Вы находитесь на странице: 1из 4

Proceedings of the 7th International and 45th National Conference on Fluid Mechanics and Fluid Power (FMFP)

December 10-12, 2018, IIT Bombay, Mumbai, India

FMFP2018–406

Turbulence measurement in Axisymmetric air jet using the Particle Image Velocimetry

Raju Murugan Bharadwaz Bollu Pankaj S Kolhe


Research Scholar M.Tech Student Assistant Professor
Combustion Diagnostics Laboratory Combustion Diagnostics Laboratory Combustion Diagnostics Laboratory
IIT Hyderabad, Kandi, Sangareddy. IIT Hyderabad, Kandi, Sangareddy. IIT Hyderabad, Kandi, Sangareddy.
Email: me14resch11006@iith.ac.in Email: me18mtech01002@iith.ac.in Email: psk@iith.ac.in

Abstract (point measurement and intrusive), Particle Image


Particle imaging velocimetry measurements were carried Velocimetry[9] (planar measurement and non-intrusive).
out in axisymmetric turbulent air jet for Reynolds number Viggiano, B. et al.[6] carried out the experimental work for
range covering transition (Re: 2350 to 4000) and turbulent different fluids such as Helium, air and argon for different
flow (Re: 5000 to 11000) regime based on jet exit Reynolds number[10] by using PIV techniques. Centerline
conditions. Endeavor was to provide near the field velocity decay profile plotted as normalized velocity
statistical information of velocity field in turbulent {centreline velocity (Us)/ jet exit velocity (Uj)} with axial
axisymmetric jet of diameter, d = 6.9 mm. Test case of Re locations from these experimental study was not following
= 11000 was used to validate the PIV measurements any particular trend for decay with respect to the Reynolds
carried out with that in the literature. Radial profiles of number, which is revisited in the present study.
normalized mean velocity and turbulent intensity are in
agreement with previous studies. Transition from top hat Present study employs particle imaging velocimetry for
velocity profile at nozzle exit to Gaussian velocity profile flow measurement in axisymmetric air jet. PIV records the
at the end of potential core region is observed. At any fixed position over time of small tracer particles introduced into
axial location downstream of potential core region, mean the flow to extract the local fluid velocity. Although the
axial velocity profile in normalized co-ordinates show the visualization technique employed is qualitative, pattern
same underlying radial distribution for Reynolds number matching algorithm yields quantitative instantaneous
in turbulent regime (Re > 4000) at nozzle exit. velocity of particles across flow field from a pair of images
taken very short duration apart. Since PIV requires,
Keywords: particle seeding, it is imperative to understand the particle
Particle Image Velocimetry (PIV); axisymmetric air jet; seeding densities in the flows and their effect on pattern
Turbulent flow, Jet half width, centre line velocity decay. matching algorithm and accuracy of velocity field
estimation The main objective of this study was to
I.INTRODUCTION characterize the velocity field statistics in the near nozzle
In general, any flow measurement involves exit region. Centreline decay of axial velocity, mean and
characterizing bulk fluid movement through differential turbulent intensity radial profiles are presented to
pressure or velocity measurements. Based on fluid motion, characterize the near field region for transition and
flows are classified as laminar flow or turbulent flow and turbulent jet exit Reynolds number
Reynolds number range is useful in demarcating the two
flows. Many real practical flows are coming under the II. Experimental setup
category of turbulent flow, so it is very important to The schematic representation of the experimental setup
understand the turbulent flow characteristics from is shown in Fig 1. The experimental setup consists of laser
perspective of entrainment, mixing and developing models diagnostics unit {Particle Image Velocimetry (PIV)} and
for simulations of such flows. Velocity field estimation the flow system with compressor, mass flow controller
gains importance from this perspective as one can easily (MFC) and particle generator/ particle seeder.
estimate entrainment, turbulent kinetic energy, Reynolds The laser diagnostic unit for PIV measurement consists
stresses etc. Such information is essential for numerical of a double cavity Nd:YAG laser (Nano PIV, Litron)
modelling. delivering pulses of 30 mJ at 532 nm and a progressive scan
interline CCD camera (Powerview Plus 4MP, TSI), and
Many researchers[1]–[7] carried out the experimental and laser sheet optics to make laser sheet out of laser beam. The
numerical work for turbulent axisymmetric jet for different scattered signals from the particles were captured using the
fluids to understand the turbulence[8] by using different CCD camera. A 12-bit image pair of 2048  2048 pixels
measuring techniques such as hot wire anemometer probes
where each pixel corresponds to 110.4 m. The field of
1
view captured by CCD camera is 226  226 mm2. Proper
calibration was carried out for the entire experimental
matrix to get the correct velocity based on the particles
displacement from one frame to another frame for a given
time interval between two laser pulse. The raw images
(sample 500 images for each experiment) are captured and
imported into Insight 4G software for pre-processing with
Gaussian filter to reduce the random noise.

Figure 2: PIV air jet images (Re = 11000) with time


difference ∆t = 20 s (inverted grey scale).

Figure 1: Schematic diagram of air jet injector with


the arrangement of PIV

III. RESULTS AND DISCUSSION


Figure 2 shows a pair of instantaneous images captured
with inverted grey scale wherein seed particles can be seen
in black color. Note that the time difference between the
pair of images to estimate instantaneous velocity field was
kept at 20 s. The instantaneous vector files thus obtained
were used to estimate mean, RMS of velocity components
in the longitudinal plane of laser. Velocity contours Figure 3: PIV velocity (m/s) contour images along with
superimposed with velocity vectors for different axial Velocity profiles at different axial location for Re =
locations (Re = 11000) are shown in Fig .3, which gives 11000
the qualitative and quantitate information about the
centreline velocity decay (refer in Fig .4) and velocity
spread in the radial direction (refer in Fig.5). Nearly linear
decay of centreline axial velocity is observed and results
show excellent agreement with measurements by Lumley
et al. [7]. This mean axial velocity decays linearly as
Ujet/Us = 0.1659  x/d, which is similar to the previous
experimental result (Ujet/Us = 0.165  x/d) by Lumley et al.
[7].
Radial profiles of normalized mean axial velocity from
the PIV results plotted in fig.5, strongly agree with Lumley
el [7] and Wygnanski and Fiedler [3]. Figure 6 shows that
Turbulence intensity across radial direction strongly agree
with previous experimental works [3, and 7]. Based on this
experimental validation, further experiments were carried Figure 4: Centreline normalized meal axial velocity
by varying Reynolds number from 2350 to 10000. Figure decay profile along the jet with the linear trend
7 shows the contour plots with superimposed velocity (Ujet/Us = 0.1659  x/d).
vectors at different axial locations for Re = 5000 to 10000.

2
Figure 5: Normalized radial velocity profile for
different r/x location, which strongly agreed the
previous experimental results.

Figure 7: PIV velocity (m/s) contour images along with


velocity profiles for different Reynolds number
Figure 6: Turbulence intensity of the axial velocity indicated in the figure.
fluctuations across the jet at 60 diameters downstream.
Figure 8 depicts radial profiles of the normalized mean
axial velocity with jet centreline velocity for different axial
location. Note that the radial co-ordinate is normalized by
jet half width (denoted as delta) which corresponds to
radial location where velocity is 50% of the centreline
velocity. These radial profiles conform to Gaussian
distribution much better for downstream axial locations.
Figure 9 depicts the normalized velocity profile across
the normalized radial locations for different Reynolds
numbers in transition and turbulent regime based on jet exit
conditions for axial location 5 diameter downstream of
nozzle exit. Note that x/d = 5 is taken as approximate
representative for end of potential core region.
For Jet exit Reynolds number in turbulent regime (Re > Figure 8: PIV results for axial mean velocity profile
4000), same underlying radial distribution of normalized across the jet for different axial location for Re=10000.
mean axial velocity was observed. However, for jet exit
Reynolds number in transition regime (2350  Re  4000) Fig.10 shows the centreline velocity decay with respect to
velocity distribution particularly at outer radial location is the normalized axial location for different Re. Magnitude
significantly different indicating difference in transport of axial gradient reduces with increasing Reynolds number,
mechanism. which could be attributed to momentum dominated jet and

3
reduced radial spread. Note that even though the centreline  At far downstream locations, there exist underlying
velocity decay is different for different Reynolds number same distribution in normalized co-ordinates for mean
the radial profile show universal shape in normalized co- velocity and turbulent intensity radial profiles.
ordinates, which makes interesting case for investigating  Beyond potential core-region, normalized centreline
velocity mixing patterns through proper orthogonal axial velocity decay profile is observed to be linear
decomposition technique. with decreasing slope at higher Reynolds number.

Acknowledgements
Authors are also grateful to MHRD for providing
research fellowship. Authors would also like to
acknowledge Department of Science and Technology
Grant No. ECR/2015/000365, for supporting the capital
equipment costs and consumables for the experimentation.

REFERENCES
[1] D. M. Kyle. and K. R. Sreenivasan, The instability
and breakdown of a round variable density jet , J.
Fluid Mech. (1993) 619–664.
[2] T. Djeridane, M. Amielh, F. Anselmet, and L.
Figure 9: PIV results for mean axial centreline velocity Fulachier, Velocity turbulence properties in the
profile for different Reynolds number at axial position near-field region of axisymmetric variable density
of x/d = 5. jets, Phys. Fluids, 86, pp. (1996)1614–1630.
[3] I. Wygnanski and H. Fiedler, Some measurements
in the self preserving jet, J. Fluid Mech.38 (1969),
577–612.
[4] P. Wang, J. Fröhlich, V. Michelassi, and W. Rodi,
Large-eddy simulation of variable-density turbulent
axisymmetric jets, Int. J. Heat Fluid Flow, 29
(2008) 654–664.
[5] M. Khashehchi, A. Ooi, J. Soria, and I. Marusic,
Evolution of the turbulent/non-turbulent interface
of an axisymmetric turbulent jet, Exp. Fluids,
54(2013).
[6] B. Viggiano, T. Dib, N. Ali, L. G. Mastin, R. B. Cal,
and S. A. Solovitz, Turbulence entrainment and
low-order description of a transitional variable-
density jet, J. Fluid Mech. ,836(2018) 1009–1049.
Figure 10: Normalized centreline velocity for different [7] N.R.Panchaprakesan and J.L Lumley, Turbulence
Reynolds number. measurements in axisymmetric jets of air and
helium. Part 1. Air Jet, J.Fluid Mech, 246(1993)
IV. CONCLUSIONS 197-223.
PIV measurements were carried out in near field region for [8] Henk Tennekes and John L.Lumley, A First Course
axisymmetric turbulent air jet in transition and turbulent in Turbulence, The MIT Press, London, UK,(1972).
flow regimes based on Reynolds number at nozzle exit. [9] M. Raffel, C. E. Willert, S. T. Wereley, and J.
Observations lead to following conclusions. Kompenhans, Particle Image Velocimetry , A
 Top hat velocity profile is observed immediate Practical Guide, Springer, (2007).
downstream of nozzle exit and the nearest axial [10] Munson, BruceR., Young, Donald F., Okiishhi,
location to nozzle exit where Gaussian velocity profile Theodore,Fundamentals of Fluid Mechanics,4th
is observed marks the end of potential core region. Edition, John Wiley & Sons, Inc, New York,
 For transition flow (jet exit Re < 4000), transport USA(2002).” .
mechanism is different as evident by varying mean
velocity profile in normalized co-ordinates in near field
region.
4

Вам также может понравиться