Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

The Lully Viaduct, A Composite Bridge With Steel Tube Truss

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Paper Number : 55

The Lully Viaduct,


a Composite Bridge with Steel Tube Truss
H.-G. Dauner 1, G. Decorges 1, A. Oribasi 1, D. Wry2
1

Dauner Ingnieurs Conseils - DIC SA, CH-1860 Aigle


Fribourg Cantonal Highway Administration, CH-1706 Fribourg

ABSTRACT
The innovative design of the Lully viaduct proposes a light and transparent structure made of a triangular
cross-section fabricated entirely from unstiffened circular tubes. The result is twin space trusses, with a typical
span of 42.75 m. Each transversal triangular cross-section is 2.9 m high and 4.0 m wide, and is supported by a
single slender pier. The largest diameters and thickness of the tubes are over 500 mm and nearly 70 mm
respectively. One major difficulty during the design of the connections was to define the stress distribution
along the complex intersecting perimeters of the tubes, and to calculate the hot spot stresses. Geometry
calculations, precision cutting and edge preparation of the tubes were necessary for performing full
penetration welds. The mobile formwork, which also ensures the stability of the tubular trusses during
concrete pouring, also required special consideration.
This paper describes the evolution of the project, from design, truss fabrication and welding, to construction
on site.

KEYWORDS
Aesthetic, bridge, composite bridge, space truss, circular hollow tubes, fatigue, stress distribution, stress
concentration factor, hot spot stress, weld.

INTRODUCTION
Located on the Swiss highway A1, the 1000 m long Lully viaduct, made of space tubular trusses, is the result
of an engineer design contest. This project was chosen by the jury for its originality and aesthetic quality. In
the past, no one dared to build a road bridge with welded tubular nodes due to the dynamic stress.

DESIGN CONTEST
Located near the village of Lully in the Canton of Fribourg, the viaduct is incorporated into highway A1
running from the East to West of Switzerland. Crossing a rural flat valley surrounded by wet land and trees,
this bridge will complete a highway link between Murten and Yverdon. The owner, the Fribourg Cantonal
Highway Office, had to choose between 3 projects submitted by selected experienced consulting firms.
The partcipants had to respect the following conditions:
Total bridge length : approximately 1'000 m
Width of the bridge deck : from 13.25 m to 16.00 m in each traffic direction
For maintenance reasons, it was decided to build two separate roadways
Longitudinal inclination : between 2.9 and 3.6 % in a concave circular arc with a radius of 40'000 m
Horizontal curve : circle of 3'000 m between 2 transition radius
Height over the valley : between 4 and 15 m.
The following projects were submitted to the 5 members of the jury :
One prestressed concrete box girder with an average span of 44.60 m and a constant depth of 2.50 m.
One prestressed concrete box girder with an average span of 42.50 m. This girder depth varied between 2
and 2.45 m.
One composite space truss with an average span of 42.75 m and a constant height of 3.75 m.
The third project was recommended by the jury for its lightness and transparency allowing it to integrate
into the countryside (Fig 1). The recommendation was approved by the responsible authority who accepted
the challenge of innovative design. This project was presented by the consulting engineers group DIC-DMA
(Dauner Ingnieurs Conseils SA and Devaud, Mongatti et Associes SA).

Fig 1 General view of the tubular composite bridge (photomontage)

CONCEPTUAL DESIGN
The shape of the surrounding trees inspired the author for the conceptual design [1]. A certain analogy can be
made with the steel construction where the elements size can be easily changed and adapted to live load
capacity [2].
Three different cross sections were considered :

Fig. 2

Cross section design evolution

The first alternative was the ideal design (Fig.2). It has slim cylindrical piers without bracing, but did not
respect the maintenance condition of two separated roadways. The third solution has been design to avoid
having piers dominating the surrounding trees. This included perpendicular truss connecting the two
longitudinal girders at the cross section on the piers. The result was a three-dimensional tubular truss
supporting structure.

PROJECT DESCRIPTION

Geometry
The dimensions of the trusses were based on equilateral triangles. Compared to a traditional box girder, the
truss depth is 50% higher. The slenderness (L/H) of this lightweight superstructure is approximately 13
instead of 20 as for regular beam girder (Fig 3).

Fig. 3 Longitudinal view and standard cross-section


The truss geometry was determined first by considering the span length as well as the maximum transportable
element. Then the tube diameters were given by preliminary calculations and the following considerations:
The diagonals governs the size of other members. Preliminary analysis leads to a diagonal diameter of 267
mm and wall-thickness between 11 and 50 mm. Adding to this the geometric conditions at the nodes, the
smallest diameter for the lower chord was 508 mm.

Fig. 4 KK-shaped joints geometry (lower chord)


It was beneficial to use the smallest possible tube in order to improve the force transfer between diagonals
due to the thicker walls . The thickness of the lower chord tubes varied between 25 and 50 mm (Fig 4). In the
support zone the thickness is increased from 50 to 70 mm. The thicker walled tube has a diameter of 559 mm
and a length of 2 m centered on the bearing. Unsightly stiffeners could therefore be avoided.
The upper chords nodes are less complicated (K-shaped joint) (Fig 5). The choice of the tube size depended
more on the considerations below than on the actual forces, which were carried mostly by the slab in the final
stage.

1. Nodal forces had to be equilibrated without taking the concrete resistance into account. Overlapping of the
diagonals was therefore chosen in order to transfer some of the vertical force directly between the
diagonals.
2. The tube diameter had to be large enough to provide adequate space for the welded shear connectors and
to allow minimum concrete cover.

Fig. 5 K-shaped joint geometry (upper chord)


One of the difficulties encountered in the fabrication of the space tubular structure was the welding at the
intersecting perimeters of the tubes and checking the penetration at the weld root. Even if the fatigue
requirements are not too severe in road bridge construction, the fear of uncontrolled or bad quality weld roots
in the tension tubes is understandable. For this reason, welding on backing shells was adopted. This increased
the width of the inner weld toe.
The brace truss at the piers were built with smaller tubes (Fig 6 and 7), as they are subject to lower forces.
The diameter of their chords varied between 219.1 and 323.9 mm, all diagonals are made of 168.3 mm
diameter tube.

Fig 6 and 7

Brace truss during erection and after completion

The deck width varied between 12.0 and 14.65 m. In order to limit the long-term deflection of the relatively
wide cantilever wings (4.0 to 5.33 m) and to minimize the weight of the deck, transversal tendons (600 mm2)
were used. The concrete deck was prestressed longitudinally with the same type of tendons. Longitudinally
the prestressing force has been chosen to insure compression in every section under dead load (concrete and
road surface).

Design
The bridge was designed according to the Swiss standard SIA 160.
Two different structural models were used for the calculation of the space truss:
1. Hinged diagonals with continuous chords for structural safety design
2. Rigid nodes for fatigue and serviceability design
The bridge deck had to be converted into a plane grid model (Vierendeel), made of two longitudinal chord
and vertical members placed at the intersection with the diagonals of the space truss (Fig 8).

Fig 8

Space truss model for static analysis during erection and final stage

Equivalent displacement theory was used in order to determine the stiffness of the longitudinal chord. Deck
stiffness, composite effect with upper chord and Saint-Venant torque resistance were considered in the
vertical member properties.

Joints
The structural safety of the members and joints was verified with the internal forces of the hinged model
(discussed in [3]).
No standard recommendations could be found to check fatigue resistance, or an empirical method for
computing local stresses at the circular hollow tube intersection, nor a fatigue category considering the
backing shell.
The following criteria were adopted after discussion with the owner and the expert engineers
1. Fatigue load according to SIA 160
2. Calculations of internal forces with
full and cracked composite section (with n = Esteel/Econcrete = 10)
rigid nodes

Fatigue
Fatigue effects are very important for this spatial tubular structure, especially at the welded joints. The
comparison between stress range and fatigue resistance is given by the well known formula :

E = fat
where

s fat

s R

:
:
:
:

R
fat

stress range under traffic loads


reduction factor for traffic loads
fatigue resistance for the considered detail
fatigue resistance factor

The problem specific to this bridge lays in the determination of the local stress range at the joints under traffic
load. It involves complicated analyses taking into consideration :
the geometry of the connection
the diameters of the elements
the width of the tube
All these factors are included in a "Stress Concentration Factor" (SCF), which multiplies the nominal stress s N
in the diagonal. It attempts to include both global geometrical stress concentration as well as the local stress
concentration associated with the weld.

fat = SCF N
with

s N = N/A

nominal stress (without bending effects)

The stress concentration factor compares the hot spot stress at a joint with the nominal stress in the diagonal.
The SCF is the ratio of the local stress at the weld toe to the nominal stress in the brace.

SCF =
with

sN
s HS
SCF

HS
N

: Nominal stress
: Hot spot stress
: Stress concentration factor

The SCF factor was obtained by comparing 3 different methods :


empirical formulas
finite element analysis
stress measurement with strain gauges

Empirical formulas
Equations for the distribution of SCF around the intersection of tubular K joint can be found in [7], Fig 9.

Fig 9 Geometry of a K-node


In the chord
0,666
SCFK = 1,506 -0,059
t 1,104 0,067 sin1,521

In the diagonal

0,157
SCFK = 0,920 -0,441
t 0,560 0,058 e1,448 sin

with

d
D

D
2T

t
T

g
D

No equations currently exist for stress distribution in KK joints [4]. For this type of connection the stress
concentration factor is higher and a corrected factor ? was applied to the K-joints stress concentration factor.
SCFKK = ? SCFK
with

? = 1,5 for chords


? = 1,1 for diagonals

This hypothesis was verified by finite element analysis.


Finite Element Analysis
Elastic finite element analysis ([8] and [9]) was performed to study the distributions of stresses along the joint
intersection. and through the thickness of the chord using thin shell element with no weld modelling. In this
analysis the SCFs were calculated from the maximum principal stresses at the mid-surface intersection.
For the K joints between diagonals and upper chords, the results of the empirical formula were consistent
with the FE analysis (Fig 10).
In the KK-joints the empirical formula results were consistent with the finite element analysis, especially for
the diagonals. The stress distribution in the chords given by the FE model, gives a lower hot spot value for the
circular tubes than for the rectangular tubes. Therefore the factor ? could be reduced (Fig 10).
These results have to be confirmed on the full scale structure, using the strain gauges.

Type of
joint

Element
Diagonal

K
Chord
Diagonal
KK
Chord

Fig 10.

Type of
joint

SCF
empirical

SCF
FEA model

(%)

3,3
2,9
2,6
2,0
4,5
4,2
3,9
3,4

3,1
2,8
2,0
1,9
3,4
3,2
1,6
1,6

6%
4%
30 %
5%

1.10
1.14
0.80
0.84

Comparison between empirical formula and finite element analysis

Element
Diagonal

K
Chord

Fig 10a

Solicitation
+ Tension
Compression

Solicitation
+ Tension
Compression

SCF
empirical

SCF
FEA model

(%)

3,3
2,9
2,6
2,0

3,1
2,8
2,0
1,9

-6 %
-4 %
-30 %
-5 %

Comparison between SCF empirical and FEA model for circular hollow sections and K joint

Type of
joint

Element
Diagonal

KK
Chord

Fig 10b

Solicitation
+ Tension
Compression

SCF
FEA model
K
3.1
2.8
2.0
1.9

SCF
FEA model
KK
3,4
3,2
1,6
1,6

?
FEA
1.10
1.14
0.80
0.84

Comparison between SCF FEA model for K and KK joint

Strain gauges measurement


To study the actual behaviour of the structure, the Fribourg Cantonal Highway Administration decided to
instrument a bridge span with about 100 strain gauges. They were placed in order to measure the nominal
stresses in the diagonals, and the stress distribution around the joints. Some gauges were also placed inside
the tubes in order to obtain the stress distribution on both sides of the tube wall.
We are currently waiting for the results of stress measurement in order to compare the empirical approach
and the finite element analysis.

Fabrication
The heart and soul of this bridge lay in the cutting edge diagonal facilities (Fig 11). The fabrication would
not have been possible without this equipment which cuts tubes lengthwise and prepares contact and welding
surfaces (Fig 12).

Fig 11.

Hollow tubes cutting process

Fig 12.

Diagonals with cut edges and backing shell

The truss was assembled on a frame enabling each segment length to be adapted according to the complex
geometry, taking into account the camber, the horizontal transition radius and the vertical circle. Two
adjacent sections could be preassembled on this frame. Diagonals joints were then welded manually and the
position of the entire section could be adapted to suit the welding conditions.
On site construction
Every pier stands on 4 bored piles at depths of 12 to 20 m. After the completion of the western abutment and
of 8 piers, the erection of the steel space truss could take place. The truss was assembled with crane trucks
from ground level. The truss sections weighted between 25 and 35 tons. These were delivered on low bed
trucks in lengths of 19 and 24 m. The building speed was of one full span every 2 weeks.
After positioning the truss up to the seventh pier, the mobile formwork was installed at the western abutment.
Every stage was half a span long and took one week. The form and reinforcement placement added to the
concrete pumping and hardening was moving at the same speed as the truss erection. The prestress of the
transversal and longitudinal tendons was applied just after the form removal.

Fig 13. The Viaduct of Lully lightness and transparency


CONCLUSIONS

The Lully viaduct using triangular tube-truss is a novelty in road bridge construction (Fig 13). Special thanks
are due to all the companies who were involved for their contribution to this high quality bridge. No major
problem occurred during the construction and some valuable experiences were gathered in tube truss
assembly and construction of variable width deck. The analysis of stress measurement will be beneficial to the
knowledge of the fatigue strength of KK shaped joint. Combined with new research in fatigue, it will certainly
promote the use of tubes for road bridge structure. Let us hope that enough financial resources will be found
in order to enable a real breakthrough in modern bridge construction.

Address of the authors


Dr. Hans-G. Dauner, civil engineer
Andr Oribasi, civil engineer
Guy Decorges, civil engineer
Dauner Ingnieurs Conseils - DIC SA, CH-1860 Aigle (Switzerland)
Tel. +41 (0)24 466 35 32
Fax +41 (0)24 466 35 33
Email : Dauner@compuserve.com
Denis Wry, civil engineer
Fribourg Cantonal Highways Administration, CH-1700 Fribourg (Switzerland)
Tel. +41 (0)26 305 38 79
Fax +41 (0)26 305 38 18
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]

Dauner, H.-G. : Der Viadukt von Lully - eine Autobahnbrcke, der Natur abgeschaut, Baukultur 6/96
Dauner, H.-G. : Stahlverbundbrcken im Aufwind, SI+A Schweizer Ingenieur und Architekt 26/97
Wardenier, Kurobane and others : Berechnung und Bemessung von Verbindungen aus Rundholprofilen
unter vorwiegend runder Beanspruchung, CIDECT/Verlag TV Rheinland, 1991
Goutie, Davies and others : Tubular Structures V/Proceedings of the Fifth International Symposium,
Nottingham, UK, 25-26 august 1993
CEDRUS 3 : Berechnung und Bemessung von Stahlbetonplatten nach der FE-Methode, CEDRUSSoftware, Zrich 1993
MICROSTRAN-3D : Berechnung von Stabwerken, Eng. Systems Limited, London/Sydney 1989
ARSEM . Guides pratiques sur les ouvrages en mer - assemblages tubulaires souds, Edition
TECHNIP, Paris 1985
NASTRAN : NASA FE-Program, Mac Neal-Schwendler Corporation (MCS), USA
ALGOR : FE Analysis System, Algor Interactive Systems, INC. Pittburgh 1990

You might also like