Why Are We Here? - International Civil Aviation Organization are we here.pdf · 2014-05-28 · •...

Why Are We Here? (The Accident Loss of Control Record) ICAO Loss of Control Symposium May 20-22, 2014

1

Dennis A. Crider Chief Technical Advisor, Vehicle Simulation National Transportation Safety Board

LOC Accident / Incident Data Set

• NTSB LOC data collection • Dick Newman (Crew Systems) and

Tony Lambregts • 2008 paper • Frequently quoted

• NASA • Wouldn’t it be nice to work together? • Broader LOC dataset

2

LOC Accident / Incident Data Set • Data Sources

• Aircraft Accident Reports on DVD (R. Dorsett, 2006) • Australian Transport Safety Bureau (ATSB) • Aviation Safety Network (ASN) • Canadian Transportation Safety Board (TSB) • Flightglobal (Ascend Database) • French Bureau d'Enquêtes et d'Analyses pour la sécurité de l'aviation civile (BEA) • German Bundesstelle für Flugunfalluntersuchung (BFU) • International Civil Aviation Organization (ICAO) • Irish Air Accident Investigation Unit (AAIU) • National Transportation Safety Board (NTSB)

• Search Criteria • “loss-of-control” • “upset” • “unusual attitude” • “stall” • “uncontrolled”

3

5

Take

off

Initi

al C

limb

Clim

bC

ruis

eD

esce

ntH

oldi

ngAp

proa

chVF

R P

atte

rnC

irclin

gFi

nal A

ppro

ach

Land

ing

Go-

arou

ndM

isse

d Ap

proa

chM

aneu

verin

g

0

20

40

60

80Ev

ents

Events

0

500

1000

1500

2000

2500

On-Board Fatalities

On-

Boar

d Fa

talit

ies

LOC by Flight Regime

6

LOC by Event Factors

Stall Examples

• Pinnacle Airlines, Flight 3701 • Bombardier CL-600-2B19 • Aircraft stalled at high altitude

• Colgan Air Flight 3407

• Bombardier DHC-8-400 • Stalled on approach to Buffalo

Presenter

Presentation Notes

Several common threads of stall upsets that are addressed by the training matrix were seen in Oct 2004 Pinnacle Airlines 3701 jet accident and the Feb 2009, Colgan 3407 turboprop accident. Pinnacle 3701, a CRJ, stalled twice at high altitude loosing both engines. Colgan 3407, a dash 8 Q400 , stalled on approach to Buffalo.

8

02:54:30 02:54:40 02:54:50 02:55:00 02:55:10 02:55:20-80

-60

-40

-20

0

20

40

60

80

-40

-30

-20

-10

0

10

20

30

40Right

Bank

Ang

le (d

eg)

Time (Hr:Min:Sec)

Roll - left system Roll - right system

Right

Left

Whe

el (d

eg)

Left Wheel Right Wheel

Left

Presenter

Presentation Notes

Wheel and bank angle for the Pinnacle accident are shown here. Note left wheel before uncommanded right roll off . lateral instability and reduced control effectiveness are common features of stall upsets and a cue to the stall.

9

95520 95530 95540 95550 95560 95570-50-40-30-20-10

0102030405060708090

100110

-100

-80

-60

-40

-20

0

20

40

60

80

100Buffalo Colgan DHC-8-400

Roll (

deg)

FDR Time (sec)

Roll

Upse

tLeft

Right Right

Whe

el (d

eg)

Captain's Wheel

Left

Presenter

Presentation Notes

Here roll is shown in green while wheel is shown in red for the Colgan turboprop accident. An abrupt left roll off occurred at the upset. The captain responded as expected by moving the wheel right. The turboprop entered a series of roll oscillations, which were ineffectively opposed by opposite wheel inputs similar to the Pinnacle Jet. Such roll dynamics and control wheel ineffectiveness are typical of stall departures and can thus be a great cue to the flight crew that the aircraft is stalled and that the pilot needs to pitch down to recover.

10

02:54:30 02:54:40 02:54:50 02:55:00 02:55:10 02:55:20-15

-10

-5

0

5

10

15

20

25

30

25

20

15

10

5

0

-5

-10

-15

-20

-25

Angl

e of

Atta

ck (d

eg)

Time (Hr:Min:Sec)

FDR Angle of Attack

Nose Up

Ele

vato

r (de

g)

FDR elevator

Nose down

Presenter

Presentation Notes

the Pinnacle CRJ crew responded to the stick pusher by pulling back on the column. causing the angle of attack to increase which again triggered the pusher which the crew again fought four times with the maximum angle of attack increasing each time. 10 deg fuselage AOA beyond stall. This “stick pusher dynamics” is in itself a cue to stall. Very common

11

95520 95530 95540 95550 95560 95570-5

0

5

10

15

0

5

10

15

20

25

30

35

40

45

50

Cont

rol C

olum

n (d

eg)

FDR Time (sec)

Left Column Right Column

Upse

t

Left Angle of Attack Right Angle of Attack

Ang

le o

f Atta

ck (d

eg)

Presenter

Presentation Notes

Control column for the Colgan turboprop is shown in red and angle of attack is shown in blue. Again, the crew pulled back against the stick pusher entering a stick pusher dynamic pattern. The upset continued to ground impact. The peak angle of attack is 22 vane aoa, approximately 13 deg fuselage AOA, beyond the stall upset.

Breaking The Accident Chain

• Addressing elements in the sequence

• Train cue recognition • Account for human nature

• Train crews for all cues

12

13 13

Precursor Sequence

Loss of Energy State

Awareness / Inappropriate

Energy Management

Stall / Departure

Improper / Ineffective Recovery

Normal Flight LOC Crew

Fatigue / Impairment

Colgan Air 3407 (2/12/2009)

Loss of Energy State

Awareness / Inappropriate

Energy Management

Stall / Departure


Normal Flight Ground

Impact

Engine failure

Pinnacle 3701 (10/14/2004)

Stall Cues

• Stick Shaker • Ignored

• Stick Pusher • Fought • Stick pusher dynamics not

recognized • Un-commanded Roll

• Not recognized

14

Icing and Contaminated Airfoils

Presenter

Presentation Notes

Contaminated airfoils and icing were also identified as factors in several accidents. Though there have been upsets, such as the October 1994 American Eagle Flight 4184 ATR-72 accident near Roselawn, Indiana, where icing has caused upsets by introducing control problems, the majority of contaminated airfoil/icing accidents are stall upsets.

National Transportation Safety Board

Upsets from Stalls due to Icing • Saab 340: Eildon Weir, Australia,

• November 1998 • 2300-foot altitude loss

• Saab 340: Albury, Australia, • June 2004 • 40-foot altitude loss

• Saab 340: San Luis Obispo, CA, USA • January 2006 • 5000-foot altitude loss

Presenter

Presentation Notes

Icing incidents involving the Saab 340 contrast the results of prompt positive stall recovery with delayed stall recovery. Talk Slide On November 11, 1998, a Saab 340A entered the holding pattern at Eildon Weir, Australia, where it slowed and stalled without a stall warning, losing 2,300 feet of altitude before recovery. On June 18, 2004, at Albury, New South Wales, Australia, a Saab 340 had just leveled off at 12,000 feet when the aircraft began to slow. The pilot, who disconnected the autopilot as the stick shaker activated briefly, recovered after only a 7-degree roll-off. On January 2, 2006, a Saab 340B+ that was operated as American Eagle 3008 slowed as the autopilot maintained the climb rate and stalled while climbing out of San Luis Obispo, CA. The aircraft recovered controlled flight at about 6,500 feet.

17

0 5 10 15 20

Lift C

oeffic

ient -

C L

Angle of Attack - α (deg)

Albury Eildon Weir American Eagle 3008 No ice lift curve

Saab 340Icing Event Comparison

18

0 5 10 15 20 25

CRJ Icing Event Comparison

C L

Angle of attack

Fredericton no ice (sim sweep) Baotou no ice (sim sweep) Fredericton Baotou

Fredericton

Baotou

Presenter

Presentation Notes

Jets can also show an early stall due to icing contamination. This plot shows the extracted lift curves for the 1997 Fredericton Canada (SOLID RED) and 2004 Baotou China (SOLID BLUE) CRJ accidents compared with the lift curves without ice contamination (DASHED LINES). The early stall with ice contamination for both accidents is evident. The December 16, 1997, Air Canada CRJ accident at Fredericton Airport, New Brunswick, occurred when the aircraft stalled and rolled to 60 degrees after initiating a go-around. The November 21, 2004, China Eastern Airlines Bombardier CRJ-200 accident at Baotou, China, illustrates the effect of frost accumulated on the ground. The aircraft stalled soon after takeoff, rolled 42 degrees left, 68 degrees right, and 62 degrees left before crashing.

19 19

Precursor Sequence Ice Contamination

Stall / Departure


Normal Flight LOC

Icing

Stall Cues with Ice

• Stick Shaker • May be after departure

• Stick Pusher • May be after departure

• Natural stall cues • Un-commanded Roll • Buffet • Stall break

20

Summary • Elements of LOC

sequences have been identified

• Stall leading element

• Need to address multiple points in accident chain for best affect.

• New training we are outlining this week addresses most LOC factors

• Some elements may be better addressed by technologies

21

Spatial Distortion

• In pitch • Typical go-around in IMC • Acceleration produces illusion of pitch up

• Pilot pitches down flies into ground

• In Roll • Slow roll develops pilot doesn’t feel • Then confused when bank angle warning

or g’s alert to roll but feel level

Presenter

Presentation Notes

Several spatial distortion accidents appear in the accident record. A spatial disorientation in pitch typically occurs when the crew executes a go-around in instrument conditions. The acceleration can give the illusion that the aircraft is pitching up. The pilot responds by pitching the aircraft down and flying into the ground. The second more common type of spatial disorientation is spatial disorientation in roll. In this scenario, a slow roll develops that the pilot does not feel. Pilots may then get a bank angle alert and be confused as their senses tell them they are flying level.

Atmospheric Disturbance

26

Sch

edul

ed A

irlin

es

Non

-Sch

edul

ed

Non

-Rev

enue

Ope

ratio

ns

Exe

cutiv

e Tr

ansp

orta

tion

0102030405060708090

100110120130140150

Even

ts

Events

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

5500

6000

On-Board Fatalities

On-

Boar

d Fa

talit

ies

LOC by Operation

Recommendation A-10-22

Require 14 Code of Federal Regulations Part 121, 135 and 91K operators and 14 Code of Federal Regulations Part 142 training centers to develop and conduct training that incorporates stalls that are fully developed; are unexpected; involve autopilot disengagement; and include airplane-specific features such as a reference speeds switch

Presenter

Presentation Notes

The lack of proper pilot response to the stall and stick pusher led the Safety Board to issue recommendations for full stall and stick pusher familiarization training. {Pause or read}


• Require all 14 Code of Federal Regulations Part 121, 135, and 91K operators of stick pusher-equipped aircraft to provide their pilots with pusher familiarization simulator training.

Presenter

Presentation Notes

Supersedes Safety Recommendation A-07-4


Define and codify minimum simulator model fidelity requirements to support an expanded set of stall recovery training requirements, including recovery from stalls that are fully developed. These simulator fidelity requirements should address areas such as required angle-of-attack and sideslip angle ranges, motion cueing, proof-of-match with post-stall flight test data, and warnings to indicate when the simulator flight envelope has been exceeded.

Public Law 11-216 (sec 208)

The Administrator of the Federal Aviation Administration shall conduct a rulemaking proceeding to require part 121 air carriers to provide flight crewmembers with ground training and flight training or flight simulator training – A. To recognize and avoid a stall upset of an

aircraft or, if not avoided, to recover from the stall;

Presenter

Presentation Notes

The U.S. Congress agreed with the National Transportation Safety Board’s stall and stick pusher training recommendations and required such training in Public Law 11-216, Section 208, IMPLEMENTATION OF NTSB FLIGHT CREWMEMBER TRAINING RECOMMENDATIONS {Pause or read} The FAA is responding to this law with revisions to simulator device standards (Part 60) and flight crew training (Parts N & O).


Define and codify minimum simulator model fidelity requirements for aerodynamic degradations resulting from airframe ice accumulation. These requirements should be consistent with performance degradations that the National Transportation Safety Board and other agencies have extracted during the investigations of icing accidents and incidents.

Presenter

Presentation Notes

A similar early stall was seen in the January 27, 2009, crash of Empire Airlines flight 8284 short of the runway at Lubbock, Texas, Preston Smith International Airport and several other accidents and incidents. The Safety Board sited the Saab 340 icing upsets in the report on the Lubbock TX ATR 42-320 accident to recommend enhanced simulator fidelity in icing in this rec [pause] And, Part 121 supplemental cargo flight from Fort Worth TX

Recommendation A-11-47 Once the simulator model fidelity requirements requested in Safety Recommendation A-11-46 are implemented, require that flight crews of all aircraft certificated for flight in icing conditions be trained in flight training simulators that meet these fidelity requirements. Such simulation training should emphasize the following: (1) cues for recognizing changes in the aircraft’s flight characteristics as airframe icing develops; (2) procedures for monitoring and maintaining appropriate airspeeds in icing conditions, including the use of icing airspeed reference indices; and (3) procedures for responding to decaying airspeed situations, stall protection system activation, and early stalls that can occur without stall protection system activation.

Presenter

Presentation Notes

READ REC Several icing accident have occurred on takeoff. Training can give flight crews better tools to deal with some of these situations as well.

Stall Upsets (no icing) Accident

Shaker

Pusher

Break

Airborne Express DC-8-63 December 22, 1996 IO IO NR

China Airlines Flight 676, A300-622, February 16, 1998 UK UK UK Thai Airways Flight 261, Airbus A310-300, December 11, 1998 UK UK UK Pinnacle Airlines Flight 3701 Bombardier CL-600-2B19, October 14, 2004

NR NR NR

United Express Flight 6291 Jetstream 4101, January 7, 1994 NR NR UK Pulkova Flight 612, TU-154M, August 22, 2006 UK UK UK West Caribbean airlines MD-82 near Machiques, Venezuela, August 16, 2005

NR UK NR

Continental Connection Flight 3407 Bombardier DHC-8-400, February 12, 2009

NR NR NR

IO = Inoperative UK = Unknown NR = Not responded to

Presenter

Presentation Notes

This slide list some stall accidents and the crews response to the shaker (POINT) , Pusher (POINT) and Stall break (POINT) cues. Note the number of cases where the crew has not responded to these cues. The ICATEE Training matrix contains elements for recognizing stall cues beyond just the shaker. Part N&O now includes full stall and stick pusher familiarization training.

Stall Upsets (icing) Accident/Incident

Shaker

Pusher

Break

Saab 340A, VH-LPI, Eilden Weir, Victoria, November 11, 1998 ES ES NR

Saab 340A, VH-KEQ, Albury, New South Wales Australia, June 18, 2004

ES ES PR

Saab 340B, VH-OLM, Bathurst, New South Wales Australia, June 28 2002

ES ES NR

American Eagle Flight 3008, Saab 340B+, San Luis Obispo, January 2, 2006

ES ES NR

Air Canada Flight 646, Canadair CL-600-2B219, December 16, 1997 -- -- -- Cessna Citation 560, Pueblo, Colorado, February 16, 2005 UK UK UK Comair Flight 3272, Embraer EMB-120RT, January 9, 1997 ES -- NR Skywest Flight 3855, Bombardier CL-600-2B19, January 17, 2004 -- -- -- ComAir Flight 5054, EMB-120, March 19, 2001 NR NR NR

Empire Airlines Flight 8284, ATR-42, Lubbock TX, January27, 2009 NR ES NR IO = Inoperative UK = Unknown PR = Proper response

NR = Not responded to ES = Early stall

Presenter

Presentation Notes

This slide list some icing stall accidents and the crews response to the shaker (POINT) , Pusher (POINT) and Stall break (POINT) cues. Note the number of ES codes in green where the early stall precluded warning from the shaker or pusher. Comair 3272– stall then stick shaker

Why Are We Here? - International Civil Aviation Organization are we here.pdf · 2014-05-28 · •...

Documents

Transcript of Why Are We Here? - International Civil Aviation Organization are we here.pdf · 2014-05-28 · •...