Advanced helicopter flight simulation with controller in the loop

ADVANCED HELICOPTER FLIGHT SIMULATION

WITH CONTROLLER IN THE LOOP

ThomasHablowetz, Thomas Mannchen, Prof. K.H.Well

Institute of Flight Mechanics and Control, University of Stuttgart

Pfaenwaldring 7a, 70550 Stuttgart, Germany

Homepage: www.ifr.uni-stuttgart.de

Abstract

Inthispaperasimulationtoolfor exible

rotor-craft and its application to helicoptervibration

controlispresented. Anintegratedaerodynamic,

structural and control simulation environment

based on proven engineering tools in

computa-tional uiddynamics,multibodysimulationand

controldesignisdeveloped. Thesimulationtool

allows aeroelasticanalysis and controller in the

loopsimulationof exiblerotorcraft. Duetothe

modularapproachtheaccuracylevelofthe

struc-turalsimulationcan beadaptedto therequired

analysistaskandtheaerodynamicmodelcanbe

choosen accordingto the aerodynamic

phenom-ena which haveto be considered. A model ofa

BO105rotor equipped with anindividual blade

root control systemisset up. Resultsof afully

coupledsimulationcomprisingstructure,

aerody-namicsandcontrolarepresented.

1 Introduction

Rotorcraftapplicationsarewidelyspreadin

aero-nauticsnowadays. Theiruniqueabilitiesin

trans-port, surveillance and air rescue have ensured

their great success in modern society. Typical

problemsregardingpassengercomfortandbroad

acceptancelieintheinherenthighvibrationand

noiselevelduetotherotatingliftproducing

mech-anism. As a result of vibration all structural

parts related to therotatingmechanicalsystem

havetobeinspectedandreplacedinrathershort

timeintervals. Duetohighmaintenancecostbut

alsotopassengercomfortitisofgreatinterestto

furtherreduce vibrationandinteriornoise.

Fur-thermorestringentrestrictionsonnoiselevelfor

yingin denselypopulatedareasrequirefurther

reductionofemittedsound pressurelevel.

Much research has been conducted to

tacklethese problems. Promising new methods

in vibrationreduction havebeenintroduced by

using active control strategies 1

like higher

har-monic control (HHC), 2

individual bade control

(IBC), 3;4;5

theactivelycontrolledtrailingedge

ap (ACF) 6

and activecontrol of structural

re-sponse (ACSR). 7

Regardingthe noiseproblem,

control studies have been focussing onthe

heli-copter interior noise-eld, 8

whereas the area of

noiseemissionresearch isconcentratingon

ana-lysis and prediction of sound pressure

experi-encedintheneighbourhoodoftherotorcraft. 9

Throughadvancesinsmartstructure

tech-nologytheactivecontrolstrategiesjustdescribed

becomemorelikelyto beput into practice. For

assessing the benet of smart structures in

ro-torcraftcontrolandaerodynamicsveryaccurate

models of the structure- uid interaction are in

demand.

The present work demonstrates the

ap-plicationofasimulationenvironmentcomprising

structuralandaerodynamicmodels 10

totheeld

of helicopter vibration simulation and analysis.

The accuracy level of both the structural and

the aerodynamic model is modularly adaptable

asrequiredbytheanalysistobeperformed(e.g.

ightsimulationoraeroelasticanalysis). Forthe

aerodynamics it can be chosen eighter a 'fast'

blade elementin owmodel 11

withaerodynamic

coeÆcients tuned by ight test result, 12

a

vor-tex latticemethod(2Ddiscretization) 13

ora3D

Eulermethod. 15

Theoverallsimulationmodelisbasedon

ageneralpurposemultibody codeSIMPACK, 17

developedatDLR(GermanAerospaceResearch

Center)allowing bodies to be modelled asrigid

or exible. Thusonthestructuralsidethereisa

choicebetweenapurelyrigidbodymodel,elastic

beammodelsfor therotorblades andarbitrary

FEMmodelsforbothbladesandfuselage. Force

elements are used to model applied forces and

torques such as aerodynamic loads, or

interac-tionbetweenthebodies,resultingfromdampers,

springs,actuatorsorcontact.

Computer oriented procedures called

'multibody formalisms'areusedtogeneratethe

equations of motionfor thesystemin a general

form. Here,theequationsofmotionareprovided

in statespace representation,i.e. aminimal set

Back to Session Subjects

linear dierential equations, in which the

con-straintforceshavebeeneliminated. 17

Tointegratethecontroller,thesimulation

engine of MATRIXx/SystemBuild is linked via

TCP/IP interface. In caseof structural

model-ingbasedonFEM,theFEmodelisincludedina

preprocessingstep. Complexaerodynamic

mod-els are provided by co-simulation via TCP/IP

standard interfaces to CFD codes developed at

the Institute of Aerodynamics at University of

Stuttgart.

13;15and16

Otherexamplesofmultibodymodelingin

helicopter simulation canbe found in 24{ 26

with

particularfocusonrotordynamics. An

applica-tion of MBS simulation to the identication of

controllerdesignmodelsisreportedinBertogalli

etal. 27

Alltheseworksaremainlybasedon

an-alyticalin owmodelsasreviewedin, 28

whichof

course is reasonableand suÆcient for real-time

simulationpurposes. Usageof MBSin

aeroelas-ticanalysiswithbasicaerodynamicwake

model-ingisintroducedbyMantegazza etal. 29{31

InthispaperaMBSmodelofaBO105

he-licopter will be developed. The simulationtool

is set up to model the BO105 rotor equipped

withindividualbladerootcontrol(IBRC)

actua-tors. Fullscalewindtunnel investigations 32

and

ight tests 33

have been successfully conducted

with this helicopter. In the rotor system the

rigid pitch link rods are replaced by hydraulic

actuators. This allows an individual control of

the pitch angle of each blade superimposed to

theconventionalcontrol viaswashplate.

Resultsofarotordynamicssimulation

comprising structure,aerodynamicsandcontrol

from the MATRIXx/SystemBuild environment

arepresented.

2 Coupled Simulation

Environment based on

Multibody Simulation Code

2.1 MultibodySimulation

Multibody systems (MBS) are models of

tech-nical systems consisting of rigid or deformable

bodies. Thebodiescontainmass,inertiaand

ge-ometricalproperties. Theyareconnectedtoeach

other or to theenvironment bymeans of joints

and forceinteraction. Theenvironmentmay be

an arbitrary moving reference frame or just

in-ertialxed. Jointsdenotetherestrictionofeach

bodyto movewithatmost6degreesoffreedom

to each other depending on restrictionsdened

by neighbouring bodies. Force interaction

de-notestheforceinterferenceofsuchbodiesorthe

Ageneralmultibodysystem,asconsideredhere,

isshowningure1. Amethodtoprovide

dier-global reference frame

(inertial or accelerated)

joint with

kinematic

excitation

prismatic joint

universal

joint

damper

actuator

coupler

spring

external force

body

(flexible)

rubber

bearing

sensor

sensor

body

(rigid)

external force

controller

input signal

friction force

body

(rigid)

Figure1: GeneralMultibody System

entialequationsto describethe MBSbehaviour

iscalled amultibodyformalism.

Multibodyformalismsarecomputer

algo-rithmstogenerateautomaticallytheequationsof

motionforsystemsofthegeneralformshownin

gure1.Thesearebasedondata,whichdescribe

the system elements and system topology, i.e.

thewaythenodesonthesystembodiesare

con-nectedbyforceelementsandjoints. Twogroups

of formalismsmay be distinguishedresulting in

basically dierent typesof equationsof motion.

Therst groupyieldstheLagrangianequations

oftype1,whichcontaintheunknowngeneralized

constraint forces in terms of Lagrangian

multi-pliers. Thesedierentialequationsare

accompa-nied by a set of algebraicconstraint equations.

The resulting representation of the system

mo-tion is sometimes called the descriptor form of

the equations of motion. It is simple to

gen-erate, but it requires the numerical solution of

dierential-algebraic equations. 18

By contrast,

thesecondgroupofformalismsprovidesthestate

spacerepresentationofmotion,i.e. aminimalset

ofrstorder(kinematicalanddynamical)

dier-ential equations, in which the constraint forces

have been eliminated. Numerical methods for

solving these equations are often considered to

be more mature with respect to computational

eÆciency. The starting point for the

develop-ment of both typesof formalismsare the

equa-tions describing the motion of a representative

systembody i,acteduponbytheapplied

exter-nal and internal forces and torques due to the

force elements and the unknown internal joint

Themotionofanarbitrarybodyisdescribedby

itsposition(x

I

)anditsvelocity(x

II )vector: x I = 2 4 r  q 3 5 x II = 2 4 v ! _ q 3 5 (1)

Thesevectorssatisfythekinematic

equa-tionsofmotion _ x I =X(x I )
x II ; X(x I )= 2 4 E ~r 0 0 X ang 0 0 0 E 3 5 (2)

whichexhibitthelineardependencyofthe

deriva-tives of the position variables with respect to

timeonthevelocityvariables.

The general intend of formulating

equa-tionsofmotionincomputationaldynamicsisnot

tondnicelyproblem-adoptedequations,

more-over it is aimed to formulate the algorithm in

a way to scope with a broad variety of model

classesinregardtocomputationaleÆciency.

BasedonHamilton'sprincipleoneyields:

Mx_ II =h a +h c (3) The matrices M, h a and h c denote the

generalized mass matrix, the applied and

con-straintforcesrespectively. Theappliedforcemay

beseparatedinto: h a =h ! +h g +h e +h p +h f (4) Inthisexpressionh !

aregeneralized

iner-tia forces dueto angularvelocity! of thebody

referenceframemotion. Theyareaswell asthe

gravitationalforcesh

g

distributedoverthebody

volume V

0

. The generalized internal forces h

e

resultfromelasticbodydeformationwhereash

p

is due to externalsurface forces. Thelast term

h

f

represents the forces and moments applied

by force elements attached to the body, they

are known functions of the system states and

possibly additional quantities as shown in

g-ure 1. Forces arising from joints are unknown,

theyyieldtheconstraintforces. Thegeneralized

massmatrixM can bepartitioned accordingto

the6+n e velocityvariablesx II : M(x I )= 2 4 mE ::: sy m: m~c I ::: M et M er M ee 3 5 (5)

The scalar m represents the body mass,

matrices c and I stand for the distance vector

ofcentreofmassfrombodyreferenceframeand

tively. The sub-matrix M

ee

contains the

gen-eralized masseswith respectto themodal

coor-dinates q, arising from Ritz approximation

ac-counting for elastic bodies. The matrices M

et

andM

er

containthecouplingtermsofreference

motionanddeformation,respectively.

The separationof body motion into

ref-erence motion and deformation leads to a

cor-respondingseparationoflinearandangular

mo-mentum vectorforthebody. Allof the

general-ized forcesand masses in equation (3) are

alge-braicexpressions, containing thestatevariables

(1)andintegralsovertheshapefunctions. 19

Havingderivedkinematicandkinetic

equa-tionsforonebodybenetisdrawnofthegeneral

tree structure of mechanical systems. Thus the

equations canbe applied to each body coupled

by constraint equation restricting relative

mo-tionamongthem. Systemswithkinematicloops

aretransformedtotreestructuredsystems. The

loop closing constraints, obtained as algebraic

equations,formtheDierentialAlgebraic

Equa-tions (DAE). Special adopted solversare

devel-opedto scopetheseproblems. 18

Application examplesas considered here

appeartobeoftreestructure. Theimplemented

recursiveequationsetupschemeyieldsthe

non-linearequationsofmotioninexplicitform

_

x=f(x;u;t) (6)

where x_ arethegeneralizedstates(positionand

velocity)ofthesystem. Thevectorudenotes

in-putstothesystem. Notethebenetin

computa-tional eÆciency oftheso called O(n) formalism

to generate explicit ODE by avoiding inversion

of theoverall systemmassmatrix 20

(processing

timeincreaseslinearwiththenumberofbodies).

2.3 MBSInterfaces via IPC Coupling

Adressingmulti-eldproblemssuchasthe

struc-ture- uidcouplingofelasticaircraft, the

under-lyingMBScodeSIMPACKoersthepossibility

to interfere when creating the equation of

mo-tion (6). This can be achieved by means of so

called UserRoutinesallowingforcodingofuser

dened functionality. Regarding equation (4),

thiscorrespondstointroduceuserdenedforces

h

f

(x;u;t) tothe system. At thesametimethe

full state vectorx is made available. These are

exactlythevaluesneededtomatchtheboundary

MBS

includes Nonlinear Kinematics & Dynamics

Rigid Blades/Fuselage

Flex. Blades (Beam) /Rig.

Fuselage

FEM Blades/Fuselage

Analytical Inflow Model

"Fast Aerodynamics"

Surface Discretisation

Vortex Lattice Method [13]

Full 3D Discretization

Euler Method [15]

Controller Model

dW

Vt Vc Vi dF

dA

θ

r

φ

α

e Trailing Edge Leading Edge Free Wake

bound vortices

Plant

Controller

IPC

IPC

Figure2: ModularModelingandIPC SimulationInterfaces

The mechanicalsystem, dened by aset

ofrigidand exiblebodiesissubmittedto loads

bynodalforcesand torquesto approximate the

continuous distributed uid forces. The uid

eldwithitscommoncontactareastothesurface

ofthebodieshastofulll thekinematic

bound-aryconditionsgivenbythebodysurfaceposition

and velocities. This requiresthe choiceof

uid-structurecontactsurfaces(wetsurface)forwhich

anodaldiscretization hastobedone. Forthese

nodes,kinematicsaremadeavailabletothe uid

solver,whichcalculatestheresultantnodalforce

andtorqueload.

Considering a typically coupled problem

such asanaircraft wingin free owconditions,

the discretization and solving of the uid grid

requires afar largeramountof processorpower

andmemoryconsumptionthantheparticipated

structuralsolver.Forthisreason,thePanel-and

Euler - uid solvers have to run on high

per-formancemultiprocessorcomputerstoguarantee

resultsinreasonabletime. TheMBScodeasthe

structural solver requires about ten percent of

overallsimulationtimeand runseasily on

stan-dard Unix orPC workstations. To enable

nec-essarysoftwarecommunication,anInterProcess

Communication (IPC) scheme had been

devel-opedand setup. 16

Itenablesplatform

indepen-dentdatacommunicationviaInternet. Another

important interface is the possibility of having

linked theMBScodeto controlsystem analysis

programs. Afarsimplercaseofacoupled

multi-eld problem isthat of controller-MBS

interfer-ence, someoutputorstatequantitiesofthe

me-chanicalsystemaremeasuredandfeedbackbya

mostly linearfeedbacklawto generateactuator

signals. Animportantissueisthepossible

intro-ductionofalgebraicloopsintotheoverall

simula-tionbydirect-feed-throughterms. Thishastobe

accountedfor choosing anumerical solver

algo-rithm. Forgeneralityandsimplicityofthe

over-allsimulationscheme,thesameinterfacemethod

via IPC hadbeenused to link control loops

es-tablishedin MATRIXx/SystemBuild.

2.4 Modular structural modeling

Depending on the application example

consid-ereddierentlevelofcomplexityarepossible(see

gure2). Thesimplestoneisthepurlyrigidcase.

All bodies in the helicopter model are selected

to be rigid. This might be suÆcient for trim

calculationsand necessarywhenusing the

over-all simulation for real time simulation purpose.

Thenextstageofcomplexityisgivenbyselection

of exibleblades. HerebyeitherEuler-Bernoulli

beams are available orarbitrary complex beam

or shell models for rened FE-modeling of the

blades. The exible bodies are hereby set in

a preprocessing step, in case of FE modeling a

modalanalysis has to be performed. The

high-est levelofmodelcomplexityisgivenbyfullFE

modelingof boththe rotorbladesandthe

fuse-lage. This mightbe necessaryfor investigating

vibrationlevelinsidethecabinatthepilotsseat

oratlocationsofsensitivepayload.

Methods ofmodeling exible bodiesin a

multibody system have been reviewed in

Sha-bana. 21

Here the oatingframeofreference

for-mulation will be used. Inthis methodologythe

motion of a exible body is subdivided into a

erenceframe,whereasdeformationisthemotion

ofthepointsofthebodywith respecttoits

ref-erenceframe. IntroducingaRitzapproximation,

oneobtains arepresentation of thebody

defor-mationbyareducedsetofmodalvariables. 22

Thedeformationsareassumedtobesmall

which holds for many applications.

Simplica-tions due to linearization can be applied to

in-crease computational eÆciency. Incase of high

acceleration, e.g. due to high rotational

veloc-ity (helicopter application), high inertial forces

act upon the body. If the stiness in direction

of inertia load is high, the system deformation

remains small. However,in this caseadditional

termsinthelinearisedequationshavetobe

con-sidered,socalled 'geometricstinessterms' 19,23

whichareaccountedforintheMBScode.

2.5 Modular aerodynamic modeling

Forthecoupledaeroelasticand ightmechanics

simulationdierentstagesofaccuracy(andalso

processingspeed)ofaerodynamicmodelscanbe

chosen(gure2).

For trim and 6 dof ight simulation an

analyticalbladeelementtheoryisavailable. The

lift coeÆcients are either given from tables or

might be adjusted (tuned) by simulation with

aerodynamic models ofhigherlevel ofaccuracy.

Thedatafordragareestimatedfromproledata.

Thebasictaskusingananalyticalapproachlies

in thedetermination of thelocal induced

veloc-ity. The method applied hereis thecalculation

ofameaninducedvelocityfortherotordisk

ful-llingmass,energyandmomentumconservation

for the rotor as an entity (momentum theory).

This results in aradial constant mean induced

velocity. The combination of momentum and

blade element theory givesaradialdistribution

ofinducedvelocityandallowstheconsideration

of local geometrical and aerodynamical

param-eters. Bymeans of userdened force elements,

this method has been directly implemented in

theMBScode. 11

Furtherwork isin progressto

includedynamicin owmodels 28

toimprovethis

fastmethod.

Thenextlevelofaccuracyisdenedusing

a panel method, the 'Rotor Free Wake Vortex

Lattice Method' (ROVLM). 13

This panel code

followslinearvelocitypotentialtheory. The

dou-bletstrength ofeachnew spanwisewakerowat

the end of each time step is obtainedfrom the

blade trailingedge panels ofeach spanwise

sec-tion. TheROVLM code had to be modied to

usethecommonIPCinterfacedatafor

kinemat-icsofthe'wetsurface'nodesaswellfortherigid

body motionof rotor and fuselage ofthe MBS.

Thus the actual rotorgeometry and its full

ve-Havingresolvedpressurefrom velocity

distribu-tion,localforcevectorsforeachbladepanelcan

becalculated. These are transferedbackto the

MBScouplingnodestoapplyaerodynamicload.

Basicsofthemethodanditsmodicationcanbe

foundin theliterature. 13,14

The highestlevelof accuracyavailable is

couplingtheMBSmodeltoanite-volumeEuler

method called INROT. 15,16

The physical laws

of conservation ofmass, momentum andenergy

constitute thefounding equationfor all

aerody-namicequation. Applyingtheseequationstoan

innitesmallcontrolvolumina,oneyieldsa

sys-tem of nonlinear partial dierential equations,

whicharewellknownastheNavierStokes

Equa-tions. Itssolutionforpracticalproblemsisquite

diÆcultintermsofprocessingtimeandmemory

requirement. Neglectingeects such asfriction

and heat transfer the equation simplify to the

Euler equations. Regard to parallel processing,

INROTusesthesocalled Chimeratechniqueto

discretisizethe3D uideld. Thistechnique

al-lowscomputationaleÆcientdiscretizationofthe

eld in case of relative motion among dierent

aerodynamicbodies. Incaseofthehelicopter

ap-plicationthere are individual gridsaroundeach

bladeandthefuselage. Allindividualbodygrids

moveinabasegridwhichcoverstheentire

com-putational domain. In contact regions of the

gridstheboundaryconditionsare fullled.

Ref-erencesimulationshavebeenconducted, 16

aeroe-lasticinvestigationsonrotaryandxedwing

ap-plications areinprogress.

3 Multibody model of BO105

helicopter

In the following an application examplewill be

presented. The helicopter considered is a four

bladed Eurocopter BO105 helicopter. A

topol-ogy map of the system with one representative

bladeisshowningure3.

The MBS model set up consist of four

ridid and four exible bodies forming a typical

chain-likestructure of the MBS.The rst rigid

bodyisadummybodywhichisdrivenin

transla-tionalx-directionbyakinematicexcitation

func-tion to maintain constantforwardvelocity. On

thisbodytherigidfuselageisattachedviaazero

dof joint. Thejointinbetweenis used to preset

fuselage pitch angle in forward ight condition.

Ontop,arigidrotormastcontinuesthebodytree

oftheMBSconnectingtherotatingrotorbodyto

thefuselage. Aconstantangularvelocityof44,4

rad/secisensuredusinganotherkinematic

exci-tationjoint. Byanaxisosetof0.25mfour

con-x

z

y

I_sys

x

z

y

(br) Fuselage

y

z

x

Fuselage__Rotor

y

z

x

(br) Rotor

y

z

x

Rotor__Fuselage

y

z

x

ROTAXIS_1

x

z

y

Rotor_DumA_Blade1

x

z

y

x

z

y

x

z

y

x

z

y

x

z

y

x

z

y

DumA_Rotor_Blade_1 (br) DumA_Blade_1 DumA_DumB_Blade_1 (br) Dummy_B_Blade_1 DumB_DumA_Blade_1 DumB_Blade_1

x

z

y

Fuselage_rot

Dummy_B_Blade1 0 DOF Dummy_A_Blade1 Rev y

x

z

y

x

z

y

x

z

y

Blade_1 Rheonom Sin (br) Blade_1_1 Blade_1_x Blade_1_xmax Bladetip

x

z

y

Fuselage 0 DOF Heli_dummy 0 DOF Rotor Const Ang. Vel.

Heli_dummy Flugrichtung

Neues Model

x

z

y

I_sys

x

z

y

$J_Heli_dummy

Kinematic

Excitation X

Flight Direction

Joint_1

0 dof

Joint_2

0 dof

LEADLAG

Damper

U(12)

U(13)

U(14)

Kinematic

Excitation X

AERO Force

AERO Force

AERO Force

Figure 3: TopologymapofMBShelicopter

nectingjointsofthedummybodiesinbetweenare

chosen to be zerodof accounting for themodel

of ahingelessrotor. Thebuild-inorientation of

the blade attachment points for the particular

rotorarerotatedby2,5degreestosetabuild-in

preconeangle. Eachbladeisattachedbymeans

ofaonedofkinematicexcitationjointimposing

the blade pitch angle. The blade apping and

lead-lag motion is accounted for in the

exibil-itydistributionoftheeuler-bernoullibeamused.

Each blade is discretisized in twoways. Firstly

thestructuraldiscretizationintosevenbeam

sec-tions dening spanwise stiness and mass

dis-tribution. Thesecond discretization means the

aerodynamiccouplingnodesdeningthe

aerody-namic center line of the blade. Figure 4 shows

these 26markerpointsand theirrepresentation

as panel grid points in the vortex lattice code.

Aconstantlinearbladetwistof-8degreesis

ac-countedforinthepanelcode.

4 Investigation of Vibration based

on Coupled Simulation

Forhelicopter vibration control a measurement

ofsomeorallofthefollowingquantitiesisneeded:

Forces and moments at the rotor hub,

acceler-ations at specic points of the fuselage

respec-tively in the cabin and/or accelerations at the

rotor blades themselves. Based on these

mea-surementsappropriatecommandsforthe

actua-torsarederivedbyacontrollerandarefedback

totheIBCinputs. Therebythepitchanglesand

consequentlythe bladeloads arechangedin

or-dertoachivetheaimofvibrationreduction.

Tobeusedinhelicoptervibrationcontrol,

asimulationtoolmustprovidetherequired

mea-surements and accept the necessary inputs for

0.2

Z

-1 -0.5

X

1 1.5 2 2.5 3 3.5 4 4.5

Y

Figure 4: Aerodynamic Discretization of the

blade

the actuators. Here appropriate interfaces are

realised as non-linear user code blocks (UCBs)

in theMATRIXx/SystemBuildenvironment.

In the process of designing a vibration

controller,theresponsesofactuationhaveto be

determined. Thusthetransferfunction from

ac-tuatorinputtosystemoutputintermsof

vibra-tional responsesattherotorhubisofinterest.

Todemonstratethecapabilitiesofthe

pre-sented simulation tool for helicopter vibration

control, an appropriate open loop control

sim-ulation has been performed. For that purpose

signal generators have been set up in the

MA-TRIXx/SystemBuildenvironmenttogeneratethe

desired IBC inputs. The resulting outputs are

saved. Forclosedloop vibrationcontrol the

sig-nalgeneratorsarereplacedbythecontrollerwhich

calculatestheIBCinputsfromthemeasured

quan-tities.

4/rev Phase Angle [deg]

270

180

90

0

360

4/rev Magnitude drag/side force [N]

300

200

100

0

400

baseline drag force

baseline side force

drag force

side force

Figure5: 4/revdragandsideforceatrotorhub

vs. IBC phase angle for single harmonic 4/rev

collective control with amplitude 0.2 deg,  =

4/rev Phase Angle [deg]

270

180

90

0

360

4/rev Magnitude roll/pitch moment [Nm]

200

150

100

50

0

250

baseline roll moment

baseline pitch moment

roll moment

pitch moment

Figure 6: 4/rev roll and pitch moment at

ro-torhubvs. IBCphaseangleforsingleharmonic

4/revcollectivecontrol withamplitude 0.2deg,

=0:26

Figures5and6exemplarilyshowthe4/rev

magnitudesofthevibrationalforcesandmoments

attherotorhubasresponsestosingleharmonic

4/revcontrolinputdependentontheIBCphase

angleincomparisonwiththevibrationswithIBC

o (baseline case). The calculations are done

forthehelicoptertrimmedinforward ightwith

=0:26 andacollective4/revIBCinput with

anamplitudeof 0.2deg.

ThesimulationshowsthattheIBCinputs

haveaconsiderableeect ondragandsideforce

vibrationswhereastheamplitudeof0.2deg

ap-pearstobetoosmalltoallowasignicant

vibra-tionreductionin the rollandpitchmoments at

therotorhub.

X

Y

Z

Figure7: Floweld fromcoupledsimulation

af-terfour revolutions,=0:26

Theaerodynamicsarecalculatedbya

vor-texlatticemethod(ROVLM),the oweldafter

fourrevolutionisshowningure7. Attheouter

boundariesofthe oweldthepanelscanbe

ob-servedtorolluptobuildtherotorwakevortex.

5 Summary and Conclusions

Needs for controller design and verication but

also for basicstudies of physical phenomena in

helicopter vibration control led to the further

development of simulation and modeling

capa-bilities. This paper presents the application of

a simulation environment for exible rotorcraft

to the eld of vibration control. The tool is a

modularensembleofprovensoftwaretools,each

of them highly specialized in its own

engineer-ing discipline. Thecenterlink toall modules is

a general purpose multibody code. Depending

on the issue to be investigated, the structural

modelcanberenedandextended bynite

ele-mentmodelsofcertainbodies. Theaerodynamic

model canbechoosen accordingto the

aerody-namicphenomenawhichhaveto beconsidered.

Havingreviewedthebasicideasof

multi-body simulation the modular concept of model

renementon boththestructural andthe

aero-dynamic part were presented. In the sense of

aerodynamics and controller link the modular

conceptisrealizedonthelevelofdataexchange

via inter process communication. In the

MA-TRIXx/SystemBuild environmenttheinterfaces

are realised as non-linear user code blocks and

thus allow maximum freedom in the choice of

controllertesting andimplementation.

Remark-able advantages ofthe modular conceptare

de-centralizedcalculation,maintainanceand

devel-opmentoftheparticipatedsoftwaretools.

The application to individual blade

con-trol of rotors shows the capability of the

pre-sentedsimulationtooltobeusedinhelicopter

vi-brationcontrolresearch. Thisestablishesabasis

for validating vibration reduction controllers in

averyaccuratesimulation,which typicallyhave

beendesignedbasedonlowerorder,lessaccurate

models.

Acknowledgement

The authorsare grateful toH. Strehlow andO.

Dieterich, Eurocopter Deutschland GmbH, for

[1] Friedmann,P.P.,Millott,T.A."Vibration

Reduc-tioninRotorcraftUsingActiveControl: A

Com-parisonofVariousApproaches",Journalof

Guid-ance,ControlandDynamics,Vol.18,No.4,1995,

pp.664-673.

[2] Richter, P., Schreiber, T."Theoretical

Investiga-tionsandwindtunneltestswithHHC-IBC",

Pro-ceedings20th EuropeanRotorcraftForum,

Ams-terdam,PaperNo.71,1994.

[3] Nitzsche, F. "Designing eÆcient helicopter

indi-vidual bladecontrollers usingsmart structures",

Proceedings AIAA/ASME Adaptive Structures

Forum,HiltonHead,SC.,1994.

[4] Dieterich, O., "Application of Modern Control

Technology for Advanced IBC Systems",

Euro-peanRotorcraftForum1998

[5] Mannchen, T.,

"Hubschrauber-Rotorblatt-regelung Entwurf mittels H

1

-Verfahren",

Diploma Thesis, IFRSR98001, Institut of

Flight Mechanics and Control, University of

Stuttgart,1998.

[6] Myrtle,T.F.,Friedmann,P.P.,"Vibration

Reduc-tioninRotorcraftUsingthe Actively Controlled

TrailingEdgeFlapand IssuesRelatedto

Practi-cal Implementation",Presented attheAHS54rd

AnnualForum,Washington,D.C.,1998.

[7] Chiu, T., Friedmann, P.P., "ACSR

sys-tem for vibration suppression in coupled

helicopter rotor/ exible fuselage model",

AIAA/ASME/ASCE/AHS/ASC Structures,

Structural Dynamics and Materials Conference

and Exhibit, 1996, Technical Papers. Pt. 4

(A96-2680106-39),1996,pp.1972-1990.

[8] Gembler,W.,Schweitzer,H.,"HelicopterInterior

NoiseReductionbyActiveGearboxStruts",54th

AHSForum,Washington,D.C.,May1998.

[9] Zerle, L., Wagner, S., "DirectBVI-Sound

Pres-sureCalculationbyUsingAerodynamicDataand

Application ofRetardedPotentials", In

Proceed-ingsofthe25.EuropeanRotorcraftForum,Rome,

Italy,September1999,PaperNo.B9.

[10] Hablowetz, T., "Advanced HelicopterFlightand

AeroelasticSimulationbasedonGeneralPurpose

Multibody Code", InProceedingsAIAA

Model-ingandSimulationConference,Denver,Colordo,

August2000,PaperNo.AIAA2000-4299.

[11] Martin, G., "Aufbau und Validierung eines

Rotor-Mehrkorpermodells", Diploma Thesis,

IFRSR97004,InstituteofFlightMechanicsand

Control,UniversityofStuttgart,1997.

[12] Lindert, H.-W., "Anwendung einer

struk-turmechanischen Methode zur Rekonstruktion

der Luftkrafte am rotierenden Rotorblatt aus

Windkanal und Flugversuchsmedaten", Ph.D.

Thesis,RWTHAachen,VDI-Fortschrittsberichte,

Reihe12,Nr.245,VDIVerlag,Dusseldorf,1994.

[13] Zerle, L., Wagner, S., "Development and V

ali-dation of a Vortex Lattice Method to calculate

theFloweldofaHelicopterRotorIncludingFree

Wake Development", In Proceedings of the 18.

European Rotorcraft Forum, Avignon, France,

September1992,PaperNo.72.

chenPanelverfahrensmiteinemMehrkorper

simu-lationsprogrammzurBerechnung derInteraktion

wischen Stromung und Strukturbei einem

Hub-schrauberrotor",DiplomaThesis,IFRSR99014,

InstituteofFlightMechanicsandControl,

Univer-sityofStuttgart,1997.

[15] Stangl, R., "Ein Eulerverfahren zur

Berech-nung der Stomung um einen Hubschrauber im

Vorwarts ug",Ph.D.Thesis,Instituteof

Aerody-namics,UniversityofStuttgart,1996.

[16] Hierholz,K.-H.,Wagner,S.,"Simulationof

Fluid-Structure Interaction at the Helicopter Rotor",

ICAS-98-2.5.1, Melbourne, Australia,September

1998,PaperNo.B9.

[17] Kortum,W.,Rulka,W.,Eichberger,A.,"Recent

Enhancements of SIMPACK and Vehicle

Appli-cations", European Mechanics Colloquium,

EU-ROMECH320,Prague,1994.

[18] Eich-Soellner, E. and Fuhrer C., "

Numeri-cal Methods inMultibodyDynamics", Teubner,

Stuttgart,1998.

[19] Schwertassek, R., Wallrapp, Dynamik exibler

Mehrkorpersysteme., Vieweg & Sohn

Verlagsge-sellschaftmbH,Braunschweig/Wiesbaden,1999.

[20] Schwertassek,R.,Rulka,W.Aspects of EÆcient

and Reliable Multibody System Simulation.,

Re-altime Time Integration Methods for

Mechan-ical System Simulation, pp. 55-96, E.J. Haug,

R.C.Deyo(eds.),SpringerVerlag,Berlin,1991.

[21] Shabana,A.A.,"Flexiblemultibodydynamics:

re-viewofpastandrecentdevelopments",Multibody

SystemDynamics1,1997,pp.189-222.

[22] Meirovitch,L.andKwak,M.K.,"Onthe

model-ingof exiblemulti-bodysystemsbythe

Rayleigh-Ritz method", AIAA DynamicsSpecialists

Con-ference,LongBeach,CA,April1990.

[23] Schwertassek, R.: "Flexible bodies in multibody

systems",ComputationalMethodsinMechanical

Systems,pp.329-363,J.AngelesandE.Zakhariev

(eds.),Springer-Verlag,Berlin,1997.

[24] Elliott, A.S., McConville, J.B., "Application of

a General PurposeMechanical SystemsAnalysis

CodetoRotorcraftAnalysisproblems",AHS

Spe-cialists'MeetingonRotorcraftDynamics,1989.

[25] Bachau,O.A.,Kang,N.K.,"Amultibody

Formu-lation for HelicopterStructural Dynamic

Analy-sis",Journalofthe AmericanHelicopterSociety,

1993.

[26] Bachau,O.A.,Lee,M.,"Multi-BodyFormulation

for Rotorcraft DynamicAnalysis", 51th Annual

Forumof the American Helicopter Society, Fort

Worth,Texas,1995.

[27] Bertogalli,V.,Bittanti, S.; Lovera,M.,

"Simula-tionandidenticationofhelicopterrotor

dynam-icsusingageneral-purposemultibodycode",

Jour-nalofthe FranklinInstitute, v336n5,pp.

783-797,1999.

[28] Gaonkar, G.H., Peters, D.A., "A review of

dy-namicin owmodelingforrotrocraft ight

lasticrotordynamicsbygeneralniteelementand

multibodyapproaches",IXWorldCongressofthe

TheoryofMachinesandMechanisms,vol.2,pp.

1650-1656,1995.

[30] Ghiringhelli, G.L.,Masarati,P., Mantegazza, P.,

"Multi-body aeroelastic analysis of smart rotor

blades, actuated by means of piezo-electric

de-vices", CEAS Int. Forum on Aeroelasticiy and

StructuralDynamics,vol.II,pp.115-122,1997.

[31] Ghiringhelli, G.L.,Masarati,P., Mantegazza, P.,

Nixon,M.W."MultibodyAnalysisofaTiltrotor

Conguration",NonlinearDynamics,vol.19,pp.

333-357,1999.

[32] Richter, P.; Blaas, A., "Full Scale WindTunnel

InvestigationofanIndividualBladeControl

Sys-temfortheBO105HingelessRotor",AHSAnnual

Forum,1994

[33] Schimke, D.; Arnold, U.; Kube, R., "Individual

BladeRootControlDemonstrationEvaluationof