Experimental Investigation of

E. Q. Tulapurkara
Assistant Professor.
Morel's Method for Wind Tunnel
V. V. K. Bhalla
Contractions
Graduate Student.
Based on a numerical study of the potential flow through contractions of chosen
Department of Aeronautical Engineering, geometry, Morel (1975) has given a method to obtain the shape of contraction
Indian Institute of Technology, which gives small adverse pressure gradients and low nonuniformity in the velocity
Madras 600 036, India distribution at the exit. Two contractions with area ratios of 12 and 3.464 designed
using this method are investigated experimentally. It is found that there is no separa-
tion of flow, the thickness of the boundary layer at the exit is small and the
nonuniformity in velocity at the exit is smaller than the predicted value.

I Introduction
The contraction or the nozzle is an important component of tion is obtained along the wall. On the other hand, Morel
a wind tunnel. As the flow passes through the contraction it (1975) starts with a chosen geometry for the contraction and
accelerates and this results in a reduction of the nonuniformity obtains a numerical solution for the flow inside the contrac-
and turbulence level of the stream. In practical contractions, tion along with suitable straight ducts at either ends of the
which are of finite length, one finds that (i) adverse pressure contraction. The treatment of the contraction as having a
gradients are present at the ends of the contraction (Bradshaw finite length does give the adverse pressure gradients and
and Pankhurst, 1964), (ii) the axial velocity is higher than the nonuniform velocity distributions at the ends. Based on a
velocity near the wall at the entry to the contraction and (iii) at criteria for boundary layer separation due to Stratford (1959),
the exit the velocity near the wall (i.e., outside the boundary guidelines are obtained for getting a contour without separa-
layer) is higher than that on the axis. Thus for a good perfor- tion. This provides a very quick method of obtaining a con-
mance the nozzle contour should give low adverse pressure tour satisfying the requirements for good performance. A
gradients at the ends of contraction so that no separation of brief outline of Morel's method is given now.
flow takes place, the boundary layer thickness at the exit The nozzle contour is obtained by two power-law curves
should be small and the nonuniformity in the velocity distribu- matched at a point xm and having their apexes at either ends of
tion at the exit (i.e. difference between velocity near the wall the contraction. During preliminary trials he finds that the
and that on the axis) must be small. A good contour should cubic curves give the best results. Potential flow inside the
achieve these with a small length (Z,) to upstream diameter contour is computed for values of contraction ratio (c) ranging
(D{) ratio. from 2 to 25, the ratio L/Dx from 0.75 to 1.25 and the ratio X
Nearly fifteen methods have been proposed to obtain the (=x m /L) from 0.12 to 0.8. From the computed wall static
shape of contraction. Klein et al. (1973), Chmielewski (1974), pressure distributions the pressure coefficients at the inlet and
and Morel (1975) give the bibliographic details of these the exit, Cpi and Cpe< defined as follows, are obtained.
methods. In a majority of these methods the governing equa-
tions for axisymmetric, incompressible potential flow are cpe = i - (v^i.-) 2 ; cPe = i - (t/ 2 ,»/^) 2
solved with an assumed axial velocity distribution (e.g., Tsien, where Vt and Ve are, respectively, the minimum and max-
1943, Chmielewski, 1974). This gives a set of streamlines. The imum velocities near the inlet and the exit of contraction, Ul_„
portion of a streamline which gives a monotonic velocity and £/2,oo are the uniform velocities far upstream of the inlet
distribution and very nearly the desired contraction ratio is and far downstream of the exit, respectively. Plots of Cpj and
chosen as the contour of the contraction. In another type of Cpe, with A'and L/Dx as parameters, are obtained for various
theoretical approach (e.g., Thwaites 1946, Bossel, 1969) the contraction ratios. Cross plotting these, one gets X and L/Dx
ends of the contraction are taken as equipotential planes. for a given c and chosen values of Cpj and Cpe. It is further
Then assuming a variable-separable solution one gets an ex- found that for Cpe < .2 the difference between the wall veloci-
pression for the velocity potential in the form of a series. From ty and the axial velocity at the exit is proportional to Cpe. A
this the streamlines are calculated. The coefficients in the value of 0.39 for Cpi is recommended so that separation near
series are evaluated such that a monotonic velocity distribu- the inlet is avoided. For Cpe a value of 0.06 is suggested. This
will ensure that the nonuniformity in the velocity distribution
Contributed by the Fluids Engineering Division for publication in the JOUR-
NAL OF FLUIDS ENGINEERING. Manuscript received by the Fluids Engineering at the exit will not be more than 2 percent, which is considered
Division June 8, 1987. to be a reasonable value. Incidentally this value of C„ would

Journal of Fluids Engineering MARCH 1988, Vol. 110/45

Copyright © 1988 by ASME
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/28/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 UmmJ c
2 39.97 3.464

C =3.464 0=3.464
Up stream end (x/L = 0) Down stream end (x/L =1.0)

0.4 0.6
y/R 2
Fig. 1 Contraction shapes as obtained by Morel's method (»>
:-?=-*•"

c I. ( m m ) D2 mm
12 214.47 7 2.16
3 . 4 6 4 219.97 134.32
-1.0 -0.8 -0.6 -0.4 -0.2
1. Centrifugal blower 2. Diffuser 3. Honey comb
A. Nylon screen 5. Settling chamber 6. Contraction
7. Test section 8. Second Diffuser
Fig. 3 Experimental velocity distribution for c = 3.464. (a) Upstream
Fig. 2 Experimental setup end; (b) Downstream end; (c) Along the contraction. (Uncertainty in
U = 2.5 percent near upstream end = 0.1 percent near downstream end)

also give separation free flow at exit. Once X and L/D{ are
known the contour is given by
3
D-D> . 1 / x \ „ x
D,~D2
-U-T)' <X 1.0
C = 12
:
-0-r)
0.0
- Up stream end ( x / L 0)

for ->X all

0.6
Q-xy
where D is the diameter at a distance x from the inlet and D2 is 0.4
the exit diameter.
Though one would expect the conclusions based on the 0.2 -
potential flow calculations to be valid in a contraction, it is felt
that it would be interesting to verify them experimentally. 1
0.2
1
0.U
1
0.6 0.8
Hence experiments are conducted on two contractions with
c= 12 and 3.464 and the parameters of interest like the axial
and wall velocity distributions along the contraction, the
velocity distributions at the ends and the exit boundary layer
thickness are measured and the results are reported in this arti- 0
"wall

cle. The contours studied are briefly described in section II. - H^ i

4 !W
Ji
This is followed by a description of the experimental setup, ex- fi
Ur.l r
perimental results and discussion. C =12
II

11 Contours Studied 4
Bradshaw and Pankhurst (1964) recommend a contraction - 4 * ' /

ratio of 12 for a good low turbulence wind tunnel. However 1 1 1 1 1 1 1 i i i

many wind tunnels in common use have smaller contraction -1.0 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

ratios of the order of 4. Hence contractions with area ratios of

12 and Vl2 i.e., 3.464 are chosen for the present investigation.
The diameter of the settling chamber ahead of the contraction Fig. 4 Experimental velocity distribution for c = 12. (a) Upstream end;
(b) Downstream end; (c) Along the contraction. (Uncertainly in 0 = 2.5
is 250 mm. Velocity in the settling chamber is 4 m/s. Hence the percent near upstream end = 0.1 percent near downstream end)
value of Cpi which would not give separation, according to
equation (12) of Morel (1975), is .323. A value of 0.35 is
chosen for Cpi, Based on experience of Tulapurkara (1980) for c= 12 and 3.464, respectively. The contours are shown in
and the recommendation of Morel (1975) an acceptable value Fig. 1.
for the exit nonuniformity is chosen as 2 percent. This requires
Cpe to be less than 0.057. A value of 0.05 for Cpe is chosen. Ill Experimental Setup and Technique
These values of Cpi and Cpe give X= 0.537 and L/Dx = .858 The experimental setup is shown in Fig. 2. The centrifugal
for c= 12, and X= 0.332 and £/£>, =0.88 for c= 3.464. For blower is driven by a 2 HP motor and delivers about 15 cubic
Dx of 250 mm we get D2 equal to 72.17 mm and 134.32 mm meters of air per minute. The blower is connected to the dif-

46/Vol. 110, MARCH 1988 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/28/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 fuser through a flexible hose. This prevents transmission of design of their contours. The velocity distributions are seen to
vibrations from the blower to other parts of the setup. The dif- become almost uniform, (within ±0.5 percent) at about 0.1 D{
fuser has an area ratio of 1:4 and a semiangle of 6 deg. The from the exit of the contraction. The heights of the boundary
settling chamber which follows the diffuser has (i) a layer at the exit are about 2 mm and 2.5 mm for c- 12 and
honeycomb of cell width 10 mm and depth 50 mm, (ii) nylon 3.464 respectively. It may be mentioned that the exit radii in
screens with wire diameter 0.36 mm and mesh width 1.25 mm the two cases are 36.1 mm and 67.2 mm and that the displace-
and a settling length of 389 mm. The contraction is followed ment thickness of the boundary layer would be only a small
by a test section and a diffuser. The velocity in the settling fraction of the boundary layer height.
chamber is 4 m/s. This would be nearly the settling chamber
velocity in most of the wind tunnels with the test section speed
between 50 to 60 m/s and the contraction ratio between 12 to V Conclusions
16.
The velocity distributions at the ends of the contraction and Two wind tunnel nozzles with contraction ratio of 12 and
along the axis are obtained from measurements of the total 3.464 are designed using Morel's method using (^, = 0.35 and
pressure and static pressure using pitot and static pressure Cpe = 0.05, and are tested experimentally. It is found that (i)
tubes. Micromanometers FC012 made by Furness Control the adverse pressure gradient along the wall is small and there
Ltd. of U.K. are used for pressure measurements. The wall is no separation of flow, (ii) the nonuniformity in the exit
velocity distribution is obtained from the measurement of wall velocity distribution is only 1 percent as compared to the
static pressure and the total pressure along the axis. Typical design value of 2 percent, and (iii) the velocity distribution
readings of the manometer during velocity measurements near becomes practically uniform in 0.1 £), behind the contraction
the inlet and exit were 1.3± 0 5 and 190 ±-5 mm of water, exit.
respectively. This gives an accuracy of ±2.5 percent near the
inlet and ±0.1 percent near the exit. References
Bossel, H. H., 1969, "Computations of Axisymmetric Contractions," AIAA
IV Results and Discussion Journal, Vol. 7, No. 10, pp. 2017-2020.
Bradshaw, P., and Pankhurst, R. C , 1964, "The Design of Low Speed Wind
The velocity distributions at the ends and the distributions Tunnels," Progress in Aeronautical Sciences, Vol. 5, Kuchemann, D. and
of axial velocity and wall velocity along the contraction are Sterne, L. H. G., Editors, Pergamon Press, pp. 1-69.
shown in Figs. 3 and 4 for c = 3.464 and 12, respectively. J?, Chmielewski, G. E., 1974, "Boundary Layer Considerations in the Design of
and R2 in these figures are the radii of contraction at the inlet Aerodynamic Contractions," Journal of Aircraft, Vol. 11, No. 8, pp. 435-438.
and the exit. From Figs. 3(c) and 4(c) it is seen that the adverse Klein, A., Ramjee, V., and Venkataramani, K. S., 1973, "An Experimental
Study of the Subsonic Flow in Axisymmetric Contractions," ZFW, Vol. 21, No.
pressure gradients near the inlet is small. In the experiments 9, p'p. 312-320.
one does not notice any evidence of separation and the loss of Morel, T., 1975, "Comprehensive Design of Axisymmetric Wind Tunnel
total pressure along the contraction is less than one mm of Contractions" ASME JOURNAL OF FLUIDS ENGINEERING, Vol. 97, No. 2, pp.
water. The adverse pressure gradient near the exit is hardly 225-233.
Stratford, B. S., 1959, "The Prediction of Separation of the Turbulent
noticeable. The velocity distribution at the inlet does show an Boundary Layers," Journal of Fluid Mechanics, Vol. 5, Part 1, pp. 1-16.
appreciable difference between the axial velocity and the wall Thwaites, B., 1946, "On the Design of Contractions for Wind Tunnels,"
velocity (Figs. 3(a) and 4(a)), but this is not of much Aeronautical Research Council, London, R & M 2278.
significance. The difference between the axial and wall Tsien, H. S., 1943, "On Design of Contraction Cone for a Wind Tunnel,"
Journal of Aeronautical Sciences, Vol. 10, No. 2, pp. 68-70.
velocities at the exit for both the contractions studied is only Tulapurkara, E. G., 1980, "Studies on Thwaites' method for Wind Tunnel
about 1 percent as compared to 2 percent assumed in the Contractions," Aeronautical Journal, Vol. 84, pp. 167-169.

Journal of Fluids Engineering MARCH 1988, Vol. 110/47

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 01/28/2016 Terms of Use: http://www.asme.org/about-asme/terms-of-use