Accepted Manuscript

Experimental dynamic characterization of a new composite glubam-steel truss

structure

Giuseppe Quaranta, Cristoforo Demartino, Yan Xiao

PII: S2352-7102(18)31094-5
DOI: https://doi.org/10.1016/j.jobe.2019.100773
Article Number: 100773
Reference: JOBE 100773

To appear in: Journal of Building Engineering

Received Date: 17 September 2018

Revised Date: 11 April 2019
Accepted Date: 12 April 2019

Please cite this article as: G. Quaranta, C. Demartino, Y. Xiao, Experimental dynamic characterization
of a new composite glubam-steel truss structure, Journal of Building Engineering (2019), doi: https://
doi.org/10.1016/j.jobe.2019.100773.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Experimental dynamic characterization of a new composite

glubam-steel truss structure

PT
Giuseppe Quarantaa , Cristoforo Demartinob,∗, Yan Xiaoc
a
Department of Structural and Geotechnical Engineering, Sapienza University of Rome Via Eudossiana 18,
00184 Rome, Italy

RI
b
College of Civil Engineering, Nanjing Tech University, 30 South Puzhu Road, 211816 Nanjing, P.R. China
c
Zhejiang University - University of Illinois at Urbana Champaign Institute, 718 East Haizhou Road,
314400 Zhejiang, P.R. China

SC
Abstract

U
The main characteristics of an original bamboo-steel composite truss structure are presented
AN
in this work. Specifically, the considered system is a spatial truss structure whose upper
chord and diagonal bars are made by glubam elements whereas its lower chord is made by
M

steel members with a hollow cross-section. This novel structural system has been conceived
to build roofs and low/mid-span bridges (for example, footbridges), in such a way to ensure
D

easy and rapid construction, efficient use of the constituent materials, low manufacturing
TE

costs and good environmental sustainability. A prototype spatial truss beam for laboratory
tests is initially described by providing details about geometry, connections and materials
EP

properties. The results obtained from dynamic experimental tests are then discussed. In
particular, the dynamic response under ambient vibrations and the free-decay response of
this truss structure have been recorded and analyzed in order to estimate its modal proper-
C

ties. Design values of the viscous damping ratio for glubam truss structures with steel bolted
AC

connections are finally recommended. The numerical assessment of the human-induced vi-
bration serviceability conditions for footbridges built by means of this structural system is
finally performed.
Keywords: Damping; Dynamic identification; Glubam; Spatial truss structure; Vibration
performance.

∗
Corresponding author.
Preprint submitted to Journal of Building Engineering April 13, 2019
 ACCEPTED MANUSCRIPT

1. Introduction

1 Nowadays, there is an increasing interest in the use of bamboo-based products for civil

PT
2 constructions. Its most appealing features are attributable to the fact that bamboo is a
3 highly renewable construction material with low embodied energy and high strength-to-

RI
4 weight ratio. Bamboo is used in rural housing and scaffolding mainly in South Asia and
5 South America for many years. Moreover, the use of small diameter culm and/or split bam-

SC
6 boo has been proposed as an alternative to reinforcing steel in reinforced concrete [4]. The
7 structural use of this material in modern light-frame buildings is also under investigation,
see for instance Wang et al. [38]. The possibility of using bamboo as building material

U
8

9 for modern structures in Western countries has been addressed in van der Lugt et al. [23].
AN
10 Herein, the authors applied the Life Cycle Analysis to the largest bamboo-made structural
11 projects in Western Europe at that time (namely, a bamboo tower, a pedestrian bridge,
M

12 two pavilions, and an open-air theater). Through a comparative analysis based on environ-
13 mental and financial aspects, they demonstrated that bamboo can compete with building
D

14 materials more commonly used in these countries. Mahdavi et al. [24] considered the lami-
TE

15 nated bamboo lumber from different perspectives and concluded that it can be economically,
16 environmentally and, perhaps, structurally valid choice. Further useful insights about the
potential of bamboo as sustainable building material have been presented by Escamilla and
EP

17

18 Habert [16].
19 Several efforts have been spent to gain a reliable appraisal of the mechanical properties of
C

20 bamboo in view of its use as a structural material. For instance, Dixon et al. [14] investigated
AC

21 the flexural properties of some species of bamboo – namely, Moso, Guadua and Tre Gai –
22 by means of three-point bending tests. As regards the elastic moduli of these species, it
23 was found that they largely depend on the density. Specifically, the elastic modulus of Moso
24 exhibited the least scatter with respect to density. On the other hand, the elastic modulus of
25 the Guadua was found higher than that of Moso and Tre Gai for a given density. The effects
26 of two processing methods (i.e., bleaching and caramelization) on the mechanical properties
27 of engineered bamboo were investigated by Sharma et al. [33]. The flexural fatigue behavior

2
 ACCEPTED MANUSCRIPT

28 of bamboo has been studied by Song et al. [35], who also proposed a Weibull function to
29 evaluate the probability of failure of bamboo strips subjected to flexural loading. Studies by
30 Amada and Lakes [3] explored at the material level the viscoelastic properties of bamboo

PT
31 in torsion and bending using the resonance half-width method at a temperature of 22 ◦ C.
32 For dry bamboo, values of the loss factor of about 0.01 in bending and from 0.02 to 0.03 in

RI
33 torsion have been found whereas they vary from 0.012 to 0.015 in bending and from 0.03
34 to 0.04 in torsion for wet bamboo. These results are comparable with those for woods: for

SC
35 instance, spruce and beech exhibit loss factor of about 0.02 at room temperature (about 27
◦
36 C). These findings were recently confirmed by Habibi et al. [18].

U
37 Besides the researches on the mechanical properties of bamboo, several studies have been
AN
38 conducted in order to develop engineered bamboo products and more efficient production
39 processes. A new glue-laminated bamboo material (trademarked as GluBam
c
) was intro-
duced by Xiao et al. [41, 42] whereas bamboo scrimber and laminated bamboo sheets have
M
40

41 been reviewed in Sharma et al. [32]. The need of proper structural details (e.g., joints and
D

42 connections) has also originated a significant deal of studies [e.g., 5, 27, 44, 22, 15, 31].
43 Conversely, there are few studies on large-scale structural systems made of bamboo.
TE

44 In this field, Albermani et al. [2] presented a double layer grid that consists of bamboo
45 culms assembled by means of special PVC joints and also built a prototype module of this
EP

46 spatial structure, which was tested under static loads. A 10 m long roadway bridge was
47 designed by Xiao et al. [41] employing glubam girders. The bridge was tested under static
C

48 loads due to a truck with a total weight of 86 kN. Xiao et al. [43] tested a roof plane truss
AC

49 system made of glubam under static loads. Experimental tests were carried out for two
50 configurations with spans equal to 5 m and 6 m. Another prototype spatial truss structure
51 has been developed in Villegas et al. [37] using bamboo slats in place of bamboo culms and
52 special joints designed for this system. This prototype structure was tested under static
53 loads. Paraskeva et al. [26] designed a bamboo footbridge for rural areas with a span of 8 m.
54 The footbridge was realized by using bamboo culms whereas the connections were realized
55 with bolts and specially designed steel plates. It was tested under static loads reaching a
56 maximum capacity of about 2.50 kN/m2 . As regards the case of dynamic loading conditions,
3
 ACCEPTED MANUSCRIPT

57 the feasibility of bamboo culms for lattice towers intended for small wind turbines has been
58 analyzed in Adhikari et al. [1] through numerical simulations only. Recently, Wu and Xiao
59 [40] introduced a new type of hybrid truss system composed by glubam (for web and upper

PT
60 chord members) and steel tubes (for lower chord members). They investigated the static
61 performances of this structure finding a good load carrying capacity and concluding that

RI
62 they are suitable for applications in roofs and canopies.
63 Notably, none of the existing studies has addressed the experimental dynamic assessment

SC
64 of large-scale bamboo structures. This inevitably precludes a proper appraisal of structural
65 systems under dynamic loads, e.g., lattice towers under wind loads, heavy roofs under seismic

U
66 accelerations as well as footbridges under human-induced vibrations.
AN
67 In order to fill the gap in the current literature devoted to the characterization of modern
68 bamboo constructions, this study presents some experimental results intended to provide
practical guidelines for the analysis and design of glubam structures under dynamic loads.
M
69

70 Specifically, an original bamboo-steel composite structure is considered. It is a spatial truss

D

71 whose upper chord and diagonal bars are made by glubam elements whereas the lower chord
72 is made by steel members with a hollow cross-section. This structure is the same reported in
TE

73 Wu and Xiao [40] that can be considered the companion paper of the present study. While
74 Wu and Xiao [40] focus on design and experimental testing under static loads, the present
EP

75 study deals with the dynamic behavior of such structural system. Initially, the prototype
76 spatial truss beam realized for laboratory dynamic tests is described by providing details
C

77 about geometry, connections and materials properties (Section 2). The results obtained
AC

78 from dynamic experimental tests are then discussed (Section 3). The main original and
79 valuable contribution of the present work is concerned with the estimation of the viscous
80 damping ratio, for which general recommendations are provided to support the analysis
81 and design of glubam truss structures (Section 4). The assessment of the human-induced
82 vibration serviceability conditions for footbridge use are also investigated (Section 5). In
83 particular, it is addressed the case of application to footbridges of minor importance, which
84 are typically characterized by short/medium span lengths, few non-structural elements and
85 occasional passage of walkers. Finally, the conclusions give a brief summary of the main
4
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Figure 1: Tested composite glubam-steel truss structure.
M

86 findings (Section 6).

D

87 2. Composite glubam-steel spatial truss structure

TE

88 2.1. Concept

89 A composite glubam-steel truss system is considered in the present study. This structure
EP

90 is the same reported in Wu and Xiao [40], in which additional information (including details
91 about the design and data about the materials strength) is provided. This structural system
C

92 is mainly intended to build roof systems and low/mid-span bridges (especially footbridges).
Its upper chord and diagonal bars are made by glubam while steel bars are adopted at
AC

93

94 the lower chord. In fact, in serviceability conditions, the most relevant limit state for the
95 diagonal bars and the bars of the upper chord is related to the instability under compression
96 forces. By using glubam members with solid cross-sections and high inertia values, the
97 buckling load is increased. On the other hand, the most relevant limit state for the bars
98 of the lower chord is due to the tension force, and thus hollow thin-walled steel elements
99 are deemed appropriate. The modular geometry of the structure facilitates its industrial
100 production and requires minimum work at the construction site for the final assembly thanks
5
 ACCEPTED MANUSCRIPT

PT
RI
SC
Figure 2: Spatial geometry of the composite glubam-steel truss structure.

U
101 to its connections, thereby allowing the reduction of the overall cost. Apart from the use of
AN
102 bamboo, another important environmental benefit is due to the use of reversible connections,
103 which allow for separating each bar of the truss structure from the others at the end of its
M

104 life-cycle without damages, so as they can be eventually reused for another construction.

2.2. Prototype composite truss structure

D

105

106 The tested composite bamboo-steel structure is shown in Figure 1. The spatial geometry
TE

107 of this truss beam consists of 2×8 identical square pyramids (the vertex of which is on the
108 bottom chord), see Figure 2. It can be noted that non-structural elements are not considered
EP

109 in this prototype structural system. The base of each module is 1200 mm × 1200 mm whereas
110 the height is 849 mm. Therefore, the in-plane dimension of this spatial structure is 2400
C

111 mm × 9600 mm. The bars made of glubam have a square cross-section whose size is 56 mm
AC

112 × 56 mm. These bars are built by gluing 9 smaller square elements (each one composed of
113 3 or 4 thin bamboo strips) through a 3×3 arrangement. The steel bars of the lower chord
114 have a hollow circular cross-section whose external diameter and wall thickness are equal to
115 42 mm and 4 mm, respectively. The total weight of the structure is about 460 kg and the
116 average weight per unit of length and unit of area are 55 kg/m and 22 kg/m2 , respectively.
117 The first two nodes on one side of the lower chord are constrained by means of two hinges.
118 On the other side, there are two rollers. The length of this truss beam can be considered
119 representative of the typical span for small roofs or footbridges. Accordingly, it can be
6
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Figure 3: Steel connections of the truss structure: bottom nodes (left) and upper nodes (right).

considered as a full-scale prototype.

M
120

121 All the members of this space truss system are assembled by means of steel connections
D

122 (Figure 3). Two types of connections have been designed for the upper and lower chords,
123 which are adapted depending on the specific number of ways at the considered node.
TE

124 In all the connections, 4.8-grade bolts with a diameter of 10 mm have been used. A total
125 of three bolts have been adopted in the ways connecting the elements of the lower chord and
EP

126 the diagonal bars while two bolts have been used for the elements of the upper chord.
127 Finally, given the innovative nature of the truss structure, it is opportune to provide
C

128 the unitary cost of the members adopted and the cost per unit of length of the materials.
AC

129 The costs per unit of volume are estimated using reference values based on small quantity
130 production, rather than mass production, which are reasonable for China at the time when
131 this research work has been carried out. The volumetric cost for glubam is about 8000-10000
132 yuan (about $1200-1500) per cubic meter while that of steel tube is about 35000-40000 yuan
133 (about $5000-6000) per cubic meter.
134 The spatial truss system (see Figure 1) is made of 106 glubam bars corresponding to a
135 volume of about 0.4 m3 and of 22 steel bars corresponding to a volume of about 0.07 m3 . It
136 can be seen that the amount of steel in terms of volume is much less compared with the
7
 ACCEPTED MANUSCRIPT

137 glubam (see Figure 1). Neglecting the joints, the indicative cost for a single member is:

1.2 m × (0.056 m)2 × 9000 yuan/m3 ∼

=34 yuan for glubam

PT
 2  2 
(1)
 0.042  0.038
∼
1.2 m × π  m − m  × 38000 yuan/m3 =11 yuan for steel

2 2

RI
138 The total cost of the material can be calculated as 106 × 34 yuan + 22 × 11 yuan ∼
=

SC
139 3850 yuan. The final cost per unit of length of the material employed in the structure is
140 about 3850 yuan/9.6 m ∼
= 400 yuan. It should be highlighted that these costs only refer to

U
141 the material, whereas manufacturing and installation costs are neglected.
AN
142 As a final note, it might be worth to mention that the cost for glubam is based on
143 small volume trial manufacturing order and the cost may be reduced in possible future mass
144 production.
M

145 2.3. Materials properties

D

146 The density of glubam and steel adopted within this structural system is equal to 737
TE

147 kg/m3 and 7850 kg/m3 , respectively. Standard tension tests for the materials were con-
148 ducted and the strain-stress curves of the materials are illustrated in Figure 4. The average
tensile strength of steel is 513 MPa and the average yield strength is 420 MPa. The average
EP

149

150 compressive strength of the glubam is 67.72 MPa and its elastic modulus is 10.1 GPa. The
151 moisture content of the glubam specimen measured before the tests ranged from 6% to 7%.
C

152 The results show that the glubam used in this study is comparable in terms of compressive
AC

153 strength with common laminated bamboo described in the literature [e.g., 21, 12].

154 3. Dynamic identification

155 3.1. Equipments and testing protocols

156 The dynamic response of the structure has been recorded by means of a network of uni-
157 axial piezoelectric accelerometers Lance-LC0115 (Lance Technologies Inc., Qinhuangdao,
158 Hebei, P.R. China) whose properties are the following: sensitivity 5 V/g, frequency range
8
 ACCEPTED MANUSCRIPT

600 80

70
500
60

PT
400
Stress [MPa]

Stress [MPa]
50

300 40

RI
30
200
20
100

SC
10
00 0.02 0.04 0.06 0.08 00 0.02 0.04 0.06 0.08
Strain Strain

U
Figure 4: Strain-stress curves of the materials: steel (left) and glubam (right).
AN
159 0.1 Hz – 1500Hz, resolution 0.000004 g, full-scale range ± 1 g. The data acquisition system
was the National Instruments NI PXI-1042Q equipped with the software NI Signal Express
M
160

161 2014. The adopted sampling rate was 1 kHz. The acceleration response of the truss structure
has been recorded under two dynamic loading conditions.
D

162

163 The first dynamic loading scenario consists of ambient vibrations basically attributable
TE

164 to the movements of small vehicles inside the laboratory and to the passage of heavy vehicles
165 on the roads near to it. In this case, the length of the time recordings was equal to 20 min.
EP

166 All the free joints were equipped with an accelerometer with the exception of the three nodes
167 at the beginning and the end of the upper chord. The vertical component and the horizontal
C

168 component orthogonal to the longitudinal axis of the truss structure have been recorded by
means of several layouts in which some sensors were moved while six accelerometers were
AC

169

170 left in their original positions to keep track of the phase.

171 The second dynamic loading scenario is the free-decay response of the truss structure,
172 which has been induced by removing suddenly a mass of 35 kg originally suspended at the
173 midspan of the bottom chord. The free-decay response has been recorded over a time-window
174 whose length was equal to 15 s. In this case, a single layout consisting of eight measurement
175 points on the upper chord was employed to record the vertical response. Specifically, three
176 accelerometers were regularly spaced on both sides of the upper chord whereas two other
9
 ACCEPTED MANUSCRIPT

Point A Point B Point C

0.02 0.02 0.02

PT
Acceleration [g]

0.01 0.01 0.01

0 0 0

−0.01 −0.01 −0.01

RI
−0.02 −0.02 −0.02

−0.03 −0.03 −0.03

SC
0 500 1000 0 500 1000 0 500 1000

Time [s]
Point D Point E

U
0.02 0.02
C
Acceleration [g]

0.01 0.01 B
AN
A
0 0 E
D
−0.01 −0.01
M

−0.02 −0.02

−0.03 −0.03
0 500 1000 0 500 1000
D

Time [s] Time [s]

Figure 5: Some samples of the acceleration response recorded under ambient vibrations (circular markers:
TE

vertical direction, triangular markers: horizontal direction perpendicular to the longitudinal axis).

accelerometers were installed along the longitudinal axis.

EP

177

178 3.2. Operational modal analysis using ambient vibrations

C

179 The response of the truss structure under ambient vibrations (see some samples in Fig-
AC

180 ure 5) has been elaborated in frequency- and time-domain in order to identify its modal
181 parameters. Enhanced Frequency Domain Decomposition (EFDD) and Stochastic Sub-
182 space Identification (SSI) have been adopted for output-only modal parameter estimation.
183 Theoretical details about these techniques can be found elsewhere [e.g., 25, 8, 7, 36].
184 In the first step, the vertical accelerations only have been considered in the dynamic
185 identification process while the horizontal components have been neglected. A total of five
186 modes of vibration have been identified. The corresponding natural frequencies and damping
187 ratios are listed in Table 1 whereas the operational modal shapes are shown from Figure 6
10
 ACCEPTED MANUSCRIPT

Table 1: Natural frequencies and damping ratios identified from the vertical response recorded under ambient
vibrations by means of EFDD and SSI (in the latter case, the standard deviation value is reported within
the brackets).

PT
EFDD SSI
Mode
Frequency [Hz] Damping [%] Frequency [Hz] Damping [%]
1 21.8 1.532 21.74 (0.2099) 1.686 (0.4141)

RI
2 33.58 0.8126 33.54 (0.0971) 0.7068 (0.315)
3 42.18 0.7421 41.56 (0.2097) 0.8128 (0.4714)
4 56.66 1.415 56.5 (0.2783) 1.4 (0.2632)

SC
5 73.25 0.6987 N/A N/A

U
to Figure 10 (herein, the nodes that were not equipped with a sensor are not shown).
AN
M
D
TE

Figure 6: First mode of vibration identified from the vertical response recorded under ambient vibrations
EP

by means of EFDD (natural frequency 21.8 Hz, damping ratio 1.532%).

188
C

189 Overall, there is an excellent agreement between the results obtained from EFDD and
AC

190 those carried out by means of the SSI technique (see Table 1). The fundamental modal
191 shape is a bending-type mode of vibration (see Figure 6). The second and third mode of
192 vibration are the first and the second torsional modal shapes, respectively (see Figure 7 and
193 Figure 8). Finally, the fourth and fifth modal shapes are the second and the third bending-
194 type mode of vibration. Deviations in the operational modal shapes from the ideal ones are
195 basically attributable to uncertainties in boundary conditions and structural details (such as
196 the preload conditions of the bolted joints, which were not monitored during the assembly
197 of the structure). The deviation is evidenced by the non-perfect symmetry in symmetric
11
 ACCEPTED MANUSCRIPT

PT
RI
SC
Figure 7: Second mode of vibration identified from the vertical response recorded under ambient vibrations

U
by means of EFDD (natural frequency 33.58 Hz, damping ratio 0.8126%).
AN
Table 2: Natural frequencies and damping ratios identified from ambient vibrations by means of EFDD
using the vertical response only or, both, horizontal and vertical responses.
M

Vertical response Vertical and horizontal response

Mode
Frequency [Hz] Damping [%] Frequency [Hz] Damping [%]
D

1 21.8 1.532 21.11 1.538

2 33.58 0.8126 33.97 0.7332
3 42.18 0.7421 41.92 0.8441
TE

4 56.66 1.415 56.49 1.739

5 73.25 0.6987 73.37 0.572
EP

198 mode shapes and non-perfect asymmetry in asymmetric mode shapes.

C

199 Including the horizontal response orthogonal to the longitudinal axis does not affect the
estimates of natural frequencies and damping ratio, as it can be inferred from Table 2. The
AC

200

201 identified modal shapes do not change when considering the horizontal response, with the
202 only exception of the second mode of vibration. In this case, the identification based on
203 both components of the dynamic response has revealed that the first torsional modal shape
204 occurs with a significant lateral swinging (see Figure 11). This highlights the existence of
205 a strong coupling between the first torsional mode and the first bending-type mode of the
206 truss structure in the horizontal plane.

12
 ACCEPTED MANUSCRIPT

PT
RI
U SC
Figure 8: Third mode of vibration identified from the vertical response recorded under ambient vibrations
AN
by means of EFDD (natural frequency 42.18 Hz, damping ratio 0.7421%).

3.3. Identification based on the free-decay response

M
207

208 The recorded free-decay vertical response of the truss structure has been analyzed us-
D

209 ing standard spectral analysis to estimate the natural frequencies whereas the logarithmic
210 decrement technique was employed in order to calculate the damping ratios. A band-pass
TE

211 filtering technique based on the Butterworth filter was used to isolate the mode of vibration
212 detected in the spectral analysis whereas the corresponding damping ratio was evaluated
EP

213 from the logarithm of the instantaneous amplitude obtained through the Hilbert transform.
214 The interested reader can find the theoretical basis of the logarithmic decrement technique
C

215 elsewhere [e.g., 34].

AC

216 The analysis of the free-decay response has allowed the identification of natural fre-
217 quencies and damping ratios for the first and second modes of vibration. The damping
218 identification for the first torsional mode of vibration from a lateral measurement point at
219 the midspan of the truss structure is shown in Figure 12.
220 Overall, the elaboration of the available recordings has provided the results reported in
221 Table 3 (as regards the outcomes related to the free-decay response, the average values of
222 the results obtained from all the analyzed time-histories are listed). It is possible to observe
223 that they are in very good agreement with the outputs of the operational modal analysis.
13
 ACCEPTED MANUSCRIPT

PT
RI
SC
Figure 9: Fourth mode of vibration identified from the vertical response recorded under ambient vibrations
by means of EFDD (natural frequency 56.66 Hz, damping ratio 1.415%).

U
AN
Table 3: Comparison of natural frequencies and damping ratios of the first two modes of vibration identified
from the vertical response recorded under ambient vibrations (by means of EFDD) and using the free-decay
response.
M

Ambient vibrations Free-decay response

Mode
Frequency [Hz] Damping [%] Frequency [Hz] Damping [%]
D

1 21.8 1.532 20.48 1.454

2 33.58 0.8126 33.57 0.744
TE

224 4. Recommended viscous damping ratio

EP

225 Since the damping ratio plays a fundamental role in the dynamic behavior of structures
226 such as roof systems and footbridges, it is very important to provide reliable recommenda-
C

227 tions in this regard to support the analysis and design stages. In this perspective, damping
AC

228 ratios (in percentage) suggested by some standards and guidelines for footbridges made of
229 steel and timber (the two materials relevant for the present study) are summarized in Table
230 4. This latter is adapted from Demartino et al. [13]. The values refer to the fundamental
231 mode of vibration. It can be observed from Table 4 that lower damping values apply to steel
232 bridges whereas larger values occur in timber bridges. In general, it is known that damping
233 may be divided into three classes [39], namely: internal friction throughout the material
234 making up the structure (material damping), energy dissipation associated with junctions

14
 ACCEPTED MANUSCRIPT

PT
RI
SC
Figure 10: Fifth mode of vibration identified from the vertical response recorded under ambient vibrations

U
by means of EFDD (natural frequency 73.25 Hz, damping ratio 0.6987%).
AN
235 or interfaces between parts of the structure (structural damping), and energy dissipation
M

236 associated with a fluid in contact with the structure (fluid damping). The additional effect
237 of the structural damping is considered by Eurocodes. In particular, for steel structures,
D

238 it is suggested 0.2% for welded connections while 0.4% for bolted connections (which are
more dissipative). On the other hand, for timber structures, it is suggested 1% for welded
TE

239

240 connections (i.e., welded steel socket or brackets) while 1.5% for bolted connections.
241 The current study found that the damping ratio at the fundamental frequency of the
EP

242 considered composite glubam-steel truss structure is around 1.5%. This result is compatible
243 with that of timber footbridges (Table 4). In particular, the measured damping ratio for
C

244 the fundamental frequency almost coincides with that suggested for timber structures by
AC

245 Eurocodes if mechanical joints are present (Table 4). Moreover, this value is compatible with
246 Hivoss [19] and FIB [17]. Larger values are suggested by Sétra [30], with a minimum value of
247 1.5% comparable with the measured one and a maximum value twice the minimum one (i.e.,
248 3.0%). According to the experimental data as well as to the values proposed in Standards
249 and Code of Practice, a viscous damping ratio equal to 1.5% can be recommended for the
250 fundamental (bending-type) mode of glubam truss structures with bolted connections.
251 It is important to remark that damping ratios identified in this study refer to a system

15
 ACCEPTED MANUSCRIPT

PT
RI
SC
Figure 11: Effects of the horizontal motion on the identification of the second mode of vibration from ambient

U
vibrations by means of EFDD: identification using the vertical response only (left) and identification using
both vertical and horizontal response (right).
AN
252 without non-structural elements, such as roof covering (for roof systems) or deck surface (for
M

253 footbridges). However, it has been recognized that non-structural elements can have a signif-
254 icant influence on the dissipative properties, i.e., non-structural elements typically increase
D

255 the damping ratios. Accordingly, in real operative conditions (i.e., with non-structural ele-
TE

256 ments in place), the damping ratios are likely to be higher with respect to those measured
257 in this study which, as a consequence, can be considered as conservative estimates.
EP

258 5. Assessment of the serviceability conditions for glubam-steel footbridges

C

259 The proposed new composite glubam-steel truss structure can be used for footbridges
of minor importance, which are typically characterized by short/medium span lengths, few
AC

260

261 non-structural elements and occasional passage of walkers.

262 To assess the performance for footbridge use of the proposed structure, the serviceability
263 conditions to a single walker crossing are studied by using the deterministic framework
264 proposed by Demartino et al. [13]. The latter procedure was chosen because suitable for
265 footbridges of minor importance characterized by the occasional passage of walkers that can
266 be modeled as a single walker crossing condition. In particular, to evaluate the human-
267 induced vibrations [29], it is necessary to define the characteristics of the walker (i.e., modal
16
 ACCEPTED MANUSCRIPT

1 0.04
a) b)

PT
0.5 0.02
Acceleration [g]

Acceleration [g]
0 0

RI
−0.5 −0.02

−1 −0.04

SC
0 1 2 3 4 5 0 1 2 3 4 5

Time [s] Time [s]

1 5
Instantaneous amplitude (log)

U
c) d)
Normalized amplitude

0.8
0
AN
0.6
−5
0.4
M
−10
0.2

0 −15
0 50 100 0 1 2 3 4 5
D

Frequency [Hz] Time [s]

TE

Figure 12: Damping identification for the second mode from the free-decay vertical response recorded at
point B (see Figure 5): a) acceleration time-history, b) acceleration time-history after band-pass filtering
with respect to the second mode, c) normalized frequency spectrum of the recorded acceleration response
after band-pass filtering with respect to the second mode, d) log-scale instantaneous amplitude and best-fit
EP

line for damping estimation.

Table 4: Damping ratios (in percentage) for footbridges realized using steel and timber, as given by different
C

standards and guidelines.

AC

Sétra [30] Hivoss1 ISO [20] Eurocodes2 FIB [17]

Type
Min Mean Min Mean Mean Mean Min Mean
3
Steel 0.2 0.4 0.2 0.4 0.5 0.2/0.4 0.5 1.0
Timber 1.5 3.0 1.0 1.5 - 1/1.54 0.8 1.4
1
See [19].
2
EC1 [9], EC3 [10], EC5 [11].
3
0.2% if welded connections are present, 0.4% for bolted connections.
4
1% if no mechanical joints are present, 1.5% otherwise.

17
 ACCEPTED MANUSCRIPT

PT
Figure 13: Dynamic model of the simply supported beam in the vertical direction: first mode and walker-

RI
induced loads.(x: beam axis with zero coordinate in one support; t: time; v: constant speed of the walker;
W : body weight; DLF : dynamic load factor; fw : walking frequency; δ: Dirac function; L: length of the
beam; u(x, t): displacement of the beam at the point x and time t; η(t): modal displacement at the time t).

SC
268 force) and the dynamic properties of the footbridge (i.e., mechanical model).
269 The modal force is calculated using the characteristics of the standard walker described

U
270 in Demartino et al. [13]. The mechanical model is a simply supported beam loaded by a
AN
271 constant-speed moving harmonic load for which only the first vertical mode is considered
272 (Figure 13). This is in agreement with the modal shape observed (see Section 3.2). The peak
M

273 of the modal response is expressed in terms of a transient frequency response function, ϕ,
274 that is the ratio between the modal peak non-stationary response induced by a given walker
D

275 crossing the bridge and the corresponding stationary response induced by the standard
walker. Being sin(π/2) = 1, the peak acceleration in the midspan (i.e., for x = L/2) and
TE

276

277 the peak modal acceleration are the same (see Figure 13):
EP

ˆ
ü(L/2, ˆ · sin(π/2) = η̈(t)
t) = η̈(t) ˆ ·1 (2)

The peak modal acceleration can be expressed in terms of ϕ as:

C

278
AC

DLF · W
ˆ
ü(L/2, t) = ·ϕ (3)
2ξ · m
279 where DLF = 0.35 is the dynamic load factor that is the harmonic load amplitude normal-
280 ized by the body weight W = 744 N, ξ is the damping ratio and m is the modal mass (half
281 of the total mass for a simply supported beam).
282 Using the assumptions reported above, it can be used the closed-form solution of ϕ
283 provided in Ricciardelli and Briatico [28]. Generally speaking, ϕ is a function of: (i ) α =

18
 ACCEPTED MANUSCRIPT

284 fw /f : the frequency ratio that is the ratio between the walking frequency fw = 1.898 Hz
285 [13], to the fundamental frequency of the footbridge, f ; (ii) L: the span of the footbridge
286 (see Figure 13); (iii ) ξ.

PT
287 ϕ is reported in Figure 14 (a) in the range of span length from L = 5 − 25 m and
288 α = 0.01 − 1.2 for a damping ratio of ξ = 1.5%. The damping ratio is assumed as that

RI
289 identified for the first vertical mode of this structure (see Table 3). It is noteworthy that the
290 viscous damping ratio estimates obtained in the present study (see Section 4) are expected

SC
291 to be very close to real conditions for such class of footbridges, because of the negligible
292 influence of the number pedestrians (due to the single walker crossing conditions) and small

U
293 impact of non-structural elements [e.g., 6].
AN
294 The asterisk in Figure 14 indicates the characteristics of the structure of this study,
295 i.e., L = 8.4 m and α = 1.898 Hz/21.8 Hz=0.0871. It can be observed that the expected
acceleration induced by the crossing of a pedestrian are quite low for this structure (low
M
296

297 values of ϕ). The proposed structure is relatively short (L = 8.4 m, see Figure 1) and
D

298 capable of withstanding high-loads as demonstrated in Wu and Xiao [40]. Consequently, it

299 is possible to adopt the same modular system for larger spans expecting lower frequencies
TE

300 corresponding to values of α closer to the unity (i.e., near to the resonance conditions) thus
301 larger accelerations. The frequency of this modular system as a function of the span can be
EP

302 predicted by calculating the stiffness of an equivalent continuous simply supported beam as:

1r
C

a 9.872 E · I
f= with a = (4)
2π m(L) · L4 2
AC

303 where m(L) is the modal mass per unit length that can evaluated as half (i.e., valid for
304 simply supported beam) of the mass of the proposed structure and considering the mass
305 proportional with the length (i.e., constant mass):

460 kg
m(L) = L = 27.4 kg/m · L (5)
2 · 8.4 m
306 The assumption of constant mass and stiffness (i.e., constantly distributed) is reasonable

19
 ACCEPTED MANUSCRIPT

PT
RI
ϕ ϕmax
1.2 0.1 1 1.2
0.2 0.03

SC
1 0.9 0.9 0.8 1
0.2
0.1 0.025
0.8 0.05 0.8
0.6 0.02
0.6 0.6
α

α
0.4 0.015

U
0.4 0.4
0.2 0.01
0.2 0.2
AN 0.005
0
10 20 10 20
L L
(a) (b)
M

ϕmax ϕ sign(ϕmax ϕ)
1.2 1.2
0
1 1
D

-0.2 P
0.8 0.8
-0.4
TE

0.6 0.6
α

0.4
0 -0.6 0.4
0
N
0.2 -0.8 0.2
EP

-1
10 20 10 20
L L
(c) (d)
C

Figure 14: Contour plots in the L-α plane of: ϕ, (a), maximum tolerable transient frequency response
AC

function, ϕmax , (b), and difference between the demand and capacity, ϕmax −ϕ, (c), and its sign (P: positive
- verified; N: negative - not verified) (d). The black asterisk indicates the characteristics of the structure
of this study. The magenta line indicates the frequency variation of an equivalent modular structure as a
function of L.

20
 ACCEPTED MANUSCRIPT

307 given that the same modular system can be employed for larger spans because of the good
308 static performances [40].
309 Using Eqs. (4) and (5) and the characteristics of the structure, a can be calculated as:

PT
a = f 2 · m(L) · L3 · (2π)2 = (21.8Hz)2 · 27.4 kg/m · (8.4m)3 · (2π)2 = 2.56e + 09 Hz2 · kg · m3 (6)

RI
310 However, the identified frequency is very high (corresponding to low values of the fre-

SC
311 quency ratio) because the mass is also very low due to the absence of the non-structural
312 elements such as floor and handrails. It is expected that the presence of such elements will

U
313 slightly decrease the frequency (large values of the frequency ratio) leading to large acceler-
ations. With the aim to provide a more realistic estimation of the serviceability conditions,
314
AN
315 the mass of the entire footbridge (including the floor and the handrails) is calculated using
316 reasonable values of the weight of non-structural elements. In particular, it is assumed a
M

317 floor made of common floor gratings for pedestrians areas (25 kg/m2 ) and common steel
318 handrails (15 kg/m2 ). The modal mass per unit length is evaluated as:
D

460 kg + (25 kg/m2 × 8.4 m × 2.4 m) + (15 kg/m × 2 × 8.4 m)

TE

m(L) = L = 72.4 kg/m · L

2 · 8.4 m
(7)
EP

319 The frequency of a modular structure with the same characteristics of the structure
320 investigated in this study but with the addition of the mass of the non-structural elements
is predicted combining Eqs. (4), (7) and Eq. (6). The predicted frequency expressed in
C

321

322 terms of α is reported in Figure 14 with a thick magenta line. As expected, increasing
AC

323 the span length, the frequency decreases and as a consequence α increases. Moreover, the
324 upward-pointing triangle in Figure 14 indicates the characteristics of the structure of this
325 study but with the addition of the mass of the non-structural elements (as in Eq. 7), i.e.,
326 L = 8.4 m and α = 1.898 Hz/12.8 Hz=0.1477.
327 A capacity model is needed to assess the serviceability conditions for footbridge use.
328 According to Demartino et al. [13], the maximum tolerable TFRF in the vertical direction

21
 ACCEPTED MANUSCRIPT

329 is evaluated from the ISO 10137 [20] base curves:


2 · ξ · m 0.21 if α ≤ 0.47


PT
ϕmax (α) = (8)
DLF · W  0.140 + 0.150 · α if α > 0.47

RI
330 ϕmax is reported in Figure 14 (b) where the modal mass, m, is taken as in Eq. 7.
331 The serviceability performance of the structure can be assessed by comparing the de-

SC
332 terministic value of the demand, ϕ, with the deterministic value of the capacity ϕmax (α).
333 Positive values of ϕmax − ϕ indicate verified conditions. ϕmax − ϕ is reported in Figure 14

U
334 (c) while its sign in Figure 14 (d).
It can be seen that for low values of L (corresponding to high frequencies and low
335
AN
336 values of α) the serviceability conditions are always verified (green areas in Figure 14 (d)).
337 In particular, it can be observed that a footbridge realized with this modular structure
M

338 (magenta line in Figure 14 (d)) is always falling in the verified for L ≤ 23 m proving its good
339 vibration serviceability performance for footbridge use. It should be highlighted that for
D

340 L ≥ 23 m (almost three times the span of the studied structure, see Figure 2) the modular
TE

341 structure should be re-designed to fulfill serviceability checks, for instance by increasing the
342 stiffness.
EP

343 6. Conclusions
C

344 This study investigated the dynamic characteristics of a new composite glubam-steel
AC

345 truss structure in which the elements of upper chords and diagonal bars are made of glued
346 laminated bamboo (glubam) while the bars of the lower chord are made of steel bars with
347 hollow cross-sections. Such a system was conceived to facilitate its industrial production
348 while reducing the overall cost and ensuring high environmental sustainability through ef-
349 ficient use of the constituent materials and structural details suitable to allow the reuse of
350 each element. Laboratory tests were performed on a prototype structural system in order
351 to estimate its dynamic properties.
352 After a critical review of the experimental evidence, a conservative viscous damping
22
 ACCEPTED MANUSCRIPT

353 ratio around 1.5% for the fundamental (bending-type) mode is suggested in glubam truss
354 structures with steel bolted connections whereas conservative values between 0.5% and 1.5%
355 (mean value equal to 1%) are recommended for all the modes. Finally, the human-induced

PT
356 vibration serviceability conditions for footbridge use of the proposed structure were assessed.
357 The numerical analyses demonstrated a good dynamic behavior of glubam footbridges of

RI
358 minor importance, thereby supporting the feasibility of this new structural typology in real
359 applications.

SC
360 Acknowledgments

U
361 The research conducted in this paper was supported by National Natural Science Foun-
AN
362 dation of China (NSFC) through a National General Project (No. 51678296), and the 2016
363 Jiangsu provincial Double-innovation plan. The authors would like to thank Mr. Yue Wu
M

364 for having conducted the laboratory tests.

D

365 References
TE

366 [1] Adhikari R, Wood D, Sudak L (2015) Low-cost bamboo lattice towers for small wind turbines. Energy
367 for Sustainable Development 28:21 – 28
368 [2] Albermani F, Goh G, Chan S (2007) Lightweight bamboo double layer grid system. Engineering Struc-
EP

369 tures 29(7):1499 – 1506

370 [3] Amada S, Lakes RS (1997) Viscoelastic properties of bamboo. Journal of Materials Science 32(10):2693–
371 2697
C

372 [4] Archila H, Kaminski S, Trujillo D, Escamilla EZ, Harries KA (2018) Bamboo reinforced concrete: a
AC

373 critical review. Materials and Structures 51(4):102

374 [5] Aschheim M, Gil-Martón L, Hernández-Montes E (2010) Engineered bamboo i-joists. Journal of Struc-
375 tural Engineering 136(12):1619–1624
376 [6] Avossa A, Demartino C, Ricciardelli F (2017) Design procedures for footbridges subjected to walking
377 loads: Comparison and remarks. Baltic Journal of Road & Bridge Engineering 12(2)
378 [7] Brincker R, Ventura C, Andersen P (2001) Damping estimation by frequency domain decomposition.
379 In: Proceedings of 19th International Modal Analysis Conference (IMAC XIX), Orlando, FL (USA)
380 [8] Brincker R, Zhang L, Andersen P (2001) Modal identification of output-only systems using frequency
381 domain decomposition. Smart Materials and Structures 10:441 – 445

23
 ACCEPTED MANUSCRIPT

382 [9] CEN (2003) Eurocode 1: Actions on structures

383 [10] CEN (2005) Eurocode 3: Design of Steel Structures
384 [11] CEN (2005) Eurocode 5: Design of timber structures

PT
385 [12] Correal JF, Echeverry JS, Ramı́rez F, Yamı́n LE (2014) Experimental evaluation of physical and me-
386 chanical properties of glued laminated guadua angustifolia kunth. Construction and Building Materials
387 73:105–112

RI
388 [13] Demartino C, Avossa AM, Ricciardelli F (2018) Deterministic and probabilistic serviceability assess-
389 ment of footbridge vibrations due to a single walker crossing. Shock and Vibration

SC
390 [14] Dixon P, Ahvenainen P, Aijazi A, Chen S, Lin S, Augusciak P, Borrega M, Svedström K, Gibson L
391 (2015) Comparison of the structure and flexural properties of Moso, Guadua and Tre Gai bamboo.
392 Construction and Building Materials 90:11 – 17

U
393 [15] Echeverry J, Correal J (2015) Cyclic behavior of laminated guadua mat sheathing-to-framing connec-
394
AN
tions. Construction and Building Materials 98:69 – 79
395 [16] Escamilla E, Habert G (2014) Environmental impacts of bamboo-based construction materials repre-
396 senting global production diversity. Journal of Cleaner Production 69:117 – 127
M

397 [17] FIB (2005) Guidelines for the design of footbridges

398 [18] Habibi MK, Tam LH, Lau D, Yang L (2016) Viscoelastic damping behavior of structural bamboo
D

399 material and its microstructural origins. Mechanics of Materials 97:184–198

400 [19] Heinemeyer C, Butz C, Keil A, Schlaich M, Goldack A, Lukić M, Chabrolin B, Lemaire A, Martin
TE

401 P, Cunha A, Caetano E (2009) Design of Lightweight Footbridges for Human Induced Vibrations –
402 Background document in support to the implementation, harmonization and further development of
the Eurocodes
EP

403

404 [20] ISO (2007) Bases for design of structures – Serviceability of buildigns and walkways against vibration
405 [21] Li Ht, Zhang Qs, Huang Ds, Deeks AJ (2013) Compressive performance of laminated bamboo. Com-
C

406 posites Part B: Engineering 54:319–328

407 [22] Li Z, Xiao Y, Wang R, Monti G (2015) Studies of nail connectors used in wood frame shear walls with
AC

408 ply-bamboo sheathing panels. Journal of Materials in Civil Engineering 27(7)

409 [23] van der Lugt P, van den Dobbelsteen A, Janssen J (2006) An environmental, economic and practical
410 assessment of bamboo as a building material for supporting structures. Construction and Building
411 Materials 20(9):648 – 656
412 [24] Mahdavi M, Clouston P, Arwade S (2011) Development of laminated bamboo lumber: Review of
413 processing, performance, and economical considerations. Journal of Materials in Civil Engineering
414 23(7):1036–1042
415 [25] Nisticò N, Gambarelli S, Fascetti A, Quaranta G (2016) Experimental dynamic testing and numerical

24
 ACCEPTED MANUSCRIPT

416 modeling of historical belfry. International Journal of Architectural Heritage 10(4):476–485

417 [26] Paraskeva T, Grigoropoulos G, Dimitrakopoulos EG (2017) Design and experimental verification of
418 easily constructible bamboo footbridges for rural areas. Engineering Structures 143:540–548

PT
419 [27] Ramirez F, Correal J, Yamin L, Atoche J, Piscal C (2012) Dowel-bearing strength behavior of glued
420 laminated guadua bamboo. Journal of Materials in Civil Engineering 24(11):1378–1387
421 [28] Ricciardelli F, Briatico C (2010) Transient response of supported beams to moving forces with sinusoidal

RI
422 time variation. Journal of Engineering Mechanics 137(6):422–430
423 [29] Ricciardelli F, Demartino C (2016) Design of footbridges against pedestrian-induced vibrations. Journal

SC
424 of Bridge Engineering 21(8):C4015003
425 [30] Sétra (2006) Technical guide - Footbridges: Assessment of vibrational behaviour of footbridges under
426 pedestrian loading

U
427 [31] Shan B, Xiao Y, Zhang W, Liu B (2017) Mechanical behavior of connections for glubam-concrete
428
AN
composite beams. Construction and Building Materials 143:158 – 168
429 [32] Sharma B, Gatóo A, Bock M, Ramage M (2015) Engineered bamboo for structural applications. Con-
430 struction and Building Materials 81:66 – 73
M

431 [33] Sharma B, Gatóo A, Ramage M (2015) Effect of processing methods on the mechanical properties of
432 engineered bamboo. Construction and Building Materials 83:95 – 101
D

433 [34] Shin K, Hammond J (2008) Fundamentals of signal processing for sound and vibration engineers. Wiley
434 [35] Song J, Utama Surjadi J, Hu D, Lu Y (2017) Fatigue characterization of structural bamboo materials
TE

435 under flexural bending. International Journal of Fatigue 100:126 – 135

436 [36] Van Overschee P, de Moor BL (1996) Subspace identification for linear systems. Kluwer Academic
Publishers
EP

437

438 [37] Villegas L, Morán R, Garcı́a J (2015) A new joint to assemble light structures of bamboo slats. Con-
439 struction and Building Materials 98:61 – 68
C

440 [38] Wang J, Demartino C, Xiao Y, Li Y (2018) Thermal insulation performance of bamboo-and wood-based
441 shear walls in light-frame buildings. Energy and Buildings 168:167–179
AC

442 [39] Woodhouse J (1998) Linear damping models for structural vibration. Journal of Sound & Vibration
443 215(3):547–569
444 [40] Wu Y, Xiao Y (2018) Steel and glubam hybrid space truss. Engineering Structures 171:140–153
445 [41] Xiao Y, Zhou Q, Shan B (2010) Design and construction of modern bamboo bridges. Journal of Bridge
446 Engineering 15(5):533–541
447 [42] Xiao Y, Yang R, Shan B (2013) Production, environmental impact and mechanical properties of glubam.
448 Construction and Building Materials 44:765 – 773
449 [43] Xiao Y, Chen G, Feng L (2014) Experimental studies on roof trusses made of glubam. Materials &

25
 ACCEPTED MANUSCRIPT

450 Structures 47(11):1879–1890

451 [44] Xiao Y, Li Z, Wang R (2014) Lateral loading behaviors of lightweight wood-frame shear walls with
452 ply-bamboo sheathing panels. Journal of Structural Engineering 141(3):B4014004

PT
RI
U SC
AN
M
D
TE
C EP
AC

26
 ACCEPTED MANUSCRIPT
• This study presents some experimental results intended to provide practical guidelines for
the analysis and design of glubam structures under dynamic loads.
• The dynamic characteristics of an original bamboo-steel composite structure are studied.
• The estimation of the natural frequencies, mode shapes and viscous damping ratio is given.
• General recommendations for viscous damping ratio are provided to support the analysis
and design of glubam truss structures.
• The serviceability conditions for footbridge use are investigated.

PT
RI
U SC
AN
M
D
TE
C EP
AC