bombardier aerospace lunch & learn flutter overview ... · bombardier aerospace lunch &...

Bombardier Aerospace - Confidential1-1

Bombardier Aerospace Lunch & learn

Flutter Overview Presentation

November 21st, 2016

Dynamics Group, Bombardier Aerospace, Montréal

Alexandre Jipa, Dynamics Section Chief


B.Eng. McGill University 1988

27+ years experience in the Bombardier Aerospace Dynamics Group

• Dynamics Section Chief since 2012

Aircraft Programs (In-Service)

• CL215T & CL415 Water Bombers

• CRJ Regional Jet Family (CRJ200 / 700 / 900 / 900LR / 1000)

• Global Express & Global 5000

• Challenger Family (CL604 / 300 / 350)

• CSeries (CS100 / 300)

• CF-18

• Missionized: GX-ASTOR / GX-ARMIS / CL604-Hong Kong

Aircraft Programs (In-Development)

• Global 7000

• SAAB SMII (Missionized GX)

Alexandre Jipa


Flutter Overview

Ground Vibration Testing

Flight Flutter Testing

Airworthiness Regulations (FAR)

Topics for Discussion


What is flutter?

Why is flutter important?

What drives flutter to occur?

What are the challenges in modeling and predicting flutter?

Flutter

“Some men fear flutter because they do not understand it,

while others fear it because they do”

Theodore von Kármán (1881-1963)


Collar Diagram

Definitions

• Mechanical Vibrations

• Static Aeroelasticity (Divergence, Control Reversal, Aeroelastic

Deformation)

• Aeroelasticity (Flutter, & Buffeting)

Types of Flutter

• Lifting Surface, Control Surface, Propeller Whirl, Stall, Vortex

Shedding, Galloping, and Panel

Examples

• Aircraft Wind Tunnel Flutter Models

• Aircraft In Flight

• Tacoma Narrows Bridge

Recent Aircraft Flutter Incidents

What Else Can Flutter?

Flutter


Static Aeroelasticity

Interactions between Elastic &

Aerodynamic Forces

• Divergence

• Control Reversal

• Static Aeroelastic Deformation

Aeroelasticity

Interactions between Inertial,

Elastic & Aerodynamic Forces

• Flutter

• Buffeting

• Dynamic Response

Related Fields

• Mechanical Vibrations

• Stability & Control

Collar Diagram

AerodynamicForces

ElasticForces

InertialForces

Aeroelasticity

MechanicalVibrations

StaticAeroelasticity

Stability &Control

A.R. Collar The Expanding Domain of Aeroelasticity

Journal of the Royal Aeronautical Society Vol. L August 1946


Mechanical Vibrations

Interactions between Inertial & Elastic Forces

M = Structural Inertia, B = Structural Damping, K = Structural Stiffness

• Vibration Modes

Eigensolution of characteristic equation yields the vibration modes. The

eigenvalues are the modal frequencies and the eigenvectors are the mode shapes.

Definitions

0

CRJ1000 FIN1T Mode at 4.3 Hz

6.35

times


Static Aeroelasticity

Interactions between Elastic & Aerodynamic Forces

Static aeroelastic phenomena arise when structural deformations induce additional

aerodynamic forces. These additional aerodynamic forces may produce additional

structural deformations which will induce still greater aerodynamic forces. Such

interactions may tend to become smaller and smaller until a condition of stable

equilibrium is reached, or they may tend to diverge and destroy the structure.

Definitions

• Divergence

Static instability of a lifting surface in flight, at a

speed called the divergence speed, where the

aerodynamic forces resulting from deformation of

the structure exceed the elastic restoring forces.

• Control Reversal

A condition occurring in flight, at a speed called the control reversal speed, at

which the intended effects of displacing a given component of the control system

are completely nullified by elastic deformations of the structure.

• Static Aeroelastic Deformation

Influence of elastic deformations of the structure on the distribution of

aerodynamic pressures over the structure.

StructuralRestoring Forces

Air Forces~ ¼ Chord

Elastic Axis

2D Air FoilSection

Divergence


Control Reversal

Consider an elastic straight wing with displaced aileron such that the wing rolls counter-

clockwise (looking from the rear). The aileron deflection produces a wing leading edge

down twisting moment due to an aft shift in the center of pressure. This causes a

reduction in the effective AOA of the wing and reduces the rolling moment normally

expected from a rigid wing. This twisting moment increases with airspeed until the

reversal speed where the rolling moment is reduced to zero. Reversal is aggravated by

wing sweep which normally produces torsional structural deformation under load.

Definitions

The term used to evaluate

reversal is the helix angle

(pl/U) or roll effectiveness

where,

p = roll rate,

l = wing semi-span,

U = flow velocity.

helix angle = zero at reversal.

The angle of attack distribution along the wing can be expressed as:-

Rigid Term

Damping Term

Flexible Twist

Distribution


Control Reversal

The rolling moment due to damping (i.e., helix angle) = zero at reversal.

Definitions

Ca

n b

e th

ou

gh

t o

f a

s …

Ro

llin

g M

om

ent

Der

iva

tiv

e

Can be thought of as … Airspeed or Dynamic Pressure 2=qceao/GJ


Aeroelasticity

Interactions between Inertial, Elastic & Aerodynamic Forces

Many important aeroelastic phenomena involve inertial forces as well as

aerodynamic and elastic forces. This field is referred to as (dynamic)

aeroelasticity.

• Flutter

A self-excited, oscillatory instability fundamentally associated with a

flexible structure and its aerodynamics. A rigid structure does not flutter,

nor will a flexible structure in still air. Sculptures, suspension bridges and

aircraft may all flutter provided the relative airspeed is high enough. At

the critical flutter speed, energy extracted from the air stream via the

aerodynamics sustains oscillation of the structure. This oscillation can

produce local damage or complete destruction.

• Buffeting

Transient vibrations of aircraft structural components due to aerodynamic

impulses produced by the wake behind wings, nacelles, fuselage pods, or

other components of the airplane.

Definitions


Lifting Surface

Classic bending / torsion coupling.

Types of Flutter

Control Surface

Main surface bending or torsion with control surface rotation.

Propeller Whirl

Precessional (pitch / yaw) motion or “whirl” of a propeller about its unperturbed axis

due to gyroscopic and propeller aerodynamic forces.

Stall

Flutter of a lifting surface in which the airfoil sections are in stalled flow during at

least part of each cycle of oscillation (propellers, turbine blades, compressors, etc.).

Vortex Shedding

For certain flow conditions, vortices are shed alternately from each side of a bluff

body (von Kármán vortex street). These vortices exert a periodic force on the body

perpendicular to the flow (telephone wires, smokestacks, pipelines, etc.).

Galloping

Elliptical ice buildup on electric transmission lines in freezing rain. Unstable when

the major axis is perpendicular to the airflow.

Panel

Flutter of aircraft skin panels in supersonic flight.


This is the classic main surface bending / torsion type flutter studied in the aerospace

industry. Stability is typically achieved by high main surface torsional stiffness.

Lifting Surface Flutter

Global Express

The SW1B and SW1T modes couple

in classic main surface flutter.


Structural Effects

• In the context of structural design, the most important objectives are high

torsional stiffness and a forward CG. A further method of improving flutter

speed, at short notice, is to introduce a local mass near the outboard leading

edge. Although expedient, this is not very efficient, and requires a mass of about

5% wing weight for 10% increase in flutter speed.

• However, the overriding influence is frequently the wing design loading which

dictates strength and hence, overall stiffness.

• Increasing aspect ratio usually allows a greater, more favourable separation

between torsion and bending modes, whilst the contrary is true if taper increases.

For a given wing planform sweep back is always favourable.

Lifting Surface Flutter


OWE Aircraft – Mach 0.25

SW1B Mode at 2.70 Hz

Modal D

am

pin

g (

G)

Modal F

requency (

Hz)

SAIL Mode at 0 Hz

Again, a classic flutter analyzed in the aerospace industry. Here, the control surface

rotation mode couples with the main surface bending or torsion.

Control Surface Flutter

Mass balance for manual controls or

high attachment stiffness for powered

controls. Flutter dampers also used.

Global Express – Zero Aileron Attachment Stiffness


The gyroscopic coupling causes the natural frequency of the higher mode to increase and that of the lower mode to decrease. These modes are characterized by precession type motion. In the lower frequency mode the propeller axis precesses in a sense opposite to that of the propeller rotation and in the higher frequency mode the direction of precession and propeller rotations are the same.

Although gyroscopic effects tend to couple the pitch and yaw DOF, gyroscopic action alone cannot lead to an oscillatory divergent type instability because the mechanism for adding energy to the system is lacking.

The mechanism for an energy transfer can, however, be found in the aerodynamic forces acting on the propeller. When the propeller axis is deflected in pitch, an aerodynamic vertical force and a yawing moment are developed which, for small deflections, are proportional to the pitch angle. In addition to these static forces, other aerodynamic forces proportional to the rate of change of the angular deflections are also present. Some of these air forces drive and others resist motion of the previously discussed whirl modes. A stability analysis can be performed.

Propeller Whirl Flutter

http://www.ae.utexas.edu/courses/ase363q/smithonian/5.ASApapers/ASA_lay/asa_lay.html

Consider a rigid engine propeller which is flexibly mounted so as to permit uncoupled pitch and yaw motion when the propeller is not rotating.

When the propeller rotates, gyroscopic moments are produced wherein an angular pitching velocity causes a moment about the yaw axis and vice versa.

http://www.ae.utexas.edu/courses/ase363q/smithonian/5.ASApapers/ASA_lay/asa_lay.html


Flutter of a lifting surface in which the airfoil sections are in stalled flow during at least

part of each oscillation. Generally, not a serious problem for aircraft. Of more interest

for propellers, helicopter rotors, turbine blades and compressors, which sometimes have

to operate at angles of attack close to the static stalling angle of the blades.

As the speed of a propeller is increased, a peculiar noise associated with propeller blade

flutter can be heard and considerable weaving of the propeller tips can be observed.

Stall Flutter

In the fine pitch region the blades

experience classical bending /

torsion flutter.

In the medium pitch region the

blades stall over part of each cycle

of oscillation. Here, the blade

settings correspond to the stalling

angles of the tip sections.

Torsional motion dominates.

In the coarse pitch region the

blades remain stalled throughout

each cycle of oscillation.

Fine Pitch Medium Pitch Coarse Pitch

L.H.G. Sterne Spinning Tests on Fluttering PropellersAeronaut. Research Council R. & M. 2022 1945


Von Kármán Vortex Street. For certain flow conditions, vortices are shed alternately

from each side of a bluff body. These vortices exert a periodic force on the body

perpendicular to the flow (telephone wires, smokestacks, pipelines, etc.).

Vortex Shedding

Based on experimental data,

the vortex shedding frequency,

n, behind a circular cylinder in

a Kármán vortex street can be

expressed as the variation of

the non dimensional Strouhal

No. with Reynolds No.

D = diameter of the cylinder

V = upstream velocity

= dynamic viscosityReynold No. = Inertial Forces / Viscous Forces

Str

ou

hal

No.

Circular Cylinder in Laminar Flow

Laminar Vortex Street

40 < R < 200


Vibrations of this type originate from

unfavorable aerodynamic configurations.

The classic example is the galloping of

electric transmission lines.

During a sleet storm a transmission line

may vibrate in a strong wind. The cable

span oscillates sometimes as a whole, but

more frequently with one or more nodes.

Galloping Flutter

When the oscillations become severe, the cables move irregularly, but freely, through

vertical distances of as much as 35 ft in a span of 500 ft. The phenomenon cannot be

observed every day, nor can it be seen at any specific place; it appears and disappears

suddenly. Once started, it is very persistent. Sometimes it may continue for 24 hours.

http://www.montefiore.ulg.ac.be/services/tde/new/recherche/recherche1/Solution/main_sol.htm

The cause has been shown to be sleet on the conductors. The ice forms

a cross section of a more or less elliptical shape, with the major axis

perpendicular to the wind direction. Such a section is unstable in an

airstream; the aerodynamic force exerts a negative damping component

so that, once the oscillation is started, it will continue to build up.

The observed frequencies are ~ the natural frequencies of the span.

The vibrations will stop when the ice is broken and thrown off the line.

http://www.montefiore.ulg.ac.be/services/tde/new/recherche/recherche1/Solution/main_sol.htm


He was talking about the infamous V2 rocket, some 500 of which rained death and

terror on British civilians during the last 8 months of World War II.

As the world's first supersonic weapon, the V2 fell from the sky silently, giving

virtually no warning of impending destruction. It had no wings to flutter, only tail-fin

stabilizers, and although Churchill was speaking metaphorically, he referred

unknowingly to a major obstacle that confronted the German technologists.

Panel Flutter

“The Angel of Death is abroad in the land, only you can't always hear the

flutter of its wings” Winston Churchill (1874-1965) in the fall of 1944.

The early V2s suffered from panel flutter. Indeed it is where

panel flutter was discovered. The rocket structure was a thin

container, and it was flexible, at least at the scale of forces

encountered in transonic flight. As the V2 went from

subsonic to supersonic speed, its metal skin shook apart.

About 70 early V2s crashed or veered off course as the rocket

arrived at supersonic speed. By September 1944, when the

Germans overcame the problem and began firing V2s at

London and other targets, Hitler's secret weapon was too late

and too inaccurate to seriously hinder the Allied advance.

http://www.psc.edu/science/2001/farhat/flutter_in_the_sky.html

http://www.psc.edu/science/2001/farhat/flutter_in_the_sky.html


The fluttering motion of a panel in airflow.

Flutter mechanisms exist between the 1N

(“plunge”) and 2N (“pitch") modes. Typically,

limit cycle in nature.

It is caused by the aerodynamic forces

induced by the motion of the plate. Normally,

one side of the panel is exposed to the airflow

while the other side is in still air.

In aircraft design panel flutter can only occur

in supersonic flight.

http://www.gl.iit.edu/wadc/bibliography/reportnumber/BiblioGraphySearch2.asp?lineNumber=123

The panel vibration depends on the dimensions, curvature, edge conditions, stiffeners,

static stresses, temperature gradients and material properties of the panel.

The unsteady aerodynamic forces depend on the flight condition, panel dimensions,

motion of adjacent panels, and the sealing efficiency of the cavity behind the panel.

To avoid panel flutter one can reduce the initial plate deformation, increase the plate

thickness or increase tension in the plate.

The consequences vary from; none noise fatigue failure aircraft loss.

Large amplitude panel flutter is only likely to occur for buckled panels.

Design Criteria for the Prediction and Prevention of Panel Flutter are available as

AFFDL TR-67-140.

Panel Flutter

http://www.gl.iit.edu/wadc/bibliography/reportnumber/BiblioGraphySearch2.asp?lineNumber=123


Examples

Aircraft Wind Tunnel Flutter Models

• Boeing 747

Under Wing Engine

• Grumman A-6 Intruder

Under Wing Store

• Lockheed C-5 Galaxy

T-Tail

• Lockheed Electra

Propeller Whirl

Aircraft In Flight

• Piper Twin Comanche

Tail Plane

• German Glider

Aileron

Tacoma Narrows Bridge

Boeing 747

A6 Intruder

C5 Galaxy

Lockheed Electra

Piper Comanche


• Two tests of this 4.6% scale

model took place at the NASA

Langley Transonic Dynamics

Tunnel (TDT) in 1967 & 1968.

• The purpose of the tests was to

determine the effects of the

large cowls surrounding the

engine fans on the flutter

characteristics of the aircraft.

Boeing 747Example

Under Wing Engine

• The video shows two views of the model experiencing antisymmetric flutter.

• The parameters varied during the tests were nacelle aerodynamics, engine pylon

stiffness, model mount system, and mass ratio (ratio of the mass of the model to the

mass of the volume of air flowing over it contained a specified volume).

• The nacelle aerodynamic effects on the flutter characteristics were determined by

replacing the nominal engine nacelles with "pencil nacelles" that simulated the inertia

and centre-of-gravity characteristics of the engine nacelles.

• Results from the tests indicated that the nacelle aerodynamic forces for the simulated

high-bypass-ratio fan-jet engines reduced the flutter-speed index about 20%.

• The flutter characteristics were greatly dependent on the o/b engine lateral frequency.http://www.airandspacemagazine.com/ASM/Web/Site/QT/B747Flutter.html

NASA Langley TDT

http://www.airandspacemagazine.com/ASM/Web/Site/QT/B747Flutter.html


Grumman A-6 Intruder

NASA Langley TDTMach = 0.89; Q = 155 lb/ft2

• Two tests of ¼ scale semi-span

models with an advanced

composite wing (1986 & 1987).

• The purpose of the tests was to

determine the transonic flutter

characteristics of the wing with

and without external stores.

• The video shows 2 views of the

first model being lost to flutter.

The first shot is real time. The

second shot is in slow motion.

• Tests of the clean wing, both with and without pylons, showed that

the flutter boundary was well outside the airplane's planned

operating envelope. However, during the first test it was found that

the flutter characteristics of the wing with external stores were

http://www.airandspacemagazine.com/ASM/Web/Site/QT/A6Flutter.html

Example

Under Wing Store

unsatisfactory and quite different from what had been predicted. To aid in

understanding these experimental results and the lack of correlation with analysis, a

pencil store configuration was tested. During these runs the model was lost to flutter.

• The second test served to verify that modifications incorporated into the new design

had acceptable flutter characteristics.

http://www.airandspacemagazine.com/ASM/Web/Site/QT/A6Flutter.html


• Lockheed C-5 model exhibits empennage flutter in the

NASA Langley TDT.

• A cable-mounted, six degree-of-freedom, 1/13th scale

empennage flutter model having a fuselage with stub

wings exhibits T-Tail flutter, which was a primary

concern on the aircraft.

• The first test showed that a potential vertical tail flutter

problem existed with the configuration. The vertical tail

was subsequently stiffened to eliminate the problem.

• The second video shows the tail having a pitch

instability due to lowering of the pitch stiffness and

perhaps some changes in mass distribution.

Lockheed C-5 Galaxy

http://www.airandspacemagazine.com/ASM/Web/Site/QT/C5Flutter.html

Example

T-Tail

Antisymmetric T-Tail Flutter

Symmetric Tail Flutterwith Reduced H-Tail

Actuator Stiffness

http://www.airandspacemagazine.com/ASM/Web/Site/QT/C5Flutter.html


• The first video is a demo model that was tested in a low speed wind tunnel to gain a

better understanding of the prop whirl flutter phenomena using a simple, low cost, and

relatively safe model. Note how the stable state is disturbed by the hand disrupting the

airflow in front of the propeller, resulting in dynamic instability.

• The second video is the Lockheed Electra configuration that was tested nine times

between May 1960 and December 1961 in the NASA Langley TDT. The tests were

aimed at investigating the reason for full-scale accidents, and conducting propeller

whirl flutter research. The wind tunnel tests showed that reduced stiffness engine

supports would cause the Electra to experience propeller-whirl flutter. The engine

mount systems were redesigned to provide "fail-safe" redundancies such that the

failure of any one component in the mount system would not cause flutter.

Lockheed Electra

http://www.airandspacemagazine.com/ASM/Web/Site/QT/PWFlutter.html

Example

Propeller Whirl

Two Aircraft Lost in Flight Circa 1960Propeller Whirl Flutter Due to Engine Mount

Failures caused by Heavy Landings?

http://www.airandspacemagazine.com/ASM/Web/Site/QT/PWFlutter.html


• Piper Twin Comanche

aircraft exhibits symmetric

tail plane flutter during a

flight test! The aircraft had

been modified from the

normal production standard.

Piper Twin Comanche

http://www.airandspacemagazine.com/ASM/Web/Site/QT/TCFlutter.html

Example

Tail Plane

Example #6

AileronGerman Glider

• High performance sail plane of German design with stiff

fibre glass covered wings exhibits classical aileron

flutter. The aileron rotation mode couples with the

antisymmetric wing first bending mode.

• The problem was solved by increasing the mass balance

of the aileron to inertially decouple the modes.

• Photographed from chase plane.

http://www.airandspacemagazine.com/ASM/Web/Site/QT/TCFlutter.html


• Mile long suspension bridge over the Puget sound.

• Solidly build with girders of carbon steel anchored in

huge blocks of concrete.

• First design of its type to employ plate girders (pairs of

deep I beams) to support the road bed.

• With earlier designs any wind could simply pass

through the truss, but in the new design the wind would

be diverted above and below the structure.

• Open to traffic 1st July 1940.

Tacoma Narrows Bridge Disaster

Washington State

Center Span 2 800 ft

Width 39 ft; Depth 8 ft

http://en.wikipedia.org/wiki/Tacoma_Narrows_Bridge#Film_and_Video_of_collapse

Example

Bridge Structure

• Nicknamed “galloping gertie” due to

bending type motion in windy

conditions.

• Collapsed 7th Nov. 1940 in 42 mph

wind.

• “Second Torsion Mode” at 0.2 Hz.

• Tubby (cocker spaniel dog) was the

only fatality. New Bridge

Truss Girders

http://en.wikipedia.org/wiki/Tacoma_Narrows_Bridge#Film_and_Video_of_collapse


Recent Aircraft Flutter Incidents #1

Air Transat Airbus A310 loses its rudder.

Flight #961 on 6th March 05 (Varadero Cuba to Quebec City).

90 nautical miles south of Miami, and in level flight at 35 000

ft, the flight crew heard a loud bang and felt some vibration.

The aircraft entered a dutch roll, and the captain disconnected

the autopilot to manually fly the aircraft. As the aircraft

descended, the dutch roll intensity lessened and then stopped

at ~28 000 ft. The aircraft returned to Varadero, where it was

discovered that the rudder was missing.

Air Transat A310

Missing its Rudder

The investigation concluded that the most probable cause was rudder flutter.

It is believed that some time earlier the composite rudder experienced a disbonding, which grew over time.

At the time of (and building up to) the incident, the aircraft was neither maneuvering nor experiencing turbulence. Therefore, the most significant load on the rudder would have been the pressure differential between the core interior and the ambient air at altitude.

When the damage reached critical size, it grew explosively with a sudden release of energy (loud noise) and loss of rudder stiffness.

The weakened rudder then fluttered.



Grob SPn business jet (prototype flight test aircraft #2) crashed on 29th November 06

killing the chief test pilot Gérard Guillaumaud.

Parts of the elevators and left hand horizontal stabilizer separated from the aircraft some

1,300 feet before the impact site, as the aircraft prepared for a high-speed pass over the

manufacturer’s private Tussenhausen-Mattsies airfield in southern Germany.

The aircraft had an increased span horizontal tail, elevator, and elevator tab relative to

first prototype aircraft, which had been used in the flutter clearance program.

The investigation concluded that the crash was the result of a flutter coupling between

the elevator torsion and tab rotation modes at ~ 20 Hz.

The tab is a gear-trim tab with two attachment rods and no mass balance.

Unfortunately, the tab rotation mode was too low in frequency due to freeplay in the

trim tab motor, and the elevator mass balance was largely ineffective in the torsion

mode due to being concentrated in the horn.

Crash Site Grob SPn Elevator Grob SPn



Boeing 737 elevator tab flutter.

In the event of a dual hydraulic failure, the tab lockout mechanism is released and the

tab act as a balance tab; i.e., rotates in the opposite direction to the elevator in order to

reduce the pilot forces required to move the elevator surface.

Numerous reports of fuselage vibrations which were found to be due to; (a) wear in the

tab linkages / hinges, (b) build up of de-icing fluid, and (c) damage caused by collisions

with ramp vehicles. Eg., Canadian Airlines (CDN 688) 5th Dec. 95.

FAA Airworthiness Directive 2003-NM-286-AD issued to inspect the tab control rod

assemblies and surrounding structure for looseness or damage which could result in

excessive freeplay in the tab control rods. This AD results from reports indicating loose

jam nuts and/or thread wear at the rod ends on the tab control rod assembly.

Boeing 737 Horizontal Stabilizer, Elevator & Tab


What Else Can Flutter?

• PML quantum leap rocket.

• The fins were built with an extra "tip to

tip" layer of 3oz fibreglass making their

final (average) thickness 3/16“.

• The fins appear to flutter at ~ Mach 1.0.

USS Bakula

• The post flight

examination revealed that

the fins were intact and

felt solid with only a

minor fracture at the base

on two opposing fins.

http://www.dph.com/vidroc/XPRS_2004/index.html

http://www.dph.com/vidroc/XPRS_2004/index.html


Summary Statement

Flutter is just one aspect of the field of aeroelasticity. Flutter is not restricted to aircraft.

It requires a flexible structure and airflow. It typically involves an energy exchange

between the airflow and the structure where the net energy transferred exceeds that

dissipated by the structure. There are many types of flutter.

• Lifting Surface Classic Aircraft Bending / Torsion Flutter

• Control Surface Control Surface Rotation with Main Surface Modes

• Whirl Propellers

• Stall Propellers, Helicopter Rotors, Turbine Blades & Compressors

• Vortex Shedding Telephone Wires, Smokestacks, Oil Pipe Lines

• Galloping Ice Buildup on Electric Transmission Lines

• Panel Aircraft Skin Panels in Supersonic Flow

It can be destructive, and therefore needs to be designed out.

An unstable coupling between inertial, elastic and aerodynamic forces

Accurate data and (aerodynamic & structural) nonlinearities.

Why is flutter important?

What drives flutter to occur?

What are the challenges in modeling and predicting flutter?

What is flutter?


What is a Ground Vibration Test (GVT)?

What scope of test is appropriate?

How is an aircraft prepared for testing?

What test techniques are employed at Bombardier?


Excite & Measure


Introduction

Principle of Reciprocity

Current Test Techniques

Suspension Systems

Shaker Attachment and Support

Excitation Signals

Response Measurement

Data Reduction


Excite & Measure

CRJ700 Regional Jet


A Ground Vibration Test (GVT) is a structural ground test where the fundamental

vibration modes (mode shapes, frequencies and damping ratios) of an aircraft are

measured. The aircraft is typically excited by an external force and transfer functions

of the structural response to the applied force are measured.

It is a mandatory test performed just prior to the first flight of a new aircraft type

design. Indeed, a GVT must be conducted for a new aircraft type design and for

modifications to an existing type design unless the modifications are shown to have an

insignificant effect on the aeroelastic stability of the aircraft.

At Bombardier, separate modal surveys of important aircraft components, such as

main surfaces, control surfaces, tabs and spoilers, are also performed as the structure

becomes available (i.e., prior to the complete aircraft GVT).

The test results are used to calibrate a dynamic Finite Element Model (FEM) of the

aircraft. Typically, the component tests are used to derive the component stiffnesses

while the aircraft test is used to derive the connection stiffnesses.

The tests are normally conducted with a free-free support condition. A support may be

considered to be free-free if the lowest flexible mode frequency is at least 2.5 times the

highest rigid body mode frequency. For small aircraft components this may be easily

achieved, but for a complete aircraft structure this can be a demanding requirement.

E.g., the GX, CRJ700, CRJ900 and Challenger 300 tests employed an elaborate

suspension system where all six rigid body modes were kept below 1.0 Hz.

Introduction


A weight breakdown of the test article is required. This is supported by actual weight,

CG and (occasionally) rotational inertia measurements.

The test article should also be as structurally complete as possible. Ideally, an aircraft

should be flight worthy. It is often necessary is accept certain compromises.

For a new aircraft type design, it is preferable to test both zero and full fuel states. The

use of a nonflammable fuel substitute such as SOLTROL is acceptable.

A complete aircraft test will typically measure:-

• The fundamental main surface bending and torsion modes.

• The primary flight control surface rotation modes.

• The fundamental modes of the ground, flight and multi-function spoilers.

• The nose and main landing gear leg bending and shimmy modes.

The scope of a test can be significantly reduced for a derivative aircraft.

• CRJ700 - New Type Design: Comprehensive test, 308 accelerometers, empty &

full fuel testing, Aileron, Elevator and Rudder FCS failure state testing.

• CRJ900 - Fuselage Extension to the CRJ700: Partial test, 273 accelerometers,

empty fuel testing only, (only) Elevator FCS failure state testing.

A mini-GVT is small scale test that focuses on a particular area (e.g., empennage or

wing tip) to assess the effect of small modifications such as primary flight control

system Power Control Unit (PCU) changes or lifting surface structural “beef-ups”.

Introduction


It is equivalent to:-

• Excite at one node and measure response at all nodes (fixed force).

• Excite at all nodes and measure response at one node (reference accelerometer).

Aircraft modal tests most often employ fixed force techniques:-

• Large number of accelerometers.

• Excitation provided by electromagnetic shakers at one or more nodes.

Principle of Reciprocity

The FRF due to input at location i and output at location j is identical to

the FRF due to input at location j and output at location i.

(i.e., Hij = Hji).

This is true for a linear system.

Reference AccelerometerFixed Force


At Bombardier Aerospace, the current test techniques, equipment used, and

responsibilities may be summarized as follows:-

Test Techniques

• Single Input Multiple Output (SIMO).

Symmetric and antisymmetric sinusoidal sweep excitation.

Single point sinusoidal sweep or random excitation.

• Multiple Input Multiple Output (MIMO)

Multiple point uncorrelated burst random excitation.

Quicker test times.

Need flexible stingers.

Equipment Used

• 64 channel Analogue to Digital Converter (ADC) “front end”, a workstation

analyzer with modal analysis software, up to 4 electromagnetic shakers, and

several hundred accelerometers.

Responsibilities

• The technical dynamics group prepares the test plan (i.e., defines the test

configurations, shaker locations, accelerometer locations, etc.), performs the FRF

data reduction, and prepares the test report.

• The experimental group (BAEX) conducts the test (i.e., instruments the aircraft,

acquires the data, calculates the FRF’s, etc.).

Current Test Techniques


Air Bags

• Large component test specimens such as a full span wing.

• CRJ700 full span wing modal test:-

3 airbags at the left and right hand jacking points and the centerline

(straddling the CG).

Suspension Systems

General Arrangement R/H Jacking Point Airbag

CRJ700 Wing free-free Modal Test


Soft Springs

• Medium component test specimens such

as a full span H-Tail, fin or rudder.

• CRJ700 fin modal test:-

Fin suspended by elastic cables at

the H-Tail pitch trim actuator

fitting and the bottom of the front

spar.

Foam Mattress

• Small components such as winglets,

ailerons, some elevators, tabs or

spoilers.

• Challenger 300 right hand aileron

modal test.

The strange orientation of the

aileron is due to the large amount

of mass balance in the leading

edge.

It is recommended to test with and

without the mass balance.

Suspension Systems

CRJ700 Fin free-free Modal Test

Challenger 300 R/H Aileron

free-free Modal Test


Pneumatic Isolators

• Complete aircraft GVT application.

• The aircraft is supported on high pressure pneumatic isolation canisters.

• Effective, but costly and time consuming to install.

• CRJ700 complete aircraft GVT:-

Single canister at the nose jacking point.

3 canisters at each MLG trunnion with interface structure.

Suspension Systems

Nose Jacking Point Canister L/H Main Landing Gear Canisters


Soft Tyres

• Complete aircraft GVT application where the frequency range of interest is > 8 Hz.

The “~2.5 times” rule is thus respected and the suspension is essentially free-free.

• The aircraft is supported on its landing gear with reduced (say, 50%) tyre pressures;

the highest rigid body mode frequencies is typically heave at ~ 3 Hz.

• Global Express (GX) ASTOR mini-GVT:-

Fuselage SATCOM and “Canoe” fairing and rear fuselage ventral fins.

Rigid body heave mode ~3.3 Hz; fuselage first bending modes ~8 Hz; and

lowest frequency ventral fin mode ~55 Hz.

Suspension Systems

GX ASTOR on Soft Tyre Suspension R/H Ventral Fin


Shakers should be attached to rigid structure (spars, ribs, etc.) with suitable stingers.

Normally, the shaker loads are reacted to earth through a rigid support stand.

Occasionally, for local tests of less accessible components such as in situ Multi Function

Spoilers (MFS) a softy mounted shaker support can be used. Here, the structure is

excited at frequencies substantially above the natural frequency of the suspension.

The excitation force and local acceleration response (in the direction of the applied

force) are measured with a strain gauge and an accelerometer, respectively.

Shaker Attachment and Support

Soft Shaker SupportRigid Shaker Support

CRJ700 Complete Aircraft GVT


Logarithmic Sinusoidal Sweep

• Preferred excitation signal for SIMO testing at Bombardier.

• For a symmetric structure, the signal can be applied at two mirror image locations

(in and out-of-phase) to excite symmetric and antisymmetric vibration modes,

respectively.

• Slow sweep up and down between pre-set frequency limits

(say, 2 or 3 octaves / minute).

• Lower frequency set just above the highest rigid body mode resonance.

• Upper frequency set to capture the fundamental vibration modes of interest.

• For the CRJ700 GVT, frequency sweep ranges of 2 32, 2 64, and 2 128

Hz were employed.

Step Sine Sweep

• May also be employed for SIMO testing.

• Signal changes frequency in pre-set steps.

• Dwell time at each step is specified.

• Signal can be tuned to search for modes in specific frequency bands.

• Dwell time allows modes to stabilize.

• Application is extremely time consuming.

Excitation Signals


Random or Burst Random

• Burst random has the added advantage (over random) that the signal and decay

are fully captured in the analysis block, thus negating the need for windowing.

Hence, more accurate modal damping estimates.

• Frequency range dictated by same considerations as for SIMO sine sweep testing.

• Uncorrelated burst random is the preferred excitation signal for MIMO testing at

Bombardier.

Manual

• Aircraft rigid body mode measurement on a soft suspension. Aircraft manually

pushed-pulled in sympathy with the low frequency rigid body suspension modes.

Response time history acquired and a frequency analysis performed.

• Step release of “stick-fixed” manual flight control surfaces.

Calibrated Hammer Impact

• Various sizes / weights of hammers are available with the possibility to add

additional mass to the hammer head to adjust the force level.

• Tip material (rubber steel) selected based on the required frequency range.

• Often employed for free-free modal testing of small components.

• Also, used on complete aircraft GVT’s to find local in-situ modes of spoilers,

ventral fins, MLG door, etc.

Excitation Signals


Structural response is normally measured by one or more accelerometers. One

accelerometer is used for reference accelerometer testing, while several hundred may

be used for fixed force testing.

For fixed force testing, the number of accelerometers is pretty much independent of the

size of the structure. It is defined based on the type of vibration modes to be measured

and the mode shape definition required.

Some accelerometers are best orientated in a local coordinate system rather than the

global aircraft (FS, BL, & WL) system. For example, the accelerometers located on a

winglet should follow the local cant and toe of the surface.

For a complete aircraft GVT, the measurement matrix should consider fuselage lateral

and vertical motion, main surface out-of-plane and in-plane motion, control surface

out-of-plane motion, and six DOF motion of engine nacelles / pylons. For manual

flight control systems the control circuit should be instrumented. For power systems,

the control column should be instrumented to measure the stick pumping mode.

E.g., the CRJ700 complete aircraft GVT employed 311 fixed accelerometers:-

• 53 on the fuselage, 33 per wing, 10 per aileron, 8 per winglet, 27 on each pylon /

nacelle, 31 on the fin / rudder, 68 on the H-Tail / elevators & 3 on the suspension.

Additional accelerometers were also used for local in-situ model tests; (i) 10

accelerometers for each left hand ground and multi-function spoiler; (ii) 24 for nose

landing gear; and (iii) 33 for the left hand main landing gear.



Node DirectionFS

(In)

BL

(In)

WL

(In)Description

241 Z - WRP1278 6 245

Front Spar at Rib #1

X perp. to the Front Spar axis242 X In-Plane

243 Z - WRP 1327 6 245 Rear Spar at Rib #1

244 Z - WRP 1296 34 242 Front Spar at Rib #3


246 Z - WRP1308 54 240

Front Spar at Rib #5







253 Z - WRP1342 109 235

Front Spar Rib #9







260 Z - WRP1372 157 231

Front Spar Rib at Fairing Joint


262 Z - WRP 1395 157 231 Rear Spar Rib at Fairing Joint

Local coordinate system (5° anhedral).

18 out-of-plane nodes (Z direction) on

H-Tail front and rear spars.

4 in-plane nodes (X direction) on front

spar (coincident with Z nodes).

12 out-of-plane nodes (Z direction) on

the elevator.

CRJ700 Aircraft GVT

R/H H-Tail

Measurement Matrix



Leuven Measurement Systems (LMS)

The LMS Dynamic Test and Analysis Software (Version 3.5.C) is the Bombardier

standard for modal test data acquisition and reduction.

• The CADA-X modular library of programs provides an integrated dynamic test

and analysis capability.

Test Modules include:- Fourier Monitor, Signature Monitor, Time Data

Processing Monitor, Sound Quality Monitor, Test Monitor and Structural

Integrity.

Analysis Modules include:- Geometry, CAE Gateway / Pretest, Modal

Analysis, Modal Design, FRF based Substructuring, Experimental SEA,

Principal Component Analysis, Transfer Path Analysis, Running Modes,

Acoustic Intensity, Real Time Animation, Link and the Flutter Suite.

User Programming and Acquisition (UPA) programs can be used to tailoring

the software to specific user needs.

• The Bombardier experimental group uses the Fourier Monitor to acquire test data

and calculate experimental FRF’s.

• The Bombardier technical dynamics group uses the Modal Analysis Module to

curve fit the FRF’s to obtain modal data (mode shapes, frequencies and damping

ratios). The Geometry Module is also used to create mode shape animation

geometry consistent with NASTRAN.

Data Reduction

http://www.lmsintl.com/

http://www.lmsintl.com/


Forcing Point Frequency Response Function (FRF)

Transfer function of the response w.r.t. the applied force where the response is at the

same location and in the same direction as the applied force (Hii).

• The presence of vibration modes in the test data can be observed by simple

inspection of the forcing point FRF. When the FRF is viewed in a Bode plot

format, modes appear as amplitude peaks and phase shifts across the peaks.

• The amount of damping in the modes can also be observed:-

Lightly damped modes have narrow peaks and sharp phase shifts.

More highly damped modes have broad peaks and shallow phase shifts.

• The height of a peak (g/lb) indicates how easily the mode is excited from the

chosen location.

• Comparing the forcing point FRF’s for different force levels is an important

check of system linearity. For aerospace structures, higher force levels typically

yield lower modal frequencies and higher damping ratios.

• Reciprocity Checks.

Reciprocity is another important linearity check which is normally carried

out to validate MIMO test data; i.e., the FRF due to input at location i and

output at location j is identical to the FRF due to input at location j and

output at location i (Hij = Hji).

Data Reduction


CRJ700 “empty1a” MIMO Test – L/H Nacelle Forcing Point FRF

Data Reduction

Forcing Point FRF

CRJ700 Zero Fuel “empty1a” MIMO Test

• 4 shakers employed simultaneously:-

Aft fuselage (vertical) Lower fin (lateral)

Right hand nacelle Left hand nacelle

(skewed 45° in the YZ plane) (skewed 45° in the YZ plane)


Multivariate Mode Indicator Function (MIF)

Real frequency domain functions that exhibit local minima at the natural frequencies

of normal modes.

• The number of MIF’s is equal to the number of exciters (or references).

• The primary MIF has a maximum value of unity and exhibits local minima at

structural modes. The secondary MIF will only have local minima in the case of

repeated roots. Similarly, a local minima in the tertiary MIF indicates the presence

of three closely spaced modes, etc. Thus, it is good practice to use at least one

more exciter than the maximum expected number of closely spaced modes.

• The primary MIF is a ratio of two sums:-

Data Reduction

Real H H

H2

-----------------------------------------------------

where, H represents an FRF and the summation is over the complete set of FRFs.

• MIF’s can be calculated from within the LMS modal analysis software module by

executing Index Table - MIF from the Application Specific Monitor (ASM) menu

bar (e.g., the Frequency domain MDOF ASM).


Data Reduction

Multivariate MIF

CRJ700 Zero Fuel “empty1a” MIMO Test

• 4 Exciter Locations and 308 Accelerometers

CRJ700 “empty1a” MIMO Test - Multivariate MIF’s


FRF Curve Fitters

• Frequency Domain Multiple Degree of Freedom (MDOF)

Provides the same level of accuracy as the time domain MDOF, but generally

works a little slower. Weakness is very lightly damped systems (< 0.3%), but

works better on highly damped systems. Since it operates in the frequency

domain, it is able to analyze FRF’s with an unequally spaced frequency axis.

ASM Graphical Interface:-

Data Reduction

CRJ700 GVT “empty1a”

Zero Fuel MIMO Test

(8 12 Hz Frequency Range)


What is a Ground Vibration Test?

A structural ground test where the fundamental vibration modes (mode shapes,

frequencies and damping ratios) of an aircraft are measured. The aircraft is excited by

an external force and transfer functions of the structural response w.r.t. the applied

force are measured. Mathematical curve fitters are used to extract the modal data.

What scope of test is appropriate?

A comprehensive test must be conducted for a new aircraft type design. Such tests

should consider both zero and full fuel states. The scope of a test can be significantly

reduced for a derivative aircraft. A “mini-GVT” may suffice for small modifications

to an existing type design, such as local structural “beef-ups” or a modified PCU.

How is an aircraft prepared for testing?

The aircraft should be be as structurally complete as possible. Ideally, it should be

flight worthy. It is often necessary is accept certain compromises. The aircraft should

be suspended or supported in such a way as to simulate a free-free condition.

What test techniques are employed at Bombardier?

Logarithmic sine sweep SIMO & uncorrelated burst random MIMO + LMS software.

Summary Statement


Introduction

Excitation


Flight Test Points

Data Acquisition

• Wichita Photo Montage

Real-Time Data Monitoring

Post Flight Data Reduction

• LMS UPA Flutter Suite

Bad Things That Can Happen

Flight Flutter Testing

CRJ900 Aircraft with Wing and H-Tail Exciters


Introduction

A flight flutter test is a flight test program where the airspeed envelope of the aircraft

is gradually expanded while confirming that the envelope is flutter free. The aircraft is

excited at discrete test points, and the modal frequencies and damping ratios of the

important flutter modes are measured, compared with the pre-test predictions, and

extrapolated to flutter onset. Could be regarded as … a series of GVT’s in the sky.

Final element in the flutter clearance process.

Prior analysis and ground vibration test work serves as a basis of support for such

testing, thus minimizing the risk of actually encountering flutter.

A flight flutter test must still be viewed as involving an element of danger.

Therefore, the test approach must be first and foremost concerned with flight safety.

The test aspects which are a prerequisite in this regard are, in order of importance:-

• The ability to excite reliably those modes shown by analysis to be the more

critical ones.

• The ability to measure response in these modes.

• Progressive test point sequencing, using small speed intervals.

• The employment of a safety chase aircraft.

All other considerations are basically related to time, cost and convenience:-

• Telemetry link or on-board tape recording.

• Real (or near real) time or post flight data reduction.


In order to measure the modal frequencies and damping ratios of the important flutter

modes the aircraft must be disturbed or excited.

There are three major means of excitation:-

• Inertial

Rotating or softly suspended masses are examples of inertial exciters. A

rotating mass can provide sinusoidal excitation, while a suspended mass can

provide impulse, sinusoidal or random excitation.

• Pyrotechnic

Pyrotechnic exciters are often referred to as “bonkers” and are an explosive

device which provides a tailored pulse.

• Aerodynamic

Aerodynamic excitation is the most common, and includes:-

Atmospheric turbulence

Flight control stick-raps

Sinusoidal or random inputs to a Fly-By-Wire (FBW) control system

Hydraulic oscillating aerodynamic vanes

Dynamic Engineering Incorporated (DEI) rotating cylinder exciters

Excitation


Excitation

DEI Rotating Cylinder Exciter

• Designed by Bill Reed. This exciter has become the Bombardier standard. Two

sizes are currently available (60 and 120 inch2 units).

• The exciter is a fixed vane with inner and outer slotted cylinders on the trailing

edge. An oscillatory aerodynamic load is generated by rotation of the outer slotted

cylinder. The vane shaft is strain gauged to measure the input force.

• Power is provided by a small computer controlled DC motor.

• As the outer cylinder rotates, the air flow is alternately deflected upwards and

downwards at twice the rotational frequency of the cylinder.

• The force level depends on the dynamic pressure and the degree of slot opening.

120 inch2 Exciter on with 25% Plug Schematic Diagram


Excitation


• The design provides two force levels. This is achieved by changing the direction

of rotation of the outer cylinder, and thereby closing the inboard slot sleeve via

centrifugal force. Unfortunately, this does not work very well. In certain flight

regimes, aerodynamic forces are able to overcome the centrifugal force and close,

or partially close the inboard slot sleeve. In these situations, the force levels are

often asymmetric and lower than demanded. The solution is to lock the slot sleeve

open or closed prior to flight. However, only one force level is then available for

the flight. Plugs are also available to blank off a portion of the slot opening.

• The exciters are capable of sinusoidal dwell and sweep (linear or logarithmic).

When an excitation program stops, there is a short time delay before the vanes

“home” (i.e., the outer cylinder realigns with the inner cylinder such that the slots

are open). This delay can be programmed, and is normally set at 2 seconds. The

following figure is a typical strip chart time history trace showing how the

“homing” is delayed until the airframe response has decayed to ambient levels.

L / H W in g D E I E x c it e r F o r c e

L / H W in g T ip F w d A c c e le r o m e t e r

D w e ll H o m in g S ig n a l

“Homing”

of the DEI

Exciter


Excitation


• The exciters are typically used in pairs (i.e., symmetric and antisymmetric

operation on the wing and tail). For an aircraft configured with a H-Tail, it is

possible to excite the fin modes via antisymmetric H-Tail excitation. The exciters

are controlled by the cockpit remote control panel (CRCP) which is normally

operated by the co-pilot. The following dials and switches define an excitation:-

F-Start (start frequency – Hz) F-Stop (stop frequency – Hz)

Duration (time - seconds) Vane (Wing, Tail or Fin)

Repetitions (2 means sweep up & down) Mode (Cont, Lin or Log; Cont means dwell)

Force (Low or High) Symmetry (0 or 180 degrees)

Cockpit Remote Control Panel



The aircraft should be instrumented with an array of sensors capable of measuring the

elastic response of the aircraft. These sensors should be located such that they are:-

• Responsive in the critical flutter modes.

• Capable of discriminating between these modes.

• Located on substantial internal aircraft structure (spars, ribs, frames, etc.).

• Suitable to support a vibration and buffet flight test program.

Bombardier normally uses ~25 accelerometers. Essentially, accelerometers are located

on the main surfaces to measure bending and torsion, and on the control surfaces to

measure rotation. It is also good

practice to measure NY & NZ at the

pilot’s station.

If an aircraft has manual flight

controls, accelerometers should also

be located on the tabs to measure

rotation. Furthermore, if the aircraft

has under wing engines, additional

accelerometers should be located in

the power plant to measure the

engine roll, pitch and yaw modes

CRJ700 FFT Accelerometers

Main Surface & Controls ± 10g

Fuselage ± 5g

Winglet ± 30g

Data Sampled at 512 Hz


Flight Test Points

CRJ700 used to demonstrate

the test points flown in a typical

Bombardier FFT program.

Airspeed envelope gradually

expanded from an initial

clearance of 250 kts CAS /

Mach 0.7 to VDF / MMO by a

series of straight and level test

points flown at a light aircraft

weight (low wing fuel state).

Expansion from MMO to MDF

accomplished by a series of

constant Mach dives.

Wind-up-Turns (WUT) to 2G

performed within VMO / MMO.

Mach 0.75 and 0.8 test points

repeated for a heavy aircraft

weight (high wing fuel state).

Centre tank used to maintain

wing fuel within range. CRJ700 Flight Flutter Test Points

Mach No.

Alt

itu

de

-1

00

0’s

ft

VD

VMO

15% Expanded Envelope

MD

MMO

Initial Clearance Envelope

1. Test points flown in order of increasing dynamic pressure.

2. Aircraft excited during Constant Mach expansions.

3. Aircraft load sheet held fixed – no water ballast pumping.


Data Acquisition

The Bombardier Wichita data acquisition system provide facilities for real-time data

monitoring and near real-time data reduction.

As the aircraft flies, data is continuously telemetered to the ground station.

Direct radio communication and “hot-mic”.

Ground and aircraft (Heim) tapes provide a permanent record of the data for post

flight data acquisition and reduction. An onboard SONY FM recorder is also

occasionally used when higher frequency data is of importance (e.g., qual. test data).

The flutter data includes:-

• The excitation signal and force measurement

• Accelerometer response sensors

• Flight parameters such as Mach, altitude, airspeed, fuel states, etc.

Real-Time Facilities

• Data displayed as electronic strip charts with LCD screens

(older style pen strip charts were better)

• Loral 500 work station display with software generated digital and analogue

gauges that can emulate some of the cockpit displays.

• Data can be read into the Leuven Measurement Systems (LMS) software.

Near Real-Time Facilities

• LMS data reduction such as PSD, FRF, Log-Dec, flutter margin, etc.


Wichita Photo Montage

Telemetry Antenna

SONY Tape Recorder Heim Tape Recorder DEI CRCP

DEI ExciterTelemetry Receiver



Most important function is to ensure safety of flight while the aircraft is accelerating

from one test point to the next.

Generally limited to visual inspection of the strip charts & associated LCD displays.

Monitoring becomes easier and safer, if the aircraft is actively excited during these

test point accelerations (e.g., repeated 1 sec dwell excitations).

In Bombardier’s experience, the strip chart responses provide the clearest indication

of any real-time variations in the critical flutter mode damping.

CRJ900LR FFT Strip Chart Layout


CRJ900LR Strip Chart Traces; Bank 1 - Wing

Symmetric Sweep 2 → 30 Hz Antisymmetric Sweep 2 → 30 Hz




Flight Test Control Room Layout

• Lead Flutter Engineer located adjacent to the key strip chart display and in direct

radio contact with the pilot.

• Second Flutter Engineer also monitors strip chart displays.

• Third Flutter Engineer at the LMS frequency analyser.

• Flight Test Engineer monitors cockpit display emulations.

Wichita Flight Test Control Room Layout


Post Flight Data Reduction

The primary objective is to establish the critical modal frequency and damping trends

versus airspeed and extrapolate to flutter onset.

Performed with a suite of LMS User Programing and Acquisition (UPA) programs.

The programs were initially developed during a military flight flutter test program in

1992 (pre-windows Version 2.8 of the software) and have been refined during

subsequent civil and military flight flutter test program (currently, windows based

Version 3.5).

Flight flutter test data reduction differs from ground vibration test data reduction in

that there are a limited number of response sensors and only a few may be responsive

in a particular mode. It is often necessary to carefully extract the data from a single

response channel. A good analogy may be … a surgeon rather than a butcher.

The main capabilities are:-

• Excitation inputs

• Time history playback

• Dwell-decay logarithmic decrement and PSD analysis

• FRF generation (from the frequency sweeps)

• Mach dive cross plots and PSD analysis

• Flutter margin prediction


Time History Playback

• Allows the user to view a replay of any section of time history data.

LMS UPA Flutter Suite


Dwell-Decay Log-Dec and PSD Analysis

• Performs Logarithmic Decrement and Power Spectral Density (PSD) analyses on

selected dwell-decays.

• User selects the Dwell # and Channel for analysis. The channel choice can be a

single channel or the addition or subtraction of 2 channels (symmetric /

antisymmetric enhancement).

Log-Dec Analysis

• Program performs several Log-Dec analyses per channel:-

Estimates of modal frequency and damping are based on 1 cycle of decay, 2

cycles of decay, … up to a maximum of 10 cycles.

Typically high amplitude cycles yield low modal frequencies and high modal

damping, while low amplitude cycles yield high modal frequencies and low

modal damping.

• Frequency estimated from the time between cycles, damping is taken from:-

• Parabolic interpolation of the peaks and troughs performed to enhance analysis.



Log-Dec Analysis



• Frequency domain

counterpart of the Log-

Dec program.

• Six analyses are

performed for each

channel.

• The analyses consider

different lengths of

decay time

(typically 1 2

seconds in 0.2 second

increments).

• Frequency resolution of 0.01 Hz obtained by using a 8192 byte analysis blocksize.

Additional accuracy by parabolic interpolation of the peak frequency.

• The half power point damping calculation is split into upper and lower damping

estimates (useful when two closely spaced modes are present in the decay).

• The half power point frequencies are linearly interpolated between adjacent

frequency records.

PSD Analysis



Combined Dwell-Decay Log-Dec and PSD Analysis


CRJ900LR

Right Hand Wing Front Spar Channel

Mach 0.75 at 17500 ft

AW1B Mode Estimate

Combined analysis and plotting

Available in windows version of

LMS software


FRF Generation

CRJ900LR; Symmetric Wing Sweep 2 → 30 Hz; Mach 0.75 at VMO



Mach Dive Cross Plots and PSD Analysis

CRJ900LR; Lag Corrected Dive Profiles

Mach 0.92 Dive PSD Analysis



Ground Vibration Test

Based On Measured Frequencies

And Decays

Last Test Point

Flutter Prediction Equation

Predicted Flutter Onset

Dynamic Pressure

Flu

tter

Mar

gin

-F

M

Flutter Margin

• Zimmerman method for predicting the flutter onset speed based on sub-critical

flight test data (modal frequencies and dampings).

• Method assumes a binary (two mode) mechanism and constant Mach data.

• Routh Stability criterion applied to the system’s characteristic equation yields the

following Flutter Margin (FM) equation:-

• Calculated for each flight test point.

• Points extrapolated to give the flutter

onset speed.



Flutter Margin

• Method is quite sensitive to uncertainty in the frequency data (especially when

the modes are close together). Therefore, linear extrapolation of FM against EAS

is used instead of parabolic.



Bad Things That Can Happen

In the real world of flight test, when a new civil aircraft design or an existing military

aircraft design with new wing mounted external stores or weapons is flown to the

limits of the design flight envelope for the first time, "bad things" can happen. These

"things" can be flutter itself, or problems with the aircraft design or experimental

equipment designs which emerge as the aircraft is flown faster and faster. The

following list of "bad things" is by no means exhaustive:-

• Turbulence

• Air Traffic Control

• Telemetry Drop Outs

• Bad Instrumentation

• Weak or Flimsy Secondary Aircraft Structure

• Structural and Aerodynamic Nonlinearities

• Aircraft Serviceability

• A Change in Management Priorities

• Bad Weather

• Politics

• Flutter


Summary Statement

A flight test program where the airspeed envelope of the aircraft is gradually expanded

while confirming that the envelope is flutter free. The aircraft is excited at discrete test

points, and the modal frequencies and damping ratios of the important flutter modes

are measured, compared with the pre-test predictions and extrapolated to flutter onset.

The aircraft is excited by 60 & 120 inch2 DEI flutter exciters located on the left &

right hand H-Tail and wing tips, respectively. The fin is excited by antisymmetric H-

Tail excitation. The excitation signal is sinusoidal dwell and logarithmic sweep. The

airframe response is measured by ~ 25 accelerometers. The data is sampled at 512 Hz.

The test points are group at constant Mach No.’s to MMO and flown in order of

increasing EAS. The expansion from MMO to MDF is accomplished by a series of Mach

dives. Real-time data monitoring is limited to the strip charts and associated displays.

Post flight data reduction employs Version 3.5 of the LMS UPA flutter suite.

How does Bombardier conduct a flight flutter test at Wichita?

What is a flight flutter test?


Airworthiness Regulations

Civil Regulations

• The regulatory authorities are:-

Canada Transport Canada

US Federal Aviation Authority

Europe Joint Aviation Authority

• Canadian Aviation Regulations (CAR) Chapter 525 and Federal Aviation

Regulations (FAR) Part 25 address transport category aircraft and are equivalent.

Indeed, Transport Canada has essentially adopted the US Regulations.

• The European regulations have been standardized as the Joint Aviation

Regulations (JAR) which differ slightly from the FAR.

• The FAR and JAR are currently in the process of being harmonized.


FAR 25.251 Vibration and Buffeting

FAR 25.305 (e) Strength and Deformation; Vibration & Buffeting up to VD / MD

FAR 25.343 (b) (3) Flutter & Vibration; with Zero Fuel

FAR 25.571 (b) Damage Tolerance Evaluation

FAR 25.571 (e) Damage Tolerance (Discrete Source) Evaluation

FAR 25.629 (a) Aeroelastic Stability Requirements; General

FAR 25.629 (b) Aeroelastic Stability Envelopes

FAR 25.629 (b) (1) Normal Conditions

FAR 25.629 (b) (2) Failures, Malfunctions and Adverse Conditions

FAR 25.629 (c) Balance Weights

FAR 25.629 (d) Failures, Malfunctions and Adverse Conditions

FAR 25.629 (e) Flight Flutter Testing

FAR 25.631 Bird Strike Damage

FAR 25.671 (c) Flight Control System Failures

FAR 25.672 (c) Power Operated System (Flight Control System) Failures

FAR 25.903 (d) (1) Turbine Engine Installations

FAR 25.1309 Equipment, Systems, & Installations

FAR 25.1309 (b) Aircraft System Failures; Extremely Improbable

FAR 25.1309 (d) Aircraft System Failures; Failures to Consider

FAR 25.1419 Ice Protection

FAR 25.1585 (c) & (d) Operating Procedures; Buffet Onset Envelope

Civil Regulations


FAR 25.629 Aeroelastic Stability Requirements

(a) General. The aeroelastic stability evaluations required under this section include flutter,

divergence, control reversal and any undue loss of stability and control as a result of

structural deformation. The aeroelastic evaluation must include whirl modes associated

with any propeller or rotating device that contributes significant dynamic forces.

Compliance with this section must be shown by analyses, wind tunnel tests, ground

vibration tests, flight tests, or other means found necessary by the Administrator.

(b) Aeroelastic stability envelopes. The airplane must be designed to be free from

aeroelastic instability for all configurations and design conditions within the aeroelastic

stability envelopes as follows:

(1) For normal conditions without failures, malfunctions, or adverse conditions, all

combinations of altitudes and speeds encompassed by the VD / MD versus altitude

envelope enlarged at all points by an increase of 15 percent in equivalent airspeed at

both constant Mach number and constant altitude. In addition, a proper margin of

stability must exist at all speeds up to VD / MD and, there must be no large and rapid

reduction in stability as VD / MD is approached. The enlarged envelope may be limited

to Mach 1.0 when MD is less than 1.0 at all design altitudes, and

(2) For the conditions described in §25.629(d) below, for all approved altitudes, any

airspeed up to the greater airspeed defined by;

(i) The VD / MD envelope determined by §25.335(b); or,


(ii) An altitude-airspeed envelope defined by a 15 percent increase in equivalent airspeed

above VC at constant altitude, from sea level to the altitude of the intersection of 1.15

VC with the extension of the constant cruise Mach number line, MC, then a linear

variation in equivalent airspeed to MC + 0.05 at the altitude of the lowest VC / MC

intersection; then, at higher altitudes, up to the maximum flight altitude, the boundary

defined by a 0.05 Mach increase in MC at constant altitude.

(c) Balance weights. If concentrated balance weights are used, their effectiveness and

strength, including supporting structure, must be substantiated.

(d) Failures, malfunctions, and adverse conditions. The failures, malfunctions, and adverse

conditions which must be considered in showing compliance with this section are:

(1) Any critical fuel loading conditions, not shown to be extremely improbable, which may

result from mismanagement of fuel.

(2) Any single failure in any flutter damper system.

(3) For airplanes not approved for operation in icing conditions, the maximum likely ice

accumulation expected as a result of an inadvertent encounter.

(4) Failure of any single element of the structure supporting any engine, independently

mounted propeller shaft, large auxiliary power unit, or large externally mounted

aerodynamic body (such as an external fuel tank).



(5) For airplanes with engines that have propellers or large rotating devices capable of

significant dynamic forces, any single failure of the engine structure that would reduce

the rigidity of the rotational axis.

(6) The absence of aerodynamic or gyroscopic forces resulting from the most adverse

combination of feathered propellers or other rotating devices capable of significant

dynamic forces. In addition, the effect of a single feathered propeller or rotating device

must be coupled with the failures of paragraphs (d)(4) and (d)(5) of this section.

(7) Any single propeller or rotating device capable of significant dynamic forces rotating

at the highest likely over speed.

(8) Any damage or failure condition, required or selected for investigation by §25.571. The

single structural failures described in paragraphs (d)(4) and (d)(5) of this section need

not be considered in showing compliance with this section if;

(i) The structural element could not fail due to discrete source damage resulting from the

conditions described in §25.571(e), and

(ii) A damage tolerance investigation in accordance with §25.571(b) shows that the

maximum extent of damage assumed for the purpose of residual strength evaluation

does not involve complete failure of the structural element.

(9) Any damage, failure, or malfunction considered under §25.631, 25.671, 25.672, and

25.1309.



(10) Any other combination of failures, malfunctions, or adverse conditions not shown to be extremely improbable. (i.e., 1.0*10-9)

(e) Flight flutter testing. Full scale flight flutter tests at speeds up to VDF / MDF must be conducted for new type designs and for modifications to a type design unless the modifications have been shown to have an insignificant effect on the aeroelastic stability. These tests must demonstrate that the airplane has a proper margin of damping at all speeds up to VDF / MDF, and that there is no large and rapid reduction in damping as VDF / MDF, is approached. If a failure, malfunction, or adverse condition is simulated during flight test in showing compliance with paragraph (d) of this section, the maximum speed investigated need not exceed VFC / MFC if it is shown, by correlation of the flight test data with other test data or analyses, that the airplane is free from any aeroelastic instability at all speeds within the altitude-airspeed envelope described in paragraph (b)(2) of this section.

[Doc. No. 26007, 57 FR 28949, June 29, 1992]

Airspeed / Mach Abbreviations:

VC / MC Design Cruise Speed

VFC / MFC Maximum Speed for Stability Characteristics

VDF / MDF Demonstrated Flight Diving Speed

VD / MD Design Diving Speed

1.15 VD / MD Fully Operative Aircraft Clearance Speed



The Global Express business jet is

used as an example of how to apply

the FAR 25.629 (b) aeroelastic

stability envelopes.

Fully Operative Aircraft:

VD / MD +15% VEAS at constant

Mach No. and constant altitude.

Limited to Mach 1.0 since subsonic.

Failure States:

VD / MD

Since the 15% VEAS expansion of VC

/ MC is smaller.

FAR 25.629 (b) Aeroelastic Stability Envelopes

Global Express Aeroelastic Stability Envelopes

Prior to 1992 the

expansion was 20%

VEAS. It was reduced to

15% to acknowledge

better modern day

analysis capabilities.

Mach No.

Alt

itu

de

–1000’s

ft.

51 000 ft Max Altitude Ceiling

MC=0.9

VC=340 KCAS

Sea Level VD=398 KCAS

MD=0.97

15% Expansion of VD / MD

FAR 25.629(b)(1)

15% Expansion of VC / MC

FAR 25.629(b)(2)(ii)

Flight Envelope

Fully Operative

Aircraft

Failure State

Calculation

bombardier aerospace lunch & learn flutter overview ... · bombardier aerospace lunch &...

Documents