Flow Visualization Techniques Experimental Methods in Energy and Environment Miguel Rosa Oliveira Panão IST 2003 Courtesy of A.S.Moita.

Flow VisualizationTechniques

Experimental Methods inEnergy and Environment

Miguel Rosa Oliveira PanãoIST 2003

Courtesy of A.S.Moita

WHY VISUALIZE A FLOW ?

Definition:“Flow visualization is the art and science of

obtaining a clear image of a physical flow field and the ability to capture it on sketch,

photograph, or other video storage device for display or further processing.”

P. Freymuth, Flow visualization in fluid mechanics, Rev. Sci. Instrum., 64(1), Jan, 1993

Obtain clear image of flow field

Ability to

capture image

Display and

further process

Why ...“Flow visualization aims at the discovery,

description and parametric investigation of new flow phenomena and at the educational

presentation of established ones.”

P. Freymuth, Flow visualization in fluid mechanics, Rev. Sci. Instrum., 64(1), Jan, 1993

HOW TO OBTAIN A CLEAR IMAGEOF THE FLOW FIELD ?

Tracers

HydrogenBubbles Tufts

Smoke

Dyes

Optical method

s

particles

Refractive index

Polarization density

EXAMPLES

Hydrogen Bubbles

Particles

K. Kerenyi, S. Stein and J. S. Jones, Advanced flow visualization techniques for the Federal Highway Administration Hydraulics Research Laboratory, ASCE 2001

Smoke

Courtesy of Sergei I. Shtork, PhD

EXAMPLES

Dyes

Tufts

HOW TO OBTAIN A CLEAR IMAGEOF THE FLOW FIELD ?

HydrogenBubbles Tufts

Smoke

Dyes

Refractive index

Optical method

s

particles

Polarization density

Tracers

HOW TO MAKE SURE TRACERS FOLLOW THE FLOW ?

Tracer Response Time

Equation of Motion for a spherical tracer

vuvu4D

C21

dtdv

m c

2

D

For low Re numbers(Stokes flow) the tracer response time is

c

2p

p 18

D

If F is a time characteristic ofthe flow field example

Flow through a Venturi

U

DT

UDT

F

With Stp as the Stokes number

F

ppSt

Stp << 1

Tracer follows the

flowC. Crowe, M. Sommerfeld and Y. Tsuji, Miltiphase Flows with Droplets and Particles, pp. 22, CRC Press, 1998

HOW TO CAPTURE AN IMAGE ?

LightSourc

e

TestSectio

n

ImageRecord

er

LaserLightSheet

Cilindrical lens

Plano-convexLens

Light dispersed by particles through Mie Scattering

LASER

Shadowgraphy

Magnifying lensVisualizes the second spatial derivative field of the refractive index

Spot LightIllumination

The light intensity is proportional to L-3

Background

Continuousor

Pulsed

HOW TO CAPTURE AN IMAGE ?

LightSourc

e

TestSectio

n

ImageRecord

er

FLOW FIELD

Spray impact onto a solid surface with Cross-flow

Single droplet impact onto a solid surface

HOW TO CAPTURE AN IMAGE ?

LightSourc

e

TestSectio

n

ImageRecord

er

Images are recorded with a camera and its type is chosen depending on the flow characteristics.

Digital or Film Camera

High speed

CCD camera

Important Features:• Resolution, (n x n) pixel• Exposure time• Aperture• Frame rate (for HS cam.)

HOW TO CAPTURE AN IMAGE ?

LightSourc

e

TestSectio

n

ImageRecord

er

Resolution = L/pixel

PIXEL

L

Exposure Time

short long

Aperture

opening

closing

Frame Rate

U

D = 3 mmU = 3 m/sTc = 0.001 sFR = 2(1/Tc) = 2000FPSNyquist

IMPORTANTSpace-time

scale factors ofFlow field

L – Characteristic length scale

Why ...“Flow visualization aims at the discovery,

description and parametric investigation of new flow phenomena and at the educational

presentation of established ones.”

P. Freymuth, Flow visualization in fluid mechanics, Rev. Sci. Instrum., 64(1), Jan, 1993

The visualization of a flow field allows to identify the large and small structures existing in a flow and futher to compare with local probe measurements, or CFD.

Identify Flow Structures

Spray impact on solid surface with cross-flow

• Upstream wall-jet vortex;• Droplet cloud over the surface after impact;• Turbulent boundary layer;• Fuel vaporization upon impact.M.Panão and A. Moreira, Visualization and Analysis of Spray Impingement Under Cross-Flow Conditions, SAE Technical Paper 2002-01-2664, 2000

Quantify Deformation Process

Spread and finger formation

d(t)

h(t)

Dimensionless time0

0*

DUt

t

Spread Factor

0

**

Dtd

t

0

0.5

1

1.5

2

2.5

3

3.5

4

0 5 10 15

t*

U0=0.22m/ sU0=0.44m/ s

U0=2.47m/ s

0

0.5

1

1.5

2

2.5

3

3.5

4

0 5 10 15

t*

U0=0.22m/ sU0=0.44m/ s

U0=2.47m/ s

U0=0.22m/ sU0=0.44m/ s

U0=2.47m/ s

(t*

)

A. S. Moita and A. Moreira,The Deformation of Single Droplets Impacting onto a Flat Surface, SAE 2002 Transactions Journal of Fuels and Lubricants 1477 – 1489, 2002.

Flow visualization compared with local probe measurementsLDA measurements Velocity and Vorticity

fields

x/d

y/d

-0.5 0 0.5-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

2.42.091.781.471.160.850.540.23-0.08-0.39-0.7

x'

x'

U/U03 U0

x/d

y/d

-0.5 0 0.5-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8543210-1-2-3-4-5-6-7-8

zd/4U0

1. Anacleto P.M., Fernandes E.C., Heitor M.V., Shtork S.I. Characteristics of precessing vortex core in the LPP combustor model. Abs. to the Int. Conf. On Stability and Turbulence of Homogeneous and Heterogeneous Flows, Novosibirsk, April, 25 - 27, 2001, Vol. 8, Kozlov V.V. (Ed.), Institute of Theoretical and Applied Mechanics SB RAS, Novosibirsk, 2001, pp. 14-15.

2. Anacleto P.M., Fernandes E.C., Heitor M.V., Shtork S.I. Characteristics of precessing vortex core in the LPP combustor model. Proc. Second International Symposium on Turbulence and Shear Flow Phenomena, June 27-29, 2001, Stockholm, Sweden. Lindborg E. et al. (Eds.), KTH, Stockholm, 2001, Vol. 1, pp. 133-138.

3. Cala C.E., Fernandes E.C., Heitor M.V., Shtork S.I. Characterization of unsteady swirling flow based on phase averaging of pressure and LDA probe signals. Presented at the 5th Euromech Fluid Mechanics Conference, EFMC-5, 24-28 August, 2003, Toulouse, France.

Q air

PVC

Q air

PVC

Extracting quantitative information by image processingLeonardo’s Vision of Flow in the Aortic

Track

Flow visualization using particles

Digital Particle Image Velocimetry

M. Gharib, D. Kremers, M.M. Koochesfahani and M. Kemp, Leonardo’s Vision of flow visualization, Exp. Fluids, vol.33, pp. 219-223, 2002

Extracting quantitative information by image processingTemperture distribution using a

thermographic camera

Temperature map of an aluminum plate, heated by an electric resistance, to show the uniformity degree of the heating

process.

Courtesy of Humberto Loureiro

139,0°C

224,3°C

140

160

180

200

220

Extracting quantitative information by image processingMeasurements of NO-molecule excitation

LIF images for different ammonia seeding concentrations

N. Sullivan, A. D. Jensen, P. Glarborg, M. S. Day, J. F. Grcar, J. B. Bell, C. J. Pope, and R. J. Kee, Ammonia Conversion and NOx Formation in Laminar Coflowing Nonpremixed Methane-Air Flame", Combustion and Flame 131(3), pp. 285-298, 2002.

Extracting quantitative information by image processingMeasurements of radical concentrations

with tomography

“Steady” “Unsteady”

Burner AUj=15.6 m/sUp=3.5 m/s = 6

UpUj

Courtesy of Prof. Edgar Fernandes

Extracting quantitative information by image processingMeasurements of radical concentrations

with tomography0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

r / R

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

z/R

Up

Uj

air

entr

ain

me

nt

UHC

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

r / R

0.0

0.2

0.4

0.6

0.8

1.0

z/R

2

4

6

8

10

12

14

16

UpUj UpUj

Radical <C2>*

Courtesy of Prof. Edgar Fernandes

Flow visualization compared with CFD

Multipoint Fuel Systems

C.X. Bai, H. Rusche and A.D. Gosman, Modeling of gasoline spray impingement, Atomization and Sprays, vol. 12, pp. 1-27, 2002

Turbulent Premixed Flames

WHEN IS SEEING BELIEVING ?

• Image out of focus• Incorrect light intensity• Incorrect imaging angle• Low SNR (Signal to Noise Ratio)• Too large exposure time• Improper synchronization system

Some problems associated with image processing

Courtesy of Leonardo Da Vinci

BG

HOW TO PROCESS AN IMAGE ?

An image is nothing more than a matrix.

What is an image?

3 x 3

3 x 3 x

3

Intensity Image Color Image

R

= R + G + B,in this case

There are two levels:1. the Graylevel

and;2. the RGB level.

0 1 or 255

Each pixel has a value between 0 and 1.

Image Processing Toolbox, MATLAB

Example:

HOW TO PROCESS AN IMAGE ?

Calibration Process

• To have a scale factor (Length/pixel).• To explore the major effects influencing measurement accuracy.

In Focus Out of Focus

Example: Illuminated hole in back lighting.

0

50

100

150

200

250

Gra

y L

evel

In Carvalho (1995), the following parameters were evaluated: Background Gray Level (BGL), r.m.s., standard deviation, SNR, average gray level along the hole.

The analysis was applied to the measurement of liquid film breakup lengths and the following empirical expression was derived:

f – aperturet – exposure timeG – electronic gain

9.023 G9.0exptf8.2102.57SNRBGL

I. Carvalho, Atomização de Líquidos em escoamentos Turbulentos com e sem Recirculação, PhD Thesis, 1995

halo

HOW TO PROCESS AN IMAGE ?

Particle Identification – boundary detection

• The general procedure is to separate the particle and the background.• This is done though boundary detection algorithms.

Gray Level Threshold

Portions with the gray level lower than a threshold value are counted as particles.

Methods

LBOM

LBLT

GGGG

C

Threshold Value

Gray Level Histogram

Depends on calibration

processB

ackg

rou

nd

Particles

HOW TO PROCESS AN IMAGE ?

Particle Identification – boundary detection

• The general procedure is to separate the particle and the background.• This is done though boundary detection algorithms.

Gray Level Gradient

Based on the assumption that the gray-level variation is the steepest at the particle boundaries.

Appropriate

threshold

S. Y. Lee and Y. D. Kim, Sizing of Sprays Particles using Image Processing Technique, 9th ICLASS, 2003

SUMMARY“Flow visualization has a long history.”

P. Freymuth, Flow visualization in fluid mechanics, Rev. Sci. Instrum., 64(1), Jan, 1993

• Flow visualization consists of:

obtaining a clear image; capture the image; process the image. • There are several techniques for flow

visualization: Laser light sheet; Shadowgraphy, Schlieren, interferometry; Flash or spot illumination.

• The image processing can be made to provide accurate information of the flow. The first steps and techniques are:

o Calibration; o Boundary detection algorithms;o Techniques: Particle Image Velocimetry (PIV); tomography; Particle Tracking Velocimetry (PTV); LIF; thermography; Exciplex ...