Experimental Thermal and Fluid Science 77 (2016) 159–166

Contents lists available at ScienceDirect

Experimental Thermal and Fluid Science

journal homepage: www.elsevier.com/locate/etfs

Experimental analysis of varied vortex reducers in reducing the pressure

drop in a rotating cavity with radial inflow
Xiang Luo a, An Feng a, Yongkai Quan a,⇑, Zhixiang Zhou b, Naibing Liao b
a
National Key Laboratory Science and Technology on Aero-Engine Aero-Thermodynamics, School of Energy and Power Engineering, Beihang University, Beijing 100191, China
b
China Aviation Powerplant Research Institute, Aviation Industry Corporation of China, Zhuzhou, Hunan 412002, China

a r t i c l e i n f o a b s t r a c t

Article history: Experimental research was performed to determine the pressure drop in a rotating cavity with radial
Received 9 September 2015 inflow. In order to decrease the pressure loss across the rotating cavity, three different vortex reducer
Received in revised form 26 February 2016 configurations were designed and fitted in the cavity. The experiment revealed that the performance
Accepted 11 April 2016
of a vortex reducer is determined by its own geometry, the inlet flow rate and the rotating speed of
Available online 19 April 2016
the cavity. For each vortex reducer configuration, a critical curve distinguishes its performance in terms
of reducing the pressure drop. On one side of this curve, the pressure drop is reduced; on the other side,
Keywords:
however, the pressure drop is even larger. The performance of each configuration was compared and pre-
Radial inflow
Rotating cavity
sented schematically. As regards the rotating cavity without vortex reducer (empty cavity), the experi-
Vortex reducer mental data fitted quite well with Farthing’s linear theory. For each vortex reducer configuration, a
Flow resistance fitting curve of the pressure coefficient was given.
Pressure coefficient Ó 2016 Elsevier Inc. All rights reserved.

1. Introduction In addition, Brillert et al. [8] studied the total pressure loss across
the rotating cavity.
In the secondary air system of modern aero-engines, in some With regard to analysis of the flow resistance reduction of the
cases air is taken radially inward from the final stages of the com- radially inflow rotating cavity, Owen and Farthing [9,10] studied
pressor. Because of the high revolving velocity, a strong vortex is how de-swirl nozzles affect the pressure difference between the
formed inside the rotating cavity and it causes high pressure loss inlet and the outlet. Their work showed that the de-swirl nozzles
across the cavity. If appropriate measures are taken successfully change the swirl fraction of the inlet flow and thus cause a drop
to reduce the pressure loss, the bleeding point can be designed at in the pressure loss across the cavity. Another investigation of
an earlier stage and a higher efficiency of the aero-engine will be the de-swirl nozzle can be found in Friedl [11]. Chew [12] showed
achieved. Two common ways to fulfil this goal are the use of de- that attaching radial fins to the inner side of one of the rotating
swirl shroud nozzles and the installation of proper vortex reducers disks moderately reduced the pressure drop.
in the rotating cavity. However, only a few people have studied the impact of the vor-
Much research has been performed on the flow structure and tex reducer fitted in the cavity. Gunther [13] may have been the
heat transfer in a rotating cavity with radial inflow. In 1968, Hide first to do so. In his experiment, he installed four different config-
[1] first suggested that the flow structure of a rotating cavity with urations of vortex reducers either with tubes or tubeless in the cav-
radial inflow includes a source region, an interior core, two Ekman ity and measured their performances. The vortex reducers
layers and a sink region. Following Hide’s work, Firouzian and successfully restricted the formation of the interior core and tube-
Owen [2,3] confirmed the theory by conducting a flow visualisa- less configuration performance was better than that of long tubes.
tion experiment describing not only the flow structure but the Liang et al. [14] and Chen et al. [15] experimentally studied tubed
velocity, pressure and other parameters of the flow field. Using vortex reducers and the influences of the shape of the inlet nozzles.
the same test rig as Firouzian, Farthing and Owen [4] conducted In their studies, tubes with different length were used and the total
a similar experiment concerning the radial outflow. Owen’s work pressure at the inlet and outlet of the cavity was measured. They
[5–7] laid the theoretical foundation of the study of radial inflow. showed that tubed vortex reducers have great capability in terms
of reducing the pressure drop compared with the traditional free
cavity. Other studies include those of Negulescu and Pfitzner
⇑ Corresponding author. [16], Peitsch et al. [17] and Du et al. [18].
E-mail address: [email protected] (Y. Quan).

http://dx.doi.org/10.1016/j.expthermflusci.2016.04.011
0894-1777/Ó 2016 Elsevier Inc. All rights reserved.
160 X. Luo et al. / Experimental Thermal and Fluid Science 77 (2016) 159–166

Nomenclature

DP
Ra outer radius of cavity Cp,c pressure coefficient, C p;c ¼ 1qX 2 2
Ra
2
Rb inner radius of cavity
kT turbulent flow parameter, kT = Cw/ReU0.8
s axial distance between disks
DP pressure difference between inlet and outlet, Pi  Po
G gap ratio of cavity, s/Rb
q density
P pressure
m kinetic viscosity
X rotating speed
i inlet
m_ mass flow rate
XR2 o outlet
ReU rotational Reynolds number, Re/ ¼ m b
_ e edge of the source region
Cw non-dimensional flow rate, C w ¼ lmRa

Compared with tubed vortex reducers, tubeless vortex reducers sure regulator and flow control valve. The compressor starts work-
have much simpler structures, and thus are more reliable and ing as soon as the pressure in the receiver reaches its lower limit.
easier to manufacture and install. In the present paper, we put An air dryer and a filter are used to remove water mist, oil drops
the emphasis on investigating the effect of different vortex reduc- and dust. Through several valves and a thermal mass flowmeter,
ers on pressure loss in the rotating cavity. The free cavity without air is then drawn into the test section and finally leaves the test
any vortex reducer also serves as a contrast. rig through the hollow shaft to the open air.

2. Experimental apparatus and methods 2.2. Test section

The experiments were conducted on the rotational heat transfer 2.2.1. Overall configuration of the test section
test rig at the National Key Laboratory of Science and Technology As shown in Fig. 2, the core part of the test section consists of
on Aero-Engines and Aero-Thermodynamics at Beihang University, two rotating disks, a shroud with 12 straight cylinder nozzles, a
China. The main components of the system consist of air supply vortex reducer fitting in the cavity and a hollow shaft. All these
system, test section and data acquisition system. The schematic parts are mounted together by bolts. The outer radius of the cavity
diagram of the experimental apparatus as well as the test section Rb is 204.5 mm, the inner radius Ra is 79.5 mm and the distance
is illustrated in Fig. 1. The test rig is driven by a 30 kW DC electric between two co-rotating disks is s = 40 mm. The corresponding
motor which allows the rotational speed to be varied up to non-dimensional parameters are: the gap ratio G = s/Rb = 0.196
3000 rpm with a maximum ReU of 8.72
105. and non-dimensional inner radius xa = Ra/Rb = 0.389.

2.1. Air supply system 2.2.2. Vortex reducer configurations

Three different vortex reducer configurations (VR 1  VR 3) are
The air is supplied by a compressor with a maximum pressure studied in this paper. As shown in Fig. 3a, the outer and inner
of 8 atm and regulated down to the required flow rate with a pres- diameter of the first configuration (VR 1) are 364 mm and

Fig. 1. Schematic diagram of experimental apparatus.

 X. Luo et al. / Experimental Thermal and Fluid Science 77 (2016) 159–166 161

magnitude of the inflow rate and the angular velocity. They are
defined as follows:
_
m
Cw ¼
mb

XR2b
Re/ ¼
m
_ is the inlet mass flow rate, X is the angular velocity of the cavity,
m
m is the kinetic viscosity.
To reproduce the working condition of a real engine in the lab-
oratory, the experimental condition is determined as follows:

inlet mass flow rate (kg/h): 98, 196, 294, 392, 490, 588;
the corresponding Cw (
104): 0.753, 1.471, 2.206, 2.941, 3.677,
4.412;
rotational speed (rpm): 375, 750, 1125, 1500, 1875, 2250, 2625;
the corresponding ReU (
105): 1.09, 2.181, 3.271, 4.362, 5.452,
6.543, 7.633.

Fig. 2. The structure of the core part of the test section. In the experiment, a mass flow rate is fixed every time and then
rotational speed is varied by changing the rotational speed of the
motor.
178.5 mm respectively, with 12 equally circumferentially spaced
cylinder holes 14 mm in diameter. The second configuration (VR 2.5. Experimental uncertainties
2) differs from the first only in the shape of the holes. As shown
in Fig. 3b, the transverse section of this configuration has 12 fan- In the current experiment, thermal mass flow meters, pressure
shaped holes separated by 12 equally circumferentially spaced baf- transducers and a photoelectric sensor are used to measure the
fles. The width of each baffle is 35 mm and the height (axial direc- mass flow rate, pressure, and angular velocity respectively. Some
tion) of the holes is 14 mm. Fig. 3c shows the third configuration uncertainties are involved and thus have to be determined in a cer-
(VR 3). The outer diameter remains unchanged but the inner diam- tain way to give an overall view on the reliability of the experi-
eter is 316 mm. Again, 12 equally spaced de-swirl nozzles are scat- ment. The mass flow is measured with a ST98 thermal mass flow
tered along the ring. The oblique angle of the nozzles is 60°. The meter, with FS of 0.55 kg/s and accuracy of ±1% reading + 0.5% FS.
performance of each configuration is compared with that of the The pressure was obtained by Kulite XTL-140 transducers, with
empty cavity (VR 0). FS of 100 kPa and accuracy of ±1%.
VR 1 has a similar design to tubed vortex reducers which have According to the test condition, the uncertainty of pressure is
been studied by Liang [14]. However, VR 1 is designed as a disk within 2% and the additional uncertainty caused by the vibration
while Liang used retain ring to hold all the tubes. Therefore, VR 1 of the experimental apparatus is estimated to be within 3%. The
is much more robust but it is also more bulky. VR 2 is a new design pressure transducers were calibrated before measurement, and
and it is given after being carefully modeled by CFD. VR 3 uses de- they have an accuracy of ±0.5%. Also, the uncertainty of the mass
swirl nozzle to enhance its performance in reducing the vortex. The flow varies from 1.7% to 5.9%, and the rotational speed from
outer and inner radius are determined using similitude theory to fit 0.23% to 0.7%. Thus, based on the error transfer theory offered by
the apparatus in the laboratory. Other sizes are given according to Kline and McClintock [19], the uncertainty of non-dimensional
real engine. parameter of pressure coefficient Cp is around 3.32%.
Kulite XTL-140 pressure transducer is insensible to acceleration,
2.3. Data acquisition system so any vibration and centrifugal forces would not cause any appar-
ent uncertainty. Also, every transducer had been carefully cali-
Some parameters including the inlet and outlet pressure, inlet brated before any measurement.
mass flow rate and the angular velocity of the rotating cavity have
to be measured in the experiment. 3. Results and discussion
As illustrated in Fig. 2, the inlet and outlet pressure are mea-
sured by pressure transducers attached to disk 2 co-rotating with 3.1. The flow resistance reduction performance of VR 1 and VR 2
the cavity. For each pressure, two transducers are used at the same
radius but separated by an angle of 180°. The rotating signal from As shown in Fig. 4 the flow resistance reduction performances
the pressure transducers is first translated into digital form by the of VR 1 and VR 2 are compared with that of the empty cavity
co-rotating ADAM modules and then transformed by a graphite (VR 0). For both configurations, the pressure difference between
slip ring into a stationary signal. inlet and outlet grows monotonously as the inlet mass flow and
The mass flow rate and the rotating speed are measured by a the rotating speed grow. The reason is that the disks with a high
thermal mass flowmeter and a photoelectric sensor respectively. rotating speed drive the air between them and increase its tangen-
All signals are collected by a PC using a Visual Basic program. tial velocity which leads to a growth of the radial Coriolis force and
this causes an increase of the pressure drop across the cavity. It is
2.4. Experimental parameters and test conditions shown in the figure that the pressure drop increases quickly as the
mass flow grows, and when the inlet mass flow is large (say,
Two non-dimensional numbers Cw (non-dimensional flow rate) 588 kg/h) the pressure drop becomes equally large. In other words,
and ReU (rotational Reynolds number) are used to indicate the the pressure drop increases disproportionately with the inlet mass
162 X. Luo et al. / Experimental Thermal and Fluid Science 77 (2016) 159–166

12 VR 0
VR 1
98 196 294 392 490 588
10 mass flowrate(kg/h)

ΔP (kPa)
6

0
0 375 750 1125 1500 1875 2250 2625 3000
(a) Ω (rpm)
(a)

12 VR 0
VR 2
98 196 294 392 490 588
10 mass flowrate(kg/h)

8
ΔP (kPa)

(b)
0
0 375 750 1125 1500 1875 2250 2625 3000
Ω (rpm)
(b)
Fig. 4. Comparison of flow resistance reduction performance (a) VR 1 vs. VR 0 and
(b) VR 2 vs. VR 0.

and pressure drag. However, in the case of high rotating speed, the
benefits of the vortex reducers are obvious and the larger the rotat-
ing speed is, the greater the reduction of the pressure loss the vor-
tex reducers can cause. To be more specific, for each group of solid
and dashed curves in Fig. 4, an intersection point can be found, left
of which is the inferior zone of the vortex reducer and right of
which is the superior zone. The intersection point grows as the
inflow rate grows.

3.2. The flow resistance reduction performance of VR 3

Fig. 3. The structure of the vortex reducers.
The performance of VR 3 differs from VR 1 and VR 2 mainly in
the relationship of flow resistance reduction with rotating speed.
flow rate. The reason is that at large inlet mass flow, the pressure In Fig. 5, the flow resistance reduction of VR 3 first decreases then
loss caused by friction is much larger than that caused by the radial increases with the increase of rotating speed. Namely, at each
Coriolis force. In the rotating coordinate attached to the cavity, the inflow rate, there is a certain rotating speed at which the flow
radial velocity of the fluid inside the flow paths of vortex reducers resistance reduction ability of VR 3 reaches its maximum (the dif-
is large and causes a large tangential Coriolis force which compels ference in the pressure between inlet and outlet reaches its
the fluid to accelerate along the tangential direction. This, to some minimum).
extent, is the same as increasing the rotational speed. Fig. 6 shows the velocity triangle of the outflow from the de-
It is also shown in Fig. 4 that vortex reducers cannot decrease swirl nozzle of VR 3. XRa is the linear velocity of the shroud, V is
the pressure drop in all conditions. Actually, at low rotational the velocity of the air out of the nozzle (inlet velocity) relative to
speed, the interior core of the empty cavity is weak, and the fitted the rotational coordinate and Vr and Vh are the axial and tangential
vortex reducer brings nothing but a quick increase of friction drag components of V respectively. Evidently, in an absolute coordinate,
 X. Luo et al. / Experimental Thermal and Fluid Science 77 (2016) 159–166 163

VR 0 the tangential velocity component in the opposite direction. In

12 VR 3 absolute frame, the tangential velocity of flow right out of the noz-
98 196 294 392 490 588
zles is opposite to that of local disk rotating tangential velocity,
mass flowrate(kg/h)
10 with larger velocity magnitude, however, and thus it causes
additional pressure loss. This loss, together with the loss caused
8
Pi - Po (kPa)

6 1000
intersection
VR 0
4 VR 1
0.735 1.471 2.206 2.942 3.677 4.413
Cw * 10 -4
2
100

Cp,c
0 375 750 1125 1500 1875 2250 2625 3000
Ω (rpm)
10
Fig. 5. Comparison of flow resistance reduction performance of VR 3 and VR 0.

1
1.090 2.181 3.271 4.362 5.452 6.543 7.633
Reφ * 10-5
(a)
100
VR 0
VR 2
0.735 1.471 2.206 2.942 3.677 4.413
Cw * 10 -4

10

Fig. 6. Velocity triangle at the inlet of the cavity.

Cp,c

the tangential component of the cavity inlet velocity is XRa  V,

noted by Vh,ab, for convenience. In the current experiment, the 1
geometry of the shroud is fixed, i.e., XRa and the transverse area
of the nozzles are fixed, and thus Vh,ab is determined only by the
inlet flow rate m._ Therefore, for every mass flow rate m, _ there is
unique X0, at which X0Ra  Vr = 0. Only in this case dose the abso-
lute tangential velocity reduce to zero, and the swirl intensity in 0.1
1.090 2.181 3.271 4.362 5.452 6.543 7.633
the interior core reaches its minimum.
Following the above idea, a series of X0 values can be deter- Reφ * 10-5
mined. Table 1 shows the value of X0 at each inlet flow rate. Com- (b)
paring the value of X0 in Table 1 and that in Fig. 5 (Xs relating to
the lowest points on the curves), reasonably, we found that they 1000
agree with each other. VR 0
A set of intersection points can be found in Fig. 5. On the right VR 3
0.735 1.471 2.206 2.942 3.677 4.413
side of the intersection points, the pressure curves for VR 0 are Cw * 10 -4
above that for VR 3, which means pressure loss for VR 3 is less than
100
that for VR 0, especially in conditions that rotating velocity is high
and mass flow rate is relatively small. This is not difficult to under-
stand because de-swirl nozzles here in VR 3 are much like that in
Cp,c

the shroud that give flow a negative swirl direction and thus
reduce the absolute tangential velocity. However, on the left side 10
of these intersection points, the pressure curves for VR 3 are appar-
ently above that for VR 0, especially in conditions of high flow rate
and low rotational speed, because the value of XRa  Vh is negative
and |XRa  Vh| > XRa, which means the de-swirl nozzles increases
1
1.090 2.181 3.271 4.362 5.452 6.543 7.633

Table 1 Reφ * 10-5

Experimental conditions. (c)
_ (kg/h)
m 98 196 294 392 490 588
Fig. 7. Comparison of non-dimensional flow resistance reduction performance: (a)
X0 (rpm) 350 701 1052 1403 1754 2105
VR 1 vs. VR 0, (b) VR 2 vs. VR 0 and (c) VR 3 vs. VR 0.
164 X. Luo et al. / Experimental Thermal and Fluid Science 77 (2016) 159–166

by friction and the abrupt change of flow path in and out of VR 3, 3.4. Pressure coefficient with turbulent flow parameter
makes the pressure drop rise dramatically.
As regards a rotating cavity with radial inflow, Farthing [10]
pointed out that for a certain geometry Cp,c is determined only by
3.3. Non-dimensional flow resistance reduction performance kT, namely, Cp,c = f(kT). Using a linear model, Farthing determined
this function:
Apart from the non-dimensional mass flow rate Cw, rotational
Reynolds number ReU, and the turbulent flow parameter kT (kT = -
C p;c ¼ ðx2e  x2a Þ þ c2eff ðx2
e  1Þ þ 23:7jkT j
5=8 3=8
ðxe  x3=8
a Þ
Cw/ReU0.8), Farthing [9] used another non-dimensional parameter
Cp,c to indicate the pressure drop across the cavity, defined as þ 7:89jkT j5=4 ðxa4=5  x4=5
e Þ and C p;c 6 c2 ðx2
a  1Þ
follows:
xe is non-dimensional radius at the edge of the source region and it
DP
C p;c ¼ 1 is given by
2
q X2 R2a
 1=2
where q is the fluid density and DP = Pi  Po. xe ¼ c  2:22jkT j5=8 x3=8
e

Fig. 7 shows the pressure coefficient with the Rotational Rey-

nolds number. For every configuration, the pressure coefficient ceff is the effective swirl ratio. It is slightly different from swirl ratio
Cp,c decreases with the growth of ReU, and Cp,c increases with the (which is mainly determined by physical geometry), because of the
increase of Cw. vortex in the recirculation region.
Another significant problem is determining the condition in For the current experiment, the shroud nozzle is along the
which the pressure coefficient of the rotating cavity with vortex radial direction, so c  1. The square dots in Fig. 8 are experimental
reducers (VR 1  VR 3) is smaller than that of the empty cavity data and the black curve is the result of Farthing’s derivation (for
(VR 0). One answer to this problem is to use quadratic interpola- c = 0.95) [10]. It is concluded from this figure that the experimental
tion in Fig. 7 at every Cw. Then a set of intersection points are data fit quite well with Farthing’s linear theory for 0.07 < kT < 0.6.
obtained (marked by a circle and a cross in Fig. 7). For clarity, for Fig. 9 shows the relationship between Cp,c with kT for VR 1, VR 2
the intersection on the far right of Fig. 7a (the corresponding Cw and VR 3 respectively. Again, the black square dots are the exper-
is 3.677
104), which is out of the interpolation interval, the quad- imental data and the red curves are exponential fittings to the
ratic interpolation is still reasonable because the intersection is not experimental data. The functions of these fitting curves are also
far from the upper limit of ReU. However, for a larger Cw given. As shown in Fig. 9, Cp,c grows quickly with kT. Therefore, in
(4.413
104), the intersection is too far from the upper limit for the secondary air system of real engines, the turbulent flow param-
accuracy and therefore should be discarded. eter is always below one.
Specifically, since the experimental ReU is between 1.09
104
and 7.633
104, the critical curves are applicable for
ReU > 1.09
104, and they attain great accuracy for 3.5. Comparison of performances of different configurations
1.09
104 < ReU < 7.633
104.
The critical curve separates the work condition into two zones – The flow resistance reduction performance for VR 1 and VR 2 is
an upper zone and a lower zone. In the upper zone, the introduc- illustrated in Fig. 10. For different rotating speed and mass flow
tion of a vortex reducer increases the pressure loss across the cav- rate, the performance of VR 2 is better than VR 1. The reason is that
ity. In the lower zone, however, the vortex reducer breaks the the flow path for VR 1 is narrower so the flow state changes
interior core, helping to reduce the pressure drop successfully. quickly, causing a large pressure drop. In addition, compared with
Therefore, the critical curve can be used to decide whether to use VR 2, VR 1 is more massive. Nevertheless, VR 1 has a simpler
a vortex reducer or not and to evaluate the potential benefit it geometry.
can bring. Fig. 10 shows a comparison of performance of VR 2 and VR 3. At
low rotating speed, the performance of VR 2 is better than VR 3, but
as the rotating speed increases, the latter gradually becomes supe-
8 rior. Furthermore, the quality of VR 3 is much poorer than that of
Farthing&Owen 1991 VR 2.
7 Experiment

6
4. Comment about thermal effects
5
Secondary air system is to bleed air from compressor to turbine
4 to maintain the latter working in a suitable environment. There-
Cp,c

fore, temperature is a significant variable and it ought to have been

3 taken into consideration here, not only because temperature at the
outlet of the cavity affect the cool efficiency directly, but it deter-
2 mines properties of the secondary air. However, in current experi-
ment, temperature variation across the cavity is pretty small,
1
ranging from 5 to 18 K in all conditions, according to the data
0
from two thermocouples installed at inlet and outlet of the cavity
0.0 0.1 0.2 0.3 0.4 0.5 0.6 respectively (this was not mentioned in this paper because we did
λT not include any thermal discussion). It has to be stressed here that
this data is doubtable since the temperature change is insignificant
Fig. 8. Variation of Cp,c with kT for VR 0 (black square dots: experimental data; black but the outer case is made of steel with no heat insulation. There-
curve: linear theory of Farthing). fore, we neglected thermal effects in our discussion.
 X. Luo et al. / Experimental Thermal and Fluid Science 77 (2016) 159–166 165

20 VR 1
experimental data (VR 1) 12
fitting curve VR 2
18
98 196 294 392 490 588
16 10 mass flowrate(kg/h)

14
y=2.316e1.524x-1.313
8
12

ΔP (kPa)
Cp,c

10 6
8
4
6

4
2
2

0 0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 0 375 750 1125 1500 1875 2250 2625 3000
λT Ω (rpm)
(a)
20
experimental data (VR 2)
18 fitting curve VR 2
12
VR 3
16 98 196 294 392 490 588
mass flowrate(kg/h)
14 10

12
y=2.314e1.442x-1.494 8
Cp,c

10
ΔP (kPa)

8 6
6
4
4

2
2
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
0
λT 0 375 750 1125 1500 1875 2250 2625 3000
Ω (rpm)
20 (b)
experimental data (VR 3)
18 fitting curve Fig. 10. Comparison of flow resistance reduction performances of different
configurations: (a) VR 1 vs. VR 2; (b) VR 2 vs. VR 3.
16

14

12
y=0.716e2.343x+0.505 performance of the vortex reducers) as well as in non-dimensional
Cp,c

10
fashion (to develop the critical curves and to deduce the functions
8 of Cp–kT). Some main results are as follows.
6 For VR 1 and VR 2, as the rotating speed grows the pressure
drop between inlet and outlet of the cavity increases monoto-
4
nously. However, for VR 3, the pressure drop decreases first and
2 then increases in the same situation. That is to say, there is a rotat-
0 ing speed at which the pressure drop reaches its minimum and the
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 tangential component of the inlet velocity reaches zero. The exper-
λT imental value agrees well with that calculated according to the
geometry and the inlet mass flow.
Fig. 9. Variation of Cp,c with kT for VR 1  VR 3 (black square dots: experimental For all three configurations, the pressure drop increases as the
data; red curve: exponential fitting). (For interpretation of the references to color in inlet mass flow rate increases.
this figure legend, the reader is referred to the web version of this article.)
By fitting the intersection points of the pressure curve of VR
1  VR 3 and that of VR 0, a critical curve for each configuration
is attained, and it can be considered a criterion whether to use a
5. Conclusion vortex reducer.
The experimental data for VR 0 agree quite well with Farthing’s
In this study, an experiment is conducted to analyze the flow linear theory. The Cp,c–kT curve is also given for VR 1  VR 3.
resistance reduction performance of three vortex reducer configu- The performances of all three configurations are given. VR 2 has
rations. Results are reported in dimensional fashion (mainly to give a better flow resistance reduction performance than VR 1 in all
a since of changing relationship of all the direct variables, such as cases. However, to determine whether VR 2 or VR 3 performs
pressure drop, rotating velocity, mass flow rate, and to evaluate the better, working conditions have to be taken into consideration.
166 X. Luo et al. / Experimental Thermal and Fluid Science 77 (2016) 159–166

Acknowledgement [9] P.R. Farthing, The Use of De-swirl Nozzles to Reduce the Pressure Drop in a
Rotating Cavity With a Radial Inflow, vol. 113(1), 1991, pp. 106–114.
[10] P.R. Farthing, J.M. Owen, De-swirled radial inflow in a rotating cavity, Int. J.
The authors gratefully acknowledge the financial support from Heat Fluid Flow 12 (1) (1991) 63–70.
China Aviation Powerplant Research Institute (Zhuzhou). [11] W.-H. Friedl, D. Peitsch, D. Negulescu, Improvement of high pressure turbine
air systems by de-swirl nozzles, American Society of Mechanical Engineers,
2002, pp. 331–336.
[12] J.W. Chew, P.R. Farthing, J.M. Owen, B. Stratford, The use of fins to reduce the
References pressure drop in a rotating cavity with a radial inflow, J. Turbomach. 111 (3)
(1989) 349–356.
[1] R. Hide, On source-sink flows in a rotating fluid, J. Fluid Mech. 32 (04) (1968) [13] A. Gunther, W. Uffrecht, E. Kaiser, S. Odenbach, L. Heller, Experimental
737–764. Analysis of Varied Vortex Reducer Configurations for the Internal Air System of
[2] M. Firouzian, J.M. Owen, J.R. Pincombe, R.H. Rogers, Flow and heat transfer in a Jet Engine Gas Turbines, American Society of Mechanical Engineers, 2008, pp.
rotating cavity with a radial inflow of fluid Part 1: The flow structure, Int. J. 1563–1570.
Heat Fluid Flow 6 (4) (1985) 228–234. [14] Z. Liang, X. Luo, Y. Feng, G. Xu, Experimental investigation of pressure losses in
[3] M. Firouzian, J.M. Owen, J.R. Pincombe, R.H. Rogers, Flow and heat transfer in a a co-rotating cavity with radial inflow employing tubed vortex reducers with
rotating cylindrical cavity with a radial inflow of fluid: Part 2: Velocity, varied nozzles, Exp. Therm. Fluid Sci. 66 (2015) 304–315.
pressure and heat transfer measurements, Int. J. Heat Fluid Flow 7 (1) (1986) [15] X. Chen, Y. Feng, L. Wu, The experimental investigations of centripetal air
21–27. bleed with tubed vortex reducer for secondary air system in gas turbine, in:
[4] P.R. Farthing, J.M. Owen, The effect of disk geometry on heat transfer in a ASME Turbo Expo 2014, Düsseldorf, Germany, 2014.
rotating cavity with a radial outflow of fluid, J. Eng. Gas Turbines Power 110 (1) [16] D. Negulescu, M. Pfitzner, Secondary Air Systems in Aeroengines Employing
(1988) 70–77. Vortex Reducers, American Society of Mechanical Engineers, 2001. pp.
[5] J.M. Owen, Source-sink flow inside a rotating cylindrical cavity, J. Fluid Mech. V003T001A072–V003T001A072.
155 (1) (1985) 233–265. [17] D. Peitsch, M. Stein, S. Hein, R. Niehuis, U. Reinmöller, Numerical Investigation
[6] J.M. Owen, R.H. Rogers, Flow and Heat Transfer in Rotating-Disc Systems, of Vortex Reducer Flows in the High Pressure Compressor of Modern
Rotor-Stator Systems, vol. 1, Research Studies Press, Taunton, England, 1989. Aeroengines, American Society of Mechanical Engineers, 2002, pp. 1125–1134.
[7] J.M. Owen, R.H. Rogers, Flow and Heat Transfer in Rotating-Disc Systems, [18] X.-Q. Du, H.-R. Zhu, Z.-W. Zhang, Numerical Study on Varied Vortex Reducer
Rotating Cavities, vol. 2, Research Studies Press, Taunton, England, 1989. Con for the Flow Path Optimization in Compressor Cavities, American Society
[8] D. Brillert, D. Lieser, A.W. Reichert, H. Simon, Total Pressure Losses in Rotor of Mechanical Engineers, 2011, pp. 905–912.
Systems with Radial Inflow, American Society of Mechanical Engineers, 2000. [19] S.J. Kline, F.A. McClintock, Describing uncertainties in single sample
pp. V003T001A087–V003T001A087. experiments, Mech. Eng. (1953).