maneuvering low above the terrain with sparse visual references in darkness and turbulent air. It is also a reminder of the fact that operators and pilots with thorough knowledge of local conditions, experience and training beyond the regulatory minimum standards would be more capable of handling critical situations like this.

The incident happened in December 2010. This investigation was opened in Mid-February 2015. The circumstances concerning the date of the incident, reporting and classification are discussed in the report, which later will be translated into English.

No safety recommendation is issued in connection with this investigation.

This report is in Norwegian only. English translation will be published at later time.

Following the discovery of the English version of the interim report on Jul 16th 2016 The Aviation Herald wrote following e-mail to the AIBN on Jul 16th 2016:

Dear Mr. Bertheussen, dear Mr. Halvorsen,

we noticed the updated report into the Svolvaer event at:



First I want to thank for the continued investigation and the efforts put into the investigation.

However, I can not avoid to annotate, that the investigation does not meet the expectations of an investigation attempting to uncover the real facts and draw unbiased conclusions.

I have certain understanding that it is hard, without cockpit voice recorder and without detailed FDR data indicating who did what (who applied the control inputs), whose perceiption has been right. Therefore I do not have any real quarrel with this part of the report trying to cover and explain the perceiptions by either pilot. I even think the clear distinction, that the PF was more or less visual and thus had to have a very different perceiption that the PM being focussed mainly on instruments is getting to the kernel of the issues here.

However, where I do very strong objection when the report in itsself is contradictory making clear the AIBN is steering the argument line into the direction to whitewash their initial decision to not investigate. The AIBN made a mistake in 2010 to 2012, the mistake was corrected in 2015, and the AIBN would be well advised to stand to that mistake and not try to cover up that mistake again.

I take extreme objection with following:

on page 6 paragraphs 3 and 4 read:

When stalling in still air, one would expect to see a noticeable reduction in G-load. It is worth noting that the values did not go below 1 G in connection with the nose dropping, and that none of the fluctuations stood out significantly.

The graph showing changes in the vertical acceleration does not, at any stage, correlate with the change pattern one would expect to see in connection with stalling in still air. AIBN's conclusion, in consultation with experts from the aircraft manufacturer Bombardier, is that the fluctuations were due to air disturbances and that the aircraft did not, at any time, stall.

In Appendix B first paragraph states:

(10) Between sample time 130201.9 and 103202.2, vertical acceleration is 0.4 g during which the pitch attitude is between 14.0 and 14.4¢ª nose down.

This is an absolute clear contradiction between the facts described and the analysis which completely ignores the facts raising the question about why are the facts ignored? To justify the initial decision to put it down to a turbulence on final approach not warranting an investigation?

It is noteworthy, that the original FDR output shows a vertical acceleration initially of +0.75G reducing to +0.4G during the upset, hence a very significant reduction below +1G, which also explains the very quick increase of vertical rate of descent and the loss of several hundred feet of altitude within only 8 seconds (average rate of descent of 2377 feet/min despite initial acceleration and deceleration to arrest the descent).

In addition, the report states: "In this preliminary report, the AIBN can only conclude that the circumstances and outcome show that the crew successfully averted crashing into the sea after being exposed to a major wind shear at low altitude."

But then, how does page 6 para 3 and 4 match this with their comparism and argument line based on "still air"?

We have covered the updated preliminary English report now at:



I would love to include an explanation for this contradiction buried within the report as lined out above and invite you to provide an explanation respective update to the report removing this (and other) contradictions in the report.

Kind regards
Simon Hradecky
The Aviation Herald


On Jul 18th 2016 AIBN's Kare Halvorsen responded:

Dear Mr Hradecky

Thank you for following our progress into this investigation and for recognizing the effort. For your information, the AIBN have not issued any update following the 21st of August 2015.

The investigation is made in accordance with ICAO annex 13 and REGULATION (EU) No 996/2010 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL. This imply that a draft report is issued for comments by all involved parties. With reference to EU 996/2010, Article 14 (f) Protection of sensitive safety information, drafts of preliminary or final reports or interim statements shall not be made available or used for purposes other than safety investigation.

Deadline for this commenting period was at the time of the tragic helicopter accident near Turøy involving an Airbus Helicopters H225 and thus the AIBN had to stall further work with the Svolvær incident investigation for a period. The final Svolvær report will be released late summer /early autumn 2016.

At his stage the AIBN will not comment on the preliminary report issued 21st of August 2015.

Best regards

Kåre Halvorsen
Director aviation department

The Aviation Herald in the meantime had done assessment and computations based on the flight data recorder graphics and wrote following e-mail in response to the mail above:

Dear Mr. Halvorsen,

thanks for your reply. From our point of view, when we discovered the release of preliminary report in Norwegian we were translating this and covering the report as we did.

We were then monitoring the website for an English version of the preliminary report for a couple of months but were not able to find such an English version.

On Jul 16th 2016 one of our readers, daily monitoring the AIBN website too, then pointed the English report out to us, and this report contains a lot more than the Norwegian preliminary report we handled in 2015. Hence from our point of view this report has been updated since Aug 21st 2015.

Anyway, that's of no real importance.

The important issue is, that the preliminary report in this version contradicts itself contributing to serious damage to the reputation of the AIBN if not corrected. I have lined out the direct contradictions already, where paragraphs of the report stand in direct contradiction.

There are more factual errors in this report however. Let me expand a bit on those errors, too:

1) The most obvious error, Angle of Attack:
-------------------------------------------

At timepoint 103200, the point where the CAS reduces through 75KCAS, we find the following scenario in the FDR graphs contained in (both) reports:

Speed: 75KCAS (38.6m/s)
Pitch: almost +10 degrees
Vertical Acc: +1.4G
V/S: -2400fpm (-40fps, -12.2m/s)

Ignoring 400 feet MSL and QNH, thus assuming just ISA conditions, I set 75KCAS=75KTAS for illustration of the main computation, the factors of 400 feet MSL and QNH different do add to the computation but don't change it significantly.

Computing the flight path we find sin (flight path angle)= -12.2/38.6 resulting in a flight path angle of -18.4 degrees.

Ignoring a possible angle of incidence of the wings attached to the fuselage and thus assuming, the wings are completely aligned with the longitudinal axis of the aircraft (as I have no actual mechanical data of the DH8A available) we thus find the Angle of Attack at that point was:

+28.4 degrees (+18.4 degrees negative flight path angle plus pitch angle +10 degrees)

This AoA is clearly beyond any critical angle (AlphaMax), the lift coefficient (CL) well into the descending part of the curve past CLMax already and the main wings thus are well into the stall region even if we assume largely reduced G-loads.

2) Oversight in the data analysis, disconnect between vertical speed and G-load:
--------------------------------------------------------------------------------

Until time point 103198 it is clearly visible, that vertical acceleration and change of altitude do correspond. A vertical acceleration above +1.0G thus results in an increase of vertical speed und a vertical acceleration below +1.0G results in a decrease of the vertical speed. The same can be observed past time point 103202.

However, within the 4 seconds from 103198 to 103202 the values no longer correspond and appear to be entirely disconnected from the rest of the values.

Differentiating the altitude between datapoint 103197 and 103198 we find the vertical speed was -30fps as well as 103198 and 103199 we find the vertical speed was -40fps (all within the accuracy available from the graphs contained in the reports). At the same point we see the vertical acceleration was +0.75G (vertical acceleration - 1G thus being -0.25G).

Crosschecking the 10fps (3.0m/s) difference of vertical speed within 1 second with the -0.25G we find 1G would equate to about 12m/s/s, which is close enough (within the accuracy of graphs) to the actual 9.81m/s/s that we know gravity on earth is being actually measured. Hence we find the relation between G, altitude change and change of vertical speed confirmed.

Subsequently we see the vertical acceleration in the graph at above +1.0G however, ranging up to +1.4G, however. In other words, the vertical speed would need to increase, or the descent would reduce. However, nothing like that can be seen. Instead the vertical speed decreases further.

We note, that at -0.25G net acceleration at the begin of the upset the vertical speed changes from 0m/s to -3m/s within 2.5 seconds in accordance with the G-load depicted in the graph. Now, at +0.4G net vertical acceleration (against gravity) a much quicker increase of vertical speed at roughly 4m/s/s (13fpss) would be expected, however, the vertical speed still decreases and the descent thus increases.

This can not be explained, and the G-values therefore can not be used to determine that the aircraft had not stalled as the preliminary report did. Quite the opposite, this disconnect of vertical G and differential of altitude (vertical speed) clearly suggests, that the main wings no longer developed (significant/sufficient) lift indicative of the mains being exposed to non-laminar airflow aka being stalled.

The pattern of the values thus is clear proof of a fully developed stall, confirmed also by the AoA computation of point 1.

3) Pitch Down:
--------------

The report analyses, that the large pitch down movement from +10 to -15 degrees of pitch were the result of inertia following the nose down control input changing the control column from +10 degrees (nose up) to -8 degrees (nose down) and then back to +30 degrees.

In the graph it can be seen, that at time point 103200.2 the elevator control reaches the maximum nose down deflection, the pitch just about begins to react at that point, the nose begins to accelerate downwards. At time point 103200.6 the control column goes through 0 degrees and begins to command a nose up again. However, at this point the nose down moment does NOT begin to slow down but still accelerates strongly until the nose gets below -10 degrees.

This is NOT consistent with inertia. If indeed the inertia was causing this continued nose down moment, it would significantly slow down immediately as the horizontal tail begins to produce a downward force (nose up moment) again. The still increasing pitch moment however suggests, that despite the nose up command the horizontal tail still produced an upward force adding additional pitch down moments to the inertia of the aircraft.

This leads to following scenario:

The wings were no longer creating lift only producing drag, however the horizontal tail had not stalled. Hence the aircraft was "weathercocking" with the tail producing an upward force pushing the nose down to put the aircraft - just like a weather cock - into the oncoming wind. Therefore the pitch down still accelerated as the horizontal tail still developed an upward force even though the elevator already commanded a downward force again just offsetting the net effect of the horizontal tail. Only when the angle of attack had reduced to below +10 degrees the offset commanded by the elevator became relevant again and was able to produce a net downward force to raise the nose against the oncoming airflow tending to produce an upward force on the tail (nose down moment). It can be seen, that the pitch moment decreases from that point on (as a combination of inertia and changes of forces on the horizontal tail developing more downward force as the pitch angle decreases further) and reverts to a nose up movement once the pitch had reached a minimum of nearly -15 degrees.

This scenario is in full alignment and agreement with observations both in point 1) and point 2).

4) Point of main wings unstalling and subsequent pitch up:
----------------------------------------------------------

At time point 103202 we observe the pitch at -14 degrees (with a dramatic reduction of nose down moment immediately afterwards), the G-load changing from +1.4G (at 103201.4) to +0.4G (at 103202), the control column was at +22 degrees, the vertical speed was -40fps and the speed had increased to 110KCAS (56.6m/s).

Computing the AoA at that point (with the same criteria used in point 1) we find, the flight path angle was -12 degrees at that point and the resulting AoA at that point therefore was:

12 degrees+(-10 degrees)= +2 degrees

clearly in the aerodynamic effective range permitting the turbulent airflow over the wings to become laminar again and the wings developing substantial lift again.

At that point we also see the altitude and G-load synchronise again, the G-Load at that point becomes +0.4G, the vertical speed decreases (descent increases), and the pitch begins to strongly increase. Shortly afterward the G-load turns significantly above +1G and the descent gets arrested and the aircraft begins to climb again.

Once again this computation fits perfectly in the pattern seen from point 1 through 3.

This portion of the graph, as a side result, permits to conclude that the aircraft was still properly trimmed (in other words, maintaining altitude was not commanded by nose up trims prior to the upset), otherwise either the horizontal tail would have stalled or the horizontal tail would have developed downward forces too early to unstall the main wings, both scenarios resulting in loss of control and definitive impact in the water.

Summary:
--------

As shown in the cross check of values we observe a fully developed stall. We observe an AoA well above AlphaMax at the begin of the upset, we observe altitude/vertical speed development clearly consistent with a stall. Initially even the G-load indicates a stall condition when the G-load reduces to +0.75G (with the AoA already above AlphaMax).

The values do not cross check between 103198 and 103202, this period of 4 seconds would need a closer look to explain how G-values clearly above +1.0G up to +1.4G, against all the other values and without any reaction by the aircraft, came together. Using the G-Value to rule out a stall condition is clearly erroneous due to this "disconnect".

All of these considerations do not need to take into account, whether the aircraft was moving in still or turbulent air.

As result of all of this, my previous mails and this one I do hope, that the AIBN is going to review the scenario and analysis indeed, and issues a final report, that does not contradict itself and is not that easily disproven in its analysis as this preliminary report, that became known to us only on Jul 16th 2016.

Kind regards
Simon Hradecky
The Aviation Herald


AIBN's Kare Halvorsen responded to this latest mail on Jul 18th 2016 as follows:

Dear Mr Hradecky

Thanks for your view and advise, they will be taken into consideration during the finalizing of the final report.

At this stage the AIBN will not debate the preliminary report of August 2015 and ask you to await the final report planned to be issued as stated in previous mail.

The English version of the preliminary report was issued a few days following the Norwegian one. For future monitoring we might recommend the possibility to subscribe for reports and other news

Best regards

Kåre Halvorsen

Following this e-mail exchange The Aviation Herald learned, that the draft final report had already been sent out for comments by the parties involved in this investigation.
Aircraft Registration Data
Registration mark
LN-WIU
Country of Registration
Norway
Date of Registration
DjkqAngAlcp jAj Subscribe to unlock
Manufacturer
DE HAVILLAND AIRCRAFT OF CANADA LIMITED
Aircraft Model / Type
DHC-8-103
ICAO Aircraft Type
DH8A
Serial Number
Pl Subscribe to unlock
Maximum Take off Mass (MTOM) [kg]
Lbhic Subscribe to unlock
Engine Count
G Subscribe to unlock
Main Owner
DAAlpAkqjedgbdjckmA mqcjlnknmnjkpcilddlcfbAjeeddAgn Subscribe to unlock
Incident Facts

Date of incident
Dec 2, 2010

Classification
Incident

Airline
Wideroe Airlines

Aircraft Registration
LN-WIU

Aircraft Type
De Havilland Dash 8 (100)

ICAO Type Designator
DH8A

Airport
Svolvær Helle Airport, Svolvær

Airport ICAO Code
ENSH

This article is published under license from Avherald.com. © of text by Avherald.com.
Article source

You can read 2 more free articles without a subscription.

Subscribe now and continue reading without any limits!

Subscribe now
Are you a subscriber? Login
Subscribe

Read unlimited articles and receive our daily update briefing. Gain better insights into what is happening in commercial aviation safety.

Compare our plans

Send tip

Support AeroInside by sending a small tip amount.

Send tip

Related articles

Newest articles

Subscribe today

Are you researching aviation incidents? Get access to AeroInside Insights, unlimited read access and receive the daily newsletter.

Pick your plan and subscribe

Partner

Blockaviation logo

A new way to document and demonstrate airworthiness compliance and aircraft value. Find out more.

ELITE Logo

ELITE Simulation Solutions is a leading global provider of Flight Simulation Training Devices, IFR training software as well as flight controls and related services. Find out more.

Blue Altitude Logo

Your regulation partner, specialists in aviation safety and compliance; providing training, auditing, and consultancy services. Find out more.

AeroInside Blog
Popular aircraft
Airbus A320
Boeing 737-800
Boeing 737-800 MAX
Popular airlines
American Airlines
United
Delta
Air Canada
Lufthansa
British Airways