1352

A simplified method for seismic analysis of rooftop

telecommunication towers
Rola Assi and Ghyslaine McClure

Abstract: A simplified method is presented in this paper for the estimation of forces at the base of telecommunication
towers mounted on building rooftops due to seismic excitation. Although some codes and standards propose simplified
methods for the evaluation of base shear forces for towers founded on ground, no method yet exists for the evaluation
of overturning moments. The proposed simplified method is based on numerical simulations using truncated modal
superposition, which is widely used for seismic analysis of linear structures. The method requires the prediction of input
seismic acceleration at the buildingtower interface, the definition of an acceleration profile along the building-mounted
tower, and the determination or evaluation of the mass distribution of the tower along its height. The method was developed
on the basis of detailed dynamic analyses of three existing towers assumed to be mounted separately on three buildings.
It was found that the method yields conservative results, especially for the overturning moments.

Key words: self-supporting towers, earthquake, horizontal excitation, dynamic analysis, acceleration, modal superposition.
Rsum : Cet article prsente une mthode simplifie destimation des forces agissant la base de pylnes de tlcommunication
installs sur les toits dimmeubles et causes par une excitation sismique. Alors que certains codes et normes proposent
des mthodes destimation simplifies des forces de cisaillement la base pour les pylnes autoportants installs sur le
sol, aucune mthode nexiste encore pour lvaluation des moments de renversement. La mthode simplifie propose est
base sur des simulations numriques utilisant une superposition modale tronque qui est couramment utilise pour lanalyse
sismique de structures linaires. La mthode requiert la prdiction de lacclration sismique linterface immeuble-pylne,
la dfinition dun profil dacclration le long et la dtermination de la distribution de la masse du pylne le long de sa
hauteur. La mthode a t dveloppe en se basant sur lanalyse dynamique dtaille de trois pylnes existants prsums
tre installs sur trois immeubles existants. La mthode a produit des rsultats conservateurs, particulirement pour les
moments de renversement.
Mots-cls : pylnes autoportants, sisme, excitation horizontale, analyse dynamique, acclration, superposition modale.
[Traduit par la Rdaction] Assi and McClure 1363

Introduction unpublished report, 1999), it was concluded that tall broadcast

To perform an adequate seismic design of telecommunica-
tion equipment, it is necessary to evaluate seismic forces
realistically. Because the design of towers on ground is usu-
ally controlled by ice and wind loads, most of the research
and information resources on the analysis of steel lattice
telecommunication towers are devoted to analysis under wind
and ice loads; therefore, research on the seismic response of
these towers has not been abundant. In a survey of earth-
quake performance of telecommunication towers (A.J. Schiff,
Received 20 March 2006. Revision accepted 26 April 2007.
Published on the NRC Research Press Web site at cjce.nrc.ca
on 2 November 2007.
R. Assi1 and G. McClure.2 Department of Civil Engineering
and Applied Mechanics, McGill University, 817 Sherbrooke
Street West, Montral, QC H3A 2K6, Canada.
Written discussion of this article is welcomed and will be
received by the Editor until 29 February 2008.
1
Present address: SNC-Lavalin Group Inc., 620 Ren-Lvesque,
Montral, QC H3B 1N7, Canada.
2
Corresponding author (e-mail: ghyslaine.mcclure@mcgill.ca).
towers and large building-supported microwave towers are
the most vulnerable to earthquakes, although none of these
towers has been a direct threat to life safety during an earth-
quake. In areas prone to earthquakes, the main issue for
strategic telecommunication towers is their functionality during
or immediately after an earthquake. Modern codes and stan-
dards have recently addressed the seismic analysis of tele-
communication towers on building rooftops by either proposing
a simplified method for the estimation of seismic base shear
forces (ICC 2000; NRCC 2005; ANSI/EIA/TIA 2005) or at
least acknowledging the importance of the problem (SAA
1994; CSA 2001). The current seismic provisions available
in codes and standards address only the input base shear
force to telecommunication towers, whereas the base over-
turning moment is not covered. In the opinion of the authors
based on detailed dynamic analyses (Assi 2006), the simpli-
fied approaches in codes and standards are often inappropriate,
especially since most telecommunication towers are consid-
ered as acceleration-sensitive components with distributed
mass and stiffness (CSA 2001).
One of the first publications discussing earthquake effects
on antenna-supporting lattice towers was authored by Konno
and Kimura (1973), who studied an instrumented tower owned

Can. J. Civ. Eng. 34: 13521363 (2007) doi:10.1139/L07-061 2007 NRC Canada
Assi and McClure
Table 1. Properties of the buildings studied.
Year of
Location Building construction
Tainan, Taiwan CHYBA9 1980
Jia-Yi, Taiwan CHYBA4 1983
Montral, Quebec, Canada 2020 University 1973
by the Nippon Telegraph and Telephone Corporation (NTT)
and mounted on a building rooftop during the 1968 Off-
Tokachi earthquake in Japan. Sato et al. (1984) analyzed the
data from strong-motion accelerographs in selected build-
ings owned by the NTT in Japan. They studied the input
seismic force to be used for the design of appendages, par-
ticularly telecommunication towers, and found that a maxi-
mum acceleration amplification of 4 at the rooftop was
appropriate. A drawback of their study was that they aver-
aged the amplifications calculated for the two main horizon-
tal directions, considering that the structural system may be
different in both principal directions of the buildings and
consequently affect their response. Hiramatsu et al. (1989)
reported the continuation of this investigation of the seismic
response of NTT telecommunication equipment mounted on
building rooftops. In general, their results agreed with the
earlier observations of Sato et al.
There have been unofficial reports of tower damage incurred
during the 1994 Northridge earthquake in California, involving
mostly localized damage in the vicinity of antenna mounts
(Madugula 2002). Similar localized damage was reported by
Pierre (1995) following a visit to Japan after the Hanshin
Awaji (Kobe) earthquake that occurred on 17 January 1995.
Kanazawa and Hirata (2000) developed a seismic response
spectrum method for the analysis of secondary systems while
considering the dynamic interactions between the primary
and secondary structures and the transient response effects.
To evaluate their proposed method, the researchers performed
timehistory analyses on a buildingtower model consisting
of a tower mounted on a single-degree-of-freedom primary
system. A similar approach had been developed at McGill
University by Khedr (1998) and Khedr and McClure (2000) for
steel lattice towers on firm ground and subjected to both
horizontal and vertical earthquake accelerations, but their
method is not applicable to towers mounted on rooftops or
other flexible supporting structures. In a preliminary study,
McClure et al. (2004) used numerical simulations to explore
the correlation between the building accelerations and the
maximum seismic base shear as well as the base overturning
moment of towers mounted on building rooftops. This study
was the precursor of the research reported in Assi (2006).
This paper presents a simplified, accurate, and yet easy to
implement static method for the analysis of telecommunica-
tion towers mounted on building rooftops and subjected to
earthquake excitation. The purpose of this method is to pro-
vide tower designers a tool to quickly determine the seismic
forces on telecommunication towers mounted on building
rooftops, namely the tower base shear force and overturning
moment. In design, these seismic effects can then be com-
pared to those of other environmental loads like wind and
ice. If this preliminary assessment indicates that seismic loads
1353
No. of
storeys above Height
Use LLRS ground (m)
Telecom Dual 4 20.0
Hospital Frame 6 24.2
Office Frame 27 115.2
govern the tower design, then a detailed dynamic analysis of
the buildingtower combination would become necessary
for a more accurate assessment of internal forces and base
reactions.
Method of analysis
The proposed method is based on numerical simulations
using truncated modal superposition analysis to determine
the maximum forces in the tower legs at the buildingtower
interface. Three-dimensional (3D) finite element models for
three towers assumed to be mounted on three buildings were
generated using the SAP2000 software (Computers and
Structures, Inc.) (Wilson and Habibullah 2003). Elastic time
history modal superposition analysis was performed for each
buildingtower combination. The 20 lowest frequency modes
were considered, and a uniform viscous damping ratio of 3%
critical was used for all modes, which is a common practice
for bolted steel lattice structures (Madugula 2002). Two of
the modeled buildings are located in Taiwan, and the third is
located in downtown Montral, Quebec, Canada. The models
of the Taiwanese buildings were calibrated using floor accel-
erations recorded during the 1999 Chi Chi earthquake, and
the fundamental periods were extracted by system identifica-
tion techniques using the same accelerograms (Assi 2006).
The model of each buildingtower combination was sub-
jected to two sets of earthquake records applied to both prin-
cipal horizontal directions (U1 and U2) of the buildings.
Description of the buildingtower combinations
Some geometric details of the buildings and towers along
with a brief overview of the main finite element modeling
assumptions are presented in the following sections. The
earthquake records applied to the base of the buildings are
also described.
Buildings
Geometric properties of the three buildings studied are
given in Table 1, and their isometric and facade elevation
views are shown in Figs. 13. The lateral load resisting
system (LLRS) is either a reinforced concrete frame system
(frame) or a moment-wall system (dual). Detailed 3D elastic
models of the three buildings were generated in SAP2000.
Rigid floor slabs were assumed, and the columns, beams,
and walls were modeled in detail. The mass of nonstructural
components and finishing was distributed to columns and
walls according to their tributary area. Table 2 summarizes
the lowest periods of vibration obtained for the building
models, corresponding to the fundamental sway modes and
the torsional mode.
2007 NRC Canada
1354 Can. J. Civ. Eng. Vol. 34, 2007

Fig. 1. Isometric and facade elevation views of building CHYBA9, Fig. 2. Isometric and facade elevation views of building CHYBA4,
Tainan, Taiwan. FL, floor. Numbers indicated on elevation view Jia-Yi, Taiwan. Numbers indicated on elevation view denote labels
denote labels of the various sensors. of the various sensors.

Fig. 3. Isometric view of building at 2020 University Street,

Montral, Quebec, Canada.
Towers
Three typical medium-height self-supporting steel towers
were studied. Table 3 lists the geometric properties of the
towers, including their height, base width, top width, and
mass. It should be noted that the mass of the towers does not
include the mass of antennas, transmission lines and other
attachments, platforms, and ladders. The towers labeled TC1,
TC2, and TC3 are three-legged steel lattice towers with an
equilateral triangular base. The towers were modeled in
SAP2000 as 3D frametruss linear elastic structures. Frame
elements were used for the main legs, and truss elements for
diagonal and horizontal members. The tower models were
assumed rigidly connected to the roof of the building models.
An example of such a connection is illustrated in Fig. 4 for
the telecom building in Tainan City, Taiwan. Figure 5 shows
the finite element meshes of the tower models, and Table 4
gives the four largest sway and torsion natural periods calcu-
lated for bare towers.
(PGA) to the peak ground horizontal velocity (PGV), includ-
Earthquake records ing 14 records with high a/v ratio labeled H, 15 records with
The generated models in this study are subjected to two medium a/v ratio labeled M, and 15 records with low a/v
sets of horizontal input accelerograms. The first set com- ratio labeled L. More details about these earthquake records
prises 44 historical records resulting from 23 events listed in can be found in Tso et al. (1992). The second set comprises
Table 5. The records are classified in three categories according three series, each including 10 synthetic timehistories com-
to the ratio (a/v) of the peak ground horizontal acceleration patible with the target uniform hazard spectra for Montral

2007 NRC Canada

Assi and McClure 1355

Table 2. Natural periods (T1, T2) of the building models. Fig. 5. Finite element mesh of the tower models: (a) tower TC1;
(b) tower TC2; (c) tower TC3.
T1, sway T2, sway T2, torsion
Building mode (s) mode (s) mode (s)
CHYBA9 0.30 0.26 0.17
CHYBA4 0.41 0.31 0.23
2020 University 2.00 1.90 1.36

Table 3. Geometric properties of the telecommunication towers.

Height Base Top Mass
Tower (m) width (m) width (m) (kg)
TC1 30 2.50 1.50 2245
TC2 20 2.50 1.50 1735
TC3 20 5.50 1.30 2920

Fig. 4. Example of tower base support.

Table 4. Natural periods of the tower models on a rigid base.

T1, sway T2, sway T3, torsional T4, sway
Tower mode (s) mode (s) mode (s) mode (s)
TC1 0.37 0.37 0.110 0.099
TC2 0.19 0.19 0.081 0.049
TC3 0.25 0.25 0.084 0.048

spectral accelerations instead of peak values, it should be

(Adams and Halchuk 2003), corresponding to probabilities
of exceedance of 2%, 10%, and 50% in 50 years. The three
series are labeled 2%, 10%, and 50% in the figures. These
timehistories were generated based on the stochastic approach
presented by Atkinson and Beresnev (1998). A total of 15
magnitudedistance (MR) scenarios were applied to cover
the frequency range of interest. Due to the randomness of
the generated records, two acceleration timehistories were
used for each MR scenario, as indicated in detail in Table 6.
The time step for each accelerogram is equal to 0.01 s.
Different sets of records were used to investigate the effects
of frequency content of ground motion on the elastic response
of the towers and their supporting buildings. It is recognized
that the intensity of an earthquake mostly affects the response
of nonlinear structures, and both the intensity and frequency
content of an earthquake affect the response of linear elastic
structures. Sewell and Cornell (1989) have discussed the
effects of ground motion characteristics on floor response
spectra in both linear and nonlinear structures, and Miranda
and Taghavi (2003) have studied the effect of frequency con-
tent on the acceleration demands in buildings.
Currently, the Canadian Standards Association standard
CSA S37-01 (CSA 2001) defines the seismicity level based
on a simple classification of the peak horizontal ground
acceleration, with three categories of high (>30% g), moder-
ate (15%30% g), and low (<15% g). Although this classifi-
cation may need revision to comply with recent changes in
the Canadian seismic hazard maps, namely to account for
kept in mind that self-supporting lattice towers founded on
ground are mostly acceleration sensitive (Khedr and McClure
1999).
Approximate response to horizontal accelerations
The proposed simplified method requires the determina-
tion of the input acceleration at the tower base rooftop
level, which is discussed in Assi (2006) and Assi et al.
(2005); the mass profile of the tower, m(x), which can be
calculated from the tower structural drawings and attach-
ments (antenna drums, transmission lines, platforms, etc.);
and a horizontal acceleration profile a(x) along the tower
height, l. The concept of the method is illustrated schemati-
cally in Fig. 6, and the main steps in the procedure are as
follows: (1) obtain the peak horizontal acceleration at the
building rooftop level; (2) obtain the tower mass distribution
along its height, including the lumped masses of localized
equipments (antennas, platforms, etc.) wherever applicable;
(3) obtain the fundamental sway mode shape and the natural
period of the self-supporting tower on a rigid base, and
consider the principal directions; (4) calculate the tower
acceleration profile using the method proposed in this paper;
(5) multiply the tower mass profile obtained in step 2 by the
horizontal acceleration profile calculated in step 4, with the
result being a lateral load profile; (6) calculate the rooftop
tower seismic base reactions using eqs. [1] and [2]; and
(7) alternatively, obtain the tower base reactions by a static
analysis of the tower on a rigid base subjected to the lateral
load profile calculated in step 5.

2007 NRC Canada

1356 Can. J. Civ. Eng. Vol. 34, 2007

Table 5. Earthquake historical records used in the study.

Earthquake and location Date Magnitude
Long Beach, California, USA 10031933 ML = 6.3
Lower California, USA 10121934 ML = 6.5
Helena, Montana, USA 31101935 ML = 6.0
Imperial Valley, California, USA 18051940 ML = 6.6
Kern County, California, USA 21071952 ML = 7.6
San Francisco, California, USA 22031957 ML = 5.3
Honshu, Japan 5041966 MJMA = 5.4
Parkfield, California, USA 27061966 ML = 5.6
Borrego Mountain, California, USA 8041968 ML = 6.5
Near east coast of Honshu, Japan 16051968 MJMA = 7.9
Lytle Creek, California, USA 12091970 ML = 5.4
San Fernando, California, USA 9021971 ML = 6.6
Central Honshu, Japan 26021971 MJMA = 5.5
Near south coast of Honshu, Japan 2081971 MJMA = 7.0
Near east coast of Honshu, Japan 11051972 MJMA = 5.8
Near east coast of Honshu, Japan 17061973 MJMA = 7.4
Near east coast of Honshu, Japan 16111974 MJMA = 6.1
Oroville, California, USA 1081975 ML = 5.7
Monte Negro, Yugoslavia 9041979 ML = 5.4
Monte Negro, Yugoslavia 15041979 ML = 7.0
Banja Luka, Yugoslavia 13081981 ML = 6.1
Michoacan, Mexico 19091985 MS = 8.1
Nahanni, N.W.T, Canada 23121985 MS = 6.9
Note: Dates are given as daymonthyear. MJMA, Japan Meteorological Agency scale; ML, local mag-
nitude; MS, surface wave magnitude.

Table 6. Characteristics of magnitudedistance (MR) scenarios considered for Montral, Quebec, Canada.
Record 1 Record 2
Magnitude, Epicentral PGA PGV PGA PGV Duration Return period
M distance (km) (g) (m/s) (g) (m/s) (s) (years)
6.0 30 0.430 0.170 0.520 0.150 8.89 2500
6.0 50 0.240 0.072 0.190 0.084 8.89 2500
7.0 50 0.510 0.190 0.630 0.290 12.39 2500
7.0 70 0.300 0.140 0.290 0.160 12.39 2500
7.0 100 0.240 0.150 0.260 0.210 20.56 2500
5.5 30 0.180 0.047 0.190 0.045 20.56 475
6.0 50 0.240 0.072 0.190 0.084 20.56 475
7.0 150 0.130 0.079 0.130 0.086 20.56 475
7.0 200 0.084 0.072 0.087 0.067 24.08 475
7.0 300 0.042 0.042 0.040 0.040 24.08 475
5.5 50 0.069 0.022 0.083 0.028 23.08 75
6.0 70 0.045 0.015 0.045 0.018 23.08 75
7.0 100 0.039 0.015 0.035 0.015 5.83 75
7.0 200 0.084 0.072 0.087 0.067 5.83 75
7.0 300 0.042 0.042 0.040 0.040 12.39 75
Note: PGA, peak ground horizontal acceleration; PGV, peak ground horizontal velocity.

The prediction of the tower acceleration profile a(x) is the base shear force, Vbase, and the base overturning moment,
key factor in this method. It was found that the acceleration Mbase, using the proposed method:
amplification profile along a telecommunication tower
mounted on a building rooftop matches reasonably well its l l
fundamental mode shape when mounted on a rigid base. [1] Vbase = Vx dx = m(x)a(x) dx
Equations [1] and [2] provide the basis for evaluating the 0 0

2007 NRC Canada

Assi and McClure 1357

Fig. 6. Concept of the proposed simplified method. Mbase, overturning moment at the base of the telecommunication tower calculated
according to the proposed simplified method; SRSS, modal combination method using the square root of sum of squares; Vbase, base
shear force at the base of the telecommunication tower calculated according to the proposed simplified method; x, position of tower
section along elevation.

where Vx is the shear force distribution along the tower; Fig. 7. Acceleration amplification profiles of tower TC1 mounted
l l on building CHYBA9, U1 direction. The prefixes H, M, and L
in Figs. 724 denote high, medium, and low a/v ratios, respec-
[2] M base = Vx x dx = m(x)a(x)x dx
tively; and 2%, 10%, and 50% denote probabilities of exceedance
0 0
of 2%, 10%, and 50% in 50 years.

Prediction of tower acceleration profiles, a(x)

In most cases studied, a strong correspondence was found
between the tower acceleration amplification profile and
its fundamental sway mode shape. The graphs shown in
Figs. 724 represent the average acceleration amplification
profiles along the height of the towers mounted on each
building. These profiles are calculated for each of the six
series of earthquake records applied separately to both prin-
cipal horizontal directions of the buildings, U1 and U2. Also
added to the graphs is the proposed acceleration amplifica-
tion profile corresponding to the fundamental sway mode
shape of each tower mounted on a rigid base, adjusted to
match the maximum acceleration amplification at the tower
top.
Figures 716 show that the proposed shape of the acceler-
ation profile is generally conservative, especially at higher
tower elevations. The only significant exceptions are shown
at low elevations in Figs. 17 and 18 for tower TC3 mounted

2007 NRC Canada

1358 Can. J. Civ. Eng. Vol. 34, 2007

Fig. 8. Acceleration amplification profiles of tower TC1 mounted Fig. 11. Acceleration amplification profiles of tower TC3
on building CHYBA9, U2 direction. mounted on building CHYBA9, U1 direction.

Fig. 9. Acceleration amplification profiles of tower TC2 mounted Fig. 12. Acceleration amplification profiles of tower TC3
on building CHYBA9, U1 direction. mounted on building CHYBA9, U2 direction.

Fig. 10. Acceleration amplification profiles of tower TC2 Fig. 13. Acceleration amplification profiles of tower TC1
mounted on building CHYBA9, U2 direction. mounted on building CHYBA4, U1 direction.

2007 NRC Canada

Assi and McClure 1359

Fig. 14. Acceleration amplification profiles of tower TC1 Fig. 17. Acceleration amplification profiles of tower TC3
mounted on building CHYBA4, U2 direction. mounted on building CHYBA4, U1 direction.

Fig. 15. Acceleration amplification profiles of tower TC2 Fig. 18. Acceleration amplification profiles of tower TC3
mounted on building CHYBA4, U1 direction. mounted on building CHYBA4, U2 direction.

Fig. 16. Acceleration amplification profiles of tower TC2 Fig. 19. Acceleration amplification profiles of tower TC1
mounted building CHYBA4, U2 direction. mounted on building 2020 University, U1 direction.

2007 NRC Canada

1360
Fig. 20. Acceleration amplification profiles of tower TC1
mounted on building 2020 University, U2 direction.
Fig. 21. Acceleration amplification profiles of tower TC2
mounted on building 2020 University, U1 direction.
Fig. 22. Acceleration amplification profiles of tower TC2
mounted on building 2020 University, U2 direction.
Can. J. Civ. Eng. Vol. 34, 2007
Fig. 23. Acceleration amplification profiles of tower TC3
mounted on building 2020 University, U1 direction.
Fig. 24. Acceleration amplification profiles of tower TC3
mounted on building 2020 University, U2 direction.
on building CHYBA4: these localized effects are largely
compensated by the overprediction of the amplifications at
higher elevations, and the global effects at the tower base
using the proposed method are conservative.
When the tower is more flexible than the building (see
Figs. 7 and 8), the former does not always experience
dynamic amplification along the complete height. This behavior
is not reflected in current code provisions, which tend to
propose a larger component force amplification factor for
flexible components without specific reference to the sup-
porting structure. The proposed method also provides con-
servative estimates of seismic forces in such a case. It is
further observed that this 30 m tower (TC1) is very flexible
and represents an upper limit case of towers mounted on
building rooftops.
The unusual behavior of tower TC1 mounted on the 2020
University building (Fig. 20) reflects the contribution of some
higher modes of the tower attributed to the influence of the
higher modes of the supporting building. This result is shown
to illustrate a case where the simplified acceleration profile
predicted by the fundamental tower sway mode is not appro-
2007 NRC Canada
Assi and McClure 1361

Table 7. Calculated tower acceleration amplification factors in Table 10. Proposed tower accel-
principal horizontal directions U1 and U2 for tower TC1 on eration amplification factors for
three buildings. towers mounted on stiff buildings
(T < 0.5 s).
U1 U2
Amplification Amplification Tp /T Amplification factor
Building Tp /T factor Tp /T factor 00.6 1.0
CHYBA9 1.43 0.40 1.24 0.80 0.91.1 4.0
CHYBA4 0.92 3.32 1.21 1.43 1.2 1.0
2020 University 0.19 1.25 0.20 3.21
Note: T, fundamental period of the building; Tp, fundamental period of
the tower mounted on a rigid base. Fig. 25. Proposed and calculated tower acceleration amplification
factors versus Tp /T for towers mounted on stiff buildings (T <
0.5 s).
Table 8. Calculated tower acceleration amplification factors in
principal horizontal directions U1 and U2 for TC2 on three
buildings.
U1 U2
Amplification Amplification
Building Tp /T factor Tp /T factor
CHYBA9 0.72 1.09 0.62 0.60
CHYBA4 0.46 1.00 0.61 0.91
2020 University 0.09 2.12 0.10 2.75

Table 9. Calculated tower acceleration amplification factors in

principal horizontal directions U1 and U2 for TC3 on three
buildings.
U1 U2
Amplification Amplification
Building Tp /T factor Tp /T factor
CHYBA9 0.97 2.70 0.84 1.31
generally conservative, in particular when the Tp /T ratio
exceeds 1.2.
CHYBA4 0.62 1.34 0.82 0.88
For a tower mounted on a flexible building (T 0.5 s) it is
2020 University 0.13 1.59 0.13 1.96
proposed to amplify the rooftop acceleration by a factor of 3
to obtain the tower acceleration profile. As discussed previ-
ously in reference to Fig. 20, a flexible tower mounted on
priate. It is recommended to perform detailed dynamic the rooftop of a flexible building would require a dynamic
analysis for similar buildingtower combinations. analysis of the buildingtower system to obtain more realistic
The adjustment factors calculated for the acceleration pro- seismic tower loads.
files of the buildingtower combinations studied are pre-
sented in Tables 79, where T and Tp are the fundamental Parametric study to verify the proposed
periods of the building and supported tower mounted on a simplified method
rigid base, respectively. These adjustment factors simply
multiply the building horizontal acceleration at rooftop level Using eqs. [1] and [2] for each buildingtower combina-
to yield the value of the acceleration at the tower top, and tion, the values of equivalent base shear forces and overturning
the shape of the acceleration profile fits the fundamental moments at the buildingtower interfaces were calculated as
sway mode calculated for the tower on a rigid base. shown schematically in Fig. 6. The shear force diagram was
Following this study, amplification factors were proposed obtained by multiplying the mass profile and the accelera-
as listed in Table 10 and illustrated in Fig. 25 to estimate the tion profile; and the base overturning moment was obtained
tower acceleration profile when mounted on a stiff building by integrating the mass profile, acceleration profile, moment
(T < 0.5 s). arm. Results were compared with the values obtained from
The limit of 0.5 s was selected to match the limit of appli- the detailed SAP2000 models using the square root of sum
cability of the simplified static force procedure for buildings of squares (SRSS) modal combination method. The average
recommended in NBCC 2005 (NRCC 2006). This limit results for all sets of records are presented in Table 1113.
relates to the assumption that the seismic response of the Detailed calculations can be found in the complete study by
building can be approximated by its fundamental sway mode Assi (2006). In these tables, Vf and Mf are the base reactions
response. Figure 25 shows that the proposed factors are calculated in the numerical simulations, and Vbase and Mbase

2007 NRC Canada

1362 Can. J. Civ. Eng. Vol. 34, 2007

Table 11. Verification of the simplified method for towers TC1, Conclusions
TC2, and TC3 mounted on building CHYBA9.
A simplified analysis method is presented for self-supporting
Mbase/Mf Vbase/Vf telecommunication towers mounted on building rooftops sub-
U1 U2 U1 U2 jected to seismic base excitation in the horizontal direction.
Tower m s m s m s m s The proposed method was verified by comparing the predic-
tions with the results of detailed numerical simulations of
TC1 1.36 0.28 1.26 0.20 1.72 0.40 1.50 0.27 nine buildingtower combinations subjected to 74 input ac-
TC2 0.98 0.08 0.98 0.08 0.97 0.07 0.99 0.08 celerograms applied to the two main building directions and
TC3 1.03 0.01 0.99 0.12 0.96 0.05 0.99 0.14 generated in the SAP2000 software. It was found that the
Note: Vbase and Mbase are calculated using the proposed method, and Vf method yields conservative results for the base shear forces
and Mf are obtained from numerical simulations. m and s, mean and stan- and overturning moments. It is suggested, however, that a
dard deviation of the calculated ratios, respectively.
detailed dynamic analysis be performed for flexible towers
mounted on high-rise buildings and for towers supporting
heavy attachments, especially in high-seismicity zones. In
Table 12. Verification of the simplified method for towers TC1, such situations, higher frequency modes of the building
TC2, and TC3 mounted on building CHYBA4. structure may excite higher modes of the tower structure.
The proposed simplified procedure will nonetheless give an
Mbase/Mf Vbase/Vf indication of the magnitude of the tower seismic base shear
U1 U2 U1 U2 and overturning moment.
Tower m s m s m s m s
TC1 0.99 0.08 1.17 0.19 1.04 0.11 1.30 0.18 Acknowledgments
TC2 0.99 0.10 0.97 0.06 1.11 0.13 1.04 0.05
TC3 1.08 0.25 1.04 0.05 1.12 0.20 0.94 0.08 Financial assistance from the Natural Sciences and Engi-
neering Research Council of Canada (NSERC) and the
Lebanese National Council for Scientific Research (LNCSR)
Table 13. Verification of the simplified method for towers TC1,
is gratefully acknowledged. We would also like to thank
TC2, and TC3 mounted on building 2020 University.
Professor George C. Yao of the National Cheng-Kung Uni-
versity in Taiwan and the Central Weather Bureau in Taiwan
Mbase/Mf Vbase/Vf for providing the instrumented building data. Financial sup-
U1 U2 U1 U2 port from the National Science Council in Taiwan (ROC) is
also acknowledged.
Tower m s m s m s m s
TC1 1.04 0.08 1.60 0.54 1.14 0.08 1.39 0.26
TC2 0.99 0.02 1.03 0.09 1.09 0.03 1.06 0.06 References
TC3 1.05 0.15 1.12 0.59 1.02 0.13 1.21 0.47 Adams, J., and Halchuk, S. 2003. Fourth generation seismic hazard
maps of Canada: values for over 650 Canadian localities
intended for the 2005 National Building Code of Canada. Geo-
logical Survey of Canada, Open File 4459.
are the base reactions calculated according to the proposed
ANSI/EIA/TIA. 2005. TIA-222-G: Structural standards for steel
simplified method. The mean and standard deviation of the
antenna towers and antenna supporting structures. Telecommu-
calculated ratios are given as m and s, respectively. It is nication Industries Association (TIA), Arlington, Va.
noted that each statistical parameter is based on 74 numeri- Assi, R. 2006. A simplified method for seismic analysis of tele-
cal results. communication towers mounted on building rooftops. Ph.D. thesis,
Department of Civil Engineering and Applied Mechanics, McGill
Discussion of the proposed simplified method University, Montral, Que.
The average values given in Tables 1113 indicate that the Assi, R., McClure, G., and Yao, G.C. 2005. Floor acceleration
proposed simplified method predicts higher values than those demands for 11 instrumented buildings in Taiwan during the
from the detailed calculation in most cases, so it is conserva- 1999 Chi Chi earthquake. In Structures 2005: Metropolis and
tive in the base forcemoment predictions. Moreover, the Beyond: Proceedings of the 2005 Structures Congress and the
2005 Forensic Engineering Symposium, New York, 2024 April
proposed method becomes more accurate as the fundamental
2005. [CD-ROM]. American Society of Civil Engineers (ASCE),
period of the tower decreases.
New York.
The small standard deviations between the predictions of Atkinson, G.M., and Beresnev, I.A. 1998. Compatible ground-motion
the different loading cases for each buildingtower combina- time histories for new national seismic hazard maps. Canadian
tion suggest that the method is suitable regardless of the Journal of Civil Engineering, 25(2): 305318.
frequency content of the input seismic excitations. It is also CSA. 2001. Antennas, towers and antenna-supporting structures.
noted that the method is more accurate for the calculation of Standard CSA S37-M01, Canadian Standards Association (CSA),
base overturning moments than for the calculation of base Toronto, Ont.
shear forces. This was also observed by McClure et al. Hiramatsu, K., Sato, Y., Akagi, H., and Tomita, S. 1989. Seismic
(2000) in relation to the predicted response of towers founded response observation of building appendage. In Proceedings of
on ground. the 9th World Conference on Earthquake Engineering, Tokyo,

2007 NRC Canada

Assi and McClure
29 August 1988. Japan Association for Earthquake Disaster
Prevention, Tokyo. Vol. 6, pp. 237242.
ICC. 2000. International building code. International Code Council
(ICC), Falls Church, Va.
Kanazawa, K., and Hirata, K. 2000. Seismic analysis for telecom-
munication towers built on the building. In Proceedings of the
12th World Conference on Earthquake Engineering, Auckland,
New Zealand, 30 January 4 February 2000. [CD-ROM]. New
Zealand Society for Earthquake Engineering, Upper Hutt, New
Zealand. Paper 0534.
Khedr, M. 1998. Seismic analysis of lattice towers. Ph.D. thesis,
Department of Civil Engineering and Applied Mechanics, McGill
University, Montral, Que.
Khedr, M., and McClure, G. 1999. Earthquake amplification factors
for self-supporting telecommunication towers. Canadian Journal
of Civil Engineering, 26(2): 208215.
Khedr, M., and McClure, G. 2000. A simplified method for seismic
analysis of lattice telecommunication towers. Canadian Journal
of Civil Engineering, 27(3): 533542.
Konno, T., and Kimura, E. 1973. Earthquake effects on steel tower
structures atop buildings. In Proceedings of the 5th World Con-
ference on Earthquake Engineering, Rome, 2529 July 1973.
Ministry of Public Works, Rome, Italy. Vol. 1, pp. 184193.
Madugula, M. (Editor). 2002. Dynamic response of lattice towers
and guyed masts. American Society of Civil Engineering,
Reston, Va.
McClure, G., Lapointe, M., and Khedr, M.A. 2000. Seismic behavior
of steel lattice telecommunication towers. In Behaviour of Steel
Structures in Seismic Areas: STESSA 2000: Proceedings of the
3rd International Conference, Montral, Que. Edited by F.M.
Mazzolani and R. Tremblay. A.A. Balkema, Rotterdam. pp. 335
337.
McClure, G., Georgi, L., and Assi, R. 2004. Seismic considerations
for telecommunication towers mounted on building rooftops. In
Proceedings of the 13th World Conference on Earthquake Engi-
neering, Vancouver, B.C., 16 August 2004. [CD-ROM]. Cana-
dian Association of Earthquake Engineering, Vancouver, B.C.
Paper 1988.
Miranda, E., and Taghavi, S. 2003. Estimation of seismic demands
on acceleration-sensitive nonstructural components in critical
facilities. In Proceedings of the ATC-29-2 Seminar on Seismic
Design, Performance, and Retrofit of Nonstructural Components
in Critical Facilities, Los Angeles, 2324 October 2003. Applied
Technology Council (ATC), Redwood City, Calif. pp. 347360.
NRCC. 2005. National Building Code of Canada, Part 4 of Divi-
sion B. Institute for Research in Construction, National Research
Council Canada (NRCC), Ottawa, Ont.
NRCC. 2006. Users guide NBCC 2005 structural commentaries
(Part 4 of Division B). Commentary J. Design for seismic
effects. Canadian Commission on Building and Fire Codes,
Institute for Research in Construction, National Research Coun-
cil Canada (NRCC), Ottawa, Ont.
1363
Pierre, J.R. 1995. Damage caused by the Hanshin-Awaji (Kobe-
Japan) earthquake to electrical and telecommunication networks
and its impact on the implementation of emergency measures.
Report RE-GEN-95-40, Hydro-Qubec, Montral, Que.
SAA. 1994. AS 3995 (1994) Design of steel lattice towers and
masts. Standards Association of Australia (SAA), Homebush,
New South Wales, Australia.
Sato, Y., Fuse, T., and Akagi, H. 1984. Building appendage seismic
design force based on observed floor response. In Proceedings
of the 8th World Conference on Earthquake Engineering,
San Francisco, Calif., 2128 July 1984. Prentice Hall Inc.,
Englewood Cliffs, N.J. pp. 11671174.
Sewell, R.T., and Cornell, C.A. 1989. Factors influencing equip-
ment response in linear and nonlinear structures. In Proceedings
of the Transactions of the 9th International Conference on Struc-
tural Mechanics in Reactor Technology. A.A. Balkema, Rotter-
dam. Vol. K2, pp. 849856.
Tso, W.K., Zhu, T.J., and Heidebrecht, A.C. 1992. Engineering
implication of ground motion A/V ratio. Soil Dynamics and
Earthquake Engineering, 11: 133144.
Wilson, E.L., and Habibullah, A. 2003. SAP2000 users manual.
Computers and Structures, Inc., Berkeley, Calif.
List of symbols
a peak ground acceleration (PGA)
a(x) acceleration profile
l tower height
m(x) tower mass per unit length at position x
M magnitude
Mbase overturning moment at the base of the telecommunica-
tion tower calculated according to the proposed simpli-
fied method
Mf overturning moment at the base of the telecommunica-
tion tower calculated in the numerical simulations
MJMA Japan Meteorological Agency scale
ML local magnitude
MS surface wave magnitude
R distance
T fundamental period of the building
Tp fundamental period of the component
v peak ground velocity (PGV)
Vbase base shear force at the base of the telecommunication
tower calculated according to the proposed simplified
method
Vf base shear force at the base of the telecommunication
tower calculated in the numerical simulations
Vx shear force distribution along the tower
x position of a section along its elevation
m mean
s standard deviation
2007 NRC Canada