1

Bulletin of the JSME

Journal of Fluid Science and Technology
Vol.9, No.2, 2014
Paper No.14-00013
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
Abstract
We developed a technique for the measurement of uctuating hydraulic-pressure at an arbitrary point in turbulent
ow. A pressure-measuring system was composed of a miniature static-pressure probe (SP-probe), a strain-gage-
type pressure transducer, and a tube connecting them. A model equation of dynamic response of the pressure-
measuring system was introduced, and a calibration technique for determining parameters in the model equation
was developed. Eects of the size of the SP-probe, length of the tube, and sensitivity of the pressure sensor were
systematically investigated using several SP-probes, tubes, and pressure transducers with dierent sizes, lengths,
and sensitivities, and uctuating hydraulic-pressure measurement was performed in a wake of a circular cylinder.
Key words : Fluctuating pressure, Flow measurement, Wake, Pitot tube, Free shear ow
1. Introduction
Fluctuating pressure is a fundamental and important property of turbulent ow as well as uctuating velocity, and ef-
forts to measure the uctuating static-pressure in turbulent ow have been made from the 1950s. However, the uctuating
pressure measurement at an arbitrary point remote from wall is still dicult as compared to velocity measurement; one
needs to use a Pitot-tube-type pressure probe to pick up the instantaneous static-pressure with sucient spatial resolution,
but the intrusive nature results in some problems such as disturbance to the ow and measurement error caused by lateral
velocity component. In addition, compressibility of air inside the pressure probe results in a non-at frequency response.
These problems in experiment in air were addressed by Shirahama and Toyoda (1993) and Toyoda et al. (1994),
who optimized geometric conguration of a miniature static-pressure probe (SP-probe) so that the disturbance to the ow
and the cross-ow error were minimized. They also proposed some techniques to calibrate the frequency response of the
pressure-measuring system. The similar techniques have been adopted in several experimental studies (Iida et al., 1999;
Naka et al., 2006; Tsuji et al., 2007; Kawata et al., 2011; Terashima et al., 2012) and successful measurement results have
been reported.
On the other hand, the uctuating static-pressure measurement in water is still limited to wall-pressure measurement,
while some aforementioned successful measurements have been reported by experiments in the air. The reason might
be diculties in manufacturing a pressure sensor applicable to hydraulic pressure measurement and, at the same time,
sensitive enough to measure small pressure uctuation produced by turbulence.
An attempt to measure fluctuating local pressure in free
turbulent flow in water
*Department of Mechanical Engineering, Keio University
3141 Hiyoshi, Kohoku-ku, Yokohama, Kanagawa, 2238522, Japan
E-mail: kwtk1113@gmail.com

Received 6 January 2014
Takuya KAWATA

, Hoshito MAEDA

and Shinnosuke OBI

In the past hydraulic wall-pressure measurement, a strain-gage-type pressure transducer was used as a pressure sensor
and connected to the pressure tap on the wall by tubing. Dynamic response of such a pressure-measuring system has been
investigated in the previous studies (e.g., Geddes et al., 1984), and it was pointed out that the dynamic response was non
at mainly due to elasticity of the pressure sensor while eect of compressibility of the uid is negligible. The methods
to correct these eects have been studied in literatures (Donovan et al., 1991; Taylor and Donovan, 1992; Donovan et al.,
1994; Aydin, 1998). In these studies, the behavior of the pressure-measuring system was modeled as that of a harmonic
oscillator with one degree of freedom, and parameters in the model equation, the natural frequency and the damping
factor, were determined by calibration.
22
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
Fig. 1 (a) Schematics of pressure-measuring system; (b) its model.
In the present study, a technique for uctuating pressure measurement at an arbitrary point in turbulent ow in
water was developed. A pressure-measuring system was composed of an SP-probe, a strain-gage-type pressure transducer
and a tube connecting them. A model equation of the behavior of the pressure-measuring system was introduced and
the parameters in the model equation were evaluated in a newly proposed calibration procedure. Dependency of these
parameters on size of the SP-probe, length of the tube and sensitivity of the pressure transducer was systematically
investigated. The uctuating pressure measurement was performed in a wake of a circular cylinder, and correction of
pressure signal based on the calibration was demonstrated.
2. Analytical Model for Frequency Response of Pressure-Measuring System
The system, which enables us the uctuating static-pressure measurement at an arbitrary point in water, is schemati-
cally shown in Fig. 1a. It comprises an SP-probe, tubing and a pressure transducer, and can be simplied as a tube which
connects the pressure sensor to a pressure source with sudden contraction at one end as shown in Fig. 1b. In the following,
an equation which relates the true value of uctuating pressure at the pressure source p
s
and the measured value p
m
is
derived.
Flow inside the tube is driven by pressure dierence between the pressure source and inside the cavity of the pressure
transducer, and it is damped by the viscosity of the uid. Hence, the behavior of volume ow rate through the cross section
of the tubing Q can be modeled as:
I
dQ
dt
+ RQ = (p
m
p
s
) , (1)
where I is the uid inertance, and R is the resistance of the system mainly caused by the viscosity of the working uid.
The volume ow rate Q can be related to the volume change in the cavity of the pressure transducer V as;
dV
dt
= Q. (2)
Introducing the elastic compliance of the pressure transducer C, which represents the elasticity of a diaphragm in the
pressure transducer;
C =
dV
dp
m
, (3)
and substituting Eqs. (2) and (3) into Eq. (1), one can obtain the equation which relates p
m
to p
s
as:
IC
d
2
p
m
dt
2
+ RC
dp
m
dt
+ p
m
= p
s
. (4)
33
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
7
.
5

2

36
1
.
2
0

11.20
2
.
4
0

2
7
.
5

1
.
5

33
0
.
9

8.5
1.4
1
.
7
7
7
.
5

1
.
0

28.8
0
.
6

5.3 1
1
.
3
5
53.9
0
.
4

0
.
8

7
.
2

19.5
1
.
0

Probe - d10 Probe - d20
Probe - d Probe - air 15
Fig. 2 Schematics of static-pressure probes.
Equation (4) has the same form as the conventional model equation of a system for the wall-pressure measurement (Dono-
van et al., 1991). Thus, the behavior of the pressure-measuring system is equivalent to that of a damped harmonic oscil-
lator, and the damping ratio and the natural frequency (in rad/s) of the system are given as
=
R
2

C
I
,
n
=

1
CI
, (5)
respectively. Solving Eq. (4) with the pressure input uctuating at a constant frequency , one obtains the amplitude ratio
A and the phase lag between p
m
and p
s
as:
A =
1

(1 u
2
)
2
+ 4
2
u
2
(6)
= tan
1
(
2u
1 u
2
)
(7)
with u being the non-dimensional frequency dened as u = /
n
. The proles of A and against the frequency of the
input pressure uctuation were measured by calibration of dynamic response of the pressure-measuring system, and the
damping ratio and the natural frequency were evaluated by tting measured variation of A and to Eqs. (6) and (7).
3. Experimental Apparatus
3.1. Pressure-Measuring System
3.1.1. Static-Pressure Probe The SP-probes employed in the present study is schematically shown in Fig. 2. There
were four types of the SP-probe: the probe-air, d10, d15 and d20. The probe-air had similar geometric conguration to
those employed in previous experiments in the air (Toyoda et al., 1994; Naka et al., 2006). It consisted of a thin tube with
a circular corn tip, and the outer and inner diameters were 1.0 and 0.8 mm, respectively. There were four pressure-sensing
holes on the surface separated by 90

in the circumferential direction with the diameter of 0.4 mm.

The probe d10, d15 and d20 were designed for hydraulic pressure measurement, and had the inner diameters of the
tube part d of 1.0 mm, 1.5 mm, and 2.0 mm. The diameters and lengths of each probe are presented in Fig. 2. They
were made of transparent material so that one could check if bubbles remained inside of the SP-probe, and the tip was
semi-spherical shape. The number of the pressure-sensing holes was eight, and we determined their diameter , so that the
impedance of the pressure-measuring system was retained low. To accomplish this, we assumed the pressure loss caused
by the eight pressure-sensing holes to be equal to that in the tube of the SP-probe for the same volume ow rate. Under
the assumption that both the ow through the tube of the SP-probe and that through the eight pressure-sensing holes are
44
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
20 10 0 10 20
0.04
0.03
0.02
0.01
0
angle of attack (

)
C
p
=
(
p

p
0
)
/

U
2


Fig. 3 Error in pressure measurement due to cross ow: blue, probe-d20; green, probe-d15; red, probe-d10;
black, probe-air.
approximated by the Hagen-Poiseuille ow, the volume ow rate in the former and that through the latter are proportional
to d
4
and 8
4
, respectively:
8
4
d
4
= 1

d
=
(
1
8
) 1
4
= 0.59. (8)
Therefore, the diameter of the pressure-sensing holes was determined to be 60% of the tube diameter of the SP-probe.
The analysis based on the Hagen-Poiseuille ow would over-estimate the pressure loss caused by the pressure-sensing
holes, as there would be only minor viscous eect. Hence the diameter of the pressure-sensing holes determined above
should be large enough not to cause a major inuence to the dynamic response of the pressure-measuring system.
Measurement error caused by the cross ow was experimentally explored. The SP-probe was placed in a free stream
of a uniform ow of a wind tunnel, and the pressure was measured with various angles of attack. Figure 3 shows the
variation of the pressure coecient dened as
C
p
=
p p
0
U
2

, (9)
with p
0
being the pressure value measured with the ow angle of 0

. It is shown that pressure is underestimated when the

ow direction is not parallel to the axis of the SP-probe. The probe-air shows better angle characteristics than those of the
SP-probes designed for use in water, which may be attributable to dierence in shape of the tip section.
3.1.2. Pressure Transducer A low-range strain-gage-type pressure transducer was used for pressure sensor in order
to sensitively measure the small pressure uctuation in a low frequency range. In order to investigate the eect of the
elastic compliance C, three pressure transducers with dierent C were used for comparison; two DP45s and one DP103
(Validyne) were employed, and their characteristics are summarized in Table 1.
3.1.3. Tubing A plastic tube was used for tubing between the SP-probe and the pressure transducer placed inside and
outside the ow, respectively. The inner diameter and the thickness were 6 mm and 0.5 mm, respectively, and the length
was 0.96 m.
3.2. Other instruments and Data Acquisition
The experiments were undertaken in a closed-loop water tunnel with maximum free stream velocity of 0.28 m/s.
For the other hardware, a carrier modulator (PA501, Validyne) was used for amplication of the signals from the pres-
sure transducer, and an analog low-pass lter (DT-6LF2, NF CORPOLATION) and a 16 bits A/D converter (PCI-6621,
National Instruments) were used for ltering and acquisition of samples.
Table 1 Characteristics of pressure transducer; s, sensitivity; V volume of pressure cavity; V
max
, maximum
volumetric displacement.
Range (Pa) s (Pa/V) V (cc) V
max
(cc) C (mm
3
/Pa)
DP45-1 84 16.7 0.16 0.016 0.192
DP45-2 167 33.2 0.16 0.016 0.0967
DP103 28 5.7 0.57 0.057 1.96
55
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
f
u
n
c
t
i
o
n
g
e
n
e
r
a
t
o
roscillator
piston pressure chamber
water channel
pressure transducer
pressure transducer
(reference)
Fig. 4 Schematics of system for calibration of frequency response of pressure-measuring system.
4. Frequency response of Pressure-Measuring System
4.1. Response to Sinusoidal Pressure Fluctuation
Response of the pressure-measuring system against pressure input oscillating at a constant frequency was experi-
mentally explored. The system for this calibration is schematically shown in Fig. 4. The pressure-measuring system to
be calibrated was set up as it was in the main measurement; the SP-probe was placed in the water channel and connected
to the pressure transducer located outside by a tube which is identical to that used in the main experiment. The reference
port of the pressure transducer was connected to a chamber, and the oscillatory pressure input was generated inside the
chamber by driving a piston using an oscillator. The frequency and the amplitude of the pressure input were controlled by
a function generator. The true amplitude and the phase of the pressure input were monitored by another pressure trans-
ducer for reference, and the amplitude ratio A and the phase delay were evaluated by comparing the pressure signals
measured simultaneously by these two pressure transducers.
The dumping ratio and the natural frequency
n
were evaluated based on the measured proles of A and . The

n
was obtained as the frequency at which is /2 by interpolating the measured variation of . The dumping ratio
was evaluated as
=

1 A
2
max
2
(10)
where A
max
is the maximum value of A. Equation (10) cannot be used in the case where the measuring system is over-
damped and the prole of A does not have a maximum value. In such case, the optimal value of was obtained by tting
Eq. (6) to the measured prole of A using fminsearch, which is function for optimization available in MATLAB
R
.
The frequency response of the pressure-measuring system was investigated for various combinations of the SP-probe
and the pressure transducer, and the values of and
n
are summarized in Table 2. The measured variation of A and
are compared for cases with DP45-1 and various SP-probes in Fig. 5 and for cases with the probe d20 and various
pressure transducers in Fig. 6. The circles stand for the measured values, and the solid lines represent the proles obtained
by Eqs. (6) and (7) with the calibrated values of and
n
in Table 2. Although only the measured values of A and at
vicinity of the natural frequency were used for evaluation of and
n
, the proles drawn by Eqs. (6) and (7) are in good
Table 2 Dumping ratio and natural frequency of various pressure-measuring systems. R, I and C are given in
mm
3
/Pa, 10
2
kg/mm
4
and mm
3
/Pa.
DP45-2 DP45-1 DP103
SP-probe air d10 d15 d20 air d10 d15 d20 air d10 d15 d20
1.5 0.41 0.20 0.16 1.9 0.52 0.24 0.20 4.0 1.1 0.55 0.46

n
19.7 17.4 19.3 20.3 12.1 11.7 13.2 13.7 2.88 3.52 4.58 4.78
R 1.6 0.49 0.21 0.16 1.7 0.46 0.19 0.15 1.4 0.325 0.122 0.098
I 2.7 3.4 2.8 2.5 3.6 3.8 3.0 2.8 6.1 3.8 2.4 2.2
C 0.0967 0.192 1.96
66
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
10
1
10
0
10
1
0
1
2
3

A
10
1
10
0
10
1
0
0.5
1
Frequency (Hz)

r
a
d
)
Fig. 5 Frequency response of pressure measuring systems with DP45-1 and various SP-probes: blue, probe-d20;
green, probe-d15; red, probe-d10; black, probe-air. Circle plots and solid line show the measured values
and the tted proles, respectively.
agreement with all the measured values throughout entire the calibrated frequency range, which indicates that the model
Eq. (4) represents well the behavior of the present pressure-measuring system.
Comparing the frequency response of the pressure-measuring systems with dierent SP-probes shown in Fig. 5, one
can obviously see that the resonance peak magnitude of A becomes smaller as the SP-probe becomes thinner. Especially
in the case with the probe-air, the system is overdamped; the prole of A does not show any peaks and becomes large
already in the frequency range lower than 1 Hz. This signicant damping is also indicated by the values of in Table 2;
the value of in the case with the probe-air is much larger than those of the systems with the other SP-probes. On the
other hand, the variations of crosses = /2 around 2 Hz regardless of the size of the SP-probe, indicating that the
natural frequency
n
is not aected by the size of the SP-probe.
The results presented in Fig. 6 shows that the compliance C signicantly changes both and
n
: as C increases, the
natural frequency is shifted toward the lower frequency region, and also the system becomes more damped. This can be
explained by Eq. (5) as C appears in both and
n
.
The uid resistance R and the inertance I were also evaluated based on Eq. (5), and the values are summarized in
Table 2. One can see that R signicantly increases as the SP-probe becomes thinner. Hence, the size of the SP-probe
mainly aects the damping ratio through the value of resistance R, and does not the natural frequency
n
because R
does not appear in
n
(see Eq. (5)). On the other hand, I is relatively insensitive to both the size of the SP-probe and the
compliance of the pressure transducer.
10
1
10
0
10
1
0
1
2
3
4

A
10
1
10
0
10
1
0
0.5
1

r
a
d
)
Frequency (Hz)
Fig. 6 Frequency response of pressure measuring systems with probe-d20 and various pressure transducers: blue,
DP45-2; green, DP45-2; red, DP103
77
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
pressure transducer
pressure box
air
water
piston
soap bubble
Pa + P
Pa +gh+ P
h
Fig. 7 Schematics of system for step-response test.
4.2. Step-Response Test
Response to a step input was examined for further investigation on the dynamic response of the pressure-measuring
systems. A sudden pressure drop was generated using a system schematically sown in Fig. 7. The SP-probe of the
pressure-measuring system to be calibrated was xed in the water phase in a pressure box. A T-shape tube was connected
to the air phase of the pressure box at one end and to a piston at another end. The other end was closed by a membrane
of soap solution. The pressure in the air phase in the pressure box was kept higher than the atmospheric pressure by the
piston, and sudden pressure drop was generated by breaking the membrane of soap solution.
The behavior of the measuring system against the step input is described by the following dierential equation:
d
2
p
m
dt
2
+ 2
n
dp
m
dt
+
2
n
p
m
= 0. (11)
Equation (11) can be easily solved, and the analytical step response of the pressure-measuring system is obtained as:
p
m
p
0
=

n
t
(
cos qt +

n
q
sin qt
)
( < 1)
p
2
p
2
p
1
e
p
1
t

p
1
p
2
p
1
e
p
2
t
( > 1)
(12)
where p
0
is the height of the step input, and q, p
1
and p
2
are respectively dened as:
q =
n

1
2
,
p
1
=
n
( +

2
1),
p
2
=
n
(

2
1).
Figure 8 presents the step response of the measuring systems with DP45-1 and various SP-probes, comparing those
measured and predicted by Eq. (12). The measured responses are indicated by the blue, green, cyan and red lines for the
probe-d20, d15, d10 and air, and it is shown that step response of the system with the probes-d10, d15 and d20 shows the
overshoot, but that of the system with the probe-air is overdamped and does not oscillate. The step responses predicted
by Eq. (12) with the values of and
n
in Table 2 are indicated by the black dotted, dashed, chained and solid lines for
the probe-d20, d15, d10 and air. The predicted step response for the case with the probes-air and d10 are in quite good
agreement with the measured step responses, and those for the case with probes-d15 and d20 also agree fairly well with
the measured step response while certain underestimate of overshoot can be seen. This results also supports consistency
of the present calibration results.
88
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst00 ] 14
Table 3 Parameters of pressure-measuring system with dierent tubing length.
L (m) /
0

n
/
n0
R/R
0
I/I
0
0.46 1.1 1.26 0.94 0.86
0.96 1 1 1 1
1.9 0.94 0.76 1.23 3.73
0 0.5 1 1.5
1
0.5
0
0.5
1
t (s)

p
/
p
0


Fig. 8 Comparison of measured and predicted step response of pressure-measuring system with DP45-1. Probe-
d20, blue and ; probe-d15, green and ; probe-d10, cyan and ; probe-air, red and .
4.3. Effect of Tube on Dynamic Characteristics of Systems
The present pressure-measuring system has tubing with certain length between the SP-probe and the pressure trans-
ducer unlike the systems used in the measurement in air, in which the transducer is directly embedded to the SP-probe.
The dynamic characteristics of such pressure-measuring systems may depend not only on length but also shape of the
tube. In this section, such eect of the tubing is addressed by the step-response test.
4.3.1. Effect of tubing length The eect of the tubing length was examined by comparing the step response for
dierent tubing lengths. Three tubes were used for comparison: 0.46 m, 0.96 m and 1.9 m. It should be mentioned that
the tube with length of 0.96 m was identical to that used in the main experiment.
Figure 9 compares the step response of the pressure-measuring systems with the DP45-1, the probe-d20 and the three
dierent tubes. It is shown that the step response is clearly aected by the length of the tubing. The parameters of the
dynamic response for each case were evaluated by tting Eq. (12) to the measured step response, and the obtained values
are summarized in Table 3. The values are scaled by those obtained in the case of the tubing length being 0.96 m. It is
shown that both R and I increase as the tubing length increases. The increase of R is attributable to increase of the area of
inner wall of the tube, where the viscous drag acts on the oscillating uid inside the tubing. The increase of I is simply
interpreted as the increase of the uid mass inside the pressure-measuring system.
The natural frequency of the system
n
moves toward the lower frequency range as the tubing length increase due
to the increasing I (see Eq. (5)). The damping ratio is the function of both R and I. It decreases as the tubing length
increases because the variation of I is more signicant than that of R. Thus, the tubing length aects both the damping
ratio and the natural frequency of the pressure-measuring system through the uid inertance I.
0 0.5 1 1.5 2
1
0.5
0
0.5
1
t (s)

p
/
p
0


L=0.96 m
L=0.46 m
L=1.9 m
Fig. 9 Step response of pressure measuring systems with probe-d20, DP45-1 and tubing with dierent lengths.
99
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
0 0.5 1 1.5
1
0.5
0
0.5
1
t (s)

p
/
p
0


straight
loop
arch
Fig. 10 Step response of pressure measuring systems with same SP-probe, tubing, pressure transducer but
dierent shape of tubing.
4.3.2. Effect of tubing shape The eect of curvature of the tubing is examined. The step responses of the systems
comprising the same SP-probe, the same pressure transducer and the same tube but with dierent shape of the tube,
straight, looped and arched, were measured. Figure 10 shows that the step responses were identical to each other regardless
of the tubing shape, indicating that the tubing shape does not aect the dynamic characteristics of the measuring system.
5. Fluctuating Pressure Measurement in a Wake of a Circular Cylinder
5.1. Experimental Condition
Fluctuating pressure measurement was performed in a wake of a circular cylinder. The measurement was undertaken
in a closed-loop water channel. A circular cylinder with the diameter D of 20 mm was placed in a free steam of the water
channel, and the free stream velocity was xed at U

= 0.28 m/s. The temperature of the water was 39

C during the
measurement. The origin of the Cartesian coordinates was placed at the center of the cylinder and the x and y axes were
taken in the streamwise and transverse directions, respectively. The Reynolds number based on U

and D, Re = U

D/,
was 7800, and the measurements were conducted at x/D = 6.
The main SP-probe was placed in the wake, and the secondary SP-probe with geometric conguration identical to
the main probe was placed at the same streamwise location but outside the wake, in order to monitor the background noise
on the uctuating pressure. The background noise on the pressure signal measured by the main probe was reduced based
on the signals measured by the secondary probe in the same manner proposed by Naguib et al. (1996). For the pressure
transducer, the DP45-1 was used with the SP-probes d15 and d20, but the DP45-2 was used for with the SP-probe d10
and air because of the better dynamic response. The sampling rate and sampling time at each measurement point were
100 Hz and 300 s, respectively.
The signals of uctuating pressure were corrected based on the calibration results of the dynamic response of the
pressure-measuring system. Applying the Fourier transform to the measured pressure signals, we evaluated the complex
Fourier coecient B
n
. The amplitude change and the phase delay at nth frequency f
n
, namely A
n
and
n
, were evaluated
o
x
y
U
8
D
sub probe
main probe
DP45-1 DP45-2
Fig. 11 Measurement condition.
10 10
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
3 2 1 0 1 2 3
0.1
0.05
0
y/D
P
/

U
2

Fig. 12 Distribution of mean pressure and at x/D = 6 in a wake of a circular cylinder: blue, d20; green, d15; red,
d10; black, air.
by Eqs.( 6) and (7) with the values of
n
and in Table 2. Using these values of A
n
and
n
, B
n
was corrected as;

B
n
=
B
n
A
n
{cos(
n
) + j sin(
n
)}
, (13)
where j is the imaginary unity, and the pressure signals were reconstructed by applying inverse Fourier transform as
p =
N

n=1

B
n
exp (j2f
n
t) . (14)
As the sampling time was 300 s, the frequency resolution of the Fourier transform was 1/300 Hz. Besides, a low-pass
ltering with the cut-o frequency of 8 Hz was applied to the corrected pressure signals, in order to lter out the high-
frequency noise. The FFT, the interpolation, and the low-pass ltering included in the above procedure were conducted
by using the functions fft, interp1, and filtfilt available in the MATLAB
R
library.
It should be mentioned that the measurement location x/D = 6 belonged to the ow region associated with relatively
strong velocity uctuation which might contaminate the pressure measurement by the cross-ow eect as depicted in
Fig. 3. It is, however, dicult to validate the present pressure measurement results because of the lack in available data of
the pressure uctuation. Instead, in the present study, a focus is put on the consistency among the proposed four dierent
pressure-measuring systems and the noise level against the signal. The accuracy of the pressure measurement under a
relatively strong velocity uctuation was addressed in a separate experiment conducted in the air by Kawata et al. (2014).
5.2. Results and Discussions
The distributions of the mean pressure P at x/D = 6 measured by the four dierent SP-probes are presented in
Fig. 12. The values are scaled by U
2

, and the mean pressure values at y/D = 3.0 are taken to be zero. It should be also
mentioned that the pressure in the range of y/D < 2.0 was not measured by the probes d10 and air due to the limitation of
the traverse equipment. The prole of P measured by all the SP-probes show symmetric proles with respect to y/D = 0
and take the minimum of about 0.1U
2

at the wake center. The agreement between these results is reasonably good.
Figure 13a presents the distributions of the root-mean-square of the pressure uctuation evaluated without correction
based on the calibration results. The values of the pressure uctuation p

are also scaled by U

2

. The proles of the

uncorrected p

measured by the various SP-probes show much dierent distributions. Those measured by the probes d10,
d15, and d20 indicate the two peaks of the pressure uctuation located beside the wake center, but the peak magnitude is
dierent from each other. On the other hand, the prole of p

measured by the probe-air shows much smaller value and

has no peak of uctuation.
Figure 13b presents the proles of the pressure uctuation evaluated with the correction. The proles of p

measured
by the four SP-probes are in good agreement with each other while the proles evaluated without the correction were much
dierent from each other. In particular, the result measured by the probe-air was drastically changed by the correction;
two peaks of the uctuation appears at y/D = 1.
11 11
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
3 2 1 0 1 2 3
0
0.02
0.04
0.06
0.08
y/D
3 2 1 0 1 2 3
0
0.02
0.04
0.06
0.08
y/D
p

U
28
p

U
2 8
(a) (b)
Fig. 13 Distribution of pressure uctuation at x/D = 6 in a wake of a circular cylinder: (a) uncorrected,
(b) corrected. Colors indicate same SP-probe as in Fig. 12
Fig. 14 Power spectra density of pressure uctuation measured at y/D = 1.0: (a) uncorrected, (b) corrected;
blue, d20; green, d15; red, d10; black, air.
Figure 14 presents the power spectrum density (PSD) of the uctuating pressure at y/D = 1.0, comparing the
results measured by the various SP-probes. The proles of the PSD of the uncorrected pressure are show in Fig. 14a. All
of the proles measured by the four dierent SP-probes have a signicant peak at f = 2.6 Hz, which corresponds to the
Strouhal number S
t
= f D/U

of 0.186. This peak was obviously caused by the Karman vortices shed from the cylinder.
The dierence between the results is signicant in the frequency range higher than 1 Hz, including the vortex shedding
frequency. The PSDs measured by the probe-d10, d15, and d20 have the broad hill, which was caused by the resonance.
On the other hand, the prole of the PSD measured by the probe-air does not show such broad increase because the system
with the probe-air is overdamped.
In Fig. 14b, it is shown that the PSDs of the measured pressure are in good agreement with each other after the
correction. The broad hill apparent in the proles of the uncorrected PSD measured by the SP-probe d10, d15, and d20
disappear although the signicant peak by the Karman vortex shedding still remains. The prole of the PSD measured
by the probe air also drastically changed by the correction. This agreement between the pressure uctuations by the
correction supports validity of the present measurement technique.
In order to further examine the measurement results, the probability density function (PDF) of uctuating pressure
was investigated at y/D = 1.0 and 3.0; the former was the location where the pressure uctuation reached a local max-
imum, while the latter was outside the wake region where the pressure uctuation is so weak that basically only the
background noise was measured. Figure 15 presents the PDFs at these locations measured by the four SP-probes after
the correction of the dynamic response was made. The results measured by the four dierent SP-probes agree well with
each other. The PDFs measured at the edge of the wake y/D = 3.0 are concentrated around p/U
2

= 0, while the PDFs

measured at the peak location y/D = 1.0 show smooth distributions with a notable negative skewness. Both distributions
are smooth and free from scratch noise of any kind. Based on these observations, we consider that the noise level of the
pressure measurement was suciently small.
12 12
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst0014]
0.2 0.1 0 0.1 0.2
0
10
20
30
40
50
p/U
2

P
D
F
0.2 0.1 0 0.1 0.2
0
2
4
6
8
p/U
2

Fig. 15 Probability density function of uctuating pressure: (a) measured at y/D = 3.0, (b) measured at
y/D = 1.0. Colors indicate same SP-probe as in Fig. 12.
3 2 1 0 1 2 3
3
2
1
0
1

S
y/D
3 2 1 0 1 2 3
3
2
1
0
1
y/D

S
(a) (b)
Fig. 16 Distribution of skewness factor at x/D = 6 in a wake of a circular cylinder: (a) uncorrected, (b) corrected.
The colors indicate same SP-probe as in Fig. 12
The appropriateness of the correction based on the present calibration results is also supported by the higher moment
of the pressure uctuation. Figure 16 presents the distribution of the skewness factor of the pressure uctuation, dened
as S p
3
/p
3
, comparing the proles evaluated from the measured pressure signals before and after the correction.
The proles of S evaluated without the correction show the negative values of skewness across the wake, and have two
negative peaks located at y/D = 2. Among the results, those measured by the probe-air shows much larger magnitude
of the skewness than those measured by the other probes in the center region of the wake. On the other hand, the proles
of S evaluated with the correction are in fairly good agreement with each other, which also supports the appropriateness
of the correction on the measured pressure based on the calibration results.
As shown in the above discussion, the measured uctuating pressure were examined from dierent aspects including
the second- and third-moments, PSD, and PDF, and these results measured by the four dierent pressure-measuring
systems showed a clear consistency with each other. These observations supports a fair reliability of the present method
of dynamic-response calibration.
6. Concluding Remarks
A technique for uctuating static-pressure measurement at an arbitrary point in a turbulent ow in water was de-
veloped using a pressure-measuring system comprising a miniature static-pressure probe, a strain-gage-type pressure
transducer, and a tube connecting them. A model equation of behavior of the pressure-measuring system was introduced,
and a calibration technique for determining the parameters in the model equation was developed. Eects of the size of
the pressure probe, length of the tube, and sensitivity of the pressure transducer were systematically investigated in the
13 13
Kawata, Maeda and Obi, Journal of Fluid Science and Technology, Vol.9, No.2 (2014)
2014 The Japan Society of Mechanical Engineers [DOI: 10.1299/jfst.2014jfst00 ] 14
dynamic-response calibration. The uctuating pressure measurement was performed in a wake of a circular cylinder, and
the pressure-related statistics measured by dierent four pressure-measuring systems, including the second- and third-
moments, power spectrum density, and probability density function, were in good agreement with each other after the
correction based on the dynamic-response calibration.
Acknowledgment
The authors are very grateful to Prof. Kuniaki Toyoda of Hokkaido Institute of Technology for his shrewd advices on
uctuating pressure measurement. We are also grateful to Assoc. Prof. Koji Fukagata of Keio University for invaluable
discussions. This research was nancially supported through Grant-in-Aid for Research Fellow (No. 23-3769) of Japan
Society for Promotion of Science.
References
Aydin, I., Evaluation of uctuating pressures measured with connection tubes, Journal of Hydraulic Engineering, Vol. 124,
No. 4 (1998), pp. 413-418.
Donovan, F.M., Mcllwain, R.W., Mittmann, D.H. and Taylor, B.C., Experimental correlations to predict uid resistance
for simple pulsatile laminar ow of incompressible uids in rigid tubes, Journal of Fluids Engineering (Transaction
of the American Society of Mechanical Engineering), Vol. 116 (1994), pp. 516-521.
Donovan, F.M., Taylor, B.C. and Su, M.C., One-dimensional computer analysis of oscillatory ow in rigid tubes, Trans-
action of the ASME, Journal of Biomechanical Engineering, Vol. 113 (1991), pp. 476-484.
Geddes, L.A., Athens, W. and Aronson, S., Measurement of the volume displacement of blood pressure transducers,
Medical and Biological Engineering and Computing, Vol. 22 (1984), pp. 613-614.
Iida, A, Otaguro, T., Kato, C. and Shimode, S., Analysis of aerodynamic sound source with measurement of static-pressure
uctuation, International Journal of the Japan Society of Mechanical Engineering, Series B, Vol. 42, No. 4 (1999),
pp. 596-604.
Kawata, T., Naka, Y., Fukagata, K. and Obi, S., Simultaneous measurement of velocity and uctuating pressure in a
turbulent wing-tip vortex using triple hot-lm sensor and miniature total pressure probe, Flow, Turbulence and Com-
bustion, Vol. 86 (2011), pp. 419-437.
Kawata, T., Naka, Y. and Obi, S., Simultaneous measurement of uctuating velocity and pressure in the near wake of a
circular cylinder, Experiments in Fluids, Vol. 55-5 (2014), DOI: 10.1007/s00348-014-1731-7.
Naguib, A. M., Gravante, S. P. and Wark,C. E., Extraction of turbulent wall pressure time-series using an optimal ltering
scheme, Experiments in Fluids, Vol. 22 (1996), pp. 14-22.
Naka, Y., Omori, T., Obi, S. and Masuda, S., Simultaneous measurement of uctuating velocity and pressure in a turbulent
mixing layer, International Journal of Heat and Fluid Flow, Vol. 27 (2006), pp. 737-746.
Shirahama, Y. and Toyoda, K., Development of the probe to measure static-pressure uctuation (application to the mea-
surements of jets), Transaction of the Japan Society of Mechanical Engineering, series B, Vol. 59-567 (1993),
pp. 3381-13387 (in Japanese).
Taylor, B.C. and Donovan, F.M., Hydraulic resistance and damping in catheter-transducer systems, IEEE Engineering in
Medicine and Biology, 0739-5175 (1992).
Terashima, O., Sakai, Y. and Nagata, K., Simultaneous measurement of velocity and pressure in a plane jet, Experiments
in Fluids, Vol. 53 (2012), pp. 1149-1164.
Toyoda, K., Okamoto, T. and Shirahama, Y., Eduction of vortical structures by pressure measurements in noncircular jets,
Applied Scientic Research, Vol. 53 (1994), pp. 237-248.
Tsuji,Y., Fransson, J. H. M., Alfredsson, P. H. and Johansson, A. V., Pressure statistics and their scaling in high-Reynolds-
number turbulent boundary layers, Journal of Fluid Mechanics, Vol. 585 (2007), pp. 1-40.