Integrating CFD and Experiment in Aerodynamics

106 Pages·2004·23.65 MB·English

Integrating CFD and Experiment in
Aerodynamics
An international symposium to celebrate the
career of Prof. Bryan E. Richards
th th
8 /9 September, 2003
Kelvin Gallery, University of Glasgow
http://www.aero.gla.ac.uk/integration
Background
Bryan Richards retires in September, 2003 following forty years of research and
teachinginaerodynamics.
His career has involved both experimental work at Imperial College and Von
Karman Institute and CFD at University of Glasgow. The symposium is
dedicatedtotheimportanttopicofhowtouseCFDandexperimentstowardsthe
goalofimproved understandingofaerodynamics.Itisintendedtobringtogether
international experts in these fields to look forward to new ways of integrating
thetwodisciplines,andintheprocesscelebratethevariedcontributionsofBryan
Richardstoaerodynamics.
Scientific Rationale
CFDpractitionersandexperimentalistshaveacommongoalofunderstandingaerodynamics.Itistherefore
surprising that the disciplines often only interact for the validation of CFD. This provides a very limited
formofintegrationbutoftenthereisnointeractionbetweentheexperimentalistandtheCFDpractitioners.
Thissituationisunsatisfactoryfrommanypointsofviewincluding(a)the needto have anappreciationof
the flow before deciding what should be measured, (b) the desirability of having checks in place on the
experimentalmeasurementsastheyaretaken,(c)thedifficultyinmakingcertainimportantmeasurements,
(d) the need to assess the influence of the experimental techniques on the measurements, (e) the abilityof
CFD to provide detailed flow information and sensitivityat a reasonable cost for some cases, (f) the large
cost of CFD calculations for other cases, and (g)the lack of credibility for the CFD results for some flow
categories.
It could be argued that the process of aerodynamic investigation would be significantly enhanced if the
integration of CFD and experiments was much stronger. In particular, the design and reliability of
experiments could be significantly enhanced by CFD, the scope of experimental measurements extended
through CFD and the credibility of the simulation results enhanced by the availability of suitable
measurementsfromexperiments.Thissortofcloserintegrationishoweverrare.Theaimofthesymposium
is to bring together leading researchers from both fields to initiate more careful consideration of these
issuesandtostimulatenewwaysofapproachingaerodynamicstudies.
Local Organisers
KenBadcock GeorgeBarakos,
DepartmentofAerospaceEngineering DepartmentofAerospaceEngineering
UniversityofGlasgow UniversityofGlasgow
GlasgowG128QQ GlasgowG128QQ
UnitedKingdom UnitedKingdom
Phone:[phone] Phone:[phone]
Fax:[phone] Fax:[phone]
[email protected] [email protected]
http://www.gla.ac.uk/Research/CFD http://www.gla.ac.uk/Research/CFD
Scientific Committee
Prof.NingQin Prof.DanielFavier Prof.RichardHillier
DepartmentofMechanicalEng. LABM DepartmentofAeronautics
UniversityofSheffield UniversityofMarseilles ImperialCollege
UK France UK
Experimentalist(cid:146)s requirements for a safe methodology in
CFD code validation
Jean DØlery
ONERA (cid:150) Centre deMeudon, France, [email protected]
predicting flows that cannot be presentable as clearly
identified solutions of well known mathematical
Keywords:experimentaltechniques,code problems.Atthisstage,comparisonwithexperimental
validation,database dataismandatory.
In the past, predictive methods were validated by
Abstract comparison of the computed results with some
measured global quantities, such as forces and
moments, and with wall properties, namely the
In spite of the spectacular progress in CFD there is
pressure and the heat-transfer for hypersonic
still a strong need to validate the computer codes by
applications. The skin friction was more rarely
comparisonwithexperiments.Thefirstvalidationstep
available, this quantity being difficult to measure
is the assessment of the code numerical safety and
(even now) in compressible flows. However, the flow
the physical models accuracy. This validation step
prediction landscape has completely changed over
requires carefully made building block experiments.
the past 40 years with the advent of numerical
To be calculable, such experiments must satisfy
methods solving the Navier-Stokes equations or
conditions such as the precise definition of the test
approaches such as DSMC to predict rarefied flows.
set-up geometry, the absence of uncontrolled
However,intheirpresentstatetheCFDcodesarestill
parasitic effects, a complete information on the flow
far from being free of critics, since many difficulties
conditions and indication on the uncertainty margins.
persist both on the numerical and physical sides.
Under these conditions, the experiment can be put
There is thus a strong need to validate CFD codes,
into a database which will be precious to help in the
more particularly from the physical point of view
development of reliable and accurate codes. The
beforetheirroutineusefordesignpurposes3.
paper provides also an overview of modern
measurement techniques allowing a precise and
A comparison restricted to the wall properties is in
thorough description of complex separated flows.
general insufficient to validate the most advanced
Recommendation for the constitution of experimental
predictive methods. In particular, information on the
databasesareprovidedasaconclusion.
Mach number, temperature, density fields is essential
to elucidate the cause of discrepancies affecting, for
example, the wall quantities distribution. Such a
Introduction
requirement is still more demanding in hypersonic
flows where one has to represent complex thermo-
Methods for verifying the capabilityofa codeto solve chemical processes and/or strongly interacting and
given equations have been the object of close shock-separated turbulent flows. In this case,
examinations. Identification and elimination of various information on turbulence quantities is also needed,
types of errors and use of precision criteria, methods which is a formidable challenge in high Mach number
for convergence testing, rules for establishing grid flows! The prediction of shock/shock interferences
convergence,areallrequiredwhenonehastoassess which can have destructive effects on a nearby
thequalityofthenumericaltool.Thevariousstagesof structure necessitates an accurate prediction of the
the general process of verification that will give to the complex structures resulting from shock intersection.
codeaconfidencelabelpermittingtouseitfortesting The problem of code validation is crucial in three-
theoreticalmodelshave beensummedupbyRoache dimensional applications where the Navier-Stokes
with references to many authors1. The ERCOFTAC approachbecomesmandatory. Due tothecomplexity
association has issued Best Practice Guidelines of such flows, it is clear that the consideration of the
giving strict recommendations to asses code quality surfacepressurealoneisinadequate,thisinformation
for industrialcomputational fluid dynamics2. The code giving a very partial view of the flow (in three-
verification process constitutes by itself a complex dimensional flows, it is no longer possible to infer
program often partially carried out but that should be separation from an inspection of the wall pressure
completelysatisfiedintheidealcases.Asecondstep distributions).
is devoted to the validation of models aimed at
1
In these conditions, the validation of computer codes solution algorithms. A further step is to run the code
requires well documented experiments providing not on a configuration for which reliable experimental
only wall quantities but also flowfield measurements. resultsareavailable.Thispointisfarlessobviousthat
Itisremarkablethatthebreakthroughinourpredictive it would appear at first sight, since the experimental
capacity has been paralleled by spectacular datashouldallowtodrawclearconclusions.
developments in measurement techniques over
approximately the same period, mainly with the Secondstep:Validationofthephysicsimplementedin
advent of laser based optical methods, optical the code on elementary configurations. This is the
techniques having operated a true revolution in our mostimportant pointfor the specialist in flowphysics,
capacity to investigate flows containing shock waves, the first step being only a preliminary step simply
concentrated expansions, thin shear-layers and aiming at verifying the tool. In the second step, the
recirculatingregions4. code is used to compute what can be considered as
the elementary components of an aerodynamic flow:
Thevalidationmethodology attached boundary layer, laminar-turbulent transition
on a flat plate, separation induced by an obstacle,
Coderequirements:reliabilityoraccuracy? flow past a base, shock wave/boundary layer
interaction, start and development of a vortex
Before considering a validation action, the aim of the structure,vortexbreakdown,shock/shockinterference
calculationmustbeclearlystated. or shock crossing, etc. Two-dimensional - preferably
(cid:1) If calculation is used to predict the performance axisymmetric - as well as three-dimensional basic
of a system or a sub-system, accuracy is situations have to be considered. For this first
mandatory. validation step, the numerical results are compared
(cid:1) In the design of devices involving complex flows withbuildingblockexperimentsfocusingonaspecific
whose experimental simulation is not possible a elementaryphenomenon.
calculation showing the flow field topology is of
greathelp.Inthiscase,accuracyisnotessential, Third step: Validation on more complex sub-systems.
but reliability is crucial since one must be Once the code and its physical model(s) have been
confident on the physical features of the validated on basic cases, a more complete
computedfield. configuration must be considered consisting in a sub-
(cid:1) The physical understanding of complex flows system of a complete vehicle, where several
must be based on a theoretical analysis whose elementary phenomena are combined. This is the
aim is to help in the interpretation of the case of a profile on which one encounters laminar-
phenomena and in the establishment of a turbulent transition, attached boundary layers,
consistent topological description. In this case, transonic shock wave/boundary layer interaction,
accuracy is not needed since theoretical separation, wake development, etc. The wing
analyses are most often derived from simplifying constitutes a three-dimensional extension with the
assumptions rendering quantitative results additional problems of the vortices emanating from
questionable. the wing and control surfaces extremities. The
(cid:1) Inthelastissue,acodeisusedasatooltotesta supersonic air-intake involves shock/shock
newphysicalmodel.Then,numericalaccuracyis interference, shock wave/boundary layer interactions,
mandatory since it would be vain to implement a corner flows with vortex development. After-bodies
goodmodelinaninaccuratecode. combine supersonic jets with complexshock patterns
(Mach disc formation), shock induced separation,
Themethodologydifferentsteps either inside the nozzle (overexpanded nozzle) or on
the fuselage (jet pluming at the exit of an
The verification/ validation procedure has to be underexpandednozzle).Manyotherexamplescanbe
submittedtoafourstepmethodology. cited: propulsive nacelle, compressor/turbine
cascade,helicopterrotor,etc.
First step: Assessment of the code numerical
accuracy.Aconclusiveassessmentofthispointisnot Fourth step: Validation on the complete vehicle or
astraightforwardissueinthesensethatthenumerics object. This is the ultimate target in which the code is
involves many aspects. When possible, a first firm appliedtoacompletevehicle.
answer can be obtained from comparison with
analytical solutions in laminar flowor well established Each of the above steps implies an iterative
empirical results. In turbulent flow, the question is not procedure between computation and experiment to
so clear since numerical and physical modelling adapt or correct the code from inspection and
problems are closely linked. Verification by interpretation of the discrepancies between the
confrontation with other codes is not always computed and experimental results. This exercise is
conclusive since the codes may use different not entirely safe for the experimentalist, since a
numerical techniques, discretization schemes and
2
persistent discrepancy may be due to measurement The first one is a double cone configuration used to
errorsorilldefinedexperimentalconditions. produced shock-induced separation at high Mach
number, the second one is a model used to validate
Requirementsforgoodtestcasesconstitution codes predicting the flow past an axisymmetric
afterbodyequippedwithapropulsivejet.
Definition of the geometry. A first condition for an
experiment aiming at code verification/validation is to Boundary conditions. The boundary conditions must
focus on a configuration whose geometry is be well identified and accurately known. This
representative of a typical situation, precisely defined concerns the upstream flow conditions (Mach
and as simple as possible while avoiding singularities number, velocity, pressure, density) when a uniform
leading to meshing difficulties. When possible, an incoming flow exists. In transonic experiments
analyticaldefinitionofthecontourshouldbeprovided. executed in a channel type arrangement, one often
It is preferable to give the dimensions in metric units considers phenomena taking place on the channel
toavoidriskofconfusioninthereferencelengthused walls, the test section itself being the model. In this
to compute a Reynolds number. When possible, a case, a well defined origin with a uniform state at
two-dimensional geometry should be adopted - even upstreaminfinitydoesnotexist.Then,thedatashould
for three-dimensional problems - since it offers many provide all the flow conditions in a section located
advantages to visualise the phenomena and to sufficiently far upstream of the region of interest,
execute measurements, in addition of the lower cost includingtheboundarylayerproperties(meanvelocity
ofthetest-setupfabrication.Furthermore,theoriginal profile, turbulent quantities). If LDV measurements
set-upmustfrequentlybemodifiedbeforearrivingata now permit to know with a good approximation the
fully satisfactory flow; such modifications are far Reynolds stress profiles in moderate supersonic
easier on a two-dimensional/ axisymmetric flows,amethodmustbeconceivedfordeducingfrom
arrangement.TwotypicalmodelsareshowninFig.1. these data the dissipation rate of two-equation
turbulencemodels.
In all cases the stagnation conditions (pressure,
temperature) and the incoming stream
thermodynamic properties must be given.
Downstream boundary conditions leading to a well
posed problem must be provided. If the flow leaving
the zone of interest is supersonic, then no-conditions
have to be imposed to perform the calculation. The
question of the downstream conditions is more
delicate if the configuration is such that the flow
leavingthetestregionissubsonic.Whentheoutgoing
flow is again uniform, most often a downstream
pressure is given, since this quantity is easily
obtained.Itisfarmoredifficulttoprovidethepressure
field in a complete plane, as some theoreticians
a(cid:150)doubleconemodelforhypersonicseparation
sometimes ask for. In transonic channel experiments
study
where a shock is produced by the choking effect of a
second throat, the best way is to provide the
geometryof the second throatand, in the calculation,
to impose downstream conditions insuring the
chokingofthisthroat.ThephotographinFig.2shows
a test set-up which has been extensively used to
analyse shock wave/boundary layer interaction in a
transonic channel5. The entrance of the test section
has a converging-diverging upper wall constituting a
first sonic throat. At this location, the flow conditions,
including the boundary layer profiles, are provided. In
the present arrangement, an oscillating shock is
produced by a rotating shaft placed in the
b(cid:150)Axisymmetricmodelforpoweredbaseflow downstreampartofthechannel.Theshafthasalsoa
Investigation choking effect, thus providing well defined
downstreamconditions.
Fig.1:Twotypicalsimplemodelsforcode
validationpurposes
3
modelshowninFig.4,whichhasbeenmuchusedto
validate both laminar and turbulent shock
wave/boundary layer interactions6. Even in this case,
the flow adopts a three-dimensional structure at a
(cid:147)microscopic(cid:148) scale, as it can be seen in the surface
flowpatterninFig.5.
Fig.2:Testsetuparrangementfortransonic
interactionstudies
Parasitic effects. Side effects or uncontrolled
perturbations must be avoided, except if they can be
takenintoaccountbythecalculation.Thesideeffects
due to the finite span of any experimental
arrangement strongly affect the flow when separation
occurs. Then, the experimented flow can be very Fig.4:Hollowcylinder-plus-flaremodelfor
differentfromtheassumedidealtwo-dimensionalflow
axisymmetricflowinvestigation
whichwouldcorrespondtotheinfinitespancondition.
Fig.3:Surfaceflowpatterninanominally2D
transonicchannel(IMP/PAMdocument)
As an illustration Fig. 3, shows a surface flow Fig.5:3Dmicrostructuresinanaxisymmetric
visualisation realised in a nominally two-dimensional reattachmentregion
transonicchannel.Inthevicinityoftheside-walls,two
foci which are the traces of two tornado-like vortices When the goal of an experiment is to reproduce
are clearly visible. Confrontation of such an closelythebehaviourofanaircraft,thecoincidenceof
experiment with a planar two-dimensional calculation the wind tunnel’s Reynolds number with that of real
can be deprived of any signification and lead to flight tests is a condition often hard to satisfy. Even if
erroneous conclusions. When the boundary layer is this condition is fulfilled, premature transition can
attached or weakly separated this effect is small and occuronthemodel.Theboundarylayersonthewalls
2D calculations remain acceptable. If one desires to of classical wind tunnels are generally turbulent and
keep the mathematical simplicity of two space unsteady perturbations radiate from the wall to the
dimensions, the best is to compute an axisymmetric whole test channel. Such perturbations lead to
flow, as the one past the hollow cylinder-plus-flare transition of the flow around the model at Reynolds
4
numberssignificantlylowerthanthosewhereitoccurs the present context, any experiment should satisfy a
in real flying tests. The conception of quiet wind calculability criterion meaning that the experiment is
tunnels, at NASA, ONERA and Purdue University7, notusefulifitcannotbecalculated.
has been undertaken to overcome this difficulty.
Inversely,whenaplainlyturbulentregimeissearched Validation of experiment by calculation. Calculation is
for code validations, it is preferable that the wind alsousedbytheexperimentalisttocontrolhisresults.
tunnel be naturally turbulent without the help of There are simple situations where theory can be
auxiliary means like transition strips. The transition in considered as safe and accurate, so that a good way
general is indeed far from being treated satisfactorily tocheckanewexperimentalmethodistocompareits
by the methods of calculation presently at our results with the theoretical ones. In situations where
disposal and transitional effects are often present the calculation cannot be considered as entirely safe,
whenturbulenceissetupartificially. an important, persistent and unexplained discrepancy
should provoke a reconsideration of the experimental
Experimental needs. The description of the flowmust results. In hypersonic flows, measurements executed
be as complete as possible to provide all the withopticaltechniqueslikeLaserDopplerVelocimetry
information useful to understand its physics and to (LDV), Coherent Anti-Stokes Raman Scattering
help in the elucidation of disagreements. Flow (CARS), Electron Beam Fluorescence (EBF) have
visualisations are desirable to give a precise idea of been in great part validated by comparison with
the flow topology. Surface flow visualisations are theoretical results, no other techniques being
mandatory in three-dimensional flows to indicate the availabletoperformsuchmeasurements.
locationofseparation/attachmentlines.
Anoverviewofmodernmeasurement
Measurements reliability and accuracy. The techniques
experimental data mustbe reliable, which means that
the experiment is not "polluted" by an extraneous The purpose of this section is to briefly present
phenomenon due to a bad definition of the test modern measurement techniques which are used (or
arrangement or an ill functioning of the facility. The could be used) to investigate complex flows, mare
riskisatitsmaximumwhensuchaninfluencehasnot particularly at high Mach number. The classical and
been identified and is interpreted as a proper basic well proven methods will not be mentioned, even if
featureoftheflowunderinvestigation.Measurements theyarestillroutinelyused!
must be safe, as for the calculations, this aspect
being distinct from the problem of accuracy which Flowvisualisations
must also be carefully assessed. Measurement
accuracy must be in proportion with the calculation Fortunately, most flow phenomena can been
purposes and the modelling present status. Precision visualised, which is a great help for their physical
has a very high cost; in most circumstances what is understanding and modelling8. Flow visualisations
importantistobeconfidentintheexperimentalresults must be attempted in all circumstances where they
and to known (even approximately) the uncertainty are possible, specially on three-dimensional
margins. configurations for which they are nearly mandatory.
Thefirststepconsistsinmakingavisualisationofthe
The physical interpretation. Constitution of safe data surface flow pattern in order to localise the critical
base is not restricted to the execution of hopefully points that it contains and the accompanying
good experiments in relation with code development. separation/ attachment lines9. The photograph in Fig.
The experimentalist must also be a physicist able to 6a shows the complex surface flow pattern produced
interpret its findings and to understand the physics of by the impingement on the central body of the 24
the investigated flow field. This interpretation, which primary jets of a plug nozzle. Flow field visualisation
must be based on theoretical arguments, is essential can be achieved by several techniques, including the
toinsurethesafetyoftheresults.Itmustprecedeany laser sheet method (both in low and high speed
numericalexploitation. streams, see Fig. 6b), shlieren, shadowgraph of
interferometry in compressible flows (see Fig. 6c),
Theaboveconditionsleadtorejectofalargequantity electron beam fluorescence (EBF) in low pressure
of existing experiments for validation purposes. In high Mach number flows (see Fig 6d). Particle Image
particular, nearly all the two-dimensional ramp flow velocimetry (PIV, see below) can also be considered
experiments have to be considered with suspicion as a sophisticated visualisation method for separated
when separation occurs. Also, in a large number of flows.
publishedexperimentssomeinformationismissingto
executethecalculationortheboundaryconditionsare
not clean in the sense that the conditions on the
frontiers of the zone of interest affect the
phenomenoninacomplicatedandunclearmanner.In
5
d - EBF visualisation of the flow past a double code
modelatMach10
a(cid:150)surfaceflowonthecentralbodyofanaerospike
nozzle Fig.6:Differenttechniquesofflowvisualisation
Wallmeasurements
Pressure Sensitive Paint. This method (known as
PSP) which allows to determine the complete
pressure distribution over a model, is based on the
fact that some compounds emit light (luminescence)
when excited by a suitable source, the emitted light
having a longer wave-length than the excitation
light10,11. The quantityof light emitted depends on the
oxygen diffused into the paint because oxygen
quenchesanddeactivatestheexcitedmolecules.The
internal concentration of oxygen being a linear
functionoftheexternalpressureofthesamegas,one
can measure the pressure acting on the paint by
detecting some of its luminescent parameters. A
difficulty with PSP(cid:146)s is their simultaneous response to
b(cid:150)lasersheetvisualisationofthevorticespastan
changeintemperature,whichcouldrestrictstheiruse
elongatedbody
in hypersonic flows. The difficulty can be
circumventedeitherbyusingapaintnearlyinsensitive
to temperature or by making a correction from
measurement of the wall temperature. Convincing
PSP measurements have been performed at high
Mach number on a plug nozzle with a PSP
component having a low sensitivity to temperature12
(seeFig.7).
Skin-friction measurement. In a recent technique, the
skin-friction is determined by measuring the rate of
deformation of a thin oil film deposited on the model
surface13-15. If the film of oil is thin compared to its
length, its surface takes the shape of a small wedge
whose thickness y at any time t can be accurately
measured by an interferometric technique (see Fig.
8). Knowing the location x of the measurement point
and the time t, the determination of the wall shear
c(cid:150)interferogramofatransonicshockwave
stressisinprinciplestraightforward.
6
Heat flux measurements. Over the past 20 years,
quantitative infrared thermography has experienced a
strong development in a large number of
laboratories16-18. As it is well known, a body emits a
radiative signal whose intensityis a strongfunction of
its temperature. In infrared thermography, the model
is observed by an infrared camera containing a
detector element sensitive to infrared radiations at a
certain wave-length. By processing a series of
picturestakenatknowntimeintervals,itispossibleto
construct the time history of the model temperature
and to deduce the surface heat flux distribution by
means of the heat equation. The example of infrared
image (in false colour) in Fig. 9 shows the heat flux
distribution on a hemisphere in a Mach 5 flow.
Infrared pictures also provide a convenient way to
detectlaminar-turbulenttransition.
Fig.7:PSPmeasurementsonaplugnozzle
centralbody
t = 0 mn
2 mn
Fig.9:Infraredmeasurementsoftheheatflux
4 mn distributionoveranhemisphereinaMach5
flow.
Fieldmeasurements
Measurement of field quantities such as velocity,
temperature,density,gasspeciesconcentration,etcis
6 mn a difficult task. However, the advent of laser sources
in the early 60s has given a dramatic impetus to the
development of non intrusivemethods allowing an in-
situdeterminationofthepropertiesofagas,including
itsvelocity
LaserDopplerVelocimetry
Fig.8:Skinfrictionmeasurementbythethinoil The basic idea underlying LDV, which is now a well
filmtechnique known technique, is to measure the velocity of tiny
particlestransportedbytheflow19-21.Iftheseparticles
aresmallenough,theirvelocityisassumedtobethat
7
of the stream and LDV provides a measure of the supersonic air-intake and in Fig. 10b, the instant PIV
local instantaneous velocity. A statistical treatment of velocityfieldinarotatingjet).
a sample acquired at one point permits to determine
the mean velocity as well as the turbulent quantities.
ThebasicpostulateofLDVisnotalwaystrueinhighly
decelerating or accelerating flows in which the
particlesdonotinstantaneouslyadjusttheirvelocityto
that of the fluid. The problem of particle lag is at the
heartofLDVandoneshouldbecautiousintheuseof
results obtained in regions where the velocity
undergoes a large variation over a short distance,
situationsfrequentlymetinhypersonicflows.Reliable
LDV measurements have been obtained in shock-
separated flows up to Mach number 5; above
measurementsbecomehazardous.
Developed since 1991, Doppler global velocimetry
(DGV) - alsocalledplanar Doppler velocimetry(PDV)
- is a particle based velocity measurement system
giving the velocity of particles injected in the flow, as
LDV. The difference is that LDV determines the
velocity at one point in space, whereas DGV has the
capacitytogivethevelocityatamultitudeofpointsin
0 50 100
a given region of space22-24. The basic principle X(mm)
consists in determining the Doppler shift of the light
a(cid:150)LDVmeasurementinthebleedregionofa
scatteredbyamovingparticle.
supersonicairintake(averageMachnumber)
ParticleImageVelocimetry
The principle of PIV is to illuminate particles injected
in the flow by a laser sheet and to observe the
scattered light24-26. In order to perform velocity
measurements, two laser pulses, separated by a
short and known time interval Dt, are emitted to
provide two images recorded on the same
photographicplate(inpractice,thephotographicplate
isreplacedbyaCCDcameraprovidingtheimageina
numerical form). During the interval Dt, each particle
has moved over a distance proportional to its velocity
(assumedtobethatoftheflow)givingtwoimageson
the plate. The velocity components contained in the
plane of the image are deduced by measuring the
displacement of the particles which is done by
automatedproceduresusingsophisticatedalgorithms.
b(cid:150)PIVvelocityfieldinarotatingjet
Particle image velocimetry is a very powerful
techniquesinceitprovidesacompletevelocityfieldin Fig.10:Velocitymeasurementsbylaser
a large number of points for a whole region of space, techniqueswithflowseeding
whereas LDV is restricted to measurements at one
point. PIV is very precious for the study of unsteady LaserSpectroscopicFlowDiagnostic
phenomena since it allows to freeze the velocity field
at a given instant. On the other hand, the access to These methods are based on fundamental physical
the averaged field quantities (mean velocity) processesrelatedtotheinteractionbetweenlightand
necessitates to operate an averaging procedure over
matter an do not need seeding by heavy particles of
a large quantities of pictures. This can become
relativelybigsize.Laserspectroscopicmeasurements
problematic for the Reynolds tensor components arebasedontheradiativeinteractionofalaserbeam
whose determination requires averaging several with spectroscopic propertiesof the investigated flow.
thousands of instantaneous values. In this case LDV
Depending on the interaction process, the laser light
if stillmoreeffective (seein Fig. 10aan average LDV
is either absorbed or scattered by those species
velocity field in the vicinity of the bleed system of a whichareradiativelyactiveatwave-lengthused.
8