"Structural analysis and design of long span suspension bridges with

regards to nonlinearities, uncertainties, interactions and sustainability"

Bontempi, Franco ; Catallo, Luciano ; Sgambi, Luca

ABSTRACT

There is today a strong interest and general commitment toward the quality of the things that engineering is
creating. Generally speaking, quality involves and concerns different aspects of the engineering products,
like the aesthetics, the ergonomics, the comfort, the structural serviceability, the structural safety and the
durability. From this point of view, quality means, at the end, sustainability in the most general sense. All
these aspects should be considered in a modern conceptual design of a bridge structure. The structure
should be defined taking into account wide criteria and the critical aspects, which give clarity and coherence
to the whole structural design, must be strongly pointed out.

CITE THIS VERSION

Bontempi, Franco ; Catallo, Luciano ; Sgambi, Luca. Structural analysis and design of long span suspension
bridges with regards to nonlinearities, uncertainties, interactions and sustainability.Second International
Conference on Bridge Maintenance, Safety and Management (Kyoto, Japan, du 19/10/2004 au 22/10/2004).
In: Proceedings of Second International Conference on Bridge Maintenance, Safety and Management, 2004
http://hdl.handle.net/2078.1/182592

Le dépôt institutionnel DIAL est destiné au dépôt DIAL is an institutional repository for the deposit
et à la diffusion de documents scientifiques and dissemination of scientific documents from
émanant des membres de l'UCLouvain. Toute UCLouvain members. Usage of this document
utilisation de ce document à des fins lucratives for profit or commercial purposes is stricly
ou commerciales est strictement interdite. prohibited. User agrees to respect copyright
L'utilisateur s'engage à respecter les droits about this document, mainly text integrity and
d'auteur liés à ce document, principalement le source mention. Full content of copyright policy
droit à l'intégrité de l'œuvre et le droit à la is available at Copyright policy
paternité. La politique complète de copyright est
disponible sur la page Copyright policy

Available at: http://hdl.handle.net/2078.1/182592 [Downloaded 2023/08/29 at 19:57:15 ]

Structural analysis and design of long span suspension bridges
with regards to nonlinearities, uncertainties, interactions and
sustainability

F. Bontempi & L. Catallo & L. Sgambi

Department of Structural and Geotechnical Engineering, University of Rome “La Sapienza”, Rome, Italy

ABSTRACT: There is today a strong interest and general commitment toward the quality of the
things that engineering is creating. Generally speaking, quality involves and concerns different
aspects of the engineering products, like the aesthetics, the ergonomics, the comfort, the struc-
tural serviceability, the structural safety and the durability. From this point of view, quality
means, at the end, sustainability in the most general sense. All these aspects should be consid-
ered in a modern conceptual design of a bridge structure. The structure should be defined taking
into account wide criteria and the critical aspects, which give clarity and coherence to the whole
structural design, must be strongly pointed out.

1 INTRODUCTION

Speaking about engineering design, one can consider two different design strategies. In the first
one, the effects of the design environment are divided into two categories: general loading system
and specific loads, for example earthquake. The design solution is searched initially for the first
class of loads, and then corrected and improved by some additions. This is the classical application
of the dissection techniques for the solution of an engineering problem, with respect to the degree
of 1) understanding, 2) uncertainty, 3) modeling, 4) predictability, 5) computability, or, in general,
with respect to simple vs. complex systems. It is important to note that, when complexity is ne-
glected, there is the danger of an oversimplification of the problem, with the possibility to find bad
solutions. In the second case, instead, one finds the control of all actions by design: all the aspects
of the structural problem are considered together, and the synthesis of the design is fully intercon-
nected. It is clear that this design process is more complex, due to the difficulties in handling the
problem, from the mechanical point of view and/or from the economical point of view. Of course
the coherence of the final design is enhanced, as shown by the lack of asymmetry.
There is today a strong effort to optimize the engineering design by the so-called soft-
computing techniques, to tackle all this complexity. This approach, pursued in this research by the
Research Unit of the University of Rome “La Sapienza” within COFIN2002, is a pervasive ori-
entation of the design (global strategy) to an optimal behavior that starts from the conception of
the structure. In the references, one will find the context where, at the same time, this research
unit has its cultural basis and toward which will evolve to consider the nonlinearities (Fig.1), the
interactions (Fig.2) and the uncertainty previously recognized (Calzona & Bontempi, 2002;
Silvestri & Bontempi, 2003; Silvestri, 2002; Smith, I. 2001). This research group in the last
three years furthermore takes effective advantage from being strongly involved in the consultant
activities related to the design and the analysis of the Messina Strait Bridge.
 Cables

Towers Geometrical nonlinearities

Hangers
960 3300 m 810
777 183 3300 183 627
+383.00 +383.00

+77.00 m +118.00
+54.00 +52.00 +63.00

Dissipative Device
Material nonlinearities
Soil Behavior

Soil-Structure contact Boundary nonlinearities

Figure 1. Some relevant nonlinearities considered in the modeling of the long span suspension bridge.

Vehicle-Structure Interaction

960 3300 m 810

777 183 3300 183 627
+383.00 +383.00

+77.00 m +118.00
+54.00 +52.00 +63.00

Wind-Structure Interaction

Soil-Structure Interaction

Figure 2. Interaction phenomena considered in the modeling of the long span suspension bridge.

The performance analysis for the evaluation of the reliability of complex structures like the
Messina Strait Bridge, generally results from three main points:
1. Formulation of the structural problem and organization of the relevant data and infor-
mation;
2. Structural modeling and numerical analysis;
3. Synthesis of the results and critical evaluations.
The inherent complexity in planning and designing this kind of structures leads to the neces-
sity of a new design statement in the framework of the so-called Performance-Based Design
(P.B.D.), which aims at defining preventively and accurately the performances requested to the
structure (Fig.4). Absolutely critic it appears the ability to predict the performance required for
such structures and the necessity to provide the qualitative and quantitative elements able to
support the following decisional processes needed to specifically define such an important struc-
ture, including both limit and service behaviors. The performance approach to the structural de-
sign is based on the verification of the exact suitability of the structure with respect to fixed req-
uisites (Catallo & Biondini & Bontempi & Frangopol, 2004; Albertini, & Sgambi, & Garavaglia, 2003).
2 STRUCTURAL MODELING

The first point to face concerns the general consideration about all the activities connected with
the construction of the model: it was just recalled the central role played by modeling in the
overall performance based design. The modeling process consists of a net of theories, decisions,
and operative aspects that make possible to extract useful information about the system studied.
The existing data and information are organized in the real system model, obtained starting from
the reality through a reduction process, maintaining the useful aspects and avoiding to consider
the other ones, as schematically quoted on Figure 3: this to end to strictly define the perform-
ance characteristics of the structure (Figure 4).
2.1 The Complexity in the design of long span suspension bridges
For a long span suspension bridge, like the Messina Bridge, which presents unique and excep-
tional characteristics, because of complexity of the system (Fig.5), in order to take into consid-
eration all aspects (specifically the critical aspects of Figure 6), it is necessary to adopt a de-
composition strategy. In this manner, the structural problem and each relevant aspetcs are
hierarchical organized in several levels that can be singularly dealt with. The decomposition can
be performed for different aspects: Performance, Structural Organism and Load (Bentley, 1999;
Simon, 1980; Perrow, 1999; Fellows, R. & Liu, A. 1997).
It is clear that the decomposition is made for several aspects of the whole structural system,
such as design quality, product quality and serviceability. The whole system can be firstly di-
vided into substructures (macro-level), then into components (meso-level), and finally into ele-
ments (micro-level). Consequently, it is possible to modify each variable and optimize the struc-
tural behavior in order to achieve a specified performance objective (Bontempi & Biondini &
Radogna, 2002; Calzona & Bontempi, 2001).
The analysis of a complex structural system hardly can be reduced to a single phase only.
Generally in fact there are inside the structure two classes of regions with strongly different me-
chanical behavior. One considers the so called B-regions, where the stress picture is congruent
to a simple strain picture; the B letter, stems from Bernoulli, and D-regions, where the absence
of a simple cinematic behavior involve complex stress picture; in this way there are regions with
diffusive stress, by which stems the letter D.

Figure 3. Modeling process from theory to global knowledge to specific practice.

 It appears necessary to divide the structural system analysis in several phases, hierarchically
organized as:
- Global analysis;
- Local analysis at a first under level;
- Local analysis at a second under level;
- ….
The layout of this multilevel analysis process, can be formalized through the substructuring
concept, with the first analysis level executed with numerical models able to gain the global as-
pect of the structural behavior (for example, with Bernoulli beam elements). It is important to
note that (Figure 6):
- All the connections of the structural elements are configured as D-regions;
- The global behavior of the structure is the result, from a macroscopic point of view, of the
integration of the local behavior of these D-regions;
- The local possible crisis, can be dangerous for the integrity (robustness) of the overall
structural organism and this must be avoided with an accurate design;
- The local behavior, particularly the local strain, can lead to an unacceptable behavior (for
example, regarding the second order effects).
It must be noticed how the complexity of structures such as the Messina Bridge results just
from the matching and the interactions peculiar of this system at local behavior: in fact, it is
possible the arise of secondary effects to jeopardize the design. The development of these
mechanisms in fact, must be identified by the modeling and opportunely dominated in the de-
sign strategy (Sgambi & Biondini & Bontempi & Frangopol, 2004).

REQUIRED LIFE TIME

FAILURE
QUALITY

REQUIRED
PERFORMANCE

TIME

RELIABILITY

Figure 4. Framework for the performance identification of the structure.

NONLINEAR
TIGHT
BEHAVIOR
COUPLING
INTERACTIONS
CONNECTIONS LINEAR

LOOSE
LOW

AMBIGUITY
UNCERTAINTY

HIGH

Figure 5. Complexity of the structural behavior as result of nonlinearity, coupling and uncertainty.

2.2 Exploring the model and handling uncertainties

For structural systems with non linear behavior a realistic description of the response under all
load levels can be obtained only by taking the actual non linearity into account. In this context,
thought the reliability of the structure as resulting from a general and comprehensive examina-
tion of all its possible failure modes, one must pay attention to the following four aspects which
define the assessment process:
- Available data;
- Modeling of the uncertainties;
- Nonlinear structural analysis;
- Synthesis of the results.
On this subject, different approaches have been developed, more of them follow a probabilis-
tic approach. The application of these methods need the definition of the problem in a probabil-
istic way and require the exact distributions modeling the stochastic random variables involving
in the structure. But, sometimes, this is not simple because of the physic of phenomenon studied
can be complex and not describing, in a correct form, as a stochastic process. Another limit of
more of these approaches is the large number of simulation required to obtain the exact result.
As a consequence a high computational effort is needed. With a limited number of simulations
just an estimation of the exact result is possible. Some methods to optimize these procedures are
proposed in literature.

GRAVITIES OF FAILURES
LEVEL OF STRUCTURAL SERVICE- SAFETY SAFETY OF
FAILURE: DAMAGE ABILITY OF THINGS PERSONS
INFLUENCES
MINOR ON MAY BE - -
STRUCTURAL
COMPONENT
MAJOR ON YES MAY BE -
STRUCTURAL
ELEMENT
SEVERE ON (LARGE) YES YES MAY BE
STRUCTURAL
PART
CATA- ON YES YES YES
STROPHIC STRUCTURAL
SYSTEM

Figure 6. Gravities of failure for a complex structure like a long span suspension bridge.

The fuzzy theory is another method to investigate the uncertainties. The fuzzy logic is able to
implement in computer code, through formal systems, also linguistic definitions of complex
concepts (Petrilli, 2003). This means that every experimental information can be assumed by the
method and the problem to be solved concerns with the exploiting of the maximum information
from the trials. Therefore, with the adoption of a strategic planning of simulation, the non linear
behavior of a structure can be investigate, this taking in account a large number of uncertainties
and with a reasonable computational effort.
The fuzzy logic is based on the definition of a membership function describing the parame-
ters of the problem. The membership function represents a set of possible values that the pa-
rameter can assume, but not the probable values; therefore, the shape of the function can be cho-
sen on the basis of an expert judgment and not on a probabilistic distribution of the parameter
values (Biondini & Bontempi & Malerba, 2000).
Let make an example: let p be a parameter belonging to the set of quantities, which define the
structural problem and λ a load multiplier. It is clear that to each set of parameters corresponds a
set of limit load multiplier, one of them for each assigned limit state. For sake of simplicity, one
starts the considerations by considering the relationship between one single parameter p and one
single limit state defined by its corresponding limit load multiplier λ. At first, it is worth noting
that, in general, such relationship is non-linear even if the behavior of the system is linear. This
is typical of the design process where the structural properties, which correlate loads and dis-
placements are considered as design variables. Thus, the non linear law λ=λ(p) can be drawn as
in Figure 7-a, which shows that for each value of p, there is a corresponding value of λ. How-
ever, from Figure 7-b it is also clear that the response interval [λmin; λmax] corresponding to
[pmin; pmax] cannot be simply obtained from λ(pmin) and λ(pmax). The problem of finding the in-
terval response can be instead properly formulated as an optimization problem by assuming the
objective function to be maximized as the size of the response interval itself. Better speaking,
the bounding of the structural response interval can be framed as an anti-optimization problem.
In particular, for the general case of n independent parameter p, collected in a vector x=[p1;
p1;…. pn], and m assigned limit states, the following objective function is introduced:
m
F ( x) = ∑ (λ i , max − λ i ,min )
i =1 (1)

(a) (b)
Figure 7. (a) Relationship between a structural parameter p and a limit state load multiplier λ;
(b) Interval of the limit state multiplier λ corresponding to an interval of the parameter p.

Moreover, a sensitivity analysis should be conducted in order to investigate the proneness of

output results to changes in the input parameters due to uncertainty and/or vagueness, as shown
in Figure 8.

(a) (b)
Figure 8. Sensitivity of the structural response: a) a small change in input parameters leads to a large un-
certainty in the output results; b) the structural response is not affected by input’s vagueness

3 STRUCTURAL ANALYSIS

Static and cinematic parameters have been chosen carefully in order to understand the real be-
havior of the structure toward the performance condition preliminarily identified as prominent.
This complex analysis can be supported by the knowledge of other experiences concerning
other bridges already built in other countries. Relating to the long span and to the traffic typol-
ogy, some of the performances have been analyzed and then adopted after a phase of correction
and improvement both during the design and the construction phases.
The functional run ability performances or the skill ness of the bridge to execute safely the
correct functions of roadway and railway connection, concern both static and dynamic aspects:
they have to be compatibles with the railway and roadway traffic safety. Adequate comfort
standards must also be guaranteed.
The roadway traffic conditions are strictly connected to the complex strain path generated by
the load combination expected in the exercise phase, varying with the length and with the
stresses generated by the roadway and railway traffic. Essentially, they are formed by the trans-
versal slope (sum of the drainage initial slope and the rotation induced), by the vertical curva-
ture radius and by the vertical and horizontal accelerations induced on the deck. Relatively to
the railway run ability, the longitudinal slope, the rail torsion and the respective variations must
be also added.

Figure 9. Model of the long span suspension bridge.

Figure 10. Deformed shape of the bridge and state of stress picture.

Each of these aspects, which influence the run ability, identifies one or more realistic and se-
vere load combinations under which the structure, modeled and calculated with reliable pro-
grams, must satisfy the performances attended. Among the actions potentially dangerous toward
the run ability, great attention must be paid in analyzing the wind effects, which can really in-
fluence the safety and the comfort of the users. Relating to the average economic damage at-
tended by every hour of closure for a year, it’s necessary to quantify the Bridges performances
with respect to this event, namely the attended number of hours in which the bridge will be
closed or open with a series of traffic limitations for the different vehicular categories.
In order to analyze the behavior of the long span suspension bridge, specifically regarding the
stress state evaluation, the model of the bridge deck close to the towers (Fig.9) is refined using
two-dimensional finite elements working according to the shell theory. This zone, built with
special elements, can be considered as a large scale D-regions, needing detailed modeling to ob-
tain the realistic description of the behavior of the deck. Figure 10 shows the state of the stress
picture in the deck (Catallo & Sgambi & Silvestri, 2003; Catallo & Di Mella & Silvestri, 2003).
4 ACKNOWLEDGEMENTS

The financial support of Stretto di Messina S.p.A. and of COFIN2002 of the University of Rome
La Sapienza is acknowledged. The opinions and conclusions presented in this paper are those of
the Authors and do not necessarily reflect the views of the sponsoring organizations.

5 REFERENCES

Barberi, V. & Ciani, M. & Catallo, L. 2003. Evaluation and results’ comparisons in dynamic structural
response of Messina cable-suspended bridge. Proceedings of Second International Conference on
Structural Engineering and Construction (ISEC-02), 23-26 September 2003, Rome, Italy.
Bentley, J.P. 1999. Reliability & Quality Engineering. Addison Wesley England.
Biondini, F. & Bontempi, F. & Malerba, P.G. 2000. Fuzzy Theory and Genetically-Driven Simulation in
the Reliability Assessment of Concrete Structures. Proceedings of 8-th Conference on Probabilistic
Mechanics and Structural Reliability. 24-26 July 2000, Notre Dame, USA.
Blockley, D.I. 1980. The nature of structural design and safety. Ellis Horwood Limited.
Bontempi, F. & Radogna, E.F. & Biondini, F. 2002. La modellazione realistica delle strutture esistenti in
c.a./c.a.p (in italian). Giornate A.I.C.A.P. 2002, Bologna, Italy.
Calzona, R. & Bontempi F. 2001. Remarks on the approval process of designs of structures provided
with innovative anti-seismic systems in Italy. 7th International Seminar on Seismic Isolation, Passive
Energy Dissipation and Active Control of Vibrations of Structures. Assisi, Italy.
Catallo, L. & Biondini, F. & Bontempi, F. & Frangopol, D.M. 2004. Robust Reliability- Based Design
Optimization of Suspension Bridges. PSAM 7 / ESREL ’04, Proceedings of the 7th International Con-
ference on Probabilistic Safety Assessment and Management, 14-18 June 2004, Berlin, Germany.
Catallo, L. & Di Mella, V. & Silvestri, M. 2003. Aspects for the determination of the complex stress
states in suspension bridge for the fatigue-analysis. Proceedings of Second International Conference
on Structural Engineering and Construction (ISEC-02), 23-26 September 2003. Rome, Italy.
Catallo, L. & Sgambi, L. & Silvestri, M. 2003. General aspects of the structural behavior in the Messina
Strait Bridge design. Proceedings of the Second International Conference on Structural and Con-
struction Engineering (ISEC-02), 23-26 September 2003, Rome, Italy.
Fellows, R. & Liu, A. 1997. Research Methods for Construction. Blackwell Science.
Frangopol, D.M. 1995. Reliability-based optimum structural design. Probabilistic Structural Mechanics
Handbook. Edited by C. (Raj) Sundararajan, Chapman & Hall, New York, 1995, 352-387.
Giuliano, F. & Taddei, D. 2003. Conceptual framework for the aerodynamic optimization of the long
span bridge deck sections. Proceedings of the Second International Conference on Structural and
Construction Engineering (ISEC-02), 23-26 September 2003, Rome, Italy.
Imai, K. & Frangopol, D.M. 2002. System reliability of suspension bridges. Structural Safety 24 (2002).
Elsevier Science.
Loreti, S. & Salerno, M. 2003. Organization and decomposition of a complex structural project. Proceed-
ings of the Second International Conference on Structural and Construction Engineering (ISEC-02),
23-26 September 2003, Rome, Italy
Perrow, C. 199. Normal Accidents. Living with High-Risk Technologies. Princeton University Press, New
Jersey.
Petrilli, F. 2003. A hybrid probabilistic and fuzzy model for risk assessment in a large engineering pro-
ject. Proceedings of the Second International Conference on Structural and Construction Engineering
(ISEC-02), 23-26 September 2003, Rome, Italy.
Sgambi, L. & Biondini, F. & Bontempi, F. & Frangopol, D.M. 2004. Handling Uncertainties in Optimal
Design of Suspension Bridges with Special Emphasis on Loads. PSAM 7 / ESREL ’04, Proceedings
of the 7th International Conference on Probabilistic Safety Assessment and Management, 14-18 June,
2004, Berlin, Germany.
Silvestri, M. & Bontempi, F. 2003. Strategy and formulation levels of the structural performance analysis
of advanced systems. Proceedings of the Second International Conference on Structural and Con-
struction Engineering (ISEC-02), 23-26, 2003 September Rome, Italy.
Silvestri, M. 2002. Gli aspetti concettuali del Performance-based Design e la loro applicazione nel pro-
getto dei ponti sospesi (in italian). Graduate thesis, Faculty of Engineering, University of Rome “La
Sapienza”, Department of Structural and Geotechnical Engineering.
Simon, H.A. 1998. The Sciences of the Artificial. The MIT Press, Cambridge.
Smith, I. 2001. Increasing Knowledge of Structural Performance. Structural Engineering International,
12 (3), 191-195.