Documente Academic
Documente Profesional
Documente Cultură
21.1 General
Recent designs of arch bridges show many variations of arrangements for arches as well as hangers, adding
aesthetic value to these structures. Some examples are given in fig. 1. It is well known that arch bridges are
conceived to stand permanent loads, while their performance under moving loads is less, requiring high flexural
deck stiffness. For high speed railway transportation, significant vibrations may take place, especially under the
conditions of “entering train” and “outgoing train”, which force the bridge to vibrate according to the “quasi-
symmetric” mode, characterised by a contra flexure point located at about mid-span. Therefore, new design
solutions started recently like multiple arch bridges that consist of an external arch connected to two internal
arches by means of the hangers.
a b
Unlike many other types of bridges, e.g. the DAF (dynamic amplification factor) for vertical displacement is
strongly non-uniform along the span but butterfly shaped, with low values at mid-span and large values at the
quarters. Therefore, the effects induced by trains are larger, compared to conventional bridges, especially if the
span length is small.
Basic nomenclature
The nomenclature of the structural elements of an arch bridge is as follows:
Type A.
The basic arch bridge, with a predominating arch and with the
thrust transmitted directly to the foundation. Therefore, the
elongation effect of the roadway supporting structure is assumed
to be absent.
The arch is subjected to bending, shear forces and axial forces.
Type B.
The basic tied arch bridge. The arch still dominates, but tying the
ends of the arch through the deck system resists the thrust. The
hangers are concentrated loaded in tension by local live load and
self-weight.
Type C.
A tied arch bridge with stiffening girder. A predominating
stiffening girder is subjected to bending moments and axial
forces induced by the arch. The arch itself is mainly loaded in
compression.
Fig. 6.
Dintelhaven railway bridge, Rotterdam.
Fig. 3-5 shows three situations in which the arch lies above the deck. Other equally valid forms exist in which the
arch may be situated either entirely below or partly below the deck.
bridges with very heavy traffic loading, a truss rib section can be used. The truss can be constructed of lighter
sections and configured to provide superior stiffness as shown in (d). Overall the steel arch bridge permits
numerous structural arrangements that can be extremely graceful and attractive.
The Pont du Gard is an aqueduct built 2000 years ago to carry water across
a valley and is still standing today.
The first metal bridge was an arch made from iron in Shropshire. This was
so famous that the place was named Iron bridge, it was opened in 1781 and
still stands today.
As shown in fig. 13, the arch bridge is very competitive with other type of bridges. As an indication, the span
length of an arch bridge varies between 50 – 500m, with a maximum of approx. 250m for railway bridges and
500m for highway bridges.
Some examples of arch bridges recently constructed in the Netherlands are shown in figures 16-18.
Fig. 16.
1 e and 2e Van Brienenoord
bridge.
Constructed: 1999
Type: Tied arch bridge
with stiffening girder
Length: 170 m
Width: 12,0 m
f/l: 0,176
Arch: Steel tubular section
Hangers: vertically,
cables; c.o.c.10,45 m
Deck system: ballast
supported by cross girders.
• arch
• hangers
• stiffening girders
The Arch
The arch can be either a truss, box girders, plate girder or some form of hollow section (e.g. circular). Ideally, for
aesthetic reasons the arch should follow a smooth, continuous curved profile. However, overall curvature of the
arch may also be achieved by means of a series of short straight chords.
Fig. 19.
Demka bridge under construction (2002) and positioned at final
location.
Fig. 20.
La Barquete bridge, Sevilla (concrete).
To support the arch laterally and to reduce the buckling length of the arch members, an alternative as frequently
used is a system of bracings. Wind forces and stability effects are transmitted to the supports by rigid end portals.
Different methods may be adopted for transmitting these forces back to the supports.
The inclined rigid end portal is composed of the two last members of each arch, the reinforced last member of
the wind bracing and the end cross girder. The portal transmits forces directly back to the supports.
Fig. 23. Example on using inclined rigid end portal (Van Brienenoordbrug).
Mostly, because of the large normal force in the arch, there exist a direct connection between bearing and arch.
An example is given in fig. 25.
Fig. 26.
Left: example of an arch constructed as a truss.
Middle: example of an arch constructed as a box girder.
Right: example of arches connected by horizontal members only.
The hangers
I-sections (welded or rolled), circular hollow sections (or massive) or cables may be used for the hangers.
Opinions differ about the optimum choice of section. The most important design criteria are in the field of:
• Static strength
• Fatigue strength
• Global stiffness criteria
• Permanent tensile loading
• Aerodynamics
• Reduction in-plane arch buckling length.
I-sections hangers can be sensitive with respect to flutter, whilst hollow sections and cables can be sensitive with
respect to vortex shedding. Many types of hanger configuration are present, like vertical, diagonal, network, etc.
For the Dutch market only the first two alternatives exist.
Fig. 27.
left: cable vertical
middle: cable diagonal
right: I-section.
Because of problems on flutter, the I-section members as shown in fig. 27 (right), are connected horizontally.
Some practical design criteria, based on measurements of existing railway bridges are:
f * Lh > 1.5 m/s
λ < 220
ft /fb > 1.3
with
f = bending or torsion first frequency
Lh = cross section width or height
λ = slenderness
As shown in fig. 28, using slender hanger members means the use of a damping system like mass dampers and
energy absorbing dampers.
The risk on rain-wind induced vibration can be limited by a certain minimum on roughness outer surface or
attached circular shaped rings, see fig. 29.
Fig. 30.
Influence lines on axial force of a hanger.
Considering the tied arch more closely, it is to be noted that arch and girder are often in separate vertical planes;
the arch is therefore connected eccentrically to the girder. Any of such eccentricity must be taken into account in
design.
Fig. 31.
Example on design model in case of different
vertical planes of stiffening girder and arch.
Fig. 32.
Example of a bridge deck under construction.
Dependent on the importance on self-weight, the material used can be steel, concrete or combination of both. The
trend exists on decreasing interest on steel only. This because of high production costs and the frequent
inspection / maintenance (durability and fatigue).
Fig. 37.
Currently frequently used alternative on composite
steel/concrete deck system.
The number and type of bridge deck used considering arch 300
bridges (steel and concrete) is given in fig. 38. The main span
250
length for concrete arch bridges varies between 37 – 68 m,
main span [m]
After 1960 a period started using steel orthotropic decks, and 150
steel-concrete
0
Fig. 38. Arch bridges, Netherlands: type of bridge deck 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
used. year
The type of arch bridge used depends largely on the type of loading (railway cq. highway). The tied arch bridge
with stiffening girder is mostly used in case of train loading, and the highway bridges are mostly constructed as a
basic tied arch bridge.
trekband
trekband
verstijfde
verstijfde
zuiver
Fig. 39. Distribution of arch bridges used; left railway bridges, right highway bridges.
An increase of loading, especially traffic loading, result for optimum design in an increase of arch height ‘f’.
Generally spoken, the ratio f/L for a railway bridge is nearly the same as for a highway bridge. For the arch
bridges constructed after approx. 1970, the ratio varies within a narrow scatter of 0.13 – 0.18, see fig. 40.
0.25
0.2
0.15
f/l ratio
0.1
railway: tied arch
railway: tied stiffening arch
Fig. 40. Ratio f/L for all type of arch bridges. highway: tied arch
0.05
highway: tied stiffening arch
railway+highway teid arch
For the hangers of highway bridges, cables are
0
used after 1960. Before this period, I-section has
1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
been used only. year of construction
Because of the large ratio on σmax / σmin, which is an important parameter for the fatigue design, for railway
bridges cables are used at a much later state, namely after 1990. Considering an arch bridge with stiffening
girder, because of the stiffness behaviour and their related influence about influence line on normal force of a
hanger, the combination of diagonal hanger configuration with tied arch bridge with stiffening girder doesn’t
exist. Three large projects on arch railway bridges constructed in The Netherlands are Demkabridge (2002),
Nootdorp bridge (2004) and Eefde bridge (2005).
Year of construction
For the analysis, there exist four important patterns of loading, see fig. 43.
Full loading
In this case the normal force (compression arch, tensile deck) reaches a maximum. As a result, for the deck a
good estimate can be obtained from
qL2
H= and for the arch at the connection arch – deck:
8f
qL2 with α is the angle between arch – deck.
N=
8 f ∗ cosα
Full loading generates a normal compression force in the arch, and a compensating
tensile force in the deck. The second loading doesn’t generate a compression force in
the middle of the arch (in case of a academic approach on fully symmetric support
conditions). As the deflection under this second loading is composed of two half
waves, the girder can be considered to be composed of two parts with a hinge at mid-span. The maximum
bending moment is therefore approx.
qL2
M=
32
This maximum bending moment occurs at L/4. This type of loading generally results into the largest deflection
of the bridge (globally).
1
F= qL
2
δ 0 = l3 − (l1 − ∆l1 ) 2 − (l2 + ∆l 2 ) 2
σ1 σ2
Ni
∆l1 = l1 ∆l2 = l2 σi =
E E Ai
17 ,75
The first eigen-frequentie (bending) is approx. fb =
δ self − weight
Full loading on one side of the bridge
Again the equally distributed loading 2q can be considered as being composed of two superimposed loadings:
Full loading: +q
Anti symmetric loading as follows:
full loading on one side of the bridge: positive loading +q
full loading on the other side of the bridge: negative loading –q
Again, the full loading primarily generates normal forces. The one side loading tends to lozenge the bridge cross-
section causing horizontal lateral forces on the arch and deck. These forces create horizontal bending moments in
the girders and deck. In a homogenous orthotropic steel deck, the bending moment may lead to considerable
stresses.
As shown in table 1, freight trains cause far more fatigue damage than high-speed trains.
Fig. 45.
Peak vehicle acceleration, multiple arch bridge, L= 70 m.
The criteria listed above show clearly the unique character of the design of a bridge. Therefore, design rules on
optimisation are very restricted by local conditions.
Fig. 47.
Cross section (left) and hanger anchorages (right).
The arches have a constant cross section (1800 x 2000 x 50mm). Five box-section cross girders are used between
the 2 arches to keep them stable (lateral stability: wind, buckling).
The vertical hangers are full-section circular bars (Ø 160 or 180mm), and the anchorages of the hangers are
knuckle joints. The knuckle plates are welded inside the arches and to the top of the main girder webs.
Two constant depth (2900mm) plate girders support the deck. The cross-girders of the composite deck are
normal beams HEA 600, bolted or welded to the vertical web stiffeners of the main girder.
Fig. 48. Left: Cross-girders of the deck, bolted or welded to the vertical web stiffeners.
Right: Fitting the hanger.
The arch – main girder junction ensures the optimal tress transmission. This is achieved by means of a central
web common to the arch and to the girder.
Fig. 51. Launching sequence, stage No 3 of the launching (left) and stage No 6 moving round the toll
building (right).
Finally, the bridge was lowered onto its definitive supports, the hanger stress was measured and the deck was
concreted.
Secondary effects
There are several secondary actions that need to be considered in the design of a bridge. An example is given on
hanger design. The initial analysis
has assumed that the connections
between the hangers and both the
arch and main girder are pinned.
In practice these connections are
semi-rigid and the displacements
in the structure lead to bending
moments in the hanger.
Fig. 52. Influence lines for the bending moments in the hangers and bending displacement and moment in
the shortest hanger.
Fig. 52 shows typical influences lines (explained in the next paragraph) and moments in the hangers. It can be
seen that the greatest bending moments arise when the arch is partially loaded. The worst cases in this example
are the shortest and longest hangers. Especially for the longest hangers, the moments are subject to full reversal
and are therefore most likely to influence the fatigue life. In general, their absolute value is, however, low,
justifying their neglect in the initial global analysis. (In the Amsterdam Rhine railway bridge, the maximum
bending stress was 80 N/mm2 ). In order to reduce the magnitude of these fluctuating stresses it is possible to
reduce the minimum second moment inertia of the hangers by e.g. reducing the flange width. Another issue is
connecting the shortest hanger after full self-weight of the bridge being present.
Tie normal forces, arch normal forces, hanger normal forces, tie bending moments, arch bending moments and
tie vertical deflection.
For the Tie and Arch, three positions are considered: nearby left bearing, ¼ span length, ½ span length.
For the hanger, the shortest and longest hanger are considered.
influence line: normal force arch, nearby left bearing influence line: normal force arch, 1/2 spanlength
[unity load 100 kN] [unity load 100 kN]
200 normal force [kN]
200
normal force [kN]
150
100
50
0
0 25 50 75 100 125
distance from left bearing [m]
50 40
normal forcet [kN]
40 30
30
20
20
10
10
0 0
-10 0 25 50 75 100 125 0 25 50 75 100 125
distance from left bearing [m] distance from left bearing [m]
1000
bending moment
500
[kNm]
[kNm]
300 500
100 0
0 25 50 75 100 125
-100 0 25 50 75 100 125 -500
distance from left bearing [m]
distance from left bearing [m]
700
bending moment
500
[kNm]
300
100
-100
0 25 50 75 100 125
-300
distance from left bearing [m]
influence line, bending moment arch, nearby left bearing influence line, bending moment arch, 1/4 spanlength
[unity load 100 kN] [unity load 100 kN]
bending moment
200 200
bending moment
100 150
[kNm]
[kNm]
100
0
0 25 50 75 100 125 50
-100
0
-200 0 25 50 75 100 125
distance from left bearing [m] distance from left bearing[m]
100
0
0 25 50 75 100 125
-100
distance from left bearing [m]
Arch bending moments
influence line: vertical deflection tie, location 1/4 span length influence line: vertical deflection tie, location 1/2 span length
[unity load 100 kN] [unity load100 kN]
20
deflection [mm]
10
deflection [mm]
10
-10 0 25 50 75 100 125
0
-30 0 25 50 75 100 125
-10
-50 -20
distance from left bearing [m] distance from left bearing [m]
Theoretical background
Research on the buckling behaviour of arches can be found in several textbooks and journals. Some interesting
articles, as published by S. Palkowski can be found in Stahlbau 7 (1987) p. 246-250 and Stahlbau 8 (1994) p.
248-251. Some information/data specifically for the design of arch bridges is given.
1
M = q *l 2 and
8
q*l2
H*f =M H=
8* f
EI z
with q ki = α
l3
The critical in-plane buckling force of an arch with tension tie and hangers
H ki a * EI p 2 * EI y p cosϕ
N ki = = = ß= 8* l * f *
cosϕ 8 * l * f * cos ϕ (ß * s)2 s a
Hki = critical horizontal component of the arch normal force [kN]
Nki = critical normal force of the arch [kN]
q = uniform distributed arch loading [kN/m]
l = span length [m]
f = height of the arch [m]
s = half of the arch length [m]
ϕ = angle between arch and deck [rad]
α = coefficient depending on number of hangers and ratio on f/l [-] (see table 3)
β = effective length factor (buckling factor); depending on parameters like number of hangers and
ratio f/l [-]
The relation between f/l and arch length s expressed by a value is given in table 2.
(a)
(b)
(c)
Fig. 54. Alternative types of hanger configuration.
As shown in table 3, the lowest a-value is obtained for configuration (a) of fig. 54 and the highest for
configuration (c). This phenomenon is explained by fig. 56.
A decrease of angle between hanger and arch means an increase in effective supporting translation stiffness, see
deflected shape of arch given in fig. 56.
For arches having a large shear cross section capacity, the number of hangers strongly influences the size of
critical buckling force. A small shear cross section capacity means not only a large reduction on critical buckling
force but also a small influence of the number of hangers on this force.
This behaviour is explained by fig. 57.
EI
k=
G ⋅ As ⋅ l
2
where
2
π
N cr = E ⋅ I y
β ⋅s
Snap through of arches may be assumed to be prevented, if the following criterion is satisfied:
EA
l∗ >K
12 EI y
A = cross sectional area
Iy = moment of inertia
K = factor taken from fig. 58.
The critical out-of-plane buckling force in arches with bracing and end portals
Considering bending in-plane, the static strength is verified by the equation as given in NEN 6771
Fig. 60.
Layout of the reference design.
Fig. 61.
Arch welded steel cross section
Left: central part
Right: nearby the deck
Number of 5 6 7 8 9
spans
c.t.c. hanger 30,4 25,3 21,7 19,0 16,9
For studying the effect of alternative hanger configurations on the design of a tied arch bridge, a certain range on
values for the parameters as given in table 5 have been included. The design of the bridge is based on Eurocodes.
1400
Demka
1200
main span / deflection [-]
1000
800
600
Fig. 65. L/δ ratio for the 2 nd -hanger location and deck maximum.
From existing literature, in case of a diagonal hanger configuration, for an optimum on the use of material the
following angles are found to be optimal, see left figure.
21.8.2 Results on relation tensile force hanger / self weight of the bridge
The design criterion on a minimum of tensile force of the hanger N > 100 kN, results into a minimum on self
weight of the bridge.
40000
diagonal 100 kN
35000 tensile
Demka diagonal tensile
30000
self weight deck [kg/m]
20000
Based on influence lines
15000 on the normal forces for
the hangers, the
10000 minimum on self-weight
have been analysed. The
5000 results obtained are
shown in fig. 67.
0
16 18 20 22 24 26 28 30
c.o.c. hangers [m]
Fig. 67. Minimum on tensile force of the hanger: self-weight of the bridge.
From figure 67, considering the design criterion on a minimum hanger force, for the self-weight of the bridge, it
is found that:
Because of these ratios, when determining the cross section of the hanger, the fatigue strength is decisive in case
of a diagonal configuration and the static strength in case of a vertical configuration.
Diagonal + + 0 0 - + - 0 -
Vertical 0 0 0 - 0 0 0 0 -
Table 7. Comparison of influence number of hangers on member forces + is positive and – is negative.
Some results on bending moments for the arch and deck are shown in figures 68-69.
450000 450000
traffic full length + self weight
400000 400000
traffic half length + self weight
350000 self weight 350000
300000 300000
moment [kNm]
moment [kNm]
250000 250000
200000 200000
150000 150000
traffic full lenght+ self wieght
100000 100000
traffic half lenght+ self weight
50000 50000
self weight
0 0
16 18 20 22 24 26 28 30 16 18 20 22 24 26 28 30
c.o.c. hangers [m] c.o.c. hangers [m]
Fig. 68. Arch bending moments at first hanger connection. Left diagonal and right vertical hanger
configuration.
70000 70000
traffic full length + self weight
60000 traffic half length + self weight 60000
self weight
50000 50000
moment [kNm]
moment [kNm]
40000 40000
30000 30000
20000 20000
traffic full length +self weight
10000 10000 traffic half length +self weight
self weight
0 0
16 18 20 22 24 26 28 30 16 18 20 22 24 26 28 30
c.o.c. hangers [m] c.o.c. hangers [m]
Fig. 69. Maximum deck positive bending moments. Left diagonal and right vertical hanger configuration.
Fig. 70. Hanger configuration used as a basic model for studying the influences of “EI” on deflection and
member forces.
N M N M
Arch ++ 0 -- + +
Deck + 0 0 0 --
Table 9. Diagonal hanger configuration.
Effect of bending stiffness on deflection and member forces; + is positive and – is negative.
The influence of ‘EI’ on arch bending moment (at the position of the first hanger connection) is shown in fig. 71.
The influence of ‘EI’ on the deflection (at hanger – deck connection) is shown in figure 72.
600000 600000
bending stiffness arch
bending stiffness arch
Demka 9.135E+16
500000 9.135E+16 500000 2.84E+17
2.84E+17 7.035E+17
7.035E+17 9.135E+17
400000 400000
moment [kNm]
moment [kNm]
9.135E+17 1.1235E+18
1.1235E+18 1.5435E+18
300000 300000 3.28E+18
1.5435E+18
3.28E+18
200000 200000
100000 100000
0 0
1E+16 1E+17 1E+18 1E+19 1E+16 1E+17 1E+18 1E+19
bending stiffness deck p [Nmm2] bending stiffness deck [Nmm2]
Fig. 71. The influence of ‘EI’ on arch bending moment (at the position of the first hanger connection).
Left diagonal and right vertical hanger configuration.
3000 3000
bending stiffness arch bending stiffness arch
9.135E+16 9.135E+16
2500
2.84E+17 2500 2.84E+17
7.035E+17 7.035E+17
2000 9.135E+17 9.135E+17
1.1235E+18 2000
delfection [mm]
1.1235E+18
1.5435E+18 1.5435E+18
1500 3.28E+18
1500 3.28E+18
Demka
1000
1000
500
500
0
1E+16 1E+17 1E+18 1E+19 0
bending stiffness deck [Nmm2] 1E+16 1E+17 1E+18 1E+19
bending stiffness deck [Nmm2]
Fig. 72. Influence of bending stiffness on deflection.
Left vertical and right diagonal hanger configuration.
The influence of ‘EA’ on the maximum deflection of the deck for both hanger configurations considered is given
in fig. 73.
350 300
Demka
300 250
deflection[mm]
200 1.76E+11
1.95E+11 150
2.14E+11
150
3.91E+11 axial stiffness arch
100 9.77E+10
100 1.76E+11
1.95E+11
50
50 2.14E+11
3.91E+11
0 0
1.0E+11 1.5E+11 2.0E+11 2.5E+11 3.0E+11 3.5E+11 4.0E+11 4.5E+11 5.0E+11 5.5E+11 1.0E+11 1.5E+11 2.0E+11 2.5E+11 3.0E+11 3.5E+11 4.0E+11 4.5E+11 5.0E+11 5.5E+11
axial stiffness deck [N] axial stiffness deck [N]
Fig. 74. Geometry of the arch bridge crossing the Merwede at Papendrecht.
For the location half of the span the cross section properties are:
For the situation q = 10 kN/m, L = 202.8 m and E = 2.1?105 N/mm2 , the deflection is expressed by:
1.04875 ⋅ 1015
§ full length loading δ max = at ½ L;
I equ
3.899 ⋅ 1014
§ half length loading: δ max = at ¼ L.
I equ
Obtaining d max from computer analyses results into a value Iequa. Subsequently, by using the relation
I equ
η = , η-numbers are obtained which tells the consequence of hanger configuration on the global
I Steiner;ref
stiffness behaviour. Results on η-numbers considering four alternative hanger configurations are shown in fig.
76.
Vertical
deflection 25
[mm]
verticaal, volbelast
diagonaal, volbelast
50
netwerk, volbelast
diagonaal,halfvol belast
netwerk, halfvol belast
75
300
max [mm]
250
200
deflection
150
100
50
0
0 0,25 0,5 0,75 1 1,25 1,5 1,75 2 2,25
hanger area 1/A [*1/ Ao mm²]
Fig. 78 Relation between hanger area and deflection (hanger system 2 and 3 of fig. 69).
Ao = 5027 mm2 .
4 Rb 43
Rh
Rh
Rb
Fig. 79. Definitions used on radial converging hanger configuration.
Left: Rb /Rh = 0.5
Right: Rb /Rh = -1.0
Some results on bridge behavior (using the Papendrechtse bridge as reference) are depicted in fig. 80 – 81.
x/L [-]
-20 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0
20
zakking [mm]
40
60
80
100
120
140
RC[24, -1] RC[12, -1] RC[6, -1]
RC[24, -0.5] RC[12, -0.5] RC[6, -0.5]
Fig. 80. Deflection of the main deck girder in case of full length loading10 kN/m and negative Rb /Rh ratio.
(RC[24,-1] means hanger c.t.c. – 24 m and Rb /Rh = -1.0).
x/L [-]
-40
-20 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0
zakking [mm]
20
40
60
80
100
120
Fig. 81. Deflection of the main deck girder in case of full length loading10 kN/m and positive Rb /Rh ratio.
(RC[24,1] means hanger c.t.c. – 24 m and Rb /Rh = 1.0).
x/L [-]
-0.6
vert.,x/L=0.14
-0.4 vert.,x/L=0.32
relatieve doorbuiging w/L
-0.20.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 vert.,x/L=0.5
Rb/Rh=-1, x/L =0.14
0
[ m m /m - ]
Rb/Rh=-1, x/L=0.32
0.2 Rb/Rh=-1, x/L=0.5
0.4 Rb/Rh=1, x/L=0.14
Rb/Rh=1, x/L=0.32
0.6
Rb/Rh=1, x/L=0.5
0.8
For the three hanger configurations vertical / positive Rb /Rh ratio / negative Rb /Rh ratio, the smallest deflection is
found for Rb /Rh = -1.0, see fig. 82. The stiffness behavior under half length loading shows for the three
alternatives not that much difference.
REFERENCES
[1] M. Otten, Influence hanger configuration on the design of an arch bridge: thesis work TU-Delft, 2000.
[2] Palkowski, S., ‘Statik und Stabilität von Zweigelenkbögen mit schrägen Hängern und Zugband’.
Stahlbau, 7 (1987), p. 246-250.
[3] Palkowski, S., ‘Über die Stabilität von schubweichen Zweigelenkbögen’. Stahlbau 57,
(1988), p. 59-62.
[4] Palkowski, S., e.a., ‘Bemerkungen über die Stabilität von Bögen mit Hängern und Zugband’.
Stahlbau 63, 8 (1994), p. 248-251.
[5] Palkowski, S., e.a., ‘Stabilität von Bögen mit Hängern und konvexem Zugband’. Stahlbau 66,
(1987), p. 508-510.
[6] http://www.struct.kth.se/research/bridges
[7] Friot, D., Bonpas Tied-Arch TGV mediteranee high-speed rail. Bridge over the A7 motorway toll plaza.
Composite Bridges, Proc. 3rd International Meeting, Madrid, 2001.
[8] Van Bogaert, P., et al., Tied Arch Bridges with Radial Oriented Hangers subject to Rail Traffic, IABSE
2003, Antwerpen.
[9] Braga, F. et al., Bowstring bridges for high-speed railway transportation, IABSE 2003, Antwerpen.
[10] Del Forno, J-Y, et al., Steel and composite bridges designed by the bureau Greisch, Proc. 3rd
International Meeting, Madrid 2001.
[11] Prof. Kuhlmann, U, et al., University.of Stuttgart, Germany.
[12] Ramondenc P., et. Al. Les viaducs de Mornas et de Mondragon, Ponts metailliques No . 19 –199, pp. 95-
129.
[13] L. Do, Network and radial converging hangersystem: thesis work TU-Delft, 2004.
APPENDIX
The bridge shown in fig. 63, can be considered as an example of close co-operation between engineering office
and architect. The span of the steel arch is approx. 150m. Two series of 28 cables ensure the deck support and
stabilise the arch transversally. The width of the orthotropic steel deck is 25m, and the deck is constructed as a
box girder.
The behaviour of this bridge is purely a three dimensional behaviour, even under self-weight. The consequence
of the concept is the necessity to construct the deck with op four cambers (large torsional stiffness); see fig. 65:
§ a longitudinal camber, to compensate the elongation of the deck under the arch effects;
§ a vertical bending camber, classical for any bridge;
§ a horizontal bending camber, in the plan of the deck to compensate the deflections under horizontal tension
of the hangers due to their non equilibrated slope;
§ a torsion camber of one percent at the middle of the main span.
Fig. 86.
Deformed shape under permanent loads.