Professional Documents
Culture Documents
www.elsevier.com/locate/engstruct
Received 16 January 2006; received in revised form 29 June 2006; accepted 29 June 2006
Available online 14 August 2006
Abstract
Highway light poles are slender structures usually characterized by low values of structural damping, a factor that can lead to large-amplitude
vibration sometimes leading to collapse. This paper is motivated by a recent investigation, conducted to identify the reason for repeated failures,
experienced by aluminum tapered light poles in the State of Illinois during a winter storm. The study combined numerical and experimental
full-scale analysis of the structural system and its response to simulated external actions. It was observed that, despite the simple structural form,
the definitive identification of the mechanism causing the oscillations was challenging due to inherent variability in the configuration as well as
the paucity of environmental and response data. However, a plausible mechanism was identified, and a mitigation technique was proposed and
evaluated for amplitude reduction.
c 2006 Elsevier Ltd. All rights reserved.
Keywords: Light poles; Dynamic loads; Vibration; Damping; Wind; Frozen precipitation; Galloping
1. Introduction
Highway light poles are frequently subjected to environmental and wind loads. Despite the simple structural system and
the availability of specific design tools (e.g., [1]), particular occurrences of unpredictable load configurations can lead to the
failure of such systems or their sub-components. This paper is
motivated by a research project managed by the Illinois Department of Transportation, initiated to understand the nature of
some recent failures experienced on light pole structures in both
serviceability and strength (collapse). One of the problems was
the occurrence of large oscillation amplitudes experienced during a wind storm (Western Illinois) in tapered aluminum-alloy
posts (Fig. 1(a) and (b)). These are typical slender systems, with
reduced mass and low damping. As a result, approximately 140
units failed during this event.
Corresponding address: Department of Civil and Environmental Engineering, Northeastern University, 400 Snell Engineering Center, 360 Huntington
Avenue, Boston, MA 02115, USA. Tel.: +1 617 373 5186; fax: +1 617 373
4419.
E-mail address: lucac@coe.neu.edu (L. Caracoglia).
822
Fig. 1. 13.7 m aluminum-alloy highway light pole. (a), (b) Example of collapsed units near Galesburg, Illinois, USA, 2003 (courtesy: Illinois Department of
Transportation). (c) Full-scale laboratory experiments.
(1)
(2)
The reference design value of wind speed (AASHTOIV, [1]) is defined as the three-second gust with fifty-year
return period at z = 10 m (basic wind speed, BWS). In the
R(z)
through an appropriate and unique Gust Effect Factor
(GEF) [11],
G R (z) = R(z)/
R,
(3)
823
824
Table 1
First-mode gust buffeting response for a 13.7 m aluminum-alloy pole with D = 5 mm and 1,Struct = 0.5%. (a) u = 1.4 m/s and variable z 0 ; (b) U (10) = 22 m/s,
variable z 0 and u
(moment, shear at the base indicated as M0C max , V0C max and
derived as equivalent static loads [10]).
For an average wind and turbulence scenario (z 0 = 0.02 m,
u = 1.4 m/s, U (10) = 22 m/s, along-wind turbulence
intensity Iu = 0.16) X eq,C (L) can vary in the range 0.45
0.60 m (Table 1(a)), corresponding to a dynamic oscillation
with 0.50 < xeq,C (L) < 0.60 m peak-to-peak amplitude
(Table 1(b)), and a mean deflection equal to 0.26 m under steady
wind for D = 5 mm. The first-mode total damping in the
along-wind direction, 1,x,tot = 1,x + 1,x,Aerod (structural
and aerodynamic components) is significant and mainly related
to the beneficial effect of the luminaire: 4%5% at U (10) =
22 m/s. A further parametric investigation, also considering
variability of C D0 and C D1 , suggests a minimum value of 3.4%.
Despite the moderate amplitudes, the moment and shear forces
recorded at the base are approximately two to three times lower
than the values necessary to induce failure in the break-aways.
In the tables, the wind conditions associated with
both failure mechanisms (dashed line for mechanism one,
dasheddotted line for mechanism two) were computed,
including the effect of the potential presence of geometric
tolerance in the installation of the anchorages to ground. It was
concluded that the wind characteristics or the extremely large
vibration necessary to exceed the tensile capacity of the anchor
bolts were not consistent with the meteorological conditions at
the time of the event, requiring much higher winds or physically
unreasonable turbulence intensity levels.
5. Influence of vortex shedding on the response during
collapse
Vortex shedding in the across-wind direction was neglected
since it usually involves higher modes, and dynamic interaction
can be excluded due to evident effectiveness of the canister
vibration damper (Section 2) tuned to the second mode. Lockin from vortex shedding was not responsible for failures,
evidenced by the observed large-amplitude oscillations and the
(4)
(fatigue limit states) either applied to the aluminum crosssection (pole base) or to the break-aways tensile strength
to simulate the effects associated with the different fatigue
categories (severity of the loading). In a recent study [21] it was
shown that fatigue life of steel poles can be as low as one year
under special wind exposure conditions and due to the effects
of drag reduction in the critical Re region. The subsequent
application of the closed-form method proposed in [21] to
this specific case suggested a life expectation variable between
two and seven years for the aluminum pole base cross-section,
although a penalizing factor relating stress variation to wind
speed was employed due to the limited information and data.
Fatigue-life predictions larger than seven to eight years
seem compatible with previous results derived for steel poles
through the method utilized in [18]. Moreover, buffetingrelated fatigue did not seem compatible with the multiple
simultaneous occurrences experienced in this particular event,
also considering the fact that most units were installed between
1998 and 2000. Specifically, a concurrent collapse of systems
located in the same area with different failure characteristics
(and service ages) exclusively associated with fatigue was
considered as statistically improbable. Fatigue issues related to
vortex shedding were also excluded due to the extremely low
value of lock-in velocity range corresponding to the first mode,
the presence of the impact canister damper and the indication
of large-amplitude vibration at the time of the incident.
7. Across-wind galloping instability
Galloping is an aeroelastic instability phenomenon involving
large-amplitude oscillation in the across-wind direction [3,
5], triggered by a decrement of the aerodynamic damping
component and involving energy transfer from the wind flow.
The necessary condition for across-wind first-mode galloping
is
1,y,tot = 1,y + 1,y,Aerod = 0.
(5)
C L0
+ CD
1
1
2
2 m 1,eq 1,y Dref
,
(6)
825
826
Table 2
First-mode galloping critical velocity, UCR , at z = 10 m, for a 13.7 m pole with M L = 18 kg as a function of wall thickness, D , and variable C D , C L , C L0
(i)
(ii)
(iii)
(iv)
(v)
(vi)
(vii)
(viii)
Case
D
(mm)
C D0
0
C L0
C D1
0
C L1
UCR
(m/s)
7
7
5
7
7
7
7
7
1.0
1.0
1.0
1.0
0.4
0.4
0.4
0.0
1.5
1.9
1.9
0.0
0.0
0.0
0.0
0.0
1.4
1.4
1.4
1.4
1.4
1.4
0.5
0.0
0.0
0.0
0.0
5.3
5.3
3.0
2.0
2.0
69
18
15
110
14
208
300
17
Fig. 2. Schematic representation of galloping threshold UCR and aeroelastically enhanced phenomenon as a function of normalized root-mean-square
across-wind response.
(7)
must satisfy the condition K S > N (see Eq. (6)). This quantity
is a measure of susceptibility to aerodynamic instability
(typically, lock-in from vortex shedding). The positive quantity
N varies from case to case; however, for masts with roughness
or longitudinal attachments in the critical Reynolds Number
regime, a value of 4.0 is suggested in [27], while a K S between
1.2 and 1.4 was computed for the analyzed case, for a reference
diameter (0.2 m), equivalent mass corresponding to first-mode
vibration and 0.5% damping.1
An increment of structural damping was strongly indicated
as a possible solution, leading to an increase of galloping
instability threshold. Different damping devices have been
proposed for light poles, traffic signals and luminaries
susceptible to wind excitation [20,28], including, for example,
aerodynamic damping devices [29], TMD/impact dampers [30]
on long horizontal arms of traffic lights or impact ball dampers
on posts to reduce Karman-vortex-induced oscillation [31]. In
1 A recent development of this research, based upon a statistical method
for the numerical simulation of the failure probability ( p f ) due to galloping
induced by simulated precipitation deposit, suggested that the original threshold
indicated in this paper may be replaced by N = 3.0, approximately equivalent
to 104 < p f < 103 (yearly value) [23].
827
828
829
Table 3
First-mode structural damping (1,x ) of a 13.7 m aluminum alloy highway pole
for large (a) and moderate (b) oscillation amplitudes
(a)
Case
f (Hz)
(note 2)
x0,L
(m)
1
(%)
M L = 0, NO chain
(test01BNC081, note 1)
M L = 0, chain
(test01BNC081)
M L = 13 kg, NO chain
(test01BC080, note 1)
M L = 13 kg, chain
(test01BC080)
1.418
(1.461)
1.417
n.a.
0.2
0.24
4.1
1.157
(1.156)
1.109
n.a.
0.1
0.25
2.1
1.418
(1.468)
1.412
n.a.
0.1
0.12
3.9
1.158
(1.158)
1.142
n.a.
0.1
0.12
2.4
(b)
M L = 0, NO chain
(test01BNC040, note 1)
M L = 0, chain
(test01BNC040)
M L = 29 Lb, NO chain
(test01BC040, note 1)
M L = 13 kg, chain
(test01BC040)
Note 1: 1,y derived from residual values in the presence of hanging chain.
Note 2: Measured f in parentheses computed during free-vibr. without chain.
830
831