LMS Conference Europe, March 22/23 2006

Landing gear shimmy

Detect, understand and fix shimmy problems early

in the design using simulation

TNO Automotive

Igo Besselink

23 March 2006 1
Contents
• introduction: what is shimmy?
• analytical stability results using a simple model
• solutions to solve shimmy stability problems
• detailed shimmy analysis
-modelling of the landing gear
-tyre modelling
• conclusions

23 March 2006 2
Introduction
Shimmy is an unstable lateral/yaw vibration…

2 examples:
helicopter NLG measurement on aircraft MLG

23 March 2006 3
Tyre marks on the runway…

23 March 2006 4
Some notes:
• frequency typically in the range of 10 to 30 Hz

• degree of instability may vary from annoying vibrations

up to structural damage or even a collapse of the
landing gear

• shimmy can occur on both nose and main landing gears

Twin wheeled cantilevered main landing gears as seen on

many commercial aircraft may experience shimmy stability
problems. Bogie landing gears and levered suspension
configurations are generally not sensitive to shimmy

23 March 2006 5
Landing gear configurations…

danger of shimmy!

23 March 2006 6
The importance of simulation models
• full scale shimmy testing on an aircraft may be very risky
and dangerous… (also indoor testing on a drum may be not
be successful)

• no room to “experiment” or create various prototypes: the

design has to be “the first time right”!

• shimmy is a complex phenomenon: testing only does not

directly lead to a solution: trial and error, unclear
“solutions” to solve the problem

• shimmy stability should be considered in the initial design

phase of an aircraft when no hardware is available
23 March 2006 7
The cause of shimmy… (1)
• interaction between the dynamic behaviour of the
landing gear structure and tyre

simplified representation
tyre lateral force (Fy) results
in a side slip angle of the
landing gear

due to the side slip angle the

tyre develops a lateral force

23 March 2006 8
The cause of shimmy… (2)
• the feedback loop may become unstable!
• both the dynamics of the tyre (e.g. relaxation
behaviour) and dynamics of the structure (e.g.
eigenfrequencies/modeshapes are important)

some other views on shimmy:

• energy is extracted from the forward motion of the
aircraft and converted into a lateral/yaw vibration
• for a rolling tyre a combination of lateral and yaw
input exists where self-excitation occurs

23 March 2006 9
Analytical solutions
• trailing wheel system with lateral flexibility (topview)

23 March 2006 10
Analytical solutions… (q=0 and zero damping)
also unstable, but generally less problematic

problem area

23 March 2006 11
What can we learn from the analytical results?
• shimmy problems can be solved by changing stiffness,
geometry and/or inertia…
• just increasing the (yaw) stiffness is no guarantee to solve the
shimmy instability effectively

two feasible solutions to avoid instability:

• small negative trail – yaw stiffness low – lateral stiffness high
– large yaw inertia
• big positive trail – yaw stiffness high – lateral stiffness low
– small yaw inertia

note: tyre characteristics will also affect shimmy stability, but

are difficult to change in practice

23 March 2006 12
One step closer to a “real” landing gear…
simple multi-body model consisting of:
• three rigid bodies
-strut
-trail body
-wheelaxle
• two wheels
• two tyres
(TNO MF-Swift)

bodies are connected by

means of revolute joints

simple,
but representative

23 March 2006 13
Animation (baseline)

23 March 2006 14
Simulation of a landing event
• velocity: 75 m/s (=270 km/h)
• asymmetrical spin-up of the wheels on touch down

tyre vertical force tyre longitudinal force

23 March 2006 15
Some more simulation results…

tyre side slip angle tyre lateral force

time scale: 2 sec.

23 March 2006 16
Design study – Yaw Stiffness Variation
red:
nominal
(4e5 Nm/rad)
green:
increased yaw
stiffness
(5e5 Nm/rad)
blue
further increase
(6e5 Nm/rad)

23 March 2006 17
Animation (increased yaw stiffness)

23 March 2006 18
modifying the geometry (negative trail)

23 March 2006 19
Animation (modified geometry)

23 March 2006 20
Modelling in an engineering environment
so far: (highly) simplified models not suitable for
designing a real landing gear
…but definitely important to develop some “feeling” for
the problem!

For an accurate, predictive shimmy analysis both landing

gear and tyre have to be modeled accurately!

Several component tests may be necessary to validate

certain aspects of the model, e.g. stiffness tests, modal
testing, tyre testing,…

23 March 2006 21
Landing gear structure

23 March 2006 22
Stiffness modelling
• stiffness is dependent on the shock absorber closure
due to changes in overlap and torque link geometry
(stiffness may change by a factor 2 or more)

• also important: flexibility of the fuselage/wing

23 March 2006 23
Modelling of the vertical air spring/damper
characteristics

23 March 2006 24
Non-linear behaviour
• free-play: reduces the effective stiffness of a landing
gear. Older, worn landing gears may have a reduced
stability margin. A simulation study may be necessary
to determine the maximum allowable free-play limits

• friction: cantilevered landing gear designs may exhibit

quite a bit of friction. To get a good agreement with
tests on the aircraft and component tests it has to be
included in the model. Friction may also “hide”
possible shimmy instability problems

23 March 2006 25
Tyre modelling
• traditionally in aircraft shimmy analysis the tyre models of
Von Schlippe, Smiley or Moreland are used. Theory of these
models dates back to the 1940’s and 1950’s

• More up to date models are available

• Over 20 years of research at the TU Delft and TNO

Automotive under the guidance of professor Pacejka has
resulted in tyre models called MF-Tyre and MF-Swift

23 March 2006 26
Experimental validation (MF-Swift)
• yaw oscillation test stand
frequency response functions of
side force and aligning torque to yaw angle up to 60 Hz

amplitude ratio Fy ψ amplitude ratio Mz ψ

106 10
6

5
10

105 10
4

110
110
3
10
20
20
4 2
10 10
100 101 10
0 1
10
Frequency [Hz] Frequency [Hz]

Experiment Simulation V [km/h]

23 March 2006 27
MF-Tyre/MF-Swift
• originally developed for passenger car tyres
• validated extensively by numerous experiments

most recent version: MF-Tyre/MF-Swift 6.0

in comparison to the older MF-Tyre 5.2 and Swift 1.2 much more
useable for aircraft simulation studies:
• very significantly reduced data requirements
• build in parameter estimation procedures
• includes turn slip, important for shimmy

23 March 2006 28
Advantages over “traditional” shimmy tyre models
• includes non-linear behaviour with side slip angle, both
forces and dynamic behaviour (e.g. reduction of the
relaxation length)
• combined slip (braking and steering at the same time)
• includes gyroscopic effects of the rotating tyre belt
• can drive over short wavelength obstacles
• includes bottoming of the tyre
• includes turnslip for steering/twisting moments at
standstill

one model which can cover all possible applications!

23 March 2006 29
Concluding remarks
• Simulation allows to identify landing gear shimmy problems
very early in the design stage, where you still have the
chance to change the design for the better…

• Modern tools as LMS Virtual.Lab Motion and TNO tyre model

MF-tyre/MF-Swift 6.0 can be applied successfully for the
analysis of aircraft shimmy stability

• Despite a long history, shimmy is still a very relevant design

issue for aircraft landing gears!

23 March 2006 30
Contact
TNO Automotive - Helmond, the Netherlands

Igo Besselink ([email protected])

PhD thesis:
Shimmy of Aircraft Main Landing Gears
TU Delft, 2000

www.delft-tyre.com
www.tno.nl

23 March 2006 31