engleski

Design and Testing of the Reconstructed Pag Bridge

1 INTRODUCTION The Pag Bridge was built in 1968 to link the island of Pag with the Croatian mainland. It is a reinforced concrete arch bridge spanning 193.2 m. The original concrete superstructure comprised a series of simply supported grillages. The columns were precast and connected to the arch by prestressed tendons. Extremely aggressive maritime environment combined with inadequate detailing, poor workmanship, exceptionally small concrete cover and damages caused by aircraft missiles led to the rapid structural degradation: cracking of concrete, peeling off of concrete cover, reinforcement corrosion and failures of bearings. Structural deterioration required the reduction in the traffic speed and volume. Over the years various repair techniques were tried on the superstructure and columns, but none proved efficient. The arch was repaired in 1991 with the removal of the damaged concrete cover, grouting all visible cracks and placing an additional reinforcement mesh on all external surfaces covered by a fine 4-cm thick concrete layer. Planned replacement of the superstructure and column strengthening had to be postponed due to the war activities in the region. 2 DESIGN OF THE PAG BRIDGE RECONSTRUCTION The reconstruction commenced in 1999. Two independent continuous steel superstructures were constructed with span lengths 5&#61620 ; ; 23.30+11.65=128.15 m and 4&#61620 ; ; 23.30+11.65=104.85 m. The structural solution comprising steel was favoured since it provided reduction in the weight of the structure. This was important since the traffic loads according to the current codes are almost double the traffic loads to which the original bridge was designed. Both superstructures comprise two 1.5 m deep steel plate girders connected by a 12-mm thick orthotropic deck with open stiffeners. Such structural solution benefited the bridge&#8217 ; ; s durability as it provided access to all steel structural members. Existing columns were encased in 12-mm thick steel casings while leaving a 12-cm wide gap between the casing and the original column. This space was filled with fine aggregate low shrinkage concrete. Disability to ascertain whether the columns are fixed or pinned at their bottom presented additional problem in the design since new pot bearings would be installed to support the new steel superstructures, thus substantially changing the overall stability of the bridge. Calculations were carried out with respect to the actions and combinations of actions as defined in DIN 1072, except for the wind loading which was anticipated larger than that required in the code. The deviation of the actual arch axis from the designed one was taken into account as a strong discontinuity with the displacement from the designed arch axis of 26 cm was detected at approximately quarter of the arch span on the side facing the island of Pag. Execution stages comprising the sequence of dismantling the old concrete superstructure and erection of the new steel superstructure had to be carefully checked, especially their influence on the arch behaviour. The new superstructure is lighter than the original one, but since the arch axis is designed as a thrust line for certain permanent load, the distribution of lighter permanent load can be unfavourable and adversely affect the arch behaviour. The calculations revealed that the arch is capable of withstanding new loading within the designated threshold level only if the arch reinforcement contributes in the compressive zone and if the actual measured compressive concrete strength corresponding to C-50 grade instead of C-35 anticipated in the original design is accounted for. 3 LOAD TESTING OF THE RECONSTRUCTED BRIDGE The load testing of the reconstructed bridge was performed in 1999, prior to the bridge&#8217 ; ; s re-opening to the service. Both the static and dynamic tests were conducted with controlled traffic loads. The static tests were performed with 27 different arrangements of various number of test vehicles (30-ton trucks). Deflections were measured at 58 points along the superstructure. 8 strain-gauges, 17 accelerometers and 5 displacement transducers were used to record the bridge&#8217 ; ; s response to vertical loads. The results of the static tests were compared with the stresses and deflections calculated using the same finite element model as in the design calculations. Linear structural analysis was used for calculating deflections and stresses in the bridge superstructure. In addition to the linear structural analysis, a non-linear analysis of a simplified FE model of the bridge taking into account the measured concrete grade of C-50 was used to assess the stresses in the arch. The numerical and experimental results agree very well, so it was concluded that the FE model used in the reconstruction design calculation was adequate. But, this also confirms that the stresses in the arch under the designed traffic loads are indeed approaching ultimate allowable values. The dynamic tests comprised single or a pair of test vehicles (30-ton trucks) crossing the bridge at different speeds (20, 40 or 60 kph) and positions, with and without braking, with and without placing an obstacle on the roadway to create an additional dynamic impulse. The same sensors setup was used as in the static tests. All the experimental natural frequencies are higher than the numerical ones, thus once more confirming the accuracy of the FE model used in the bridge reconstruction design. The recorded dynamic properties were compared with the results of the dynamic tests performed in 1993, prior to the reconstruction. The recorded natural frequencies of the mode shapes in the vertical plane are higher for the reconstructed than for the original bridge. This is a consequence of the 20% lighter mass of the overall structure, but also of an increased modal stiffness of the reconstructed bridge in the vertical plane. This is especially evident in the comparison of the frequencies of the mode shapes which are characteristic for the superstructure. 4 CONCLUSION Extreme severity of the exposure conditions combined with the exceptionally small concrete cover, inadequate detailing, poor workmanship and maintenance led to the rapid structural degradation of the arched Pag bridge. In 1999, the bridge was thoroughly reconstructed with the replacement of the original simply-supported bridge superstructure with the new continuous steel superstructure and columns strengthening by steel and concrete casings. This has reduced the weight of the overall structure and since the arch axis was designed as a thrust line for certain permanent load, it had to be checked to what degree does the redistribution of the total permanent load in relation to the original design change stresses in the arch. The static and dynamic testing of the reconstructed bridge proved the accuracy of the assumptions incorporated in the reconstruction design calculations. But, this also confirms that the stresses in the arch under the designed traffic loads are indeed approaching ultimate allowable values. This issue is of the utmost importance in planning repair of an arch bridge.

Reinforced concrete arch bridge; reconstruction; load testing; non-linear analysis

nije evidentirano