Bridge strengthening by conversion to network arch.

Design criteria and

economic validation

Matías A. Valenzuela* and Joan R. Casas

Public Works Department, Santiago, Chile – UniversitatPolitecnica de

CatalunyaBarcelonaTech, Barcelona, Catalunya-Spain

Morandé 71, [email protected], phone: +56956374031

Bridge strengthening by conversion to network arch. Design criteria and
economic validation

The paper presents a new strengthening methodology in bridges over river beds affected by
scour and erosion at piers. The proposal is part of an innovative concept, implementing a
structural change of the original bridge, without the need to strengthen the substructure,
thanks to the construction of an upper arch with network and vertical hanger's. The vertical
hangers are responsible to lift the deck from the piers, generating a single span (simply-
supported tied arch). The paper describes the construction phases, considering the
conditions and difficulties in their implementation due to multi-objective targets and
additional boundary conditions and requirements. Additionally, it gives a description of the
structural behavior and a layout of each element, to use as criteria or guideline for future
applications of the method. Finally, a validation is made through a comparative analysis in
a Chilean bridge, between a traditional method of strengthening by additional piling in the
foundations and the proposal here presented. Additionally to the feasibility of application,
the example shows that the new solution is cheaper.

Keywords: network arch; strengthening; scour;hanger; lifting

1.Introduction

One of the major problems in bridges over riverbeds is the scour in the foundations. The

consequences are of varying severity, but certainly the most important is the total collapse of the

structure, generating high human and economic costs. As a result, they have been implemented

several measures of inspection, monitoring and maintenance, not properly applied at underwater

infrastructure conditions. In addition, the action of accidentals loads (earthquake) and the poor

support conditions produce a highly vulnerable and risky scenario for these bridges.

In Andean countriessuch as Chile, one of the most common problems is the undermining

of the piers (Muñoz, 2006; Seaurz, 2006). The repair and strengthening are frequent and generate

high economic losses. The damage caused by the flow streamin Andean rivers is important

because they are of torrential nature, able to move large and heavy rocks. These factors induce a
high risk in the piers, which are eroded in their foundations and can be themselves fully

destroyed by the passage of the water flow.

Figure 1. Scour at bridge piers.

The condition of the foundation because of the scour, leaving the piles out of the ground

as shown in Figure 1, is extremely dangerous in countries with high seismic activity, requiring a

constant monitoring, repair and strengtheningbyadditional piles (steel or concrete) and screeds.

For this reason, the Ministry of Public Works of Chile, through the Department of Bridges has

promoted the study and development of an alternative strengtheningmethod for this type of

damage.

Mitigation mechanisms have been developed, which often extend the service life of the

structure. However, many of them present difficulties in their implementation (require qualified

personnel and are very expensive).Among the traditional repair methods are those used for the

protection of the channel and bridge elements. Frequent techniques correspond to: the armoring

protection around the piersand the river bed (Figure 2), dam construction, the increase of the

span of the bridge and constant monitoring of the scour. These techniques, in emergency cases,

improve the conditions of the bridge. However, they present the disadvantage of physical

changes to the river flow, and the requirement to work in complex access conditions, deriving on

high costs of materials and labor.

Figure 2. Protected armoring. [Courtesy of B. Moya]
Other protection techniques are the improvement of resistance of elements and structural

support. As an example, the concrete screed, mortars using mixed cement with sand, gravels and

aggregates (bagged concrete or fluid micro-concrete) improve the strength and structural support

at the base level (Figure 3). When the foundation is under water, the application is through pipe

systems or injecting a concrete grout to fill any existing hole. The main drawback is the

geometrical increase of the area exposed to the river current and the changein flow patterns,

inducing future erosion problems.

Figure 3. Concrete screeds.[Courtesy of B. Moya]

Finally, to improve the bearing capacity, the most commonly technique used is the

strengthening by additional piles (Figure 4). These enable to transfer the loads to a capable layer

of soil when it is at a considerable depth. It is a common technique which requires a detailed

study of soil stratigraphy and geology, looking for a good support layer. Despite its good
performance and wide variety of techniques (kneeling, percussion, excavated) and materials

(steel, concrete, and wood) this method presents a number of drawbacks, including:

• Difficulty of implementation (limited access of the equipment, vertical clearance,

inclined piles).

• Depth of foundations (Placing reinforcement and concrete joints, false rejections in the

driving, presence of thin hard layers).

• Incompatibility between construction method and soil type (e.g. driving in the presence

of coarse gravel, on-site method in locations not permitting shoring and movement of

material).

• Durability of the material. (Steel corrosion or wood rottenness in unsaturated zone).

Figure 4. Strengthening by piles.[Courtesy of B. Moya]

Another technique is to strengthen by micro-piles. They are used in areas where is not

possible to implement conventional piles by difficulty in the access. Among its advantages, is the

fast implementation and minimal discomfort at worksite.However, the micro-pile bearing

capacity is lower than that of the piles, requiring a large number of them, thus increasing the

cost. Also, it is not usable in very deep foundations, and does not provide a good response to

lateral actions in the presence of a seismic event. Normally, the strengthening sequence is as
follows: After the collapse of the piers and placement of provisional supports, the superstructure

is released from the piers, and it is lifted by jacks. Then, the damaged part is demolished, and the

foundations strengthened by piles or driving caissons. Finally, the pier is rebuilt and the

superstructure is placed again into place (Valenzuela, 2012). This technique improves the

bearing capacity of the piers and locates the foundation level below the general and local scour,

delivering anoptimum configuration. However, the interference of current flow is not eliminated,

requiring future inspections and monitoring of the area.

In Malerba (2014), some examples of problems in bridges due to scour and the

techniques used to strengthen the foundations are presented.

2. New strengthening methodology

A new strengthening method of short andmedium-span bridges, with several continuous spans is

proposed. The method performs the modification of the static scheme (from continuous beam to

tied-arch) by building an upper-deck arch with a network hanger arrangement, in which the

existing deck becomes the tie of the system. This allows the removalof the intermediate supports,

generating a single span, thereby eliminating anyfuture scour problems (Valenzuela, 2012;

Valenzuela & Casas, 2013). The main characteristic of network arch bridges is that the inclined

hangers cross each other at least twice. They have demonstrated an excellent performance for the

construction of new bridges (Tveit, 2007; Valenzuela, 2012; Pircher et al., 2013). In the present

case, the network arch is used in the strengthening of an existing bridge.

This new method of strengthening is sustainable due to the decrease of the repair costs,

avoiding repetitive and expensive repairs and strengthenings in the foundations and piers.

Additionally, it delivers a new significance to the repair concept, providing an enhanced

aesthetic solution. The proposal is part of a philosophy of repair and strengthening of bridges

with emphasis on the reuse of components, allowing structural and economic benefits and

providing additional aesthetic value.

 Based on the analysis of the construction process, the method proposes a verification

associated to the definition of and upper and lower bound and an acceptance bandwidth based on

these bound values. The method can be applied to bridges with little initial information, i.e., lack

of detailed drawings, steel reinforcement, and types of materials, through the study of admissible

ranges of stress (MAB) (Valenzuela, 2012; Valenzuela & Casas, 2013).The range corresponds to

threshold values obtained by the analysis of theactual stressesin the bridge under analysis. To

this end, based on the actions present in the existing bridge, the actual state of stressin the deck is

obtained. These maximum and minimum limits in the service state can not be overpassed by the

existing deckduring the different phases of the strengthening application and later on in the new

service phase of the strengthened bridge.

The initial step in the strengthening methodology corresponds to the calculation of the

maximum stress state in the original deck over piers due to the actual loading conditions and

feasible load combinations. If the information gathered is insufficient, a geometric survey and

material coredrilling and testing may be required to up-date the limited available information.

The proposed newstrengthening procedure is divided into three phases, as follows:

2.1. Preparation Phase.

During this phase is evaluated the need to improve the deck geometry according to current

standards (e.g., incorporating an extension of the carriageway, enlarging the width of the existing

lanes,…) or because of the additional space required by the strengthening process itself, for

instance, the reconstruction of the carriageway with the necessary dimensions to deploy the arch

at the center of the cross-section. It should be also assessed the scour condition and resistance

capacity of the abutments to accommodate the additional vertical force transmitted in the new

static scheme. (Figure 5.a).

Figure 5.a.Modification in superstructure (a)
The changesthat will be applied to the deck require the use of temporary towers on the

original deck, located over each pier to accommodate the arch segments. The arch segments are

placed over the towers through the use of cranes. The number of arch segments will depend on

the crane capacity (Figure 5.b.).

Figure 5.b. Construction of the upper arch (b).

After the deployement of the arch segments, the longitudinal beams composing the deck

cross-section are unlinked from the piers by cutting them (e.g., through diamond wire), leaving

the superstructure simply-supported on the piers. Then, it is necessary to incorporate an element

acting as the tie of arch between the two abutments of the bridge. This is accomplished by the

use of external prestressing with straight tendons from abutment to abutment (Figure 5.c.).

Figure 5.c.Externalprestressing in the deck (c)

 After post-tensioning the deck with the external tendons, the temporary towers can be

removed, avoiding excessive strain in the arch. Later on are deployed two groups of passive

hangers (network and retention), and the vertical hangers without applying any tensioning

(Figure 5.d).

Figure 5.d.Deployement of network hangers (d).

The network hangers follow a configuration such that they cross each other at least twice

(Tveit, 2007). The retention hangers correspond to the set of hangers placed near the junction

between deck and arch, to regulate the negative bending moments in the deck at this point.

After the placement of the passive hangers, the first phase of the construction is

completed. It should be noted that in these early stages, the requirement to be fullfilled is: "avoid

that the stress at any point, in any of these stages, exceeds the limits set by the MAB". Following

this recommendation, the critical point is normally located in the deck, since this is the zone

where the reference stressbandwidth (MAB) is more restrictive.

2.2. Lifting Phase.

The second phase corresponds to the tensioning of thethe vertical hangers, which are responsible

for gradually lifting the deck from the piers, as shown in (Figure 5.e).
Figure 5.e. Tensioning of vertical hangers (e)

The proposed sequence of tensioning considers the introduction of the total force in the

vertical hangers by only one jacking operation, reducing the number of tensioning operations to a

minimum to optimize the cost. As a basic arrangement, the vertical hangersare placed over each

original pier. At this point, the sequence and magnitude of the tension in each vertical hanger to

achieve adequate lift of the deck should be defined. Thestructural analysis starts from the stress-

strain state obtained at the end of the preparation phase, and follows the guidelines of the

criterion of MAB (Valenzuela, 2012; Valenzuela & Casas, 2013).

The lifting phase of the deck is the most complex and delicate operation throughout the

process of bridge strengthening. It involves a large number of parameters and variables that

affect the response of the structure and define the success of the operation, considering all the

stress changes in the deck due to the introduction of the force in the hangers. In this stage is

when, in fact, the change of the structural longitudinal configuration (from continuous beam to

arch) occurs. For this reason, this phase is considered fundamental in the strengthening process.

The major computational efforts are present in this phase, as the tension introduced in the

hangers must lift the deck and leave it in an allowable stress state during both the construction

and the in-service phases. Therefore, it is necessary to obtain safe and acceptable values of the

jacking sequence and the magnitude of the jacking force in the hangers, through an optimization

tool, solving a multi-objective optimization problem by means of genetic algorithms

(Valenzuela, 2012; Valenzuela & Casas, 2013). The algorithm is formulated through the

combination of discrete and continuous variables, without an explicit objective function. It

optimizes two variables: order of tensioning and magnitude of the force to be introduced in the

vertical hangers. The optimization seeks to reduce the differences between the original stress on

the deck and those after each tensioning phase to lift the superstructure, allowing the lifting of

the deck over each pier and maintaining a state of allowable stresses in all bridge elements (old

and new). In addition, the optimization also searchs to use the minimum amount of materials and

minimum number of jacketing operations (economical constraints) (Valenzuela & Casas, 2011;

Goldberg, 2002).

2.3. Network phase.

If necessary, in this phase the tensioning process in each hanger network is defined. Only the

network hangers that require an initial tensioning will be stressed. This is defined according to

the structural analysis of the bridge during the in-service operation. The location of hangers

corresponds to a spacing not greater than 4 meters at the arch and a symmetrically configuration.

Considering this, the tension of these hangers meets the needs of an efficient structural

performance in the service stage (Valenzuela, 2012).

Finally, when the deck is completely supported by the hangers, the piers are demolished

(Figure 5.f) and the new network arch configuration is obtained.

Figure 5.f.Removalof the piers (f).

3. Design criteria

The basic criteria to define the elements used inthe strengthening method include the arch,

hangers, and the modification of the deck. Based on a great number of calculations and
parametric studies (Valenzuela, 2012), it has been possible to deliver a summary of these design

criteria. According to the significance coefficients obtained from multivariate studies

(Valenzuela, 2012; Valenzuela & Casas, 2013), the greatest impact upon the correct performance

of the strengthening comes from the type of profile used in the arch, the external prestressing

layout and the weight of the deck. In the following, the proposeddesign criteria based on a

parametric study are presented. Further information on how they are derived can be seen in

(Valenzuela, 2012; Valenzuela & Casas, 2013).

3.1. Arch

(1) Behavior

The arch is the main design element. It is proposed the use of a single arch centered and placed

over the longitudinal beam, seeking for a reduction of material and number of operations. The

internal forces in the arch are composed of staggered axial forces and bending moments in the

plane, similar to the behavior of a continuous beam,i.e., it is composed by negative peaks at the

points of application of the vertical hangers. However, these peaks have a decrease along length

due to the secondary effects of alternation of the bending moments provoked by the network's

hangers.

The bending moments in the plane of the arch are similar or slightly greater than the out-

plane ones. Thus, it is recommended to use symmetrical profiles. The minimum square profile

recommended is higher than the profiles used on a new construction projects. From this, the

relationship between edge arch profile and total span is higher than for new construction projects

(Valenzuela, 2010; Rongish, 2011).

When the deck has a significant width, it is necessary to use a double arch. In these cases,

the inclination of the arches is a subject to study in the future due to the fact that it induces an

axial effort in the transverse elements of the deck.

 The rise-to-span ratio is relatively high, with a moderate participation in the lifting

process and in the reduction of the effects of embedment between arch and edge of the deck.

(2) Design Criteria

Based on the explained behavior and the results of the parametric study, the following design

criteria were obtained:

• When using a single centeredarch centered, it is recommended a symmetrical profile

(same inertia in the two principal directions). When using a double arch solution, it is

recommended to use arch profiles with greater inertia in the plane of the arch.

• The ratio between the inertia modulus of the arch and the longitudinal beams of the deck

should not exceed 75. This improves the ability to lift the deck without inducing a failure.

Inertia Beam
RB / A = ≤ 75
Inertia Arch (Equation 1)

• A circular arch shape is recommended.

• A rise-to-span ratio in the order of 0.14 is recommended in those cases where the wind

effects are very important. However, in the cases where the criterion has been the profile

optimization, the rise-to-span ratio should be in the range 0.16 to 0.17. It is important to

never exceed the value 0.2.

• Due to the selected profiles, the relation cross-section depth / span should not be less than

1/127.

With these design criteria, the buckling of the arch both in and out of plane is avoided, as shown

in the analysis carried out by a Finite Element Model of the bridge (Valenzuela & Casas 2013).
3.2. Deck

(1) Behavior

In the context of strengthening, the deck is not a design variable, as it corresponds to the unique

active reused element of the original bridge, considering a static system that includes

intermediate diaphragm at mid-span. The deck section studied corresponds to continuous

longitudinal beams with transverse diaphragms, thus, the strengthening should preserve the stress

states and internal forces to those of a continuous beam bridge, with negative moments in the

areas of vertical hangers and positive in bays, with slight modification provoked by the network

hangers. This induces a decrease of maximum values in the lifting phase respect to established

limits.

The deck acts as edge beam or tie element of the system, thanks to the addition of the

external prestressing. The transverse beams are actively involved in the stress distribution

between the longitudinal beams and to face the transverse moments.

The embedment effect is a consequence of implementing the network structure

conditions, where the arch and deck are linked generating negative moment at the longitudinal

edge of the deck. Since the adequate behavior is defined as similar to continuous beam bridges,

these effects are not desired. Therefore, it is recommended to modify the geometry of the arch or

the curvature at the base of the arch; to use retention hangers or consider a new hinge joint

between arch and deck, trying to reduce or eliminate these efforts.This connection is composed

by a horizontal plate under the pin conection, in order to distribute the stress, and a vertical plate

for prestress tendon joint (Figure 6).

The incidence of the embedment effect is negligible in the beams that do not support the

arch, and when the arch has small profiles. Initially, the negative moment at the edge of the deck

is not a problem for the admissibility. However, it is advisable to use mechanisms that reduce the

effects of embedment. When the requirement of maximum allowable internal force is not
complied using a hinge joint at the base of the arch, it requires a study of the ductility of the deck

and its ability to form plastic hinges.

Figure 6. Arch-deck pinned connection.

The support system in the abutment corresponds to a simple support (fixed and

unidirectional sliding), with "pot" bearing devices (neoprene confined). These are used due to the

need to replace the original bearings, because of the increase of the vertical loads (due to

removal of the piers and the additional weight of the arch). Additionally, the problem of the

dimensions of the bearings can be reduced by the placement of more than one device under the

arch support, considering as minimum one at each beam. This also improves the behavior

because it reduces the transverse moments of the arch and produces a greater stability of the

bridge against lateral loads. The analysis of the strengthened bridge, carried out by a Finite

Element Model,shows a low distribution of the vertical loads to the beams that do not support the

arch (4 - 7%). This requires to choose between the use of a stiffening system at the edge

transverse beam to homogenize load on the bearing devices; or simply, to use bearings with

different characteristics.

(2) Design Criteria

• It is recommended to check the ductility of the deck in the link with the arch to see the

feasibility of development of plastic hinges at this pointand the ability to redistributethe

 stresses in the structure. If enough redistribution capacity is guaranteed then it is

recommended to place a hinge connection between the arch and the deck.

• Internal forces at the edge of the deck.

(a) It is not recommended the change in geometry of the arch based on the efforts at the

edge of the deck.

(b) However, if the stresses at the edges exceed the MAB, and the bearing capacity and

ductility of the deck is poor, it is justified the application of passive solutions (change of

the curve of the arch, retention hanger, etc.).

• For the external prestressing it is recommended to use astraight layout placed at the

center of gravity of the deck. It is allowed an eccentricity within the range -0.1 to 0

meters, looking for reactions and adequate reductions of the axial distribution of the

network and vertical hangers.

In some cases, if the transverse beams (diaphragms) are not sufficiently designed to

accommodate the internal forces in the new static system, the local strengthening of these

elements should be also carried out.

3.3. Vertical hangers.

(1) Behavior

The vertical hangers are the active elements in the process of lifting the deck and responsible to

induce a stress state similar to the original state of the deck, replacing the original piers. They are

placed over each original pier. These hangers are fundamental in the lifting stage and the in-

service state too. They never should lose tension, nor reduce it to levels that create stress states

outside the MAB. In the same manner, the maximum tension should not exceed their material

resisting capacity or induce stress states, in the deck and arch, higher than the design criteria.
 From the jacking operation in the lifting stage, the normal behavior of these hangers is in

tension. The maximum tension occurs in the service condition due to traffic load. The minimum

stress occurs associated with vertical seismic loads in the cases when the top and bottom

anchorages approach themselves.

(2) Design Criteria

• The parameter "Ad" representsa limit criteria defined by the ratio between deck weight

and its flexural resistance. This parameter is defined in the following way:

Pt EQ
Ad = (Equation 2)
M (−) MAX

Where:

Pt i + Pti +1
PtEQ: Weight per stage equivalent, defined as Pt EQ =
2

Pti: Weight per span (i)

M (-) MAX: Maximum negative bending moment.

• The value of this parameter is used prior to the analysis of the strengthening

process, such that if Ad > 1.2 m-1 the iteration process of tensioning must begin with a

higher number of vertical hangers or deferred lifting loads must be considered (by a

second group of tension hangers or a multi-step tension operation after verification of the

ultimate limit state).

• Tension force and order of tensioning of the vertical hangers.

• In the multi-objective optimization, the algorithm used in this study is based on

the theory of genetic algorithms (Valenzuela, 2012) being the tension force and the order

of tensioning the variables. Therefore, a set of possible solutions (value of the tension
 force and order of tensioning) is obtained at the end of the optimization (Pareto front).

The following criteria for selecting the final solution should be adopted:

(a) Choose solutions with a low-tension value of the central hanger, and a homogeneous

tension at all the vertical hangers. This achieves a better performance in the arch, deck

and in the tension of the vertical hangers.

(b)For lifting of very heavy decks (Ad > 1 m-1), it is recommended to choose a process

where all vertical hangers are acting and the lifting is produced just when the last vertical

hanger is jacked. This avoids the appearance of non admissible peak stress in the arch

and deck.

(c) Take a jacking tension of the hangers at least 20% lower than the maximum allowable

stress (fatigue criteria). Furthermore, it is recommended a jacking tension that produces a

deflection in the permanent state 25% higher than the maximum allowable deflection.

3.4. Network hangers.

(1) Behavior

The application of a set of network hangers as part of the strengthening process solves the need

to stiffen the system and get a truss-beam system, where the arch acts as top chord and the deck

as lower chord. The network hangers are responsible to distribute the efforts from a chord to the

other. In the case of the strengthening process, the behavior of these hangers varies depending on

the construction phase and the influence of the vertical hanger.

Initially, the collaboration of the network hangers is passive, collaborating in the lifting

process, i.e. they act when the process requires it. The forces in network hangers show a

reduction of tension with the increase of the participation of the vertical hanger, due to the

deformation in the arch. This provokes a compression in some of the network hangers. Normally,

there are a reduced number of active network hangers after the end of the lifting stage. The
maximum tension (which is always checked to be less than the maximum tensile force in the

material under permanent load to avoid fatigue and also less than the material strength) is

obtained in this stage, corresponding to the hangers placed at the ends of the bridge.

Although their passive participation in the lifting process, their performance is essential

to preserve admissible stress in the deck. After the lifting phase,a tensioning process of the

network hangers is performed with the goal that all of them participate actively in the service

state, i.e. to face vertical seismic loads (without loss of tension) and traffic loads (below of

failure load). In this way, the structureperforms avoiding structural changes and assists in the

dynamic response of the bridge. Additionally, the involvement of these hangers reduces stress on

the arch and the deck against the effects of service loads.

On the other hand, there is a number of network hangers, beyond whichan increase does

not generate an improved static behavior, showing an asymptotic decrease of the benefits. When

the tension of the vertical hanger is low, it is recommended to increase the distance of the

network hangers.

The radial arrangement is the best fit of hangers. It is favorable to the process of lifting

and service, respect to rhomboid-type arrangement.This behavior is confirmed by observing the

spacing between hangers. The rhomboid-type arrangement with spacing of 4 meters obtains

worse results than radial arrangement with spacing between 2 and 2.6 meters (respect to the

performance of the deck and the percentage used of the arch profile). In relation to the angle, it is

proposed a range of 45 ° to 65 ° (similar to a new design).

(2) Design Criteria.

• It is advisable to have areas of hangers (network and vertical) in a ratio between 0.67 and

1.

• Network hanger’s arrangement.

(a) Always use passive hangers during the lifting process.

 (b) When choosing a sequence of lifting process that begins at the center and continues to

the ends, then it is possible to use a vertical passive hanger’s arrangement.

(c) The recommended spacing of network hangers at the arch and deck is between 2 to 4

meters.

(d) It is recommended the use of a radial arrangement with individual placement of the

hanger, thus increasing the efficiency.

(e) The process of tensioning of the hangers network following the lifting phase, should

produce very favorable effects in front of in-service loads (truck loading and earthquake).

(f) The range of maximum and minimum axial force recommended for tensioning of the

hangers is defined byRn, the ratio between the l magnitude of the force in vertical and

network hangers. This is a key parameter justified by the high redundant condition of the

bridge as a consequence of the vertical and network hanger arrangements (similar to a

cable-stayed bridge). In this sense, when a hanger is jacked, the whole arrangement

redistributes its stress producing, in some cases, a final non-tension force in some

hangers. This criterion allows to control this problem.This ratio should be: 6.5 <Rn <10.

Avi + Avi +1
Rn = (Equation 3)
An j

Where:

Av,i: Axial force in the vertical hanger (i)

An,j: Axial force in network hanger in the span j (part of the deck between vertical

hangers (i) and (i+1)

(3) Connections

As mentioned before, it is recommended the use of symmetrical profiles of the arch (squares).

The connection of the hangers (vertical and network) is by welding a plate at the bottom of the
profile, joined by bolts between the plate and the end of the hanger (Figure 7).

Figure7. Connection: arch-hanger.

The connections between bars and concrete deck cannot follow the recommendations of

new construction (Brunn & Schanack, 2003; Schanak, 2008); i.e. a connection by anchoring of

the steel bar by welding, (Figure 8) or bars connected by its own thread in the hanger, anchored

to a pitchfork, (Figure 9) (Sasek & Faler, 2006).

Figure 8. Connection embedded: hanger-deck.

Figure 9. Pitchfork connection hanger-deck.

In the strengthening solution this is impossible, due to the requirement to drill the deck at

each of the positions of the hangers, with a sufficient dimension to introduce the rods with the

plates and to join the connector elements, looking for the collaboration in the transmission of

forces between the old and new concretes.

The connection is also not feasible (analysis in San Luis bridge, Chile (Valenzuela,

2012)), because the width of the longitudinal beams (40 centimeters) is not enough to placea

connection system and perform the physical union between the old and new concrete (because of

drilling and the reinforcement, Figure 10).

Figure10. Strengthened connection: hanger-deck.
For this reason, it is proposed a system similar to the KarolienenBridge (Unterweger,

2008), where the bar goes through the concrete element, connecting with a horizontal plate

located below the longitudinal support beams. The position of the rods depends on the

eccentricity between them (Figure 11).For the inclined hangers, the connection is located at each

side of the girder, passing through the slab with a penetration equal to the length defined by each

hanger slope. For that reason the depth of the drilling is equal to the thickness of the slab plus the

projection of the slope. Then the hanger follows the longitudinal girder from the bottom of the

slab reaching the bottom flange tied to the sides of the girder. Finally when arriving to the flange,

it is joined to a plate (bolted), providing the support to the girder and the hanger.

Figure11.Proposed connection: hanger-deck.

The same system is considered for tensioning the vertical hangers. However, the

application of these hangers requires additional work because of their location in the area where

the beam rests on the pier. For this reason, during the construction process, when un-linking the

beam from the pier, it is necessary to carry out a cut in the pier in the points of support of the

longitudinal beams (if the gap between pier and transverse beam is small). In this position,

vertical jacks are placed (the same that are subsequently used for the replacement of bearing

devices in the abutments), applying a controlled force to enable increase the gap to place the

lower plate and fix the bar (Figures12 and Figure 13).
Figure 12. Lifting mechanism to placed the hangers.

Figure 13. Final scheme: placed the hangers.

As a summary, Table 1 presents some of the design criteria previously analyzed.

Table 1.Element characteristic criteria

Elements Strengthened bridge

One centered (deck< 8m)
Position
Double (deck > 8m)
Inclination No
Bracing Vierendeel
Arch
Layout Circular
Rise/span ratio 16-17%
Profile Square
Size Profile/Length < 1/127
Type of hangers 2
N° of groups 3
Network
Distribution Radial - Vertical
Hangers
Distance hangers 2 to 3 meters
Angle 55° a 65°
Material Concrete reinforcement
Deck
Type Long. &Transv. Beams
 Increase high Transversal beam
Use transversal beam Yes
Model Support Simple support

Bearings Type Pot Bearing

N° devices Original support
Connection Coupling hanger No
Arch-hanger Connection plate and screw
Hanger-deck Bar with inferior plate
Active hanger-deck Bar with special inferior plate
Connection
Double plate embedment
Arch-deck + exterior prestress

Verified pinned alternative

4. Economic validation

The San Luis road bridge, Chile (Figure 14) was used to perform an economic cost comparison

between the traditional and new proposed strengthening methods. . The bridge is composedof a

cross-section with three continuous beams and a reinforced concrete slab (Figure 15).

Diaphragms (transversal beams) are provided over the piers, abutments and at mid-span of each

span. The deck is supported on three wall-piers, generating four spans (13.5 - 16.5 meters)

covering a total length of 60 meters.

Figure 14. General view San Luis Bridge.

Figure 15. Transverse section (cm units).
The main damage observed (MOP, 2007), corresponds to a problem of scour of the three

piers. Consequently, the foundations composed by double track railway piles present a high

degree of erosion and rust. As solution to the problem, a strengthening sequence by driven steel

piles at each pier was designed, despite the scarcity of vertical clearance.

The implementation of steel piles was done by a pile driver, located on an adjacent

embankment, to avoid overloading the bridge. It uses tube piles ("Yoder") of 30 cms in diameter

and 12 m in length, which are driven through holes at the slab to a depth of 15 meters. These

piles are linked to the piers through a concrete pile cap embracing the old foundations. This

process requires a diversion channel of the estuary through a parapet. These were the main

activities. Other activities comprise: repair of the the concrete slab, sealing of cracks in the slab

of the carriageway and sidewalk, replacement of bearing systems, expansion joints and drains,

and channel improvement (construction of flood defense works) (MOP, 2007). The costs of the

repair is summarized in Table 2.

Table 2. Budget of the original strengthening project

Item Total cost [euro]

Infrastructure 148.936,40

Superstructure 104.149,40

Various repair 6.728,60

Preparation of the workspace 6.057,60

Earth moving 360,90

Drainage and protection of the deck 9.607,90

Controls and security elements 15.195,40

Additional works 31.774,40

Environmental measures 2.958,60

Works in the riverbed 54.311,40

Industrial benefit (6 %) 27.137,80

Indirect cost (13 %) 58.798,50

19% VAT (Value Added Tax) 88.543,20

The total cost of the implemented repair was 554.560 Euros.

The alternative strengthening through tied arch includes some activities similar to the

original project, specifically associated with repairs to the superstructure and additional

activities. On the other hand, the biggest change is associated with the reduction intervention in

the infrastructure, avoiding changes in the riverbed. Thus, based on the presented criteria, the

strengthening of the bridge was designed with the following solution: s a single centered arch

with a square profile with dimensions 75x75x3,6 centimeters, the use of 26 network hangers and

three active vertical hangers (one over each pier). Furthermore, it includes a widening and

strengthening of the deck road according to the current bridge regulations in Chile. The complete

design is available in Valenzuela (2010). The predicted budget for this alternative solution is

presented in Table 3.

Table 3. Budget of the new strengthening project

Item Total cost [euro]

Infrastructure 5.394,83

Superstructure 12.745,16

Various repair 4.419,60

Preparation of the workspace 3.695,88

Earth moving 183,60

Control and security elements 49.776,94

Environmental measures 2.958,61

new strengthening at superstructure 236.292,72

Industrial benefit (6 %) 18.928,00

Indirect cost (13 %) 41.010,80

19% VAT (Value Added Tax) 71.327,20

The total budget obtained corresponds to 446.734 Euros.

The proposed project is economically feasible, with a budget of 20% less than the

original project, showing this new methodology as a feasible alternative to traditional

strengthening.

Besides this difference on construction budgets may bealradyconsidered as relevant

enough to propose the alternative strengthening solution, we should take into account additional

benefits, some of them not easy to deal with in economic terms, as:

1.- The alternative solution considers an improvement (widening) of the deck, avoiding

actual traffic limitations and enhancing safety to the users

2.- The elimination of the piers, will avoid future under water inspections, reducing the

life-cycle cost of the bridge. Additionally, the vulnerability of the bridge to scour and seismic

hazards will decrease which will translate in reduced feasibility of bridge closure and therefore

in less expected user costs from a life-cycle cost perspective. By contrast, future inspection lines

will be focused on the hanger system maintenance and corrosion of the metallic elements.

3.- The arch solution will completely eliminate the scour problem and therefore the

necessity of any repetitive and expensive additional repair work in the future, thus decreasing the

total expected maintenance and repair costs.

 4.- The aesthetics of the proposed alternative is much more enhanced compared to the

original bridge.

In summary, the total cost from a life-cycle cost perspective will be much lower for the

proposed strengthening solution.

6. Conclusions.

It has been presented and verified the feasibility of a new method of strengthening of bridges

with scour problems at piers. The method attempts to eliminate the problem by lifting the deck

from the original piers and support it with an upper arch, using the original deck (supplemented

with an external post-tensioning) as a tied element of the arch system. This allows to eliminate

the original piers.

The method requires the incorporation of steel arches, with an arrangement of vertical

hangers responsible to lift the deck by jacking, and an arrangement of network hangers to

modulate the stress states, providing stability in the lifting stages and improving the in-service

performance to asymmetric loads.

The method is defined with a minimum of three construction phases: preparation

(placement of elements), lifting (active stressing of vertical hangers) and network (tensioning of

the network hangers to allow optimal in-service conditions).

The design criteriais controlled by the requirement to keep the stress state in the deck as

close as possible to the original condition of the deck (a continuous beam). Thiscontrols the

design of other elements and avoids possible failure of the structure. Therefore, the arch becomes

the primary design parameter. Similarly, the study of vertical tensioning of the hangers (both in

value of the tension force and the order of tensioning) is fundamental to obtain an efficient

solution.

The design criteria derives on a set of condition to the arch and the hangers. First of all,

the arch profile is higher than in the case of a new network arch. In the case ofthe hangers, it is
recommended a close relationship of areas between vertical and network hangers. The multi-

objective optimization problem seeking to obtain the optimal tension in each hanger and order of

tensioning is derived by a genetic algorithm.

The conditions at the longitudinal edges of the deck require a particular study on the joint

between arch and deck.

The application of the method to the San Luis Bridge confirms that this strengthening

technique is economically sustainable, durable and improves the bridge aesthetics. In the paper,

the method is applied to a reinforced concrete bridge composed of longitudinal beams and upper

slab. However, the methodology can be easily extended to composite (steel+concrete) bridge

decks and also to slab and box-girder concrete bridges. The main limitation could be the

dimensions of the cross-section and the total length of the bridge after the removal of the original

piers.

7.References

Brunn, B. &Schanack F. (2003), "Calculation of a double track railway network arch bridge
applying the European standards".Diploma Thesis, TechnischeUniversitat Dresden,
Germany
Goldberg, D.E. (2002) “ The design of innovation: lesson from and for competent genetic
algorithms”. Ed. Kluwer Academic, Boston, USA
Malerba, P.G. (2014) “Inspecting and repairing old bridges: experiences and lessons”. Structure
and Infrastructure Engineering, 10:4, 443-470
MOP (2007). Informe de Estructuras “Reparacion Puente San Luis” [Structure Reporte –
Strengthening of San Luis Bridge] (in Spanish). Ministerio de Obras Públicas de Chile,
Direccion de Vialidad Santiago, Chile.
Muñoz, E. & Valbuena, E. (2006), “Socavacion de puentes” [Scour of bridges] (in Spanish).
Infraestructura Vial, Vol. 8 N° 15, pp. 25-39.
Pircher, M., Stacha, M., Wagner, J. (2013) “Stability of network arch bridges under traffic
loading”. Proceedings of the Institution of Civil Engineers (ICE), Bridge Engineering,
166, BE3, 186-192.
Rongish, J. (2011) “Refuerzo del Puente Puangue en Chile mediante configuracion tipo network.
Comportamiento sismico” (in Spanish), Master Thesis, UniversitatPolitecnica de
Catalunya, Barcelona, Spain.
Sasek, L. & Faler, R. (2006) “Network Arch Bridge in Bechyne”, 6th International Symposium
on Steel Bridges, Prague, Czech Republic.
Schanack, F. (2008) Network Arch Bridges, Ph.D Thesis, University of Cantabria, ACHE.
Spain.
Seaurz, A. (2006) “Dimensionamiento hidraulico optimizado de puentes con terraplenes”,
Ph.D.Thesis (in Spanish). University of Piura, Peru.
Tveit, P. (2007) “The Network Arch. Bits of Manuscript after Lectures in 44 Countries” 140
pages. Retrieved from: http://pchome.grm.hia.no/~pchome/
Unterweger, H. (2008) “Remarkable Strengthening of an old Steel Highway Bridge”, Creating
and Renewing Urban Structures, 17th Congress of IABSE, Chicago, USA.
Valenzuela, M.A. (2010) “Refuerzo de puentes existentes por cambio de esquema estatico.
Aplicacion al puente San Luis en Chile” (in Spanish). Master Thesis,
UniversitatPolitecnica de Catalunya, Barcelona, Spain.
Valenzuela, M.A. & Casas, J.R. (2011) “Bridge strengthening by structural change: Optimization
via genetic algorithm”, Proceedings of the IABSE-IASS symposium, London, UK.
Valenzuela, M.A. (2012) “Refuerzo de puentes de luces medias mediante conversión en arco
atirantado tipo network” (in Spanish). Ph.DThesis, UniversitatPolitecnica de Catalunya,
Barcelona, Spain.
Valenzuela, M.A., Casas, J.R. (2013) “Refuerzo de puentes: método del arco atirantado. Estado
del arte y criterios de diseño” (in Spanish). Editorial Publicia. Saarbrücken (Germany).