NTSB Board Meeting

AA Flight 587

Airplane Motion and

Vertical Stabilizer Loads
John OCallaghan

Location of Wake Turbulence Encounters

4.5

WIND

4.0

FDR
accelerations
were typical of
wake encounters
Crew commented
on wake
turbulence
Simulation
indicates wake
encounter
NASA wake
study supports
encounter
Wake was similar
in each
encounter

Distance north of impact site, nm

3.5
3.0
2.5
0915:36 Start of first wake turbulence encounter

2.0
1.5

0915:51 Start of second wake turbulence encounter

1.0
0.5
0.0
-0.5
-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

Distance east of impact site, nm

3.0

3.5

4.0

Effect of the Wake

Encounters on the Airplane
Motion
NASA study indicates nothing unusual about
wake.
NTSB simulations determined that the effect
of wake on airplane motion was minor.
The airplane was not in or at risk of an
upset.

NTSB Board Meeting

Control Inputs Following Start of First Wake Encounter

Column, degrees

10

AFT COLUMN

First officer responded

with column & large
wheel inputs

5
0
-5
-10

FWD COLUMN

09:15:34

Wheel, degrees

-100

09:15:36

09:15:38

09:15:40

09:15:42

09:15:44

LEFT WHEEL

-50
0

Small changes in pitch

and roll angles

50
100

RIGHT WHEEL

09:15:34
-3

09:15:36

09:15:38

09:15:40

09:15:42

09:15:44

LEFT PEDAL

Airplane motion was

unremarkable

-2

Pedal, inches

First officer did not use

the rudder pedals

-1
0
1
2
3

RIGHT PEDAL

09:15:34

09:15:36

09:15:38

09:15:40

Time

ATC Time, HH:MM:SS EST

09:15:42

09:15:44

Control Inputs Following Start of Second Wake Encounter

Column, degrees

10

AFT COLUMN

Time = 09:15:51

Start of second wake

encounter

0
-5
-10

FWD COLUMN

09:15:50

Wheel, degrees

-100

09:15:52

09:15:54

09:15:56

09:15:58

LEFT WHEEL

Controls approximately
neutral

-50
0
50
100

RIGHT WHEEL

09:15:50
-3

09:15:52

09:15:54

09:15:56

09:15:58

09:15:54

09:15:56

09:15:58

LEFT PEDAL

Pedal, inches

-2
-1
0
1
2
3

Airplane in climbing left

turn

RIGHT PEDAL

09:15:50

09:15:52

Time

ATC Time, HH:MM:SS EST

Control Inputs Following Start of Second Wake Encounter

Column, degrees

10

AFT COLUMN

Time = 09:15:52

Large right wheel input

0
-5
-10

Full right pedal input

FWD COLUMN

09:15:50

Wheel, degrees

-100

09:15:52

09:15:54

09:15:56

09:15:58

LEFT WHEEL

-50
0
50
100

RIGHT WHEEL

09:15:50
-3

09:15:52

09:15:54

09:15:56

09:15:58

09:15:54

09:15:56

09:15:58

LEFT PEDAL

Pedal, inches

-2
-1
0
1
2
3

RIGHT PEDAL

09:15:50

09:15:52

Time

ATC Time, HH:MM:SS EST

Pedal used to help control

roll
Pedal not necessary
Wheel alone sufficient to
control roll
Full wheel and pedal
inputs unnecessary and
excessive

Control Inputs Following Start of Second Wake Encounter

Column, degrees

10

AFT COLUMN

Time = 09:15:53.1

Full left wheel input

(78)

0
-5
-10

FWD COLUMN

09:15:50

Wheel, degrees

-100

09:15:52

09:15:54

09:15:56

09:15:58

LEFT WHEEL

-50
0
50
100

RIGHT WHEEL

09:15:50
-3

09:15:52

09:15:54

09:15:56

09:15:58

09:15:54

09:15:56

09:15:58

LEFT PEDAL

-2

Pedal, inches

Full left pedal input

-1
0
1
2
3

RIGHT PEDAL

09:15:50

09:15:52

Time

ATC Time, HH:MM:SS EST

First full
alternating
rudder pedal
input

Control Inputs Following Start of Second Wake Encounter

Column, degrees

10

AFT COLUMN

Time = 09:15:54.2

Full right pedal input

0
-5
-10

FWD COLUMN

09:15:50

Wheel, degrees

-100

09:15:52

09:15:54

09:15:56

09:15:58

LEFT WHEEL

-50

Second full
alternating rudder
pedal input

0
50
100

RIGHT WHEEL

09:15:50
-3

09:15:52

09:15:54

09:15:56

09:15:58

09:15:54

09:15:56

09:15:58

LEFT PEDAL

-2

Pedal, inches

Growing oscillation in
column inputs

-1
0
1
2
3

RIGHT PEDAL

09:15:50

09:15:52

Time

ATC Time, HH:MM:SS EST

Control Inputs Following Start of Second Wake Encounter

Column, degrees

10

AFT COLUMN

Time = 09:15:55.6

Full right pedal

input maintained

0
-5
-10

FWD COLUMN

09:15:50

Wheel, degrees

-100

09:15:52

09:15:54

09:15:56

09:15:58

LEFT WHEEL

Large nose-down
column input

-50
0
50
100

RIGHT WHEEL

09:15:50
-3

09:15:52

09:15:54

09:15:56

09:15:58

09:15:54

09:15:56

09:15:58

LEFT PEDAL

Pedal, inches

-2
-1
0
1
2
3

Wheel moves to large

right deflection

RIGHT PEDAL

09:15:50

09:15:52

Time

ATC Time, HH:MM:SS EST

Control Inputs Following Start of Second Wake Encounter

Column, degrees

10

AFT COLUMN

Time = 09:15:57

Full left wheel

input

0
-5
-10

FWD COLUMN

09:15:50

Wheel, degrees

-100

09:15:52

09:15:54

09:15:56

09:15:58

LEFT WHEEL

-50

Third full alternating

rudder pedal input

0
50
100

RIGHT WHEEL

09:15:50
-3

09:15:52

09:15:54

09:15:56

09:15:58

09:15:54

09:15:56

09:15:58

LEFT PEDAL

-2

Pedal, inches

Full left pedal

input

-1
0
1
2
3

RIGHT PEDAL

09:15:50

09:15:52

Time

ATC Time, HH:MM:SS EST

Control Inputs Following Start of Second Wake Encounter

Column, degrees

10

AFT COLUMN

Time = 09:15:58.4

Wheel moves right

0
-5
-10

FWD COLUMN

09:15:50

Wheel, degrees

-100

09:15:52

09:15:54

09:15:56

09:15:58

LEFT WHEEL

-50

Fourth full
alternating rudder
pedal input

0
50
100

RIGHT WHEEL

09:15:50
-3

09:15:52

09:15:54

09:15:56

09:15:58

LEFT PEDAL

Vertical stabilizer
separates from
airplane

-2

Pedal, inches

Full right pedal

input

-1
0
1
2
3

RIGHT PEDAL

09:15:50

09:15:52

09:15:54

09:15:56

Time

ATC Time, HH:MM:SS EST

09:15:58

Sideslip Angle Buildup Resulting From First Officers Control I

Sideslip Angle

Sideslip angle, degrees

10

Vertical stab. separation

Airflow

NOSE LEFT OF AIRFLOW

-5

-10

NOSE RIGHT OF AIRFLOW

09:15:50

09:15:52

09:15:54

09:15:56

09:15:58

Airplane flew as
commanded
until vertical
stabilizer
separation
09:16:00

Airflow

Calculation of Vertical
Stabilizer Loads
Loads dependent on airspeed, sideslip angle, and
rudder deflection
Aerodynamic loads determined by wind tunnel
testing during airplane development
No wind tunnel data available at the extreme sideslip
angle corresponding to vertical stabilizer separation
Other methods required to compute loads at time of
separation

NTSB Board Meeting

Computational Fluid Dynamics

(CFD)
CFD is the use of computers to mathematically
determine the aerodynamic characteristics of
airplanes.
CFD is used increasingly in the industry to
supplement wind tunnel data and optimize airplane
designs.

NTSB Board Meeting

Computational Fluid Dynamics

(CFD)
CFD is the use of computers to mathematically determine the
aerodynamic characteristics of airplanes.
CFD is used increasingly in the industry to supplement wind tunnel data
and optimize airplane designs.
Airbus CFD code has demonstrated capability for solving flow problems
such as flight 587 vertical stabilizer loads.

CFD studies directed by NTSB and reviewed by

NASA Langley Research Center.

NTSB Board Meeting

Pressure Coefficient

CFD Results: Pressure Distribution Over Vertical Stabilize

Chordwise Distance (mm)

CFD Results: Streamlines of Flow at High Sideslip Angle

Left Side
Flow Separation

Right Side

Bending Moment History During Second Wake

2.5
Encounter
TAIL BENDS LEFT

Bending Moment / Limit Load

2.0

Ultimate Load (1.5 x Limit)

1.5

Limit Load

1.0
0.5
0.0
-0.5
-1.0

Limit Load

-1.5

Ultimate Load (1.5 x Limit)

-2.0
TAIL BENDS RIGHT

-2.5
09:15:50

09:15:52

09:15:54

Time

09:15:56

09:15:58

09:16:00

Bending
Moment

Base of
Vertical Stabilizer

Bending Moment History During Second Wake

2.5
Encounter
TAIL BENDS LEFT

Bending Moment / Limit Load

2.0

Ultimate Load (1.5 x Limit)

1.5

Limit Load

1.0
0.5
0.0
-0.5
-1.0

Limit Load

-1.5

Ultimate Load (1.5 x Limit)

-2.0
TAIL BENDS RIGHT

-2.5
09:15:50

09:15:52

09:15:54

Time

09:15:56

09:15:58

09:16:00

Bending Moment History During Second Wake

2.5
Encounter
TAIL BENDS LEFT

Bending Moment / Limit Load

2.0

Ultimate Load (1.5 x Limit)

1.5

Limit Load

1.0

Limit load:
Highest load expected in lifetime
Determined by conditions specified in FARs

0.5
0.0
-0.5
-1.0

Limit Load

-1.5

Ultimate Load (1.5 x Limit)

-2.0
TAIL BENDS RIGHT

-2.5
09:15:50

09:15:52

09:15:54

Time

09:15:56

09:15:58

09:16:00

Bending Moment History During Second Wake

2.5
Encounter
TAIL BENDS LEFT

Bending Moment / Limit Load

2.0

Ultimate Load (1.5 x Limit)

1.5

Limit Load

1.0

Ultimate load:
Equal to limit load times safety factor of 1.5
Structure must not break up to ultimate load

0.5
0.0
-0.5
-1.0

Limit Load

-1.5

Ultimate Load (1.5 x Limit)

-2.0
TAIL BENDS RIGHT

-2.5
09:15:50

09:15:52

09:15:54

Time

09:15:56

09:15:58

09:16:00

Bending Moment History During Second Wake

Encounter
2.5
TAIL BENDS LEFT
Range of
loads at
separation
based on
wind tunnel &
CFD analysis

Bending Moment / Limit Load

2.0

Ultimate Load (1.5 x Limit)

1.5
1.0

Wind tunnel analysis

Limit Load

0.5
0.0
-0.5
-1.0

Limit Load

-1.5

Ultimate Load (1.5 x Limit)

-2.0
TAIL BENDS RIGHT

-2.5
09:15:50

09:15:52

09:15:54

Time

09:15:56

09:15:58

09:16:00

Conclusions
Airplane encountered wake turbulence twice
Indicated by FDR, CVR, simulation, and wake analysis

First officers control inputs following second

encounter were unnecessary and excessive
Simulation indicates wake had minor effect on motion
Airplane was never in an upset condition

Airplane responded to control inputs as expected

until vertical stabilizer separation
Simulation indicates large sideslip angles were the result of
control inputs

Vertical stabilizer separated at a bending moment

load well above ultimate load
Determined by wind tunnel and CFD analysis

NTSB Board Meeting

National Transportation Safety Board

American Airlines Flight 587

Belle Harbor, New York
November 12, 2001
NTSB Board Meeting
October 26, 2004