See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/268468798

Design of a Lattice Wind Tower and a Comparison of the Structural Response

With and Without a TMD

Conference Paper · January 2013

DOI: 10.4203/ccp.102.45

CITATION READS

1 659

2 authors:

Jorge Henriques Rui Barros

University of Porto University of Porto
6 PUBLICATIONS   3 CITATIONS    183 PUBLICATIONS   322 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Latticed towers View project

Traffic vibrations View project

All content following this page was uploaded by Rui Barros on 12 January 2018.

The user has requested enhancement of the downloaded file.

 ©Civil-Comp Press, 2013
Proceedings of the Fourteenth International Conference on
Paper 45 Civil, Structural and Environmental Engineering Computing,
B.H.V. Topping and P. Iványi, (Editors),
Civil-Comp Press, Stirlingshire, Scotland

Design of a Lattice Wind Tower

and a Comparison of the Structural Response
With and Without a TMD
J.M. Henriques and R.C. Barros
Department of Civil Engineering
Faculty of Engineering of the University of Porto, Portugal

Abstract
This paper considers the issues relating to the modelling and design of a latticed
wind tower 150 meters high. Then a comparative study of the response of a latticed
wind tower is presented when subjected to natural winds with and without a TMD.
The actions considered during this type of tower design were the wind actions,
combined with the action of ice, and the seismic actions. The lattice tall tower
geometry was defined by the authors interacting with the metallic company
Metalogalva (Trofa, Portugal). The geometry was to satisfy the required
characteristics for the tower to support a wind turbine. The metallic calculations
were achieved with by iteratively using Autodesk Robot Structural Analysis
Professional, with a program to calculate the wind actions. The latter program was
developed by the authors. As expected with this type of tower, the actions that
conditioned the design were the wind action combined with ice.
A simplified method for quantifying the dynamic action on these structures is
adopted, with the purpose of studying techniques for vibration control of the along-
wind response in terms of displacements and accelerations.

Keywords: lattice, tower, wind, response, dynamic, tuned-mass damper, vibration,

control.

1 Introduction
With fossil fuels becoming increasingly scarce and increasingly expensive [1], the
world seeks a solution that serves the interest of economic development and the
preservation of nature. This concern of the World is leading to an increasing demand
for renewable energy sources such as wind energy [2]. Wind energy assumes a very
important role in the global panorama of energy, as it is a source of renewable
energy that has the least impact on nature. The rising demand of wind energy has

1
lead to a huge development of the related technologies, for example on the type of
tower to be used.
This new idea of performing latticed towers of great height, since they have less
costs in construction [3][4], poses new challenges for structural engineers with
regard to dynamic effects. Lattice towers are sensitive to the dynamic environments
generated by wind, ice, earth-quakes, impact, blast, explosions and mechanical
failures of some of their components. The vibrations excited by these environments
cover an ample spectrum of frequencies, which affect the towers in different ways,
ranging from serviceability problems, to fatigue and collapse [5].
To mitigate the dynamic effects can be installed several types of damping
devices, one such device are the TMD's. A tuned mass damper (TMD), consisting of
a mass, damping, and a spring, is an effective and reliable structural vibration
control device commonly attached to a vibrating primary system for suppressing
undesirable vibrations induced by winds and earthquake loads. The natural
frequency of the TMD is tuned in resonance with the fundamental mode of the
primary structure, so that a large amount of the structural vibrating energy is
transferred to the TMD and then dissipated by the damping as the primary structure
is subjected to external disturbances. Consequently, the safety and habitability of the
structure are greatly enhanced. The TMD system has been successfully installed in
slender skyscrapers and towers to suppress the wind-induced structural dynamic
responses [6].

2 Design
2.1 General Considerations
The geometry of the tower (Figure 1) was developed and used by the first co-author
of this paper, in his Master of Science thesis [3]. The geometric and technical
requisites have been: 150 m of height and dimensions of turbine (FL2500 of 2,5
MW and rotor diameter of 100m).

Figure 1: Tower Geometry [3]

2
 The elements of the structure were disposed based on rules of triangles so as to
shorten the buckling lengths of the structural system members. The construction of
the geometry of the beams and diagonals was based on the Eurocode EN 1993-3-1
[7] and on dispositions of existing towers.
The sections of the bars used in the tower are angles and association of angles
(Figure 2). The steel used in the design is S235 and S355.

Figure 2: Association of angles [3]

The model of the tower was introduced in Autodesk Robot Structural Analysis
Professional 2012, using model bars linked through rigid connections; the
foundations were modeled with supports that restrict all displacements and rotations.
The non-structural elements (stairs, cables) were not modeled, however were
introduced two additional nodes in the bar elements of support to non-structural
elements, so as to introduce the forces resulting from the actions on the non-
structural elements.
The modeling of the wind turbine (by itself) was not performed. However were
introduced bars with great rigidity and null weight to simulate the rigidity of the
wind turbine on the top of the latticed tower structure. The weight of the wind
turbine was considered at the top the tower by adding four vertical forces in the top
of the tower with 362.60 kN each.
For definition of wind action this tower was divided into 14 panels

2.2 Actions

2.2.1 Wind Action

Wind is an air movement initiated by the transport of air masses in the atmosphere,
and its development is related with the variations of air pressures which, in turn, are
originated thermally by means of radiation. The movement of the wind when it
encounters an obstacle causes a loading in the obstacle. It is therefore necessary, in
some structures, consideration of the wind for their analysis and design [8].
The study of the wind characteristics is in the field of meteorology. It is therefore
this science that provides to the designer the information of the wind flow
characteristics, which are necessary for the determination of the wind action [9].
At the same time, the problem of a given load due to wind action also is a subject
in aerodynamics. This discipline is used for many purposes in aeronautics and in the
car industry, and ever since was applied in structural engineering by helping the
designer in the determination of loading from wind actions [10] [11].
Joining the information supplied by the meteorology and aerodynamics, the
designer can begin the resolution of the problem.

3
 However it will be still necessary to take into account the vibrations due to wind
actions. This additional effect will be achieved through the application of laws of
aeroelasticity.
Therefore calculation of the dynamic effect of wind action into slender structures,
which includes the case of this study, is composed by three stages: description of the
wind, description of physical and aerodynamics properties and the combination of
these factors for determination of the structural response. These procedures are the
basics of Eurocodes EN 1991-1-4 [12] and EN 1993-3-1 [7], which were applied in
this study as well as in [11].
The wind force in the direction of the wind on the tower, according to the
Eurocodes, is determined by of the following expressions.
For the mean wind load:

, . (1)
1 7.

For equivalent gust wind load:

1 7. . 1
, , . 1 1 0,2. . (2)

where
is the turbulence intensity at height ze and is defined as the standard deviation of the
turbulence divided by the mean wind velocity. The Iv depends of the basic wind velocity,
terrain factor, turbulence factor, orography factor and roughness length
is peak velocity pressure at height ze and includes mean and short-term velocity
fluctuations. Depends of the air density, turbulence intensity and mean wind velocity
is the reference area of the structure or structural element.
is the structural factor
is the force coefficient for the structure or structural element.

The structural factor (cscd) determines the dynamic response of structures in the
fundamental mode of vibration, and can be divided into its components: the size
factor (cs) and the dynamic factor (cd). The structural factor combines effects of non-
simultaneous action of peak wind pressures over faces of the structure (generally
called the ‘size effect’) and vibration of the structure in its fundamental mode due to
the action of turbulence (generally called the “dynamic response”) [11].
The calculation procedure assesses the dynamic response of a structure in the
along-wind direction as the root-sum-square of a “background” and a “resonant”
component. The background component represents the quasi-steady (i.e. not
amplified) response of the structure to the atmospheric turbulence, while the
resonant part represents the dynamic oscillation of the structure at its natural
frequencies. This is usually called the Davenport method [11].
The structural factor is calculated by the following expression:

1 2. . .√
(3)
1 7.

4
where
zs is the reference height for determining the structural factor
kp is the peak factor defined as the ratio of the maximum value of the fluctuating part of the
response to its standard deviation
B2 is the background factor, allowing for the lack of full correlation of the pressure on the structure
surface
R2 is the resonance response factor, allowing for turbulence in resonance with the vibration mode.

In Table 1 are represented the required parameters to define the wind action on
the tower. The tower under study will be deployed in a zone type B, with the terrain
category 2.

Designation
Basic wind velocity 30 m/s
The reference height for determining the structural factor 87,996 m
Mean wind velocity at a height zs 42,60 m/s
Turbulence intensity at height zs 0,134
Turbulent length scale 195,72 m
Non-dimensional power spectral density , 0,115
Non-dimensional frequency , 1,103
Background factor 0,544
Resonance response factor 0,103
3,801
0,720
Aerodynamic admittance 0,228
Aerodynamic admittance 0,653
Peak factor 3,057
Up-crossing frequency 0,096
Logarithmic decrement of damping 0,827
Logarithmic decrement of aerodynamic damping 0,777
Logarithmic decrement of structural damping 0,050
Equivalent mass per unit length 12271,4 Kg/m
Force coefficient for wind action in the wind direction 3,093
Structural factor 0,856

Table 1: Parameters for the definition of wind action

Taking into account the constitution and the symmetry of the tower were
analysed two directions of wind: 0º and 45º.

2.2.2 Ice Action and Combination of Wind Action with Ice Action

The ice action and the combination of the wind action with the ice action are
determined as described in EN 1993-3-1 [7] and ISO 12494 [13], and as were used
in [10]. For exposed locations, the atmospheric ice on towers can grow considerably
in thickness; when combined with the wind action, this total action on the structure
can increase considerably conditioning the design.
The magnitude, the density, the placement and shape of the ice on the towers
depend on the local weather conditions, topography and structure shape. For
engineering design it is usually considered that all members of a tower are covered

5
with uniform thickness and uniform density of ice. To be able to express the amount
of ice that may be formed at a given site, the term Ice Class (IC) was introduced in
ISO 12494 [13]. In the design of this tower was chosen the glaze ice type, with 900
kg/m3 of density. Within the glaze ice type was chosen IG2 class, where the
thickness of ice to be used is 20 mm and the k parameter is 0.45.

The combinations of ice and wind to be taken into consideration have been:

, , (4)

, , (5)

The factors γ are given in EN 1993-3-1 [7] and additional information in EN

1993-1-1 [14].
Table 2 shows the forces of the wind action on the final structure, in the direction
at 0º.
Wind Wind with Ice
vm(ze) qp Iv Aref Fm,W FT,W Aref Fm,W FT,W
Panels ze (m) cf cf
(m/s) (ze) (ze) (m2) (kN) (kN) (m2) (kN) (kN)
1 18,60 33,74 1,553 0,169 48,28 3,355 115,26 215,36 56,99 3,257 132,02 246,68
2 30,95 36,64 1,753 0,156 28,28 3,386 80,35 143,69 33,56 3,289 92,59 165,59
3 40,98 38,24 1,868 0,149 20,74 3,370 63,87 111,72 24,87 3,263 74,18 129,76
4 51,00 39,49 1,959 0,144 26,63 3,123 81,05 139,50 32,00 2,978 92,86 159,82
5 62,20 40,62 2,044 0,140 24,60 3,160 80,16 136,03 29,57 3,020 92,07 156,25
6 71,90 41,44 2,107 0,138 20,48 3,091 67,96 114,19 24,19 2,956 76,78 129,01
7 81,00 42,12 2,159 0,135 17,69 3,048 59,81 99,71 21,03 2,903 67,71 112,86
8 89,70 42,71 2,205 0,133 15,96 2,973 54,10 89,58 18,97 2,819 60,96 100,94
9 96,23 43,11 2,236 0,132 11,89 2,849 39,33 64,83 13,99 2,692 43,73 72,09
10 105,00 43,60 2,276 0,131 17,53 2,582 53,78 88,17 20,07 2,437 58,11 95,27
11 115,93 44,17 2,321 0,129 19,22 2,510 58,81 95,83 21,91 2,366 63,23 103,02
12 126,93 44,68 2,362 0,128 18,96 2,335 55,26 89,55 21,61 2,190 59,06 95,70
13 136,13 45,08 2,395 0,126 15,73 2,125 42,47 68,53 17,92 1,989 45,27 73,06
14 146,66 45,51 2,429 0,125 16,95 1,941 42,56 68,38 19,47 1,823 45,96 73,84

Table 2: Wind forces and wind+ice combined forces

2.2.3 Other Actions

The self-weight of the superstructure was calculated from the specific weight of
steel of 78.5kN/m3. Was even considered a self-weight increase of 15%, to simulate
the weight of the connecting elements and galvanization.
For the seismic action the tower was considered installed in zone A and having a
behavior factor of 2.5, satisfying the regulatory conditions of the Eurocodes EN
1998-1-1 [15] and EN 1998-6 [16]. This action was not the controlling conditioning
action for the design of the tall latticed wind turbine tower [3].
It was still considered the wind action on the non-structural elements and on the
turbine in the most unfavourable position.

6
2.3 Steel Design
The regulatory provisions needed to design this type of structures – steel lattice
towers – are given in Eurocodes EN 1993-1-1 [14] and EN 1993-3-1 [7]. In these
references more specific aspects discussed are related to: classification of the
structure, analysis method, incorporation of imperfections and safety verification of
the structure (sections and bars).
According to EN 1993-3-1 [7], lattice towers may be analysed using the initial
geometry (first order theory). Due to their structure, the lattice towers mainly present
compressive forces on the bars. The elements used in the construction of lattice
towers are slender thus when subjected to big compressive forces, characterized by
the occurrence of large transversal deformations, are susceptible to instability. This
phenomenon is named buckling and, more simply, consists in the appearance of
secondary bending moments (and successive larger deformations) due only to axial
compressive forces.
The critical load, i.e., the load for which the structural element starts to develop
undetermined lateral deformations when only axially loaded, is defined by the
equation (6). This formulation is based on several hypothesis, which are: material
with linear elastic behavior, bars without initial geometric imperfections and
residual stresses, and load perfectly centred applied at the nodal points.

. .
(6)

Thus buckling resistance of a given element depends on the bending stiffness of

the cross section, of its length and of the supporting conditions. According to the
Eurocode EC3 (EN 1993-1-1 [14]) the design of elements subjected to simple
compression is based on the buckling curves. The use of the curves lets reproduce
the effect of imperfections of real bar members (lack-of-straightness, eccentricity of
the loads, residual stresses, etc) replacing these by a deformed equivalent initial
configuration.
In the Eurocode EN1993-1-1 [14], the resistance of the cross sections of elements
axially compressed is given by the following condition (7):

1,0 (7)
,

where Ned is the design normal force and Nc,Rd is the design resistance to normal
forces of the cross-section for uniform compression and is given by the expression
(8).

, (8)

where A is the total area of the section and fy is yield strength.

7
 The compression members shall additionally verify the condition NEd ≤ Nb,Rd
which in general is more conditioning. The design buckling resistance of a
compression member (Nb,Rd) is calculated by:

, (9)

In expression (9) χ is the reduction factor for the relevant buckling mode and is
calculated by the following expression

1
but such that 1 (10)
̅

with 0.5 1 ̅ – 0,2 ̅ where ̅ is a non-dimensional slenderness

given by the following expression

. 1
̅ (11)

where

A is an imperfection factor
Ncr is the elastic critical force for the relevant buckling mode based on the gross cross sectional
properties.
Lcr is the buckling length in the buckling plane considered
I is the radius of gyration about the relevant axis, determined using the properties of the gross
cross-section
λ1 is given by ⁄ 93,9 and 235⁄
Some existing compound members are designed according to the EN 1993-1-1
[14] as built-up compression members with: (a) chords in contact or closely spaced
and connected through packing plates; (b) star battened angle members connected by
pairs of battens in two perpendicular planes. These built-up compression members
should be checked for buckling as a single integral member, ignoring the effect of
shear stiffness (SV = ∞) provided that connections are performed along its length
with a maximum spacing of 15 imin (or 70 imin , in the case of bars connected by
pairs of battens, as shown in Figure 3).

Figure 3: Star-battened angle members

8
 In this work were defined several buckling lengths for the different bars and
different directions. Figure 3 indicates the buckling lengths used.
Buckling length
1L
0,5 L
0,33 L
0,25 L
0,20 L

Table 3: Buckling lengths

The rational for selecting and using buckling lengths is explained herein. The
elements which only are locked at the ends have a buckling length equal to the real
length of the bar. The elements which are locked by one, two, three or four lockups,
beyond the ends locked, have a buckling length of 0.5L, 0.33L, 0.25L and 0.20L,
respectively.
The metallic calculation was achieved with the Autodesk Robot Structural
Analysis Professional 2012, being adequately verified by calculating the most
significant bars.
For the design process an Excel application was developed, using Visual Basic
programming, which calculates the wind action and the ice action according to the
disposals established by the Eurocodes (EN 1991-1-4 [12] and EN 1993-3-1 [7]).
This application works jointly with Robot Structural Analysis 2012; so with the
geometric properties of the tower and with the mechanical properties of the bars
available from the Robot, the program develops the force value to be applied in each
of the structural panels.
The design process of lattice structures subjected to wind action (and of wind
action combined with ice action) is an iterative process, because for each structural
design the exposition area to wind is changed.
A preliminary design of the tower foundations was also performed, following the
Eurocode EN 1997-1-1 [17] and the DNV rules [18]. Since it is beyond the scope of
the present contribution, no more reference to such analysis and design will be
mentioned herein.
The weight results from the design process are presented in summary on Table 4.

ton
Tower (with links and galvanization) 350,422
Foundations 1028,75

Table 4: Design Summary

3 Numerical Modeling Tower and Dynamic Wind Action

This chapter begins by describing some aspects and simplifications of the
computational modeling to perform dynamic analysis. It is presented the modal

9
Analysis of the tower. The dynamic wind action was modeled to obtain the
structural response of the tower. Mathematical modeling of turbulent flow is rater
complex and the possibility of interaction between the flow and the tower may lead
to changes in dynamic pressure and in the response of the tower along time. This
chapter also addresses the simplifications used to consider this dynamic action [11].

3.1 Modeling a Lattice Tower

The model of the tower used is the model of the Design described in chapter 2.1.
In this work were assumed the some simplifications; it was considered that during
the dynamic action the rotor is stopped in its most unfavorable position and the mass
of the structure is lumped at the structure nodal points, the masses are assumed to
have only one degree of freedom (in direction X).
Autodesk Robot Structural Analysis includes damping in its dynamic time history
analysis application, through the classic formulation of Rayleigh damping with user-
defined quantities. A target damping ratio equal to ξ=5 % was used for every mode
of vibration [4] [19].
In the following modal analysis only the first three modes are discussed. The
vibration modes are shown in Figure 4 and the values for each mode are presented in
Table 5.

Mode f (Hz) Mx=450161,80 kg

1 0,47 53,57 %
2 2,25 25,44 %
3 3,95 14,62 %

Table 5: Natural frequencies and modal mass from the modal analysis

Figure 4: First, second and third modes of vibration (from left to right) for the
numerical model

10
3.2 Modeling wind dynamic action
Research has been done in the past in order to develop a spectrum that would
accurately predict the dynamic characteristic of wind. Although it is recognized the
great complexity in modeling turbulent flow around lattice towers, even with scaled
physical models in wind tunnels, some simplifications will be considered herein
with regards to the quantification of dynamic pressures and generalized forces due to
wind action along the time. For that, the fluid structure interaction FSI is considered
negligible and the correlations of the velocity fluctuations along the height of tower
are considered in a simplified manner.
Firstly it is addressed the methodology for generating time series of wind to be
used latter in the calculation of the instantaneous dynamic pressures and therefore in
the quantification of the generalized wind forces acting at every floor level of the
tower (diaphragms of tower).
The methodology used to generate synthetic time series is usually referred as the
Method of Shinozuka, which bases the generation of time series in the calculatingthe
inverse function of the Fourier Transform of the amplitude of the random process
(given by a spectral density function of the energy of a process) [11]. Such
generation of synthetic series of wind occurs in the range of wavelengths
corresponding to fluctuations of wind velocity with an approximately Gaussian
distribution of the atmospheric wind flow [20]. The purpose of the method is to
obtain a realization of a stochastic process (for example: a time series of the
fluctuations of the longitudinal component of wind velocity) from the spectral
density function that characterizes the process [11].
The method uses this function to perform a weighted sum of sinusoidal functions
(in this case of cosines). The contribution of each of the N waves is given by the
amplitude of the spectrum (SL (z, n)) (real function) for each corresponding natural
frequency (n). The phases are obtained (for the case of one-dimensional spectrum of
simple non correlated series) by pseudo-random number generation in the interval
[0, 2π].
According to the Method of Shinozuka, in the simplest case of one-dimensional
univariate stochastic processes, a realization of the random process may be obtained
[20] by equation (12). This method has been evaluated by Ianuzzi and Spinelli [21],
and it has been found to generate accurate fluctuations of longitudinal wind velocity
histories when compared to measured wind records.

√2 ∑ , Δ . cos 2 (12)

with

Δ (13)

11
 In the previous expression N is the number of frequencies of the discretization of
the spectrum, and n is frequency. To generate the synthetic time series of wind
velocity it is necessary to define a spectral density function of the fluctuations of
longitudinal velocity of the wind; the wind spectral density function (SL) given in
EC1 is used herein in the general dimensionless form of equation:
, 6,8 ,
, /
(14)
1 10,2 ,
with

, (15)

where:

L(z) is turbulent length scale represents the average gust size for natural winds.
With a reference height of zt = 200 m, a reference length scale of Lt = 300
m, and power α = 0,67+ 0,05 ln(z0), where z0 is the roughness length.
Sv(z,nk) is the one-sided variance spectrum.
fL(z,nk) is a non-dimensional frequency.
σv is standard deviation of the turbulence.

For the generation of the synthetic series to be considered an ergodic process,

according to [20] the number N of frequencies for discretization of the spectrum
should be sufficiently high. Taking into consideration the results in [11][19] a value
of N=1000 was shown a good compromise. The frequencies nmin and nmax must be
determined accordingly with the fL limits of power spectral density function of the
EN 1991-1-4 [12], that finally allow the wind turbulence effect will be clearly
characterized.
For the instantaneous wind velocity U(t) at any height given by the sum of a
constant mean component Ū with a dynamic fluctuation component u(t), the
instantaneous wind force F(t) on any surface A is given by:

1
. . . . (16)
2

1 1
. . . . . . . . . . . . (17)
2 2

The fluctuations of wind velocity along time also have a spatial variability, which
for a first approximation is herein neglected. For the case tall slender tower under
study, whereas the response is majorly due to the contribution of the first mode of
vibration (which is also a condition imposed by EC1 for the calculation of the
structural factor), modeled as a structural system with one degree of freedom, the

12
passage or conversion of the power spectrum of the wind velocity fluctuations into
structural response spectrum is given by:

4
. . . (18)

where [H(n)]2 represents the mechanical admittance and χ2 (n) represents an

aerodynamic admittance function given approximately by equation (19).

1
⁄
2. . √ (19)
1

According to Holmes [22], in a frequency domain analysis for very tall tower
structures, it is this latter function that takes into account the non-simultaneous
occurrence of the fluctuations of wind velocity. It is explicitly stated that: “For
larger structures, the velocity fluctuations do not occur simultaneously over the
windward face and their correlation over the whole area must be considered. To
allow for this effect, an aerodynamic admittance is introduced”.

According to EC1 for tall structures with the shape and conditions equivalent to
the case study under consideration, the parameters of the spectral density function
for calculating the structural factor should be determined for a reference height of
approximately 0.6 of the height of the structure. Given this indication, for generating
sets of time series, the height chosen was 90 meters that is about 60% of the height
of lattice wind tower.

The applied wind generated forces were obtained through equation (16) taking
into account the acting dynamic pressures and the influence area for each floor,
considering the mean wind velocities depending on the height (given by the
expression (20), taken form EC1) and the fluctuation velocities given by the random
series generated [11]. Supporting the procedure used in Ferreira et al. [11]
appropriate simplifications were performed so that the wind power spectrum was
multiplied by the aerodynamic admittance; it was with this new spectrum that the
turbulent velocities were calculated.

. . (20)

As an example, Figure 5 presents one series for the fluctuations of wind velocity
generated in these conditions at a height of 90 m and using wind power spectrum of
EC1. Figure 6 shows the same series, that is adopting the same phase angles for the
harmonics, but generated from the wind power spectrum multiplied by the
previously mentioned aerodynamic admittance function.

13
4 Modeling a TMD for passive control vibrations
The tuned mass dampers (TMD) can be used to control one or more vibration modes
of structures excited by external actions. However, in many cases, control of the first
mode is sufficient to reduce significantly the level of vibrations recorded. Except for
cases in which it is intended to simultaneously monitor the contribution of more than
one mode of vibration, the use of a single TMD may be satisfactory [11].
20

15

10

0
0 100 200 300 400 500 600

‐5 T (s)

‐10

‐15

‐20

Figure 5: Fluctuation velocity time series for: height of 90 m, basic velocity 30 m/s,
terrain of category I, using EC1 wind power spectrum (wind power spectrum not
multiplied by the aerodynamic admittance function)

20

15

10

0
0 100 200 300 400 500 600

‐5 T (s

‐10

‐15

‐20

Figure 6: Fluctuation velocity time series for: height of 90 m, basic velocity 30 m/s,
terrain of category I, using EC1 wind power spectrum (wind power spectrum
multiplied by the aerodynamic admittance function)

The design of a TMD for application to structures without damping is based on

two parameters – mass ratio μ and frequency ratio q – as detailed in Kelly [23]. The
optimum frequency ratio qopt (corresponding to the fixed points at the same level or
with the same displacement amplitude), the maximum amplitude of the controlled
principal system, and the inherent optimal damping ξ2,opt of the TMD, are given in
the set of equations (21).

14
 1 2 3
, , and , (21)
1 , 8 1

For the design of a TMD tuned for application to structures with damping, it is
still possible to use these equations provided the damping of the principal primary
system is less or equal to 1%. For higher damping of the primary system, the use of
such equations will lead to a non-optimized tuning of the TMD. For such cases, the
design of the TMD can be done with design graphs associated with the numerical
solution of the expression giving the maximum amplitude of the controlled principal
system [24].
Kwok and Samali [25] also studied the behavior of TMD's in tall buildings
subjected to the action of wind and, according to them, the considerations presented
about the effectiveness of a TMD in response of a system of one degree of freedom
can also be extended to other solid structures - such as in the case of tall buildings -
leading to a modal analysis. Kwok and Samali [25] indicated that occurred large
decreases in response for the modes controlled by the TMD’s installed, while the
higher order modes were not affected. For such higher modes to be also contributive
to the structural response, would require implementing new TMD’s tuned according
to their frequency. Thus, using modal analysis, for each vibration mode whose
contribution to the overall response of the structure is important, and that one wishes
control, it is necessary to determine the corresponding values of stiffness, of mass
and of modal damping [11].
Since the fundamental frequency of the tall wind tower under study is very low
(0.47 Hz) and because the wind action has a spectral density function with strong
content for low frequencies, it is possible that the response is conditioned by the
harmonic of the fundamental frequency. For control of vibrations purposes it is
assumed herein that the response is only dependent on the first vibration mode, with
which the TMD solutions were designed with the expressions available for harmonic
vibration with frequency equal to the first vibration frequency of the overall
structure.
Accordingly, the value of the modal mass corresponding to the first mode of
vibration was determined as 126,25 ton. For the case study wind tower structure
with the deployment of a TMD, only one mass ratio is considered, μ=0.01, for
which with design charts [24] it was possible to determine the optimal parameters to
be adopted for each TMD situation. In Table 6 the values adopted are systematized.

Size (cm) of
TMD μ m TMD ω TMD k TMD
qopt ξTMD,opt square section
(mass ratio) (ton) (rad/s) (kN/m)
steel bar, L=2 m
0.01 0,987 0,046 1,7626 2,915 14,974 3,89

Table 6: Optimal parameters of TMD

15
 Since the structural software used does not have an intrinsic function that allows
the direct introduction of dampers, herein for the simulation of a TMD were
determined the dimensions of a square section bar with a lateral stiffness equivalent
to that required for the damper placed on top. Acting as a vibrating bar (built in end–
free end) with a concentrated mass that would give the frequency obtained for the
sizing of the TMD with the damping introduced in the material parameters
constitutive of the bar [11].

Assuming a bar length L=2 m, made of steel with elasticity module E=210 GPa,
from the bar stiffness 3EI/L3 is obtained the equivalent inertia I of the square section
bar. Table 6 also indicates the dimensions required for such bar, for the mass ratio
considered in the design of the TMD. Table 3 shows the first four natural
frequencies of the vibration modes of the case-study wind tower structure
incorporating the TMD solution.

Mode f (Hz)
1 0,45
2 0,49
3 2,25
4 3,95

Table 7: Natural Frequencies of the first four modes with TMD

5 Results for the Tower with and without TMD

Based on the methodology adopted for consideration of the dynamic wind action
(using a set of 4 time series and for frequencies in the wind spectral density function
evaluated with 1000 frequency intervals), the results in terms of top displacements
and accelerations were evaluated and compared for the computational structural
model, without and with installed TMD vibrating bar (with an hypothetical vibrating
mass with appropriate stiffness and damping properties). Using the mentioned
structural software with modal superposition, a damping ratio of 5 % and an
integration time step of Δt = 0.2 seconds, the four series of wind dynamic loads were
applied and their average results obtained in terms of top displacements and
accelerations. As an example, Figure 7 and Figure 8 shows the time variations of
acceleration and displacement on the top of tower for the wind loads evaluated using
equation (16), with velocity fluctuations corresponding to wind series 1. Table 8
presents a summary of maximum values of displacements and accelerations, for
each of the time series.

16
 a (cm/s2)
250
200
150
100
50
0
‐50
‐100
‐150
‐200
‐250 T (s)
0,0 100,0 200,0 300,0 400,0 500,0 600,0

Figure 7: Accelerations on top of tower, for wind loads corresponding to wind

series 1
d (cm)
80

70

60

50

40

30

20

10

0 T (s)
0,0 100,0 200,0 300,0 400,0 500,0 600,0

Figure 8: Displacements on top of tower, for wind loads corresponding to wind

series 1

Series 1 2 3 4 Average
Maximum displacement (cm) 70,39 66,07 68,99 65,77 67,80
Maximum acceleration (cm/s2) 184,15  179,49  179,89  222,36  191,47

Table 8: Maximum displacements and accelerations on node 3, for each of the wind
time series (without TMD)

As regards to the use of a TMD on the top of the wind tower, as used in [11] in an
earlier comparison associated with a tall structure subjected to a harmonic excitation
in resonance with the fundamental frequency, the Figure 9 and Figure 10 shows such
comparison of top displacements and accelerations, along the time, without and with
TMD with mass ratios of 1%. As can be seen in Figure 9 and Figure 10, if the
structure is acted upon by a harmonic action in resonance with fundamental

17
frequency, the implementation of TMD can attenuate the response of structure
considerably.

600 d (cm)

400

200

T (s)
0
0 20 40 60 80 100 120 140 160 180 200
‐200

‐400

‐600
Structure Without TMD Structure With TMD (u=0.01)

Figure 9: Displacements of the top, under a harmonic fundamental resonant

excitation, without and with TMD’s

5000 a (cm/s2)
4000
3000
2000
1000
0 T (s)
‐1000 0 20 40 60 80 100 120 140 160 180 200

‐2000
‐3000
‐4000
‐5000

Structure Without TMD Structure With TMD (u=0.01)

Figure 10: Accelerations of the top, under a harmonic fundamental resonant

excitation, without and with TMD’s

Figure 11 and Figure 12 shows the time variations of top acceleration and
displacement of the tower, equipped with the TMD modeled before with a 1% mass
ratio, for the wind loads evaluated using equation (17), with velocity fluctuations
corresponding to wind series 1. Table 9 presents a summary of maximum values of
top displacements and accelerations of the tower, for each of the time series. It also
presents the average of such maximum values.

18
 a (cm/s2)
250
200
150
100
50
0
‐50
‐100
‐150
‐200
T (s)
‐250
0,0 100,0 200,0 300,0 400,0 500,0 600,0

Structure Without TMD Structure With TMD (u=0.01)

Figure 11: Acceleration on the top of tower, equipped with the TMD modeled with
mass ratio of 1%, for the wind loads corresponding to wind series 1
d (cm)
80

70

60

50

40

30

20

10
T (s)
0
0,0 100,0 200,0 300,0 400,0 500,0 600,0

Structure Without TMD Structure With TMD (u=0.01)

Figure 12: Displacement on the top of tower, equipped with the TMD modeled with
mass ratio of 1%, for the wind loads corresponding to wind series 1

Series 1 2 3 4 Average
Maximum displacement (cm) 70,30 63,42 70,62 63,96 67,07
Maximum acceleration (cm/s2) 139,35  151,72  156,57  180,69  157,08

Table 9: Maximum displacements and accelerations on node 3, for each of the wind
time series (with TMD)

The efficiency on the use of the modeled TMD in the tower structure can be
interpreted by the results of Table 10, here associated with mass ratio of 1%:
reduction of top maximum displacements and accelerations on the order of 1% and
18%, respectively.

19
 Structure Structure Reduction relative
Without With TMD to structure
TMD (u=0.01) without TMD
Maximum displacement (cm) 67,80 67,07 1%
Maximum acceleration (cm/s2) 191,47 157,08 18%

Table 10: Efficiency of using the modeled top TMD for mass ratio of 1%

6 Conclusions
For modeling the dynamic wind action reference is made to a method of generating
sets of synthetic wind – called the method of Shinozuka – and for which the number
of discretization intervals to adopt is discussed; the greater the number of intervals
to adopt, the better the process, but with divisions over 1000 intervals results are
already quite acceptable.
The simplified methodology adopted for the evaluation of the effects of the
dynamic wind action, consisted of varying forces over time at each stiffening floor,
following the same law of variation. This law is obtained, for each generated time
series, from the Eurocode 1 wind power spectrum multiplied by the aerodynamic
admittance function.
As regards to the implementation of the TMD in the tall wind tower structure
under study, it was concluded that it proved to be very effective in terms of both top
displacements and top accelerations, when the tower is subject to a harmonic action
in resonance with fundamental frequency of vibration of the tower.
The application of these devices for vibration control can therefore be very
effective for control of resonance phenomena, however, the attenuations found for
the tower equipped with device of vibration control (in this case TMD, on the top of
tower) when subjected to the natural wind action modeled as representative, depends
greatly on the wins series generated, it is not as effective as before, under perfect
harmonic action at resonance.
With the implementation of the TMD, it was concluded that this device is proving
to be more effective in terms of accelerations reductions when the structure is
subjected to the artificially generated natural wind.
For the TMD modeled with the parameters calculated, were observed maximum
accelerations reductions of the order of 18%, while the achieved reduction of
maximum displacements was only of the order of 1%.

Acknowledgements
This work was co-participated by funds from the project “VHSSPOLES-Very High
Strength Steel Poles” (Faculty of Engineering of the University of Porto, reference

20
21518) sponsored by the European Fund for Regional Development (FEDER)
through COMPETE (Operational Program Competitiveness Factors - POFC). The
Authors acknowledge the fi-nancial support and the opportunity to contribute to the
development of the transmission towers testing site of Metalogalva (Trofa,
Portugal).

References
[1] J. R. Ferreira, F. R. Martins, Ventos de mudança. A energia eólica em Portugal
(Centro de Estudos de Geografia e Planeamento Regional Faculdade de
Ciências Sociais e Humanas - Universidade Nova de Lisboa, Portugal, 2009).
[2] F.R. Martins, R.A. Guarnieri, E.B. Pereira. O aproveitamento da energia
eólica. Instituto Nacional de Pesquisas Espaciais, Brasil, 2007.
[3] J. M. Henriques, Projecto de torres eólicas reticuladas de grande altura (150m)
-Modelação e análise comparativa com estruturas tubulares auto-suportadas
(Faculdade de Engenharia da Universidade do Porto, Portugal, 2012).
[4] J. Henriques, R.C.Barros, Design of Lattice Wind Towers and Comparison
with the Typical Self-Supported Tubular Towers (To be published). 4th
International Conference on Integrity, Reliability & failure, Funchal, 23-27
June, 2013.
[5] M.K.S. Madugula (editor). Dynamic Response of Lattice Towers and Guyed
Masts. Task Committee Report, Structural Engineering Institute, American
Society of Civil Engineers (ASCE), Reston VA, 2002.
[6] C.-L. Lee, Y.-T. Chen, L.-L. Chung, Y.-P. Wang, Optimal design theories and
applications of tuned mass dampers, Engineering Structures 28 (1) (2006) 43-
53.
[7] Eurocode 3 – Design of steel structures - Part 3-1: Towers, masts and
chimneys. European Norm 1993-3-1, Brussels.2006
[8] R.C. Barros. On the Structural Design of a Wind Offshore Tower. Proceedings
of M2D’2006: 5th International Conference on Materials to Design, FEUP,
Portugal, 2006.
[9] T. Burton, D. Sharpe, N. Jenkins, E. Bossanyi. Wind Energy Handbook. John
Wiley and Sons Ltd., First edition, 2001.
[10] R.F. Almeida, R.C. Barros. Análise do efeito dinâmico do vento em torres
metálicas. V Congresso de Construção Metálica e Mista, Portugal, 2005.
[11] N.A.C. Ferreira, R.C. Barros, R. Delgado, Comparisons of a tall building wind
response with and without a TMD. 3rd ECCOMAS Thematic Conference on
Computational Methods in structural dynamics and Earthquake Engineering
(COMPDYN2011), Corfu, Greece, 25-28 May, 2011.
[12] Eurocode 1 – Actions on structures Part 1-4: General actions — Wind actions.
European Norm 1991-1-4, Brussels, 2005.
[13] Standardization I.O.f., ISO 12494, in Atmospheric icing of structures, 2001.
[14] Eurocode 3 – Design of steel structures – Part1-1: General rules and rules for
buildings. European Norm 1993-1-1, Brussels, 2005.

21
 [15] Eurocódigo 8 – Projecto de estruturas para resistência aos sismos – Parte 1:
Regras gerais, acções sismicas e regras para edifícios. Norma Europeia EN
1998-1-1, Bruxelas, 2010.
[16] Eurocode 8 – Design provisions for earthquake resistance of structures – Part
6: Towers, masts and chimneys. European Norm EN 1998-6, Brussels, 2003.
[17] Eurocódigo 7 – Projecto geotécnico – Parte 1: Regras gerais. Norma Europeia
EN 1997-1-1, Bruxelas, 2010.
[18] DNV/Risø. Guidelines for design of wind turbines. Copenhagen, 2002.
[19] J. Henriques, R.C.Barros, Comparisons of a tall latticed wind tower response
with and without TMD. 4th ECCOMAS Thematic Conference on
Computational Methods in Structural Dynamics and Earthquake Engineering
(COMPDYN2013), Kos Island, Greece, 12-14 June, 2013.
[20] Saraiva, J., and Silva, F., A Interacção do Vento com Grandes Estruturas,
Métodos Computacionais em Engenharia, 31 May- 4 June 2004, Lisbon,
Portugal.
[21] A. Iannuzzi, P. Spinelli, Artificial Generation and Structural response. Journal
of Structural Engineering, 113(12), 2382–2398, 1987.
[22] Holmes, J.D., Wind Loading of Structures, Spon Press, London; 2001.
[23] Kelly, S., Fundamentals of Mechanical Vibrations, McGraw-Hill International
Editions, Singapore, 1993.
[24] Barros, J.E., Moutinho, C., Barros, R.C., Utilização de TMDs de Grandes
Dimensões na Atenuação da Resposta Sísmica de Estruturas, Sismica 2010 -
8º Congresso Nacional de Sismologia e Engenharia Sísmica, Universidade de
Aveiro, Sociedade Portuguesa de Engenharia Sísmica, Portugal, 2010.
[25] Kwok, K., and Samali, B., Performance of tuned mass dampers under wind
loads. Engineering Structures, 1995, pp. 655-667, Elsevier Science Ltd., Great
Britain.

22

View publication stats