You are on page 1of 6

Available online at www.sciencedirect.

com
Available
Available online
online at www.sciencedirect.com
at www.sciencedirect.com

ScienceDirect
Procedia
Procedia Engineering
Engineering 00 (2017)
199 (2017) 000–000
1713–1718
Procedia Engineering 00 (2017) 000–000 www.elsevier.com/locate/procedia
www.elsevier.com/locate/procedia

X
X International
International Conference
Conference on
on Structural
Structural Dynamics,
Dynamics, EURODYN
EURODYN 2017
2017
Comparison
Comparison of
of TMD
TMD designs
designs for
for aa footbridge
footbridge subjected
subjected to
to human-
human-
induced
induced vibrations
vibrations accounting
accounting for
for structural
structural and
and load
load uncertainties
uncertainties
a,b,∗ b b a,b
Klaus
Klaus Lievens
Lievensa,b,∗,, Geert
Geert Lombaert
Lombaertb ,, Guido
Guido De
De Roeck
Roeckb ,, Peter
Peter Van
Van den
den Broeck
Broecka,b
a KU Leuven, Department of Civil Engineering, Technology Cluster Construction, Structural Mechanics and Building Materials Section,
a KU Leuven, Department of Civil Engineering, Technology Cluster Construction, Structural Mechanics and Building Materials Section,
Technology Campus Ghent, Belgium
Technology Campus Ghent, Belgium
b KU Leuven, Department of Civil
b KU Leuven, Department of Civil Engineering, Structural Mechanics Section, Leuven, Belgium
Engineering, Structural Mechanics Section, Leuven, Belgium

Abstract
Abstract
A vibration serviceability assessment of footbridges is required in design stage to evaluate the response to human-induced exci-
A vibration serviceability assessment of footbridges is required in design stage to evaluate the response to human-induced exci-
tation. If the calculated response does not meet the desired criteria for vibration comfort, a Tuned Mass Damper (TMD) can be
tation. If the calculated response does not meet the desired criteria for vibration comfort, a Tuned Mass Damper (TMD) can be
installed as a vibration mitigation device. The TMD mass, stiffness and damping constant must be tuned to the modal parameters
installed as a vibration mitigation device. The TMD mass, stiffness and damping constant must be tuned to the modal parameters
of the structure to obtain the desired response reduction.
of the structure to obtain the desired response reduction.
The response prediction and TMD design rely on a good knowledge of the modal parameters of the footbridge and usually
The response prediction and TMD design rely on a good knowledge of the modal parameters of the footbridge and usually
assume that the response is governed by a perfect harmonic loading causing resonance. Differences between the predicted and
assume that the response is governed by a perfect harmonic loading causing resonance. Differences between the predicted and
actual modal parameters may lead to both an unreliable response prediction and a suboptimal TMD performance. Therefore,
actual modal parameters may lead to both an unreliable response prediction and a suboptimal TMD performance. Therefore,
a robust design of the TMD is proposed accounting for uncertainties in the modal parameters. Realistic walking scenarios of
a robust design of the TMD is proposed accounting for uncertainties in the modal parameters. Realistic walking scenarios of
continuous pedestrian traffic are simulated to obtain an effective response reduction by the TMD.
continuous pedestrian traffic are simulated to obtain an effective response reduction by the TMD.
In this contribution, the robust design of a TMD is demonstrated for a slender footbridge accounting for reasonable levels of
In this contribution, the robust design of a TMD is demonstrated for a slender footbridge accounting for reasonable levels of
uncertainties in the modal parameters. Numerical optimisation is therefore applied and vibration serviceability requirements are
uncertainties in the modal parameters. Numerical optimisation is therefore applied and vibration serviceability requirements are
imposed as design constraints. The obtained TMD parameters are compared to those obtained from the formulae proposed by
imposed as design constraints. The obtained TMD parameters are compared to those obtained from the formulae proposed by
Asami, determining the stiffness and damping constant as a function of its mass. It is found that the TMD mass can be further
Asami, determining the stiffness and damping constant as a function of its mass. It is found that the TMD mass can be further
reduced when the TMD parameters are tuned independently from each other compared to the classical design according to Asami
reduced when the TMD parameters are tuned independently from each other compared to the classical design according to Asami
but a higher computation cost is needed. Furthermore, an increasing degree of robustness against variations in the modal parameters
but a higher computation cost is needed. Furthermore, an increasing degree of robustness against variations in the modal parameters
is obtained by increasing the TMD mass and damping.
is obtained by increasing the TMD mass and damping.
c 2017 The Authors. Published by Elsevier Ltd.

©c 2017
2017 The
TheAuthors.
Authors.Published
PublishedbybyElsevier
ElsevierLtd.
Ltd.
Peer-review under responsibility of the organizing committee of EURODYN 2017.
Peer-review under responsibility of the organizing committee of EURODYN
EURODYN 2017.
2017.
Keywords: vibration mitigation, robust design, human-induced vibrations
Keywords: vibration mitigation, robust design, human-induced vibrations

1. Introduction
1. Introduction
For slender footbridges subjected to human-induced vibrations, a vibration serviceability assessment is required
For slender footbridges subjected to human-induced vibrations, a vibration serviceability assessment is required
to evaluate the dynamic response to human-induced loading. If the predicted vibration levels exceed a predefined
to evaluate the dynamic response to human-induced loading. If the predicted vibration levels exceed a predefined
threshold for vibration comfort, a Tuned Mass Damper (TMD) can be installed as a vibration absorber by tuning its
threshold for vibration comfort, a Tuned Mass Damper (TMD) can be installed as a vibration absorber by tuning its
mass, stiffness and damping constant to the modal parameters of the footbridge. A frequently used method therefore is
mass, stiffness and damping constant to the modal parameters of the footbridge. A frequently used method therefore is

∗ Corresponding author. Tel.: +32-9-335-25-00.


∗ Corresponding author. Tel.: +32-9-335-25-00.
E-mail address: klaus.lievens@kuleuven.be
E-mail address: klaus.lievens@kuleuven.be
1877-7058 c 2017 The Authors. Published by Elsevier Ltd.
1877-7058 c 2017 The Authors. Published by Elsevier Ltd.
Peer-review under responsibility of the organizing committee of EURODYN 2017.
Peer-review©under
1877-7058 2017responsibility
The Authors. of the organizing
Published by committee of EURODYN 2017.
Elsevier Ltd.
Peer-review under responsibility of the organizing committee of EURODYN 2017.
10.1016/j.proeng.2017.09.381
1714 Klaus Lievens et al. / Procedia Engineering 199 (2017) 1713–1718
2 Klaus Lievens / Procedia Engineering 00 (2017) 000–000

proposed by Asami [1–3] and minimises the acceleration response for lightly damped structures assuming a perfectly
harmonic loading. Analytical expressions are given to find the TMD damping ratio and frequency ratio as a function
of its mass.
The vibration serviceability assessment and the TMD parameter tuning both require a good knowledge of the modal
parameters of the structure. The modes and natural frequencies can be estimated by a finite element model (FEM) and
the modal damping is often estimated based on prior knowledge of similar structures. However, natural frequencies
calculated by a detailed FEM of a structure may be up to 10% off from the actual measured values [4]. Furthermore,
considerable shifts in the natural frequency and modal damping during the structure’s life-time due to temperature
fluctuations, long time effects on material properties, changing cases of usage are reported in [5–7]. Even relatively
small variations on the natural frequency may lead to large discrepancies in both the predicted response and TMD
parameters [8] resulting in an inefficient TMD design.
Therefore, this contribution proposes an approach for the robust TMD design for footbridges. Compared to a
conventional design in [1–3], both the effect of uncertainties on the modal parameters of the structure and random
variations on the load are accounted for. The effect of uncertain modal parameters is accounted for using a multi-
interval approach as proposed in [9]. The contribution of multiple modes in the response calculation is considered
simultaneously while multiple realistic walking scenarios are simulated accounting for variation in the walking char-
acteristics of the pedestrians and considering a realistic distribution of step frequencies. The robust assessment then
consists of an evaluation of the ranges of calculated instantaneous peak response accounting for the effect of structural
uncertainties and for variations in the loading.
The outline of the paper is as follows. First, the strategy of the robust vibration serviceability assessment and
TMD design is presented. Second, a case study of a real footbridge is introduced and the robust assessment is
performed. Then, a TMD is designed by solving an optimisation problem where constraints for vibration serviceability
are imposed according to the proposed strategy. A comparison is made between a conventional design and a global
parameter tuning. To end, a number of conclusions are made summarising the main findings in this contribution.

2. Methodology for robust response prediction and TMD design


2.1. Uncertain natural frequency and modal damping
To deal with uncertainties in the natural frequency and modal damping of the structure, a multi-interval approach
is used as proposed in [9]. Therefore, multiplication factors β f j (α) and βξ j (α) are defined as a convex fuzzy numbers
to account for the uncertain natural frequency and modal damping, respectively, of mode j, at different levels of
uncertainty α. Its value varies in an interval depending on the level of uncertainty α between its minimal value β− (α)
and maximal value β+ (α). The boundaries of these intervals increase linearly for a higher level of uncertainty. At
the level α = 0, the interval bounds of these factors correspond to an uncertainty of 10% for the natural frequency
corresponding to the findings in [4] and of 50% for the modal damping. The level α = 1 corresponds with no
uncertainty. The membership functions of both multiplication factors are visualised in figure 1.
For structures where the contribution of multiple modes is considered, it is assumed that uncertainties in the general
stiffness equally affect the natural frequencies of all modes (β f j = β f ). Since for the modal damping it is difficult to
predict the effect of uncertainties, a similar approach is assumed (βξ j = βξ ).

(a) (b)

Fig. 1. Multi-interval approach for multiplication factor at different levels of uncertainty α for (a) natural frequency and (b) modal damping.
Klaus Lievens et al. / Procedia Engineering 199 (2017) 1713–1718 1715
Klaus Lievens / Procedia Engineering 00 (2017) 000–000 3

2.2. Crowd-induced force and variations in loading

The model of Li et al. [10] is used to describe the vertical footstep force of a single pedestrian as a summation
of five harmonic components, each with an amplitude depending on the pedestrian’s step frequency f s . Then, the
step-by-step walking force is derived considering a continuous sequence of forces alternatingly induced by left and
right foot and respecting the contact time.
To model a moving crowd, a large number of single pedestrians are considered, each characterised by their tra-
jectory, weight G, step frequency f s and walking speed v s . In this model, a continuous uni-directional pedestrian
flow is assumed. The walking paths of the individual pedestrians are straight lines in the longitudinal direction of the
bridge. For each pedestrian, a weight G of 700 N is assumed. Different pedestrian densities are considered. Generally,
for higher densities, the movements of a pedestrian are influenced by the other pedestrians resulting in a decrease of
the overall walking velocity v s . An empirical relation based on observations linking the average pedestrian velocity,
pedestrian density d [pers./m2 ] and average step frequency f¯s , is given by [11,12]. The flow now is composed by sam-
pling the strictly positive step frequencies of the pedestrians from a (truncated) Gaussian distribution f s ∼ N µ fs , σ fs
with µ fs the mean value and σ fs the standard deviation of the step frequency. No intra-subject variability is considered.
In nominal conditions, the mean value of the step frequency µ fs is given by the relation in [11,12] depending on
the pedestrian density. An average value of 0.175 Hz for the standard deviation σ fs is assumed [13,14]. To account
for variation in the mean step frequency, an interval for µ fs is defined. Corresponding to the findings in [13–16], a
variation on the average step frequency of 0.1 Hz is proposed. Then, it is proposed to compose a flow for each value
µ fs in the interval [ f¯s (d) − 0.1, f¯s (d) + 0.1]. In the simulations, the arrival times of the pedestrians are sampled from
a Poisson distribution [16]. The total modal load of the flow is then found as the superposition of the modal load of
the individual pedestrians, respecting the arrival times and is used in the response calculation for a system using a
discrete-time state space model. Alternatively, Casciati et al. in [17] propose to model the load of a small group of
pedestrians as a stochastic field accounting for the unpredictability of the human behaviour.

2.3. Robust vibration serviceability assessment

To account for both varying load characteristics and uncertainties in the modal parameters of the structure in the
response evaluation, multiple calculations at each level of uncertainty α must be performed. The response is calculated
for different values of the multiplication factors β f and βξ and for each value of µ fs in its corresponding interval. Then,
for each calculation, the time history of the acceleration response is subdivided in time windows with length T and
the maximum absolute acceleration in each window is retained as instantaneous peak value. The length of the time
window T is set to twice the time a single pedestrian needs to cross the footbridge corresponding to [18,19]. The linear
combination of the mean value and 2 times the standard deviation of the instantaneous peak value (µümax (T ) + 2σümax (T ) )
is retained as the quantity to test to comfort criteria. The relevant statistic ümax (α) of the response prediction then is
expressed as:
ümax (α) = max (µümax (T ) + 2σümax (T ) ) (1)
β f (α),βξ (α),µ f s

Similarly, the minimal instantaneous peak value is extracted to obtain a range for the instantaneous peak value at level
α. For the nominal values of the modal characteristics (α = 1), only variations on the load thus are considered.

2.4. Robust TMD parameter tuning

The TMD parameters are tuned by solving an optimisation problem. The objective is to minimise the total TMD
cost taking the mass of the TMD mtmd as a measure of its cost. The TMD mass, stiffness ktmd and damping ctmd
are implemented as decision variables. A design constraint is added to ensure a definite probability of exceeding a
threshold for vibration comfort ücomfort . Multiple non-linear constraints are defined such that the criterion is satisfied
for each possible value of the mean step frequency of the flow µ fs , depending on the pedestrian density d. Similarly,
the constraint must be satisfied for each value of the multiplication factor of the uncertain frequency β f and modal
damping βξ at the predefined level of uncertainty α. The TMD parameter tuning problem at level α is formulated as
follows:
1716
4 KlausLievens
Klaus Lievens/ Procedia
et al. / Procedia Engineering
Engineering 199
00 (2017) (2017) 1713–1718
000–000

minimise mtmd
mtmd ,ktmd ,ctmd
(2)
subject to µümax (T ) + 2σümax (T ) ≤ ücomfort
∀β f (α),∀βξ (α),∀µfs

In this paper, two different TMD designs are compared. In the first method (D1), the TMD parameters are tuned
independently from each other. In a second design (D2), the relations given by Asami [1–3] are used to calculate
the TMD stiffness ktmd and damping ctmd from the TMD mass. The optimal values for the TMD frequency ratio
ρ∗tmd and damping ratio ξtmd

for minimising the acceleration response of lightly damped structures are related to the
dimensionless TMD mass ratio µ = mmtmdj and found by the serial solution proposed in [1–3] as follows.
  
∗ 1 ∗ 3µ 27
ρtmd = ; ξtmd = 1+ µ (3)
1+µ 8(1 + µ) 32
 2 √
The TMD stiffness and damping then are obtained as ktmd = 2π f j ρ∗tmd mtmd and ctmd = 2ξtmd

ktmd mtmd , respectively
with m j and f j the modal mass and nominal value of the natural frequency respectively of mode j for which the TMD
is tuned to. For a design D2, the mode for which the TMD is tuned to must be determined before the design while for
a design D1, the mode for which the TMD is tuned to results from the optimisation problem.

3. Robust TMD design for the Eeklo footbridge


3.1. The Eeklo footbridge
The steel footbridge in Eeklo (Belgium) has a U-shaped cross section and is characterised by one main span and
two side spans with a total length of 96 m. A detailed description of the bridge and the FEM is given in [20]. The
modal parameters of the first five calculated modes are summarised in figure 2. A modal damping of 0.4% is assumed
for all modes as prescribed in [19] for steel footbridges. The modal mass of mode j is calculated using the maximum
1
displacement vector sum of the mass-normalised mode shape as m j = max(|φ j |)
2 . Only the contributions of modes with

a frequency lower than 5 Hz are accounted for in the response calculation since they cover the range of the first and
second harmonic component of the walking load.

(a) f1 =1.95 Hz, (b) f2 =3.02 Hz, (c) f3 =3.71 Hz, (d) f4 =4.99 Hz, (e) f5 =5.81 Hz,
m1 =36.0 ton m2 =22.3 ton m3 =43.3 ton m4 =45.6 ton m5 =25.2 ton
Fig. 2. Top and side view of the first five calculated modes (FEM): (a) mode 1, (b) mode 2, (c) mode 3, (d) mode 4 and (e) mode 5.

3.2. Response without TMD


A robust vibration serviceability assessment of the Eeklo footbridge is performed to evaluate the dynamic behaviour
following the method from section 2. The instantaneous peak value of the response, µümax (T ) + 2σümax (T ) , is evaluated
at different levels of uncertainty α for a pedestrian density d of 0.5 pers./m2 . For that density, the average walking
speed v s is 1.30 m/s and the average mean step frequency f¯s equals 1.89 Hz following [11,12]. Continuous flows of
pedestrians now are composed for each value of the multiplication factors β f and βξ and for each value of µ fs , each in
their corresponding interval (see section 2). The predicted range of accelerations for different α-levels is given in table
2. Minimal to unacceptable levels for vibration comfort are predicted according to [19]. At the level α = 1, the range
of calculated acceleration levels considers only variations on the load. For α = 1 → 0, the range of the instantaneous
peak acceleration, considering both variations in the loading and uncertain modal parameters, increases.

3.3. TMD design and effect on response evaluation


The optimisation problem, defined in equation (2), is solved for the Eeklo footbridge with the limit for vertical
vibrations for a pedestrian density of 0.5 pers./m2 of ücomfort = 1 m/s2 corresponding to a medium level of vibration
comfort according to [19]. The TMD is located eccentrically in the section at midpoint of the footbridge.
Klaus Lievens et al. / Procedia Engineering 199 (2017) 1713–1718 1717
Klaus Lievens / Procedia Engineering 00 (2017) 000–000 5

The proposed methodology allows to investigate the sensitivity of the total cost as a function of the level of ro-
bustness (α-levels of uncertainty). When no information is available about variations or uncertainties in the modal
parameters, for instance in design stage, the largest level of uncertainty in the modal parameters can be considered
(α = 0). If measurements get information about the expected level of uncertainty or variation on the modal parameters,
the design can be performed for a higher α-level resulting in a more cost-effective design. For the Eeklo footbridge,
the problem is solved twice, once for α = 0 and once for α = 0.6. Gradient-based techniques are used to solve the
optimisation problem.
The results of the optimisation problem (TMD constants) are summarised in table 1 both for the global parameter
tuning (D1) and parameter tuning according to Asami [1–3] (D2), each for both levels of uncertainty: α = 0 and
α = 0.6. The values of the TMD constants are given in absolute as well as dimensionless terms. The TMD frequency
ftmd , TMD frequency ratio ρtmd and damping ratio ξtmd are defined as following:

1 ktmd ftmd ctmd
ftmd = ; ρtmd = ; ξtmd = (4)
2π mtmd fj 2mtmd ωtmd
with ωtmd the circular natural frequency of the TMD. j is the number of the mode to which the TMD is tuned to. A
lower TMD mass and damping is found for a design at α = 0.6, as expected for a lower level of uncertainty. From
design D1, it follows that the TMD frequency is mainly tuned to the first mode of the footbridge with an uncertain
frequency in the range between β−f (α) × 1.95 Hz and β+f (α) × 1.95 Hz. Therefore, the design D2 also tunes the TMD
parameters to the modal parameters of the first mode. The TMD mass ratio and frequency ratio in table 1 are therefore
given considering the nominal values of the modal mass and natural frequency of the first mode (fig. 2). The TMD
frequency ratio is higher than 1 which is not the case in conventional designs (D2). A strongly reduced TMD mass is
found for the designs D1 compared to D2. Although the higher computation time needed for D1 (in this case factor 6
compared to D2), the design D1 results in a strong decrease of the TMD mass and thus total cost.

Table 1. TMD optimal parameters for design at α = 0 and at α = 0.6 such that design constraints are satisfied (ücomfort = 1 m/s2 for d = 0.5
pers./m2 ) with global parameter tuning (D1) and parameter tuning according to [1–3] (D2).
mtmd ktmd ctmd ftmd µtmd ρtmd ξtmd
[kg] [N/m] [Ns/m] [Hz] [%] [-] [%]
D1 α=0 2064.0 358922.8 9188.3 2.08 5.73 1.07 16.88
D1 α = 0.6 1548.6 232823.5 6574.3 1.95 4.30 1.00 17.31
D2 α=0 3400.7 438841.3 8889.8 1.81 9.45 0.93 11.51
D2 α = 0.6 2138.3 283352.2 4527.7 1.83 5.94 0.94 9.20

The numerical values of the predicted range of response are given in table 2 for the different TMD designs as a
function of the level of uncertainty for the considered pedestrian density of 0.5 pers./m2 . A strong reduction of the
response is found compared to the results for the footbridge without TMD. For both designs (D1 and D2), at a level of
uncertainty α = 0, the maximal predicted acceleration equals 1 m/s2 as was imposed as design constraint. Similar, for
the TMDs designed at α = 0.6, the maximal predicted acceleration equals 1 m/s2 at the level α = 0.6. The assessment
considering nominal values (α = 1) results in the smallest range of accelerations where only the effect of variations
on the loading is present. The higher the level of uncertainty considered in the simulations, the wider the range of
vibration levels.
Table 2. Range of instantaneous peak acceleration at different α-levels for the Eeklo footbridge with(out) TMD (parameters given in table 1).
No TMD D1 (α = 0) D1 (α = 0.6) D2 (α = 0) D2 (α = 0.6)
α β−f -β+f [-] β−ξ -β+ξ [-] µümax (T ) + 2σümax (T ) [m/s2 ] µümax (T ) + 2σümax (T ) [m/s2 ]

1.0 [1.00-1.00] [1.00-1.00] 1.69-1.85 0.81-0.88 0.88-0.92 0.73-0.7 5 0.85-0.89


0.8 [0.98-1.02] [0.90-1.10] 1.57-2.10 0.78-0.89 0.83-0.96 0.69-0.79 0.80-0.95
0.6 [0.96-1.04] [0.80-1.20] 1.36-2.27 0.73-0.92 0.79-1.00 0.65-0.84 0.76-1.00
0.4 [0.94-1.06] [0.70-1.30] 1.25-2.57 0.71-0.94 0.74-1.04 0.61-0.88 0.71-1.05
0.2 [0.92-1.08] [0.60-1.40] 1.12-2.77 0.69-0.99 0.73-1.11 0.60-0.98 0.70-1.17
0.0 [0.90-1.10] [0.50-1.50] 1.11-3.05 0.67-1.00 0.70-1.15 0.58–1.00 0.67-1.21
1718 Klaus Lievens et al. / Procedia Engineering 199 (2017) 1713–1718
6 Klaus Lievens / Procedia Engineering 00 (2017) 000–000

4. Conclusions
For slender footbridges subjected to human-induced vibrations, the installation of a TMD is often required to satisfy
vibration serviceability conditions. A robust approach for the design of the TMD is here proposed to account for the
effect of uncertainties in the modal parameters of the structure and for variations in the loading. A multi-interval
approach is used to deal with the effect of uncertainties in the dynamic properties of the structure. The expected range
of uncertainty is determined as a function of robustness for the TMD design. By simulating realistic walking scenarios
of continuous flows of pedestrians on the structure, the response is calculated. The TMD mass, stiffness and damping
then are tuned aiming at a minimal mass by solving an optimisation problem where design constraints are imposed
for the instantaneous peak acceleration.
For a steel footbridge, a TMD is tuned for different levels of uncertainty and according to different methods. It is
found that for a decreasing range of uncertainty on the modal parameters of the footbridge, the minimally required
TMD mass also decreases without violating the design constraints in both nominal and uncertain circumstances il-
lustrating the trade-off between cost and robustness. A similar conclusion holds for the TMD stiffness and damping.
Compared to the TMD parameters as found according to conventional methods, a strong reduction of the TMD mass
is found if all TMD parameters are tuned independently from each other but a higher computation cost is needed.

Acknowledgements
The financially support of the Agency for Innovation by Science and Technology in Flanders (IWT) funding this
research is gratefully acknowledged.
References
[1] T. Asami, O. Nishihara, A. M. Baz, Analytical solutions to H∞ and H2 optimization of dynamic vibration absorbers attached to damped linear
systems, Journal of Vibration and Acoustics 124 (2) (2002) 284–295.
[2] T. Asami, O. Nishihara, Exact solution to H∞ optimization of dynamic vibration absorbers (application to different transfer functions and
damping systems, Journal of Vibration and Acoustics 125 (3) (2003) 398–405.
[3] F. Weber, G. Feltrin, O. Huth, Guidelines for structural control, Structural Engineering Research Laboratory, Swiss Federal Laboratories for
Materials Testing and Research, Dubendorf, Switzerland.
[4] K. Van Nimmen, G. Lombaert, G. De Roeck, P. Van den Broeck, Vibration serviceability of footbridges: Evaluation of the current codes of
practice, Engineering Structures 59 (2014) 448–461.
[5] W.-H. Hu, E. Caetano, A. Cunha, Structural health monitoring of a stress-ribbon footbridge, Engineering Structures 57 (2013) 578 – 593.
[6] A. Poncela, A. Lorenzana, A. V. Poncela, J. Sebastián, N. Ibán, M. V. Istrate, One year of the structural health monitoring of Pedro Gómez
Bosque footbridge One year of the structural health monitoring of Pedro Gómez Bosque footbridge (April 2015).
[7] P. Moser, B. Moaveni, Environmental effects on the identified natural frequencies of the Dowling Hall Footbridge, Mechanical Systems and
Signal Processing 25 (7) (2011) 2336 – 2357.
[8] H. Werkle, C. Butz, R. Tatar, Effectiveness of ”Detuned” TMD’s for Beam-Like Footbridges, Advances in Structural Engineering 16 (1) (2013)
21–32.
[9] K. Lievens, G. Lombaert, G. De Roeck, P. Van den Broeck, Robust design of a TMD for the vibration serviceability of a footbridge, Engineering
Structures 123 (2016) 408–418.
[10] Q. Li, J. Fan, J. Nie, Q. Li, Y. Chen, Crowd-induced random vibration of footbridge and vibration control using multiple tuned mass dampers,
Journal of Sound and Vibration 329 (19) (2010) 4068–4092.
[11] U. Weidmann, Transporttechnik der Fussgänger: Transporttechnische Eigenschaften des Fussgängerverkehrs (Literaturauswertung), ETH, IVT,
1993.
[12] L. Bruno, F. Venuti, The pedestrian speed–density relation: modelling and application, in: Proceedings of Footbridge, 2008.
[13] S. Živanović, A. Pavic, P. Reynolds, Vibration serviceability of footbridges under human-induced excitation: a literature review, Journal of
Sound and Vibration 279 (1-2) (2005) 1–74.
[14] G. Piccardo, F. Tubino, Equivalent spectral model and maximum dynamic response for the serviceability analysis of footbridges, Engineering
Structures 40 (2012) 445–456.
[15] L. Pedersen, C. Frier, Sensitivity of footbridge vibrations to stochastic walking parameters, Journal of Sound and Vibration 329 (13) (2010)
2683 – 2701. doi:http://dx.doi.org/10.1016/j.jsv.2009.12.022.
[16] S. Živanović, Benchmark footbridge for vibration serviceability assessment under the vertical component of pedestrian load, Journal of Struc-
tural Engineering 138 (10) (2012) 1193–1202.
[17] F. Casciati, S. Casciati, L. Faravelli, A contribution to the modelling of human induced excitation on pedestrian bridges, Structural Safety 66
(2017) 51–61. doi:10.1016/j.strusafe.2017.01.004.
[18] Association Française de Génie Civil, Sétra/AFGC, Sétra: Evaluation du comportement vibratoire des passerelles piétonnes sous l’action des
piétons (2006).
[19] Research Fund for Coal and Steel, HiVoSS: Design of Footbridges (2008).
[20] K. Van Nimmen, G. Lombaert, I. Jonkers, G. De Roeck, P. Van den Broeck, Characterisation of walking loads by 3D inertial motion tracking,
Journal of Sound and Vibration 333 (20) (2014) 5212–5226.

You might also like