Applied Mathematical Modelling 36 (2012) 5492–5505

Contents lists available at SciVerse ScienceDirect

Applied Mathematical Modelling

journal homepage: www.elsevier.com/locate/apm

Construction stage analysis of Humber Suspension Bridge

Süleyman Adanur a, Murat Günaydin b,⇑, Ahmet Can Altunisßik a, Barısß Sevim b
a
Karadeniz Technical University, Civil Engineering Department, 61080 Trabzon, Turkey
b
Gümüsßhane University, Civil Engineering Department, 29000 Gümüsßhane, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: Suspension bridges are widely used as engineering structures to across long spans and give
Received 23 June 2011 rise to the usage of domains under the bridge. In the finite element analyses of suspension
Received in revised form 2 January 2012 bridges, it is assumed that the structure is built and loaded in a second. However, this type
Accepted 4 January 2012
of analysis does not always give the reliable and healthy solutions. Because, construction
Available online 13 January 2012
period of this type of the structures continue along time and loads may be changed during
this period. Therefore, construction stages and time dependent material properties should
Keywords:
be considered in the analysis to obtain the reliable and healthy results.
Construction stage analysis
Humber Suspension Bridge
This paper presents the construction stage analysis of suspension bridges using time
Finite element analysis dependent material properties. For this purpose, Humber Suspension Bridge built near
Time dependent material properties Kingston upon Hull, England is chosen as an example. Finite element model of the bridge
is constituted using SAP2000 program considering project drawings. Geometric nonlinear-
ities are taken into consideration in the analysis using P-Delta large displacement criterion.
The time dependent material strength of steel and concrete and geometric variations are
included in the analysis. Time dependent material properties are considered as compres-
sive strength, aging, shrinkage and creep for concrete, and relaxation for steel. The struc-
tural response of the bridge at different construction stages has been examined. Two
different finite element analyses with and without construction stages are carried out
and results are compared with each other. As analyses result, variation of the displacement
and internal forces such as bending moment, axial forces and shear forces for bridge deck
and towers are given with detail. It can be seen from the study that there are some differ-
ences between both analyses (with and without construction stages) and the results
obtained from the construction stages are bigger. So, it is thought that construction stage
analysis using time dependent material properties and geometric nonlinearity should be
considered in order to obtain more realistic structural response of suspension bridges.
Ó 2012 Elsevier Inc. All rights reserved.

1. Introduction

Suspension bridges are widely used as engineering structures to across long spans and give rise to the usage of domains
under the bridge. For this reason, the use of suspension bridges has increased recently. Suspension bridges are very impor-
tant engineering structure due to the high costs and logistical importance. Hence, the analysis of suspension bridges must be
done on the best possible analytical model since structural elements such as deck, towers and cables show different struc-
tural behaviour.

⇑ Corresponding author. Tel.: +90 462 377 40 20; fax: +90 462 377 26 06.
E-mail addresses: sadanur@ktu.edu.tr (S. Adanur), gunaydin61@hotmail.com.tr (M. Günaydin), ahmetcan8284@hotmail.com (A.C. Altunisßik),
bsevim18@hotmail.com (B. Sevim).

0307-904X/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.apm.2012.01.011
 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505 5493

Finite element analysis is used in order to obtain the structural behaviour of the suspension bridges under variable loads.
But, in the analytical solutions based on finite element models, it is assumed that the structure is built and loaded in a sec-
ond. However, this type of analysis does not always give the reliable and healthy solutions. Because, construction period of
this type of the structures continue along time and loads may be change during this period. Therefore, construction stages
and time dependent material properties should be considered in the analysis to obtain the reliable and healthy response. So,
analysis of suspension bridges is carried out considering construction stages and time dependent material properties.
In the literature, some papers exist about the construction stage analysis of the long span bridges considering time depen-
dent material properties. Ko et al. calculated the dynamic characteristics such as natural frequencies and mode shapes of
suspension deck in construction stages. The Tsing Ma suspension bridge with a main span of 1377 m and an overall length
of 2160 m is performed [1]. Kwak and Seo determine the time dependent behaviour of precast prestressed concrete girder
bridge. To analyse the long-term behaviour of bridges, the effects of creep, the shrinkage of concrete, and the cracking of
concrete slabs in the moment regions is considered [2]. Cheng et al. carried out the wind induced load capacity of a long span
steel arch bridge during two construction stages. The Lupu Bridge which has 550 m central span length and 100 m side spans
is selected as a case study [3]. Wang et al. analysed a cable stayed bridge during construction using the cantilever method.
Two computational processes, one is a forward process analysis and the other is a backward process analysis are established
[4]. Pindado et al. investigated the influence of the section shape of box girder decks on the moments during construction
stages experimentally [5]. Karakaplan et al. performed the construction stage analysis of a cable supported pedestrian bridge
considering time dependent material strength variations. Analysis results such as changing of cable forces, internal forces,
and support reactions are compared with the conventional finite element analysis and the differences are determined [6].
Cho and Kim carried out probabilistic risk assessment for the construction stages of the Hanbit suspension bridge. The bridge
is under construction and will be one of the longest suspension bridges in Korea in 2010. The main span is designed to be
850 m with two side spans of 255 and 220 m each. Tensile forces for main cables and deflections for stiffening girders are
controlled for each construction stages [7]. Somja and Goyet studied about nonlinear finite element analysis of segmentally
constructed cable stayed bridge. Time dependent effects including load history, creep, shrinkage and aging of the concrete
are considered in the analyses. Modification of the bridge topology has been carried out using an efficient procedure for cre-
ating/removing elements [8]. Altunısßık et al. performed the construction stage analysis of Kömürhan Highway Bridge. The
bridge is a reinforced concrete box girder bridge and constructed with balanced cantilever method, located on the 51st
km of Elazığ-Malatya highway [9]. Adanur and Günaydın studied about construction stage analysis of Bosporus Suspension
Bridge. Bosporus Suspension Bridge connecting the Europe and Asia in Istanbul is selected as an example. Two different finite
element analyses with and without construction stages are carried out and results (displacements and internal forces for
deck and towers) are compared with each other [10]. Ates studied about analytical modelling of continuous concrete box
girder bridges considering construction stages. Budan Bridge is selected as a numerical example. The Bridge constructed with
balanced cantilever method and located on Artvin-Erzurum highway, Turkey, at 55 + 729  56 + 079.000 km. The structural
behaviour of the bridge at different construction stages has been examined. After analyses, variation of internal forces such
as bending moment, shear forces and axial forces, and displacements for bridge deck and pier are given with detail [11]. Soy-
luk et al. carried out time dependent nonlinear analysis of segmentally erected cable-stayed bridges. The analysis phase is
divided into two phases: The construction phase and the service phase. In the analyses, while 33 stages which cover 970 days
are considered for the construction phase, 3 stages which lasts up to 10 years are used for the service phase. The analytical
models of the selected numerical example are solved by considering the self weight and the time-dependent nonlinear ef-
fects. The bridge responses are then compared with respect to the time-dependent effects. The study shows that time-depen-
dent effects can have important effects on cable-stayed bridges [12]. Brownjohn et al. carried out ambient vibration re-
testing and operational modal analysis of the Humber Bridge. The paper describes the equipment and procedures used
for the exercise, compares the operational modal analysis technology used for system identification and present modal
parameters for key vibration modes of the complete structure [13].
As seen in literature, there is not sufficient research about the construction stage analysis of suspension bridges. To this
end, this paper presents construction stage analysis of suspension bridges using time dependent material properties. Time
dependent material properties are considered as compressive strength, shrinkage, creep and aging for concrete, and relax-
ation for steel.

2. Description of Humber Suspension Bridge

The Humber Bridge (Fig. 1) located near Kingston upon Hull, England. It spans between Barton on the south coast and
Hessle on the north coast, connecting the East Riding of Yorkshire and North Lincolnshire. Construction of the bridge started
in 1972 and completed in 1981. The bridge has a main span of 1410 m and two side spans of 280 m and 530 m on the
Hessle and Barton sides, respectively. The bridge has reinforced concrete towers of 155.5 m high above the foundation, a
steel box deck and inclined hangers. The connection of hangers and deck are hinged in both directions in the middle of main
span.
The horizontal distance between the cables is 22 m and the roadway is 28.5 m wide, accommodating two four-lane high-
ways. The roadway at the mid-span of the bridge is approximately 30 m above the water level. Schematic representation of
Humber Bridge including dimension is given in Fig. 2.
5494 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505

Fig. 1. Humber Suspension Bridge [17,18].

Fig. 2. Schematic representation including dimension (dimensions as m).

The deck was constituted considering aerodynamic form to reduce of the wind affect along the bridge deck. The aerody-
namic steel box girder deck (Fig. 3) of the bridge consist of 124 box girder deck pieces of 18.1 m long 4.5 m deep prefabri-
cated sections 28.5 m wide, including two 3 m walkways. The top of each box section constitutes an orthotropic plate on
which 37 mm thickness mastic asphalt surfacing is laid.
The bridge has reinforced concrete towers of 155.5 m above the caisson foundations and carries the two main cables
which sag 115.5 m. ground level. These cables each have sectional area of 0.29 m2 and consist of almost 15000 each of
5 mm 1.54 kN/mm2 UTS wires grouped in strands [13]. Vertical tower legs are connected by four horizontal portal beams.
Dimension of towers are given in Fig. 4.
Cables are one of the most important elements in the design of suspension bridges. Inverse analyses are generally used to
determine the initial shapes and locations of main cables. Firstly, finite element model of suspension bridges at the final

2200
1900 9 1 00 9100 1 9 00
37
%2 %2

2 lanes(7300) 2 lanes(7300)
%2 %2
4500

3250 7400 72000 7400 3250

28500

Fig. 3. Dimensions of aerodynamic steel box girder deck (dimensions as mm).

 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505 5495

2200

8280

39800
39800
40550
23850

24400

Fig. 4. Dimensions of towers (dimensions as mm).

Fig. 5. The schematic view of initial and final shapes of cables in suspension bridges.

stage are constituted and analyses are performed. At the end of the analyses, maximum displacements on deck due to the
dead loads of the bridge are obtained. Vertical coordinates of cables at the final stages are increased as obtained maximum
displacement value (Dd). It must be considered that initial and final shapes of the cables are second degree of parabola. Fig. 5
related to schematic view of initial and final shapes of cables is given.
Cable saddles support the main cable at the towers and at the splay bents in the anchorages; the former is called the
tower saddle and the latter is called the splay saddle. But the saddles are not taken into consideration in the finite element
analyses of suspension bridges. The towers and cables are modelled using frame elements and they connect at the top of the
tower on a nodal joint. At this nodal joint, displacements at the longitudinal direction (u1) and rotations at the vertical and
transverse directions (r2 and r3) are restricted. These elements do not affect the static and dynamic responses of the
structures.

3. Finite element analysis

Finite element models are commonly considered in the design and project phase of the important engineering structures
such as bridges using some special software. In this study, SAP2000 [14] finite element program which is used for linear and
5496 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505

Fig. 6. Two-dimensional finite element model of Humber Suspension Bridge.

Fig. 7. Type of suspension bridges according to the stiffening girder.

Table 1
Material and section properties of the elements of Humber Bridge.

Element Material Properties

Modulus of elasticity (kN/m2) Poisson’s ratio (–) Section areas (m2) Inertia moment (m4)
Towers 2.00E7 0.30 40.75 132.046
Deck 2.00E8 0.30 0.73 1.94
Main Cable 1.93E8 0.30 0.58 0.0134
Side span Hessle 1.93E8 0.30 0.62 0.0153
Side span Barton 1.93E8 0.30 0.58 0.0134
Hanger 1.40E8 0.30 0.0042 –

non-linear, static and dynamic analyses of finite element models of structures is used in the analysis. To investigate the con-
struction stage response of the Humber Suspension Bridge, two-dimensional finite element model are used for calculations.
The finite element models of Humber Suspension Bridge are shown in Fig 6. As the deck, towers and cables are represented
by beam elements, the hangers are represented by truss elements in the model. Finite element model of the bridge with ver-
tical hangers has 291 nodal points, 287 beam elements and 236 truss elements and the model is represented by 855 degrees
of freedom.
Suspension Bridges can be classified into two groups according to the type of stiffening girders: Two-hinged stiffening
girder and continuous stiffening girder (Fig. 7). Humber Suspension Bridges is contained in the two-hinged stiffening girder
type. So, the response of the middle span independent from the response of side spans.
More information about the material and section properties of the elements of Humber Bridge is given in Table 1.

3.1. Modelling of the construction stages

In the construction stage analyses of Humber Bridge, a total of 41 construction stages are considered. Total duration from
the beginning to ending of construction is considered as 1190 days. Maximum total step and maximum iteration for each
step are selected as 200 and 100, respectively. Some construction stages using SAP2000 finite element analysis program
is shown in Fig 8.
In the construction stage analysis, some special points given in below should be considered;

 Geometric nonlinearities should be taken into consideration in the analysis using P-Delta large displacement criterion.
 All construction stages and their details should be determined from design to opening the traffic of the bridge.
 Working plan including construction durations of main structural elements (tower, deck and cable) of the bridge should
be prepared.
 Added and removed loads for each construction stages should be determined.
 To obtain the reliable solution, each stage results should be added to end of the each stage and next stage analysis is done.
 Non-linear solution parameters should be selected depending on the literature.

4. Time dependent material properties

In the construction stage analysis of bridges, time dependent material properties such as elasticity modulus, creep and
shrinkage for concrete and relaxation for the prestressed steel should be considered, because they are variable due to the
climate during construction [9,15]. For example, strength of the concrete increase continuously at 7th, 28th and 1000th days
of concreting. If these properties are not considered in the analysis, analysis of the bridges may not give the reliable results.
Time effects and cracking make analysis even more complex for concrete bridges. Creep strains develop at early stages of the
construction process and continue to evolve significantly after the structure is built. Depending on the construction method,
 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505 5497

Fig. 8. Some construction stages of Humber Suspension Bridge.

restrained creep can appear and induce important stress redistribution in the structure [8]. To accurately analyse structures
both during their construction and along their entire life, engineers must have at their disposal appropriate design methods.
The effects of geometry changes occurring during construction of the structure cannot be taken into account using standard
finite element codes since structural elements are added and removed at certain time instants [8].
The iterative calculations at each construction stage considering added stiffness from the initial equilibrium state. The
matrix form of finite element method is given the following equation
fFg ¼ ½KfUg; ð4:1Þ
where [K] is the stiffness matrix including elastic stiffness matrix and geometric stiffness matrix. The finite element analysis
is performed at each construction stages of the bridge by using SAP2000.

4.1. Compressive strength

The compressive strength of concrete at an age t depends on the type of cement, temperature and curing conditions. The
relative compressive strength of concrete at various ages may be estimated by the following formula [16].
5498 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505

fcm ðtÞ ¼ bcc ðtÞfcm ; ð4:2Þ

in which bcc ðtÞ is a coefficient with depends on the age of concrete and is calculated by
( "  1=2 #)
28
bcc ðtÞ ¼ exp s 1  : ð4:3Þ
t=t 1

fcm ðtÞ is the mean concrete compressive strength at an age of t days, fcm is the mean compressive strength after 28 days, t is
the age of concrete in days and s is a cement type coefficient.

4.2. Aging of concrete

The modulus of elasticity of concrete changes with time. For this reason, the modulus at an age t – 28 days may be esti-
mated as below equation
pffiffiffiffiffiffiffiffiffiffiffiffi
Eci ðtÞ ¼ Eci bcc ðtÞ; ð4:4Þ
where Eci ðtÞ is the modulus of elasticity at age of t days, Eci is the modulus of elasticity at an age of 28 days, bcc ðtÞ is a coef-
ficient which depends on the age of concrete.

4.3. Shrinkage of concrete

The CEB-FIP Model Code gives the following equation of total shrinkage strain of concrete;
ecs ðt; ts Þ ¼ ecso bs ðt  ts Þ; ð4:5Þ
where ecso is notional shrinkage coefficient, bs is the coefficient to describe the development of shrinkage with time, t is the
age of concrete in days and t s is the age of concrete in days at the beginning of shrinkage. The notional shrinkage coefficient
may be obtained from
ecso ¼ es ðfcm ÞbRH ; ð4:5aÞ
  
fcm
es ðfcm Þ ¼ 160 þ 10bsc 9  ; ð4:5bÞ
fcmo
where fcm is the mean compressive strength of concrete at the age of 28 days in MPa; fcmo is taken as 10 MPa; bsc is a coef-
ficient ranging from 4 to 8 which depends on the type of cement
bRH ¼ 1:55bsRH ; 40% 6 RH < 90%;
ð4:6Þ
bRH ¼ 0:25; RH P 99%;
where
 3
RH
bsRH ¼ 1  ð4:7Þ
RHo
with RH is the relative humidity of the ambient atmosphere (%) and RHo is 100%. The development of shrinkage with time is
given by
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðt  t s Þ=t 1
bs ðt  t s Þ ¼ ; ð4:8Þ
350ðh=ho Þ þ ðt  t s Þ=t 1

where h is the notional size of member (mm) and is calculated by h ¼ 2Ac =U in which Ac is the cross-section and u is the
perimeter of the member in contact with the atmosphere; ho = 100 mm and t1 = 1 day.

4.4. Creep

The effect is calculated using CEB-FIP Model Code (1990) creep model. For a constant stress applied at time to, this leads
to;
rc ðto Þ
ecc ðt; to Þ ¼ /ðt; t o Þ; ð4:9Þ
Eci
in which rc ðto Þ is the stress at an age of loading to, /ðt; to Þ is the creep coefficient and is calculated from
/ðt; t o Þ ¼ bc ðt  to Þ/o ; ð4:10Þ
where bc is the coefficient to describe the development of creep with time after loading, t is the age of concrete in days at the
moment considered, to is the age of concrete at loading in days. The creep coefficient is explained by
 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505 5499

/o ¼ /RH bðfcm Þbðt o Þ; ð4:11aÞ

 
RH
1 RH0
/RH ¼ 1 þ  1=3 ; ð4:11bÞ
h
0:46 ho

5:3
bðfcm Þ ¼ qffiffiffiffiffiffi ; ð4:11cÞ
fcm
fcmo

1
bðto Þ ¼  0:2 : ð4:11dÞ
0:1 þ tt1o

All parameter is defined above. The development of creep with time is given by
 
ðt  t o Þ=t1
bc ðt  t o Þ ¼ ; ð4:12:aÞ
bH þ ðt  t o Þ=t 1

( 18 )
RH h
bH ¼ 150 1 þ 1:2 þ 250 6 1500; ð4:12:bÞ
RHo ho

where t1 = 1 day; RHo = 100 and ho = 100 mm.

Table 2
Selection of analysis parameters to consider time dependent material properties in SAP2000.

Parameters Main structural elements

Deck Tower Prestress steel
Material Properties Tendon Concrete Tendon
Isotropic Isotropic Uni-axial
Nonlinear material data
Hysteresis type Kinematic Kinematic Kinematic
Stress–strain diagram User defined User defined User defined
Time dependent properties
Elasticity modulus U U U
Creep U U –
Shrinkage U U –
Creep analysis type Full Full –
Cement type coefficient 0.25 0.25 –
Relative humidity % 60 60 –
Notional size 0.6 1.3 –
Shrinkage coefficient 5 5 –
Shrinkage start age 0 0 –
Steel relaxation U U U
Relaxation analysis type Full Full Full integration
CEB-FIP class 1 1 1

3.0 2.0
(a) (b)
0.0
Stress (kN/m²) E3

Stress (kN/m²) E6

-3.0 1.0

-6.0

-9.0 0.0

-12.0

-15.0 -1.0

-18.0

-21.0 -2.0
-6.0 -4.5 -3.0 -1.5 0.0 1.5 -40.0 -20.0 0.0 20.0 40.0
Strain (m/m) E-3 Strain (m/m) E-3

Fig. 9. Stress–strain diagrams used for concrete (a) and prestressed steel (b).
5500 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505

4.5. Relaxation of steel

According to CEB-FIP Model Code (1990), relaxation classes referring to the relaxation at 1000 h are divided into three
groups for prestressing steels. The first relaxation class is defined as the normal relaxation characteristics for wires and
strands, the second class is defined as improved relaxation characteristics for wires and strands, and the last one is defined
as relaxation characteristics for bars.

36.0 24.0

Modulus of Elasticity (kN/m²) E6

Mean Strength (kN/m²) E3

27.0 18.0

12.0
18.0

6.0
9.0

0.0
0.0
0.0 0.4 0.8 1.2 0.0 0.4 0.8 1.2
Time (Days) E3 Time (Days) E3
(a) Time dependent mean strength (b) Time dependent modulus of elasticity
2.4
1E-3

1E-4
Creep Coefficient

Shrinkage Strain

1.6
1E-5

1E-6

0.8 1E-7

1E-8

0.0 1E-9
0.0 0.4 0.8 1.2 0.0 0.4 0.8 1.2
Time (Days) E3 Time (Days) E3
(c) Time dependent creep coefficient (d) Time dependent shrinkage strain

Fig. 10. Variation of time dependent material properties for concrete.

120
Relaxation Coefficient E-3

80

40

0
0.0 0.4 0.8 1.2
Time (Days) E3

Fig. 11. Variation of time dependent material properties for prestressed steel.
 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505 5501

For an estimate of relaxation up to 30 years the following formula may be applied

 k
t
qt ¼ q1000 ; ð13Þ
1000
where qt is the relaxation after t hours; q1000 is the relaxation after 1000 h; k  logðq1000 =q100 Þ in which k to be 0.12 for relax-
ation class1, and 0.19 relaxation class2; q100 is the relaxation after 100 h. Normally, the long-term values of the relaxation are
taken from long-term tests. However, it may be assumed that the relaxation after 50 years and more is three times the relax-
ation after 1000 h.
Selected analysis parameters to consider time dependent material properties are given in Table 2.
Variation of time dependent material properties used for concrete and prestressed steel is given in Figs. 9–11. These
parameters are selected from CEB-FIP design code (CEB-FIP 1990) in SAP2000. According to the parameters given in Table
1, these graphics may be changed automatically. Total duration from the beginning of construction to ending of construction
is considered as 1190 days.

Fig. 12. Deformation of Humber Suspension Bridge during some construction stages.
5502 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505

5. Construction stage analysis

For the construction stage analysis of suspension bridges considering time-dependent material properties, Humber Sus-
pension Bridge is selected as an example. This bridge has a main span of 1410 m and two side spans of 280 m and 530 m on
the Hessle and Barton sides, respectively. The bridge has reinforced concrete towers of 155.5 m high above the foundation, a
steel box deck and inclined hangers. The roadway at the mid-span of the bridge is approximately 30 m above the water level.
Analysis is performed using SAP2000 program. Nonlinear staged construction and P-Delta plus large displacements options
are selected as analysis type and geometric nonlinearity parameters, respectively.

5.1. Load cases of analyses

In the analyses of the bridge, the following load cases are considered:

 Dead load: Weight of all elements. They are calculated from the finite element software directly.
 Additional mass: Weight of the asphalt, bordure, cobble, pipeline and its supports, scarecrow. 40 kN/m distributed load is
added to each segment.

5.2. Deformation shapes

The deformations of the bridge at some construction stages are plotted and the maximum vertical displacements of the
bridge deck and maximum horizontal displacements of the bridge tower are also given in Fig. 12. It is seen that displace-
ments increase along the middle of the bridge deck and reach a maximum of 16.71 m at the 20th stage for the analysis
including the construction stage. When the construction of the bridge is completed at the 41st stage, maximum displace-
ment is obtained as 15.98 m at the middle point of the bridge deck. In addition to this, variation of the displacement in-
creases along the height of the bridge towers and reach a maximum of 94 cm at the 41st stage.

6. Numerical results

Distributions of vertical displacements and bending moments along the bridge deck are given in Fig. 13. It is seen that
displacements have an increasing trend towards to the middle of the bridge deck and reach a maximum of 15.98 m at
the middle for the analysis including the construction stages. But, maximum displacement is 13.20 m at the middle for
the analysis not including construction stage. The values of bending moments are changeable along the bridge deck and
reach a maximum of 5E4 kN m at the middle for the analysis not including the construction stage. On the other hand, the

0
Displacement (cm)

-5

-10
Construction Stage

-15 Not Considered

Considered
-20
-1000 - 750 -500 -250 0 250 500 750 1000 1250
Distance (m)
(a) Displacements
Bending Moment (kNm)

0E+0

-3E+5 Construction Stage

Considered
Not Considered
-5E+5
-1000 - 750 -500 -250 0 250 500 750 1000 1250
Distance (m)
(b) Bending moments
Fig. 13. Changing of maximum displacements (a) and bending moments (b) along the bridge deck.
 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505 5503

values of bending moments are more changeable along the bridge deck and reach a maximum of 4E5 kN m at the middle for
the analysis including the construction stage. It is seen from Fig. 13 that the displacements and bending moments obtained
from the analyses including construction stages are bigger than those of not including the construction stages.

160 160

120 120
Tower Height (m)

Tower Height (m)

80 80

40 Construction Stage 40 Construction Stage

Not Considered Not Considered
Considered Considered
0 0
0 40 80 120 0 40 80 120 16 0
Displacement (cm) Displacement (cm)
(a) The tower on the north side (b) The tower on the south side

Fig. 14. Changing of displacements along to the height of the bridge towers.

Construction Stage
Not Considered Considered
160 160

120 120
Tower Height (m)
Tower Height (m)

80 80

40 40

0 0
-5E+5 -3E+5 0E+0 -2000 -1000 0 1000 2000
Axial Force (kN) Shear Force (kN)
(a) The tower on the north side
Construction Stage
Not Considered Considered
160 160

120 120
Tower Height (m)

Tower Height (m)

80 80

40 40

0 0
-5E+5 -3E+5 0E+0 -2000 -1000 0 1000 2000
Axial Force (kN) Shear Force (kN)
(b) The tower on the south side

Fig. 15. Changing of internal forces along the height of the bridge towers.
5504 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505

Variation of maximum displacements along the height of the towers is shown in Fig. 14. It can easily be seen that the
horizontal displacements increase with the height of bridge towers and reach a maximum of 94 cm on the north side and
145 cm on the south side at the top for the analysis including the construction stage. The value of the horizontal displace-
ment with the height of bridge towers is 86 cm and 130 cm on the north and south sides respectively for the analysis not
including the construction stage.
Fig. 15 points out the internal forces such as shear and axial forces of the bridge towers corresponding to the two analyses.
The values of the axial forces are nearly equal along the height of the north and south towers for both analyses. Axial forces
decrease from the base (4E5 kN) to the top of the point (2.5E5 kN) on the north-side tower and from (3.5E5 kN) to
(2.1E5 kN) on the south-side tower. The values of the shear forces are nearly equal along the height of the bridge tower
as 277 kN and 147 kN on the north and south sides respectively for the analysis not including the construction stage, but
the values of the shear forces are changeable along the height of the bridge towers for the analysis including the construction
stage. Shear forces decrease non-linearly from the base (1.6E3 kN) to the middle point (1.5E2 kN) and increase non-linearly
from the middle point (3.3E2 kN) to the top point (1.3E3 kN) on the north side. Besides, shear forces decrease non-linearly
from the base (1.9E3 kN) to the middle point (9.4E kN) and increase non-linearly from the middle point (5E2 kN) to the
top point (1.6E3 kN) on the south side. It can be easily seen from Fig. 15 that construction stage analysis is more effective
than the other one for both internal forces. The reason of the ladder-shape distribution of shear force along the main tower
is two different responses at each nodal point before and after construction of each segment.

7. Conclusion

Suspension bridges are widely used as engineering structures to across long spans and give rise to the usage of domains
under the bridge. Construction period of this type of the structures continue along time and loads may be change during this
period. Therefore, construction stages and time dependent material properties should be considered in the analysis to obtain
the reliable and healthy results.
In this study, it is aimed to perform the construction stage analysis of suspension bridges using time dependent material
properties. Humber Suspension Bridge is selected as an example. The time dependent material strength variations and geo-
metric variations are included in the analysis. In the analyses, total duration from the beginning of construction to ending of
construction is considered as 1190 days. Comparing the results of the study, the following observations can be made:

 The vertical displacements increase towards to the middle of the bridge deck and reach a maximum of 15.98 m at the
middle for the analysis including the construction stages. On the other hand, maximum displacement is 13.20 m at the
middle for the analysis not including construction stage. The difference is reached to 2.78 m at the middle of the bridge
deck. The horizontal displacements increase with the height of bridge towers and reach a maximum of 94 cm on the north
side and 145 cm on the south side at the top for the analysis including the construction stage. Yet, the value of the hor-
izontal displacement with the height of bridge towers is 86 cm and 130 cm on the north and south sides respectively for
the analysis not including the construction stage.
 The values of bending moments are changeable along the bridge deck and reach a maximum of 5E4 kN m at the middle
for the analysis not including the construction stage. On the other hand, the values of bending moments are more change-
able along the bridge deck and reach a maximum of 4E5 kN m at the middle for the analysis including the construction
stage. The values of bending moments obtained from the analyses including construction stages are significantly bigger
than those of not including the construction stages.
 The values of the axial forces are nearly equal along the height of the north and south towers for both analyses. Axial
forces decrease from the base (4E5 kN) to the top of the point (2.5E5 kN) on the north-side tower and from
(3.5E5 kN) to (2.1E5 kN) on the south-side tower. The values of the shear forces are nearly equal along the height of
the bridge tower as 277 kN and 147 kN on the north and south sides respectively for the analysis not including the con-
struction stage, but the values of the shear forces are changeable along the height of the bridge towers for the analysis
including the construction stage. Shear forces decrease non-linearly from the base (1.6E3 kN) to the middle point
(1.5E2 kN) and increase non-linearly from the middle point (3.3E2 kN) to the top point (1.3E3 kN) on the north side.
Besides, shear forces decrease non-linearly from the base (1.9E3 kN) to the middle point (9.4E kN) and increase non-
linearly from the middle point (5E2 kN) to the top point (1.6E3 kN) on the south side.
 There are some differences between the results with and without the construction stages. It can be stated that the anal-
ysis without construction stages cannot give the reliable solutions.
 In this paper, both of the construction stages and time dependent material properties are considered in the finite element
analysis of Humber Suspension Bridge. The analyses can be divided into three groups as construction stages analyses,
time dependent material properties and, construction stages analyses with time dependent material properties. At the
end of the analyses, the differences can be obtained and which analysis has an important effect on the structural elements
of suspension bridges (concrete girder, hanger, cable and tower) can be investigated.
 To obtain real behaviour of engineering structures, construction stage analysis using time dependent material strength
variations and geometric variations should be done. Especially it is very important for suspension bridges, because con-
struction period continue along time and loads may be change during this period.
 S. Adanur et al. / Applied Mathematical Modelling 36 (2012) 5492–5505 5505

References

[1] T. Cho, T.S. Kim, Probabilistic risk assessment for the construction phases of a bridge construction based on finite element analysis, Finite Elem. Anal.
Des. 44 (2008) 383–400.
[2] H.G. Kwak, Y.J. Seo, Numerical analysis of time-dependent behaviour of pre-cast pre-stressed concrete girder bridges, Constr. Build. Mater. 16 (2002)
49–63.
[3] J. Cheng, J.J. Jiang, R.C. Xiao, M. Xia, Wind-induced load capacity analysis and parametric study of a long-span steel arch bridge under construction,
Comput. Struct. 81 (2003) 2513–2524.
[4] P.H. Wang, T.Y. Tang, H.N. Zheng, Analysis of cable-stayed bridges during construction by cantilever methods, Comput Struct 82 (2004) 329–346.
[5] S. Pindado, J. Meseguer, S. Franchini, The influence of the section shape of box-girder decks on the steady aerodynamic yawing moment of double
cantilever bridges under construction, J. Wind Eng. Ind. Aerodyn. 93 (2005) 547–555.
[6] A. Karakaplan, A. Caner, Ö. Kurç, A. Domaniç, A. Lüleç, New strategy in the structural analysis: construction stage, in: 1st Symposium of Bridges and
Viaducts, Antalya, 2007, pp. 141–153.
[7] J.M. Ko, S.D. Xue, Y.L. Xu, Modal analysis of suspension bridge deck units in erection stage, Eng. Struct. 20 (1998) 1102–1112.
[8] H. Somja, V.V. Goyet, A new strategy for analysis of erection stages including an efficient method for creep analysis, Eng. Struct. 30 (2008) 2871–2883.
[9] A.C. Altunısßık, A. Bayraktar, B. Sevim, S. Adanur, A. Domaniç, Construction stage analysis of Kömürhan Highway Bridge using time dependent material
properties, Struct. Eng. Mech. 36 (2) (2010) 207–244.
[10] S. Adanur, M. Günaydın, Construction stage analysis of Bosporus suspension bridge, in: 9th International Congress on Advances in Civil Engineering,
Trabzon, CD. SEE-161, 2010.
[11] S
ß . Atesß, Numerical modelling of continuous concrete box girder bridges considering construction stages, Appl. Math. Model. 35 (8) (2010) 3809–3820.
[12] K. Soyluk, T.G. Diri, E.A. Sıcacık, Time dependent nonlinear analysis of segmentally erected cable-stayed bridges, in: 9th International Congress on
Advances in Civil Engineering, Trabzon, CD. SEE-053, 2010.
[13] J.M.W. Brownjohn, F. Magalhaes, E. Caetano, A. Cunha, Ambient vibration re-testing and operational modal analysis of the Humber Bridge, Eng. Struct.
32 (2010) 2003–2018.
[14] SAP2000, Integrated Finite Element Analysis and Design of Structures, Computers and Structures Inc., Berkeley, California, USA, 2008.
[15] A.C. Altunısßık, Determination of structural behaviour highway bridges using analytical and experimental methods, Ph.D. Thesis, Karadeniz Technical
University, Trabzon, 2010, Turkey (in Turkish).
[16] CEB-FIP Model Code, Thomas Telford, ISBN: 0727716964, 1990.
[17] <http://en.structurae.de/photos/index.cfm?JS=668>, 2011.
[18] <http://en.structurae.de/photos/index.cfm?JS=666>, 2011.