ARTICLE IN PRESS

Journal of Wind Engineering

and Industrial Aerodynamics 93 (2005) 451460
www.elsevier.com/locate/jweia

Short note
Drag coefcients and Strouhal numbers of a port
crane boom girder section
A. Scarabino, J. Maranon Di Leo, J.S. Delnero, F. Bacchi
Laboratorio de Capa Limite y Fluidodinamica Ambiental, Facultad de Ingeniera UNLP, Dept. Aeronautica.
Calle 1 y 47, (1900) La Plata, Argentina

Received 16 October 2004; received in revised form 15 March 2005; accepted 30 March 2005
Available online 24 June 2005

Abstract

This work reports the drag coefcients (Cd) for three wind directions measured in low
turbulence ow and in turbulent ow with characteristics of the atmospheric boundary layer
wind, and the Strouhal numbers (St) of an approximately trapezoidal anged section, used in
the boom girder of a 100 m high port crane. These experimental results were obtained at the
Boundary Layer Wind Tunnel of the Dept. Aeronautica at the Universidad Nacional de La
Plata, in the range of Reynolds numbers between 30,000 and 180,000. The drag coefcients in
the three studied directions showed a reduction for turbulent ow. Further measurements were
carried out for the model with an inclination of 801 relative to the ow direction, the position
of the crane boom when out of service, giving practically the same drag coefcients.
r 2005 Elsevier Ltd. All rights reserved.

Keywords: Bluff bodies; Drag coefcient; Strouhal number; Turbulence

1. Introduction

One of the hazards of port crane operation is the effect of wind on the crane
structure. Static and dynamic wind loads are important design considerations, both
in terms of the structure and mechanisms to withstand the wind loads and for the

Corresponding author. Tel./fax: +54 221 4236679.

E-mail address: [email protected] (A. Scarabino).

0167-6105/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jweia.2005.03.004
 ARTICLE IN PRESS
452 A. Scarabino et al. / J. Wind Eng. Ind. Aerodyn. 93 (2005) 451460

crane operator to be able to take whatever precautions may be necessary. Container

crane accidents have been caused by the effect of strong gusts of wind (such as
microbursts) during normal operation. The approximately two-dimensional (2-D)
geometry and the low natural frequencies of the long crane booms cause that even
light winds may be able to generate dangerous dynamic loads through vortex
shedding mechanisms.
The fact that turbulence inuences the wind loads on bodies has been widely
reported [13]. Therefore, drag coefcients obtained in wind tunnel tests under ideal
conditions are not always representative of the real wind loads on a structure.
The aim of this work is simply to report the drag coefcients and Strouhal
numbers of a characteristic section of a port crane boom girder in smooth ow and
in ow with turbulence intensity and spectral characteristics representative of
atmospheric wind.

2. Methodology

The experiments were carried out at the Boundary Layer and Environmental
Fluid Dynamics Laboratory (LACLYFA) at the Faculty of Engineering at the
Universidad Nacional de La Plata, Argentina. The wind tunnel, equipped with an
electronic speed control, which allows speeds up to 20 m/s, is described in Boldes et
al. [4]. It is a closed section tunnel with a test section 1.40 m wide, 1 m high and 7.2 m
long, powered by a 50 HP DC electric motor with an axial ow, adjustable pitch
blade propeller. The natural turbulence intensity of the wind tunnel in the range of
velocities for these experiments was below 2%. The conditions for the measurements
in atmospheric turbulence were generated by means of an array of vertically
distributed equally spaced horizontal airfoils at the entrance of the test section,
which induce different vertical deections of the ow. Each airfoil could be
individually rotated 3601 around its longitudinal axis. Additionally, upstream of the
section model, the ow was processed by an arrangement of triangular spires.
A model of the boom section, of 0.194 m width, 0.091 m depth and 0.7 m height,
was mounted vertically in the test section, on an aerodynamic balance (load cell and
V-Shay 2310 bridge) placed under the tunnel oor. No corrections were made for
blockage ratio, which was less than 10%. The model scale was 1:20. The balance was
calibrated with an accuracy of 70.01 N. A at plate was placed slightly above the
top of the models, without touching them, in order to reduce any 3-D effects.
Fig. 1 shows the cross section, which is trapezoidal with anges at the top, at the
oor and on the sides, as shown; the model dimensions and the different studied
wind directions.
Fig. 2 sketches the crane boom girder for which this section is proposed. In its
operating condition the boom is horizontal, and wind direction 3 is perpendicular to
its axis. The out-of-service position is close to vertical (801 to the ground). In this
condition, wind directions 1, 2 and 3 must be considered. For generality, and for
comparison with other shapes, we studied rst the 2-D characteristics of the section
with a model mounted vertically on the wind tunnel oor. Then, we measured the
 ARTICLE IN PRESS
A. Scarabino et al. / J. Wind Eng. Ind. Aerodyn. 93 (2005) 451460 453

Fig. 1. Model cross section, and wind directions.

Out of Service
Config

10

Wind Dir 3

Plan View

Fig. 2. Sketch of the boom girder.

drag coefcients of the section for wind directions 1 and 2, with a second model
inclined 801, as shown in Fig. 3. The aerodynamic drag for direction 3 should not be
affected for this inclination, which implies a rotation in the plane normal to the wind.
The mean velocity was measured with a portable hot wire anemometer Dantec
Flowmaster. Instantaneous velocities were acquired at a sample rate of 600 Hz with a
two-channel hot wire anemometer Dantec 90C10, with an X lm probe, and data
were postprocessed in order to obtain values of turbulence intensity and velocity
spectra, for the characterization of the incident turbulence, and for the determina-
tion of the frequency of vortex shedding in the wake of the model.
 ARTICLE IN PRESS
454 A. Scarabino et al. / J. Wind Eng. Ind. Aerodyn. 93 (2005) 451460

Fig. 3. Model with 801 inclination, in the wind tunnel. The upper at plate forces the ow to be 2-D.

Through trial and error, we were able to reproduce the desired characteristics of
the turbulence (intensity and spectrum) at the model mid height, and for an average
Reynolds number of the tested range. By means of variable opening of vanes in the
entrance of the test section, an approximately homogeneous mean velocity prole
was achieved, in spite of the perturbations due to the triangular spires.
For the measurements in turbulent ow, the modeled wind was representative of
ow over open terrain, with a roughness height z0 between 0.001 and 0.005 m, at
target heights between 40 and 100 m. The turbulence intensity for the longitudinal
velocity component, Iu, at a height z is given by ([5])

1
Iu .
lnz=z0

For these heights, this expression gives

0:09 9%oIuo0:11 11%.

The turbulence intensity obtained with the spires and airfoils was Iu 10%,
considered representative of the desired conditions.
 ARTICLE IN PRESS
A. Scarabino et al. / J. Wind Eng. Ind. Aerodyn. 93 (2005) 451460 455

Fig. 4. Normalized spectrum of longitudinal velocity uctuations (continuous)Von KarmanHarris

spectrum (dashed). U 10 m=s.

The integral time scale was approximated with the time lag t1=e , for the
1=e value of the autocorrelation function [6]. The integral length scale
for the longitudinal uctuations, Lu was estimated from the integral
time scale, Tu, using frozen ow theory, as Lu UT u , giving values close to
0.08 m.
Normalized spectra were computed for the longitudinal uctuation u0 . As shown
in Fig. 4, spectra were in very good agreement with Von KarmanHarris analytical
approximation [5]

f Su 4f Lu=U
,
2
su 1 70:8f Lu=U2 5=6

f being the frequency in Hz, Su the spectral power density of the longitudinal
uctuations u0 , su the longitudinal uctuations rms value, Lu the longitudinal
turbulence length scale and U the mean velocity.
The drag coefcients were then measured for wind directions 1 and 2, with the
second model inclined 801 with respect to the wind direction, the angle of the crane
boom when out of service.
In order to determine the different Strouhal numbers, instantaneous velocity data
were acquired in the shear layer in the wake of the section, as sketched in Fig. 5, and
the vortex shedding frequency was deduced from the marked spectral peaks of the
velocity uctuations. Due to the wider spreading of the spectra in turbulent ow,
Strouhal numbers were only computed for low turbulence ow (indicated in the
results as smooth ow).
 ARTICLE IN PRESS
456 A. Scarabino et al. / J. Wind Eng. Ind. Aerodyn. 93 (2005) 451460

PROBE

Fig. 5. Location of the anemometric probe for the determination of St.

3. Results

Table 1 summarizes the computed mean drag coefcients and Strouhal numbers.
The balance registered only the mean drag force. Numbers in brackets indicate the
minimum and maximum values and the standard deviation of the mean drag
coefcient for each wind direction over the range of Reynolds numbers covered in
our experiments. Cd for the inclined model showed a very slight reduction.
Both the Reynolds Re LU =n, with n the kinematic viscosity) and the Strouhal
St f s L=U numbers are computed with L being the projected width of the section
(0.194 m for wind directions 1 and 2, and 0.091 m for wind direction 3) and U the
mean velocity. The vortex shedding frequency, f s , is identied from the normalized
frequency at the peak in the spectrum of the velocity uctuations.
The drag coefcients for wind directions 1 and 3, and the Strouhal numbers for all
wind directions compare well with those of other sharp-edged sections shown in
Table 2.
Fig. 6 shows the experimental results. The drag force D was measured, and the
drag coefcients C d D=12rU 2 LH are presented as a function of the Reynolds
number based on the mean velocity U and the projected width of the section in the
direction of the wind, L. H is the height of the model and r the air density.
Whereas the drag coefcients for low turbulence ow are practically constant
through the range of tested Reynolds numbers, for turbulent ow all show a small
decreasing trend for larger values of Re. The drag coefcients obtained in smooth
ow are higher, and could therefore be used in a conservative design.
 ARTICLE IN PRESS
A. Scarabino et al. / J. Wind Eng. Ind. Aerodyn. 93 (2005) 451460 457

Table 1
Drag coefcients Cd and strouhal numbers St of the section

Cd (Min.Max, Std. dev.)

Model perpendicular to the wind

Wind direction 1 1.87 (smooth) (1.841.91, 0.027)
1.71 (turbulent) (1.611.79, 0.065)
Wind direction 2: 1.28 (smooth) (1.261.30, 0.012)
1.05 (turbulent) (0.951.13, 0.054)
Wind direction 3: 1.57 (smooth) (1.521.61, 0.027)
1.31 (turbulent) (1.261.43, 0.090)
Strouhal number: 0.14 (Wind direction 1)
0.175 (Wind direction 2)
0.07 (Wind direction 3)
Model inclined at 801 (out-of-service position):
Wind direction 1: 1.88 (smooth) (1.831.92, 0.023)
1.61 (turbulent) (1.571.64, 0.028)
Wind direction 2: 1.24 (smooth) (1.181.27, 0.029)
1.07 (turbulent) (1.021.12, 0.043)

Table 2
Cd and Strouhal number of other sharp-edged shapes, from Refs. [5] and [7]

Cross-section Cd St

1.2 N/A

2 0.14

2 0.145

L L/h

2 0.06
h
0.5

2 1.5 N/A

Fig. 7 illustrates two spectra obtained for wind direction 3, at different Reynolds
numbers, in the shear layer of the near wake of the section. The frequency in the
abscissae has been made non-dimensional as fL=U, in order to identify the Strouhal
 ARTICLE IN PRESS
458 A. Scarabino et al. / J. Wind Eng. Ind. Aerodyn. 93 (2005) 451460

Smooth flow
2
1.8
1.6
1.4
1.2
Cd

1
0.8
0.6
Wind 1 Cd 1
0.4
Wind 2 Cd 2
0.2
Wind 3 Cd 3
0
0 20000 40000 60000 80000 100000 120000 140000 160000 180000 200000
Re

Turbulent flow
2
1.8
1.6
1.4
1.2
Cd

1
0.8
0.6
Wind 1 Cd 1
0.4
Wind 2 Cd 2
0.2
Wind 3 Cd 3
0
0 20000 40000 60000 80000 100000 120000 140000 160000 180000 200000
Re

Fig. 6. Drag coefcients for wind directions 1, 2 and 3. The horizontal lines represent the mean values over
all the tested Re.

number at the spectral peak. In all tested cases, the spectral peaks were easily
identied and their non-dimensional frequencies did not depend on the Reynolds
number.

4. Conclusions

Aerodynamic drag coefcients for smooth and turbulent ow and Strouhal

numbers of a anged trapezoidal section used in a port crane boom girder are
reported in this work. For smooth ow, the drag coefcients and Strouhal numbers
compare acceptably with reported values for other sharp-edged sections.
For turbulent ow, the drag coefcients showed a reduction between 4% and
11%, compared with those measured in smooth ow and also a small decreasing
 ARTICLE IN PRESS
A. Scarabino et al. / J. Wind Eng. Ind. Aerodyn. 93 (2005) 451460 459

600

400

S(f)
Re = 54000
200

0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2

300

200
S(f)

Re = 30000
100

0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
fL/U

Fig. 7. Example of spectra in the model wake, wind direction 3.

trend with Reynolds number. Further work is needed in order to investigate the
possible reasons for this behavior, testing this and different sections with turbulence
of different intensities and scales.
For sharp-edged bluff bodies, separation of ow occurs at the sharp edges and
corners for practically all values of Re of interest in wind engineering, in wind
tunnel or full-scale problems. This leads to the fact that the dependence of
pressure coefcients and consequently of drag coefcients on Reynolds number for
sharp-edged bluff bodies (but not for rounded edges) is often negligible, at least in
smooth ows [5,7]. The same happens for the non-dimensional vortex shedding
frequency for sharp-edged sections. Our results conrm this independence. On the
other hand, in turbulent ows, some characteristics, as the turbulence small scales,
can affect the pressure distributions and, in consequence, the resultant aerodynamic
forces [8,9].
Pressure and force coefcients in turbulent ows show a larger scatter than
in smooth ows, depending on the characteristics of the incident turbulence, making
difcult the choice of the suitable coefcient for a given problem. Furthermore,
while an approximately homogeneous turbulent ow can be reasonable for
computing the forces on the boom in its horizontal operating condition, a realistic
analysis of the wind loads in the quasi-vertical out-of-service position requires the
simulation of the atmospheric boundary layer on a model of the whole crane
structure, and not only the characteristics of the ABL at a certain height. In this
context, and in a design stage, the use of conservative values seems reasonable.
The Cds obtained with smooth ow are at most 20% higher than those measured
with turbulent ow, and this is much less than the safety factors usually employed
for design.
 ARTICLE IN PRESS
460 A. Scarabino et al. / J. Wind Eng. Ind. Aerodyn. 93 (2005) 451460

Acknowledgments

The authors are thankful to Mr. Geronimo Marinangeli for his valuable help in
the acquisition of the anemometric data.

References

[1] H. Tieleman, Z. Ge, M. Hajj, T. Reinhold, Pressures on a surface-mounted rectangular prism under
varying incident turbulence, J. Wind Eng. Ind. Aerodyn. 91 (2003) 10951115.
[2] H. Tieleman, Z. Ge, M. Hajj, T. Reinhold, Importance of turbulence for the prediction of surface
pressures on low-rise structures, J. Wind Eng. Ind. Aerodyn. 6971 (1997) 519528.
[3] T. Tamura, T. Miyagi, The effect of turbulence on aerodynamic forces on a square cylinder with
various corner shapes, J. Wind Eng. Ind. Aerodyn. 83 (1999) 135145.
[4] U. Boldes, J. Colman, V. Nadal Mora, The Boundary Layer Wind Tunnel at the Faculty of
Engineering, University of La Plata, Argentina, Latin Am. Appl. Res. 25 (1995) 7585.
[5] J.D. Holmes, Wind Loading of Structures ed., Taylor and Francis, London, 2001.
[6] J.C. Kaimal, J.J. Finnigan, Atmospheric Boundary Layer Flows, Oxford University Press, Oxford,
1994.
[7] P. Sachs, Wind Forces in Engineering, Pergamon Press, Oxford, 1978.
[8] H.W. Tieleman, Pressures on surface-mounted prisms: the effects of incident turbulence, J. Wind Eng.
Ind. Aerodyn. 49 (1993) 289299.
[9] H.W. Tieleman, R.E. Akins, Effects of incident turbulence on pressure distributions on rectangular
prisms, J. Wind Eng. Ind. Aerodyn. 36 (1990) 579588.