Advanced methods for assessing the seismic vulnerability of existing motorway bridges

Advanced methods for assessIng

the seIsmIc vulnerabIlIty

of exIstIng motorway brIdges (VAB)

Projects ENV4-CT97-0574 (DG 12 – EHKN)+ ENV4-CT98-0717 (DG12 – EHKN)

Duration: May 1998 – April 2001

Partners

ARSENAL, ÖFPZ ARSENAL Ges.m.b.H, (Arsenal Research), Austria,

ENEL Hydro/ ISMES, Italy,

ICTP, The Abdus Salam International Center for Theoretical Physics, Italy (ENV4-CT98-0717 (DG12 – EHKN)),

SETRA, France

UPORTO, Univerdidade do Porto, Portugal,

CIMNE, International Centre for Numerical Methods in Engineering, Spain,

JRC, Joint Research Centre, Ispra, EU.

1. Original research objectives

Recent seismic events all over the world have shown that bridge structures are particularly sensitive to earthquake loading. There are several reasons for such sensitivity. First many existing bridges were designed without adequate consideration for seismic risk. This has resulted in inadequate detailing of confining steel and insufficient shear reinforcement in the bridge piers, insufficient seat length of bearings, and inadequate strength and stiffness of the superstructure-abutment connection. Furthermore, there are many open questions concerning the ductile behaviour of large bridge piers, in particular those with rectangular hollow cross-section. Also, the seismic zonation map of many European countries has been revised recently, prescribing now higher horizontal ground accelerations in several regions. Finally, local soil conditions and the possibility of asynchronous motion at the base of the piers of long bridges are factors which can cause additional difficulties in properly designing irregular bridges.

There is therefore a need for reliable methods for assessing the seismic vulnerability of existing bridges, in particular large and irregular motorway bridges having lifeline character.

The whole project is centred on an Austrian bridge, "Talübergang Warth", which was built 20 years ago. The bridge was designed for a horizontal acceleration of 0.04 g using the quasi static method. Now, according to the new Austrian seismic code the bridge is situated in zone 4 with a horizontal design acceleration of about 0.1 g. Hence a detailed seismic vulnerability assessment is necessary.

2. Expected deliverables

Towards the above objectives, an international team has been set up covering the following disciplinary tasks contributing to the seismic vulnerability issue:

1. Dynamic insitu structural testing to identify the actual bridge properties including soil-structure interaction effects and to characterise the surrounding soil (Task leader ARSENAL). The expected deliverable consisted in the measured dynamic parameters of Warth bridge.

2. Development and calibration (fitting to test results) of numerical models for predicting the linear bridge response (Task leader ISMES). The expected deliverable was the linear model of the bridge, which is very close to reality. It represents the linear behaviour of the bridge deck and the “starting point” of the behaviour of the piers during an earthquake.

3. Numerical modelling for simulating the nonlinear behaviour of bridge piers under severe earthquake loading using damage mechanics concepts (Task leader UPORTO). The expected deliverables were the very detailed non-linear models for the rectangular hollow piers, which are finally also used for the pseudodynamic tests.

4. Physical testing of realistically large bridge piers with rectangular hollow cross section to calibrate numerical models and assess the ductility demand and capacity (Task leader JRC Ispra). The expected deliverables were the development of the continuous PSD testing with non-linear substructuring and asynchronous excitation and its application to Warth bridge.

5. Analysis of the effects on the bridge seismic response of asynchronous motion at the base of bridge piers. The expected deliverable was the computation of realistic synthetic broad-band records, taking into consideration source, path and local effects (Task leader ICTP). This time histories are used for the PSD tests.

6. Development of simplified analysis tools for assessing the vulnerability of complete bridges (Task leader CIMNE). The expected deliverables were simplified models, which are calibrated by the sophisticated “overall” procedure outlined above and which can be used as engineering tools for every day application.

7. Development of retrofitting measures to improve bridge reliability, including an evaluation of intervention costs (Task leader SETRA). The expected deliverable was a guideline for practical application, which should help the bridge owners to select the appropriate method and give also the necessary details for application.

3. Project’s actual outcome

Task 1: The in-situ test of bridge Warth showed once more the efficiency of dynamic in-situ testing. It is obvious that linear models are obtained, which reflect the actual status of the structure including also the influence of the boundary conditions. For the assessment of vulnerability this means the “linear starting point” for further structural modelling. An experienced earthquake engineer can predict the structural locations, where nonlinear mechanism will occur under a severe earthquake and he will consider this non-linearities during the next modelling step. Several possibilities to improve the test technique were shown.

Task 2: Linear modelling is a standard procedure. But still many problems exist for model updating using measurements. The methods available up to now do not cover all civil engineering needs. One promising software is HISTRIDE, but too less experience is available up to now.

Task 3: The development of a refined constitutive model devised for simulating the nonlinear behaviour of reinforced concrete bridge piers (rectangular hollow sections included) was a full success. Comparing the results of investigations with 2D- and 3D variants clearly demonstrated that 2D modeling is sufficient even in the case of rectangular hollow sections. It was found by the calculations of partner UPORTO, that important variations on the longitudinal reinforcement in almost all the piers will force the plastic hinges of the higher piers induced by the seismic loading preferably to occur in correspondence with the cross sections where reinforcement is reduced. This observation was clearly verified by the experiments.

Task 4: The JRC accomplished the objectives of the project programme in what concerns testing of large-scale models of bridge piers and pseudo-dynamic testing of a bridge model using non-linear substructuring, continuous PSD testing and non-synchronous earthquake input motions. This activity represents a step forward in advanced testing techniques. In fact, the PSD tests carried out at ELSA are pioneer at worldwide level.

Task 5: The synthetic signals, to be used as seismic input in a subsequent engineering analysis, have been produced at a very low cost/benefit ratio taking into account a broad range of source characteristics, path and local (geological and geotechnical) conditions. Lateral heterogeneities can produce strong spatial variations in the ground motion even at small length scales, that can be hardly accounted for by stochastical models. A general result of our modelling is that the effect of the differential motion can cause an increment greater than one unit in the seismic intensity experienced by the bridge, with respect to the average intensity affecting the area where the bridge is built. Different ground motions at the Warth site have been studied, in order to consider the maximum excitation in both longitudinal and transverse direction.

Task 6: CIMNE started work on Task 6 with a detailed bibliographic review to complete a state-of-the-art on the seismic estimation of the vulnerability of bridges. Next a simplified non-linear analysis model to estimate the damage produced by seismic action in this structures was elaborated. Monte Carlo simulations were carried out in order to account for the uncertainties of the seismic input, of the structural dimensions and of the mechanical material properties. Concerning the seismic input, families of artificial time histories were generated using some “mother histories” elaborated with the deterministic procedure used for Task 5.

Task 7: The assessment of the vulnerability and retrofitting of bridges has been presented, based on bibliographical research and engineering experience. The following steps have been looked into: a) The setting of a prioritization scheme, b) The performing of detailed analysis, c) The assessment of the vulnerability of the bridge, d) The application of adequate retrofitting techniques when the capacity of the bridge is not sufficient.

General: Several results are shown for bridge Warth. It is underlined that bridge Warth was designed considering a low seismicity zone with near field earthquakes (energy content only at high frequency ranges). However, similar bridges exist in medium/high seismicity zones with other earthquake scenarios (e.g. Italy, Greece, Portugal, etc.). Numerical simulations of this last situation, with models calibrated from the test results, will demonstrate that the similar structures will reach collapse for earthquake intensities far below the nominal ones.

R. Flesch