Content uploaded by Anibal Tafur

Author content

All content in this area was uploaded by Anibal Tafur on Nov 20, 2017

Content may be subject to copyright.

Effectiveness of Rubber Isolators for the Seismic Retrofitting of a

Peruvian Highway Concrete Bridge

Anibal Tafur1*, Thomas Swailes2

1* Department of Civil Engineering, Pontifical Catholic University of Peru, Peru (atafur@pucp.edu.pe)

2 School of Mechanical, Aerospace and Civil Engineering, The University of Manchester, UK

(thomas.swailes@manchester.ac.uk)

Abstract – This paper investigates the effectiveness of

rubber isolators for retrofitting highway concrete

bridges. The Huamani Bridge is analysed as a case

study. Located on the central coast of Peru, this bridge

was constructed around 1950 with obsolete seismic

provisions, using fixed steel bearings. It suffered from

severe damage during the 2007 Pisco Earthquake (8.0

Mw). As-built and isolated models were analysed and

their results were compared.

For the isolated option, the fixed steel bearings are to

be replaced with Lead Rubber Bearings (LRB), in order

to uncouple the movement of the super and sub-

structure, thus increasing the flexibility of the system

and decreasing the forces transmitted to the lateral

stability elements (piers and transverse beams). Flexible

links having the isolators’ properties were introduced

into the as-built model. The analysis of the isolated

model was performed following the AASHTO GSID

Simplified Method.

It was found that implementing the isolation system

increased the structure’s overall damping ratio by more

than 3 times, as well as its flexibility. This effectively

reduced the solicitations in the sub-structure and

transverse beams by an average of 75%. It is concluded

that the use of rubber isolators is effective for

retrofitting highway concrete bridges and that this

analytical procedure can be applied to other bridges of

similar characteristics in Peru.

1. INTRODUCTION

Peru is a country seismically high active due to its

location on the Pacific ‘Ring of Fire’, suffering from

disastrous seismic events in the past years such as the

1970 Ancash Earthquake (7.9 Mw; 70,000 deaths), the

2001 Arequipa Earthquake (8.4 Mw; 145 deaths); and

more recently the 2007 Pisco Earthquake (8.0 Mw,

519 deaths) [1]. These events caused high personal

and material losses.

Nevertheless, the seismic protection technologies

in Peru are not in wide use. In the last years, isolation

systems have started to be implemented in buildings,

but bridges remain overlooked. In January 2016 the

Peruvian Government published the current version of

the Peruvian Seismic Code (“Norma Tecnica Peruana

NTE E030”) making mandatory the implementation of

seismic protection systems for essential buildings

(schools, hospitals, etc.), without further specific

requirements for bridges (NTE E0.30, 2016). The

Peruvian Bridge Design Manual (MTC, 2016) [3]

provides minimal and vague provisions for the

analysis and design of isolated bridges, leaving local

engineers with insufficient information and guidelines

about this type of structures.

Considering that in Peru many bridges built long

time ago with obsolete seismic provisions are still

operational, and given the high seismic activity in the

country (especially on the Peruvian Coast); it is then

observed the necessity for vulnerability assessment

and retrofitting of these bridges. It is crucial to

improve their seismic performance, seeking

compliance with current seismic criteria applied in

other seismically high active and more developed

countries, such as USA and Japan.

According to Siqueira et al. (2014) [4], seismic

isolation has been recently used for the retrofitting of

existing bridges by increasing their flexibility, and

thus, their natural periods of vibration. This effectively

reduces the seismic energy input into the bridge

structure and the solicitations in the lateral stability

system [4]. Furthermore, the implementation of

isolators is generally easier to execute and less

expensive than the retrofitting of the sub-structure

(piers, columns or foundations) [5]. Even though to

the author’s knowledge there is no record of bridges in

Peru retrofitted using this technique, in other parts of

the world this has been successfully implemented.

Some notable examples are the retrofitting of the

Golden Gate Bridge in San Francisco (USA), the Bolu

Viaducts in Turkey; and the Chemin des Dalles Bridge

in Canada [4].

Figure 1 illustrates the principle of this retrofitting

method. In the case of an isolated bridge system (b),

the deformation takes place at the isolators during an

earthquake, preventing the damage of the sub-

structure; whereas in a conventional bridge system (a),

this deformation will occur at the sub-structure, which

is not desirable. The sub-structure elements (columns

or pylons) are the most demanded during an

earthquake, and it is crucial to prevent their damage. It

is clear that the most widely used isolation system for

bridge applications is the Lead Rubber Bearings

(LRB), as stated by Buckle et al. (2006) [6].

Fig. 1: Comparison of a conventional and seismically isolated

bridge [6]

The aim of this paper is to evaluate the efficiency of

rubber isolators for the retrofitting of a multi-span

concrete highway bridge, namely the Huamani Bridge

in Peru, by comparing the responses and solicitations

(displacements and forces) obtained from linear

dynamic analyses of as-built and isolated models.

Thus, the reduction of the seismic forces induced in

the structure due to the implementation of the isolation

system is evaluated.

2. CASE STUDY: THE HUAMANI BRIDGE

The Huamani Bridge (see Figure 2) is a reinforced

concrete continuous bridge that spans over the Pisco

River, in the Ica Department of Peru. It was built

around 1950 and suffered from considerable damage

during the 2007 Pisco Earthquake (8.0 Mw), located at

approximately 60 km from the epicentre. It is located

close to San Clemente Town in the Pisco Province, at

coordinates S13°41’13” W76°09’31”, and it is one of

the most important bridges on the Pan-American

South Highway, the main road network on the

Peruvian Coast.

This bridge is representative of many old concrete

bridges still operational in Peru, which were designed

with none or obsolete seismic provisions. The seismic

design criteria used for this bridge was to apply as

base shear force equivalent to 4-8% of the bridge

weight, with no provisions regarding ductility [7].

Fig. 2: The Huamani Bridge over the Pisco River after the Pisco

EQ (2007) [7]

The bridge consists of 5 spans, 3 typical of 30 metres

each and 2 spans of 23 metres on the sides, giving a

total length of 136 metres (see Figure 3, top). The

super-structure is supported by 3 hunched beams

which are 700 mm wide and 1.80 m high (away from

the supports), and at the supports the beams are 3.5 m

high. The bridge deck is a 300 mm thick reinforced

concrete slab. There are also several diaphragm

(transverse) beams that restrain the structure laterally.

Fig. 3: Layout of the Huamani Bridge (dimensions in mm)

The substructure consists of two abutments and four

interior pier walls measuring 1.10 m by 8.20 m, with

shear keys implemented on their sides as shown in

Figure 3 (bottom), with the objective to control

excessive lateral displacement. These pier walls are

founded on deep foundations (caissons) [7].

Elevation view:

Cross section:

At the intermediates pillars, the super-structure is

supported on steel pot bearings, fixed in both

directions. On the abutments, simple steel rollers were

used, which are free to move in the longitudinal

direction, but restrained in the lateral direction by

means of steel stoppers. According to the survey

reports, the bearings were found in very bad state,

corroded and with rust and dirt covering [7].

Both structural and geotechnical damage was

observed during the surveys at the Huamani Bridge

after the earthquake. The most complete reports on

this matter are the works of Tang & Johansson (2010)

[7], Johansson et al. (2007) [8] and Taucer et al.

(2009) [9], these being the main sources consulted for

this paper. Some examples of critical structural

damage are shown in Figure 4.

Fig. 4: Structural damage observed in the Huamani Bridge after

the 2007 Pisco EQ [7], [8].

Severe cracking was observed in one pier, at the

junction with its shear key. This was due to loss of the

lateral restriction in the bearings which allowed the

superstructure to move freely at this point, pounding

the shear lock and causing severe damage to it (see

Figure 4.a). The shear lock, although damaged,

prevented the bridge superstructure from suffering

excessive displacement and overturning (100 mm

permanent displacement was observed) [7]. Also,

severe cracking was observed in the transverse beams,

which provide lateral stability to the super-structure

(see Figure 4.b).

The steel roller bearings at the abutments were

meant to behave as laterally restrained by means of

steel stoppers welded to their base plates.

Nevertheless, the steel rollers at the southern abutment

suffered severe damage due to excessive displacement,

losing their lateral restraints (see Figure 4.c). On the

other hand, at the northern abutment the damage was

less severe, with the lateral restraint being still present,

but with evidence of permanent displacements (See

Figure 4.d). Severe corrosion was observed in all the

bearings [7].

3. METHODOLOGY

As-built and isolated models of the Huamani Bridge

are analysed and their results are compared. For the

isolated option, the fixed steel bearings present in the

as-built model are to be replaced with flexible links

having the isolators’ properties. The modelling process

is discussed more fully in Section 4 of this paper.

The efficiency of the isolation system is assessed

by comparing the following results from the as-built

and isolated model analyses: super and sub-structure

displacements, base shear forces, forces in key

structural elements, periods and damping ratios.

Furthermore, the piers and transverse beams’

capacities (shear and flexural) are calculated, in order

to illustrate the beneficial effects of the isolation

system on the demand-capacity ratios of these

elements.

The response of an isolated structure is very

complicated to analyse, due to the high-non linearity

induced by the isolators. Non-linear models for the

structure elements have to be combined with hysteretic

curves obtained from laboratory tests of isolators, and

time-history functions for the seismic loads.

Nevertheless, for the purpose of this paper only linear

analyses are carried out, following the AASHTO

GSID simplified method as described in Section 7.1 of

the AASHTO Guide Specifications for Seismic

Isolation Design GSID (2010) [13].

LRB isolators dissipate energy due to internal

friction inside the rubber material (also called viscous

damping), and deformation in the lead core. Figure 5

(top) shows a typical section and the displacement—

force diagram of an LRB isolator, where its hysteretic

curves and non-linear behaviour can be appreciated.

a) [7]

c) [7]

d) [8]

b) [8]

The cyclic behaviour of these isolators can be

approximately represented by a simplified bi-linear

model as shown in Figure 5 (bottom).

Fig. 5: Top: Section of an LRB and its force—displacement

diagram [10], [11]. Bottom: Bilinear simplified model of a rubber

isolator [12]

Where:

K1 = Isolator initial stiffness

K2 = Isolator post-yielding stiffness

Kisol = Isolator effective stiffness

Qd = Isolator characteristic strength

Fy = Isolator yield force

Fisol = Isolator maximum force

dy = Isolator yield displacement

disol = Isolator maximum displacement

And K2 may be estimated as follows [12]:

r

T

GA

K

2

× (1.15 to 1.20) (3.1)

Where Tr is the total thickness of rubber (addition of

layers), G is the rubber shear modulus and A is the

section area. K1 is in the range of 15 to 30 times K2.

Note that the effective stiffness Kisol is dependent on

the isolator maximum displacement disol; therefore, it is

not possible to know its exact value beforehand. The

isolator effective stiffness Kisol is useful when only

linear methods are considered. The value of Kisol is

input into the software SAP2000 to define the

properties of the link elements used to represent the

isolators in the model.

The methodology presented is this paper to determine

the isolation system properties and isolated structure

response consists of three stages:

First, the Simplified Method presented in AASHTO

GSID is applied. For the first iteration a

displacement is assumed and a preliminary set of

characteristic isolator resistance properties is

defined. The displacement response of the isolated

structure is calculated taking into account the

reduction by damping. Then, convergence is checked

comparing the obtained displacement with the initial

assumed value.

In the next stage, a Multimodal Spectral Analysis is

carried out, similarly to what is done with the as-

built structure, but with a modified response

spectrum (reduced by damping). It is used the same

as-built models, but replacing the rigid links with

flexible links having the representative isolator

properties determined in the previous stage. The

objective is to obtain similar response to that

estimated by the GSID Simplified Method. After

reaching this second convergence, displacement

response and forces are obtained for the sub-

structure and isolators.

The final stage is the Design of the Isolators in order

to secure their viability, using the forces and

displacements obtained from the spectral dynamic

analysis. The requirements and limit checks provided

by the AASHTO GSID are followed. The output of

the design includes sizes and material properties,

which should be checked against the existing bridge

dimensions.

The AASHTO simplified method reduces the isolated

structure to an equivalent single degree-of-freedom

model with equivalent properties, in order to estimate

its response (displacement of the super-structure) [12].

The super-structure displacement d is estimated as

follows:

L

effD

B

TS

d1

250

(3.2)

Where:

SD1 = Spectral acceleration coefficient

Teff = Effective period of the isolated structure

BL = Damping reduction factor given by Eq. 3.7

SD1 is defined as 0.62 from the seismic hazard study.

Since the values of Teff and BL are unknown

beforehand, for the first iteration Teff may be taken as

1.0 seconds and BL may be taken as 1.0 (5% of

damping). To estimate the system

totald

Q,

and

total

K,2

required for the first iteration (addition of all the

isolators), it is recommended by Buckle et al. (2011)

[12] that

totald

Q,

should be at least 5% of the bridge

effective weight. Since for this bridge the piers are

very stiff (no vibration), it is assumed that only the

mass of the superstructure contributes to the vibration,

therefore the effective weight

eff

W

is taken as the

super-structure weight. For

total

K,2

it is recommended

to take 10% of

dWeff /

. For this case, since the mass

of the existing structure is very large (relatively

compared with bridges designed considering isolation

from the beginning), 15% of

dWeff /

is taken. The

assumed

totald

Q,

and

total

K,2

are distributed according

to the percentage of the dead load acting at each

support (abutments and piers).

Fig. 6: Calculation of the combined effective stiffness at the

supports [12]

It is then necessary to calculate the combined effective

lateral stiffness at each support

eff

K

, accounting for

the stiffness of the sub-structure and the isolators

combined. As shown in Figure 6. The stiffness

distribution factor

is defined as follows:

dsub

d

QdK QdK

2

subeff KK

1

(3.3)

The displacements of the sub-structure

sub

d

and

isolators

isol

d

at each support (abutment and piers),

are calculated as follows:

1d

disol

isolsub ddd

(3.4)

The effective period of the isolated structure Teff is

calculated using the effective weight

eff

W

and the

summation of the effective stiffness’ at each support

eff

K

, as follows:

eff

eff

eff Kg

W

T

2

(3.5)

The damping ratio of the isolated structure

and the

damping reduction factor BL are calculated using the

following expressions:

2

2

subisoleff

yisold

ddK

ddQ

(3.6)

3.0

3.0

70.105.0

3.0

L

B

Eq.7.1-3 AASHTO GSID (3.7)

Using these values of Teff and BL, the expected super-

structure displacement is calculated using Eq. 3.2.

After a series of iterations, convergence should be

found between the initial assumed displacement and

the calculated after the iteration. Once convergence is

reached, the effective stiffness of the isolators Kisol are

obtained using the following expression, in order to be

used for the multimodal spectral analysis in the next

stage.

2

K

d

Q

K

isol

d

isol

(3.8)

In order to carry out the dynamic analysis of the

isolated model, the response spectrum is modified to

take into account the higher damping present in the

fundamental modes introduced by the isolators. The

acceleration values are divided by the damping factor

BL, at periods above 0.8Teff. Figure 10 in next section

shows the reduced spectrum resulting from the

Huamani Bridge analysis, and is used for both

directions of analysis.

K2

In order to define the properties of an isolation system

used for retrofitting, a controlling parameter has to be

defined. For existing bridges supported on columns,

the shear force in these elements is likely to control

the design. For the Huamani Bridge, the existing piers

are of massive dimensions; therefore the shear force in

these will hardly control the design. The size of the

isolators, which will have to properly fit between the

existing piers and beams, is likely to control the

design. Other parameter to take into account is the

maximum displacement of the super-structure, which

will have to be controlled in order to avoid excessive

pounding on the existing abutments [12]. Figure 7

shows a flowchart summarising the methodology

described in this section.

Fig. 7: Flowchart summarising the methodology for the analysis of

the isolated bridge

4. BRIDGE MODEL SIMULATION

Figure 8 shows a perspective view of the model

defined using finite elements and analysed using the

software SAP2000. The seismic solicitations applied

to the as-built model are represented by Response

Spectrum and Time History linear cases. The

Response Spectrum (RS) is defined according to the

AASHTO specifications and using seismic hazard

information from a similar project in Peru [16]. The

Time History (TH) cases were defined using 10

signals, including the accelerogram of the 2007 Pisco

EQ, properly scaled to PGA 0.45g, which is the

characteristic value of the peak ground acceleration

specified for the bridge location in the Peruvian

Seismic Code (Seismic Zone 4). The most

unfavourable of these cases will be considered to

obtain the design resistance demanded on the

structural elements of the as-built case (envelope).

Due to the limitations of the AASHTO Simplified

Method, for the isolated model only the Response

Spectrum case is carried out, considering the reduction

by damping.

Fig. 8: Model perspective view as seen in the software interface

For the purpose of this paper, the piers were assumed

as fixed at their base for both models. According to the

reports consulted, the foundations are caissons

embedded 8 metres into the soil [7], [8]. The material

properties input into the software were: for the

concrete f’c = 30 MPa, Young’s Modulus E = 24.6x109

MPa, specific weight γ = 25 kN/m3 and 5% damping

ratio. For the reinforcement steel it was considered

ASTM A615 Gr60 with fy = 420 MPa.

Regarding the bearings at the supports, to simulate

the fixed bearings of the as-built bridge, they were

idealised using rigid links. For the isolated model, the

rigid links where replaced with flexible links having

the isolators’ characteristic properties, in this case, the

isolators’ effective stiffness Kisol. The element meshing

was defined in the software by establishing a

maximum characteristic dimension of 300 mm. The

Assume displacement, characteristic strength Qd

and post-yielding stiffness K2

Find the effective period and the equivalent

dampingratioof the isolatedstructure usingthe

GSID Simplified Method

Find the theoretical displacement, reduced by

damping. Is it the same as the assumed?

Use the assumed displacement, Qdand K2to

find the isolators' effective stiffness Kisol to be

used in the model.

Modify the 5% damped response spectrum used

for the as-built model, reducing it by the damping

factor BL.

Run a dynamic linear analysis of the isolated

model. Are the displacements and periods

obtained similar to those calculated with the

simplified method?

Obtain the isolators' design forces and

displacements from the dynamic analysis results.

Design the isolators. If suitable dimensions cannot

be reached repeat the process from the beginning.

NO

YES

NO

YES

software has an option to automatically decrease the

size of the mesh in areas where higher discretization is

required, such as corners and curved parts.

The AASHTO Seismic Code, Section 3.10.4

‘Seismic Hazard Characterization’ [14], presents the

procedure to define the response spectrum for 5%

damping (typical concrete bridges). This spectrum is

defined as elastic, i.e. without considering ductility

reduction factors for design. Figure 10 shows the

spectra defined for the bridge analyses, for both as-

built and isolated models. Note the reduced response

due to damping for the isolated model (solid line).

The accelerograms used for the Time-History

Cases were recorded by many seismic monitoring

stations in Peru. These stations are part of the large

seismic observation network managed by the CISMID

(Peruvian-Japanese Centre of Seismic Investigation

and Disaster Mitigation), and their records are

available on their website (www.cismid-uni.org). It is

notable the case of the 2007 Pisco EQ accelerogram,

which affected the Huamani Bridge. The closest

accelerometer to the bridge is located at the ICA002

station (63 km away) and its records are shown in

Figure 9.

Fig. 9: Accelerograms recorded at the ICA002 station during the

2007 Pisco Earthquake (CISMID)

Fig. 10: AASHTO Response Spectra (as-built and isolated cases)

defined for the Huamani Bridge analysis

5. RESULTS

After the iterative process of the isolated bridge

analysis, convergence was found for the

following values:

d

=241mm,

totald

Q,

=1251 kN,

total

K,2

=15573 kN/m,

eff

T

= 2.203 s,

= 0.16 and

BL = 1.42. Table 1 summarizes the required

properties found for each isolator after the last

iteration, considering a ratio K2/K1 = 0.1 and dy =

0.1dt. These values were obtained diving by 3 the

properties found at each support.

Table 1: Required properties of the isolators obtained with the

GSID simplified method

Location

K1

(kN/m)

K2

(kN/m)

Qd

(kN)

dy

(mm)

Fy

(kN)

Kisol

(kN/m)

Abutments

2701

270

22

8.93

24

360

Piers

11627

1163

93

8.93

104

1551

Figure 11 shows a comparison of the results from the

modal analyses of the as-built and isolated models,

where important increment in the flexibility due to the

implementation of the isolation system is appreciated.

The periods increased 15 times for the longitudinal

and 10 times for the transverse direction. This caused

a very important increment in the super-structure

displacement as well, 60 times in the longitudinal and

14 times in the transverse direction. Since the as-built

model was found to be very stiff (small

displacements), theses increments are as expected.

The overall structure damping ratio increased from

the initial 5% estimated for the existing concrete

structure, to 16% for the isolated structure (more than

-400

-300

-200

-100

0

100

200

300

400

025 50 75 100 125 150 175 200 225

Acceleration (cm/s2)

Time (seconds)

North-South direction

-300

-200

-100

0

100

200

300

025 50 75 100 125 150 175 200 225

Acceleration (cm/s2)

Time (seconds)

Eat-West direction

0.00

0.20

0.40

0.60

0.80

1.00

1.20

0.00 1.00 2.00 3.00 4.00

Elastic Seismic Coefficient Csm (g)

Period Tm (seconds)

Reduced response

1.0g

PGA = 333.66 cm/s2 (0.34g)

PGA = -272.82 cm/s2 (0.28g)

0.50g

0.8Teff

(1.76s)

3 times). This is a very important feature that allows

the structure response to be reduced. The forces in the

piers were reduced by 80% in average. The forces in

the transverse beams were reduced by an average of

75%.

Fig. 11: Comparison of the analyses modal results for the as-built

and isolated models

To evaluate the significance of the seismic force

reduction in the structural elements, demand-capacity

ratios of the piers and transverse beams are calculated

(shear, flexural, and flexural-compressive capacities),

following the guidelines provided by the ACI-318

Code (ACI, 2008) [15]. Two Ultimate Limit State

(ULS) loading scenarios are considered: for the as-

built bridge the envelope of the Response Spectrum

and Time-History cases, and for the isolated bridge

the Response Spectrum reduced by damping.

Table 2 shows that the shear demand-capacity

ratios in the piers decreased significantly for all cases.

In the case of the transverse direction of Pier 2, the

reduction is crucial since the as-built demand is 1.16

(larger than 1), indicating that strengthening of the pier

would be needed. The implementation of the isolation

causes this ratio to decrease to only 0.20.

Tables 3 and 4 show the shear and moment

demand-capacity ratios for the transverse beams. It

can be appreciated the important reductions for the

isolated case. The isolation system causes the ratios to

decrease from values above one in both piers

(strengthening needed), to very low values, in the

order of 0.30 to 0.40.

Table 2: Comparison of pier’s shear demands and capacity

Direction

Element

ϕVn

(kN)

AsB

(kN)

Isol

(kN)

Ratio

AsB

ϕVn

Ratio

Isol

ϕVn

Longitudinal

X-X

Pier 1

5731

5003

1113

0.87

0.19

Pier 2

5731

3545

1113

0.62

0.19

Transverse

Y-Y

Pier 1

5446

3579

1110

0.66

0.20

Pier 2

5446

6320

1109

1.16

0.20

Notes: AsB = As-Built demands (Envelope of RS and TH cases)

Isol = Isolated Structure demands

Table 3: Shear demand – capacity ratios for the transverse beams

Case

Location of

Beam

Shear

Demand

(D) (kN)

Shear

Capacity

(C) (kN)

Ratio

D/C

As-Built

@ Pier 1

403

360

1.12

@ Pier 2

708

360

1.97

Isolated

@ Pier 1

120

360

0.33

@ Pier 2

120

360

0.33

Table 4: Moment demand –capacity ratios for the transverse beams

Case

Location of

Beam

Moment

Demand

(D) (kN-m)

Moment

Capacity

(C) (kN-m)

Ratio

D/C

As-Built

@ Pier 1

569

414

1.37

@ Pier 2

999

414

2.41

Isolated

@ Pier 1

172

414

0.42

@ Pier 2

172

414

0.42

Regarding the flexural-compressive capacity of the

piers, the most critical combination is ‘0.9 D ± S’.

Figure 12 shows the positions of these combinations

within the Piers’ Design Interaction Diagrams. It is

observed that for Pier 1 (top) the as-built design forces

combinations are very close to the design critical

curve. This indicates that the element is highly

demanded in the as-built case, possibly resulting in

necessity for strengthening due to bending.

Furthermore, it is seen that the isolation system

helps to significantly decrease this demands (4 times

in average). A similar reduction in the demands is seen

for Pier 2 (bottom), although the as-built forces are not

located on the critical curve. For this case, the

reduction due to the isolation implementation is

Pier 1

Isolated Y-Y direction

Max disp. = 240 mm

Period = 2.213 s

Mode 1

(94% participation)

Isolated X-X direction: Max disp. = 240 mm, Period = 2.206 s

(94% participation) Mode 2

Pier 2

As-built X-X direction: Max disp. = 3.91 mm, Period = 0.148 s

(49% participation), Mode 6

As-built Y-Y direction

Max disp. = 16.90 mm

Period = 0.222 s

Mode 1

(75% participation)

Pier 2

Pier 1

efficient but not critical. In the transverse direction of

the piers the demands are located far away from the

critical curves for both the as-built and the isolated

cases, due to the massive bending resistance in this

direction.

Note: As-built combinations

Isolated combinations

Fig. 12: Interaction diagrams and seismic demands for the piers in

the longitudinal direction

6. CONCLUSION

This paper investigated the effectiveness of rubber

isolators as a retrofitting measure for highway

concrete bridges. The object of the study was the

Huamani Bridge, a continuous concrete highway

bridge built around 1950 with obsolete seismic

provisions. The methodology consisted of comparing

key parameters from the as-built and the isolated

bridge models, leading to the following conclusions.

It was found that the implementation of rubber

isolators would substantially increase the modal

parameters of the bridge. It was observed that the

natural periods increased more than 10 times (from

0.15 s and 0.22 s to 2.21 s). The effective damping

ratio increased from 5% (initially estimated) to 16%

(obtained with the AASHTO Simplified Method).

The increment in the flexibility and damping ratio

of the structure caused the seismic force solicitations

to be reduced by an average of 80% for the piers and

75% for the transverse beams. It is concluded that the

implementation of rubber isolators is very efficient to

decrease the demands in the lateral stability elements.

Calculating the capacities of piers and transverse

beams it is observed that these would be exceeded by

the seismic demands of the design earthquake (1000-

yr return period). Furthermore, it was observed that

the isolation system would help to reduce substantially

the demand-capacity ratios (e.g. down to 0.20 for piers

and 0.33 for beams), in such way that no strengthening

of these elements would be needed; thus confirming

the efficacy and worth of the isolation system

implementation.

ACKNOWLEDGEMENTS

The authors acknowledge the economic support of the

Peruvian Ministry of Education that funded this

investigation. The authors would also like to thank

Julio Arias for providing his Master’s thesis, from

which essential information was obtained about the

seismic hazard parameters for this bridge.

REFERENCES

[1] Storchak, D. A., Di Giacomo, D., Bondár, I., Engdahl,

…& Bormann, P. (2013). Public release of the ISC–

GEM global instrumental earthquake catalogue

(1900–2009). Seismological Research Letters, 84(5),

810-815.

[2] NTE E.030 (2016). Peruvian Seismic Code published

by SENCICO. (Norma Técnica E.030 Diseño

Sismorresistente, in Spanish)

[3] MTC (2003) Bridge Design Manual. Ministerio de

Transportes y Comunicaciones del Peru (Title in

Spanish: Manual de Diseño de Puentes)

[4] Siqueira, G. H., Tavares, D. H., & Paultre, P. (2014).

Seismic fragility of a highway bridge in Quebec

retrofitted with natural rubber isolators. Revista

IBRACON de Estruturas e Materiais, 7(4), 534-547.

[5] DesRoches, R., Choi, E., Leon, R. T., & Pfeifer, T. A.

-40

-20

0

20

40

60

80

100

120

-30 -20 -10 0 10 20 30

ϕPn (MN)

ϕMn (MN-m)

Pier 1 Interaction Diagram for Longitudinal

Direction

-40

-20

0

20

40

60

80

100

120

-30 -20 -10 0 10 20 30

ϕPn (MN)

ϕMn (MN-m)

Pier 2 Interaction Diagram for Longitudinal

Direction

(2004). Seismic response of multiple span steel

bridges in central and southeastern United States. II:

Retrofitted. Journal of Bridge Engineering, 9(5), 473-

479.

[6] Buckle, I. G., Constantinou, M. C., Diceli, M., &

Ghasemi, H. (2006). Seismic isolation of highway

bridges (No. MCEER-06-SP07).

[7] Tang, A.K. & Johansson, J. (2010). Pisco, Peru,

earthquake of August 15, 2007: lifeline performance.

American Society of Civil Engineers (ASCE).

[8] Johansson, J., Mayorca, P., Torres, T., & Leon, E.

(2007). A Reconnaissance Report on The Pisco. Peru

Earthquake of August, 15, 2007.

[9] Taucer, F., Alarcon, J. E., & So, E. (2009). “2007

August 15 magnitude 7.9 earthquake near the coast of

Central Peru: analysis and field mission report”.

Bulletin of Earthquake Engineering, 7(1), 1-70.

[10] Zhang, J., & Huo, Y. (2009).’ Evaluating effectiveness

and optimum design of isolation devices for highway

bridges using the fragility function method’.

Engineering Structures, 31(8), 1648-1660.

[11] FHWA (2014). LRFD Seismic Analysis and Design of

Bridges Reference Manual. US Department of

Transportation. Federal Highway Administration,

Washington.

[12] Buckle, I.G., Al-Ani, M. & Monzon, E. (2011).

Seismic isolation design examples of highway

bridges. NCHRP Project, pp.20-7.

[13] AASHTO (American Association of State Highway

and Transportation Officials) (2010) “Guide

Specification for Seismic Isolation Design” Third

Edition, AASHTO, Washington, D.C.

[14] AASHTO (American Association of State Highway

and Transportation Officials) (2014) “Guide

Specification for LRFD Seismic Bridge Design”

Second Edition, AASHTO, Washington, D.C.

(Interim).

[15] ACI Committee, American Concrete Institute, &

International Organization for Standardization.

(2008). Building code requirements for structural

concrete (ACI 318-08) and commentary. American

Concrete Institute.

[16] Arias, J.C. & Vargas, J.O. (2015). Evaluación y

comparación del desempeño sísmico de cuatro

tramos de estructura de la línea 1 del metro de Lima

diseñados en épocas diferentes (Master’s dissertation,

Pontificia Universidad Católica del Perú, Escuela de

Posgrado. Mención: Ingeniería Civil).