Investigation of seismic behaviour of steel railway bridges in Turkey with probabilistic aproaches

Turkey is located on three different active fault lines including North Anatolia, South Anatolia and West Anatolia earthquake zone, and a large part of the country is under high earthquake hazard risk. The earthquakes that occurred in the past and the loss of properties and lives that have occurred have shown that the seismic behaviour of existing structures needed to be examined in detail. During an earthquake, houses, public institution buildings, factories and transportation systems can be damage. The damages on the transportation systems prevent the humanitarian aid from being delivered to the side after an seismic event as well as have negative effects on the country's economy because of the failure of the transportation problems. Bridges are the most fragile elements in transportation systems. Although there are detailed studies on the determination earthquake performance for road bridges in the literature, a limited number of studies on railway bridges can be reached. Fragility analyzes are an effective method used to predict the earthquake performances of structural and non-structural elements. This study obtained fragility curves for the steel railway bridges in Turkey. Production of railway lines in Turkey began in 1856 and Izmir-Aydin in a railway line having the length of 130 km are obtained was built by a British company. Currently, there are over 12,500 km of railway lines in Turkey. The existing railway lines are divided into 7 regions to ease the maintenance, repair and operation. Railway lines have been extensively used in mine transportation and in recent years they have been increasingly used for passenger transport thanks to high speed and accelerated train lines. Thus, it contributes significantly to the economy of the country. Determination of the earthquake performances of the railway lines, which were mostly carried out in the late 19th century and early 20th century, is of great importance for the reduction of possible earthquake hazards and losses. This study also examines the steel railway bridges in Turkey which are categorized into four different classes. Steel railway bridge classes: simply supported steel beam bridge (DGBMÇ), simply supported steel truss bridge (TABMÇM), multi span continious steel truss bridge (CASÇMK) and multi span simply supported curved steel truss bridge (ÇABMÇM). Bridges representing each class of steel railway bridges are identified and mathematical models are created with the help of design calculation and drawings. The material properties of the bridges are determined as ST37 stell, taking into account the material properties considered in the design calculation and drawings and the material properties used for steel bridges in the past and published in the literature. Since the bridge supports have similar properties to the short type steel bearings, the short type steel bearing models given in the literature are used. The geometric conditions of the bridge elements are taken into consideration and are reflected in to the mathematical model with the help of link elements. Bridge first modes and modal mass participation ratios are determined. Nonlinear time history analysis methods have been utilized to be able to realistically determine the nonlinear behavior of bridges. 30 real earthquake records are selected and scaled according the maximum ground acceleration (PGA) of the earthquake record from 0.1 g to 1.0 g to perform THA. As a result of the analyzes made, the displacements of the bridge supports, displacement of the middle of the span and internal forces acting on the bridge elements are recorded. In order to obtain bridge fragility curves, damage limit states need to be determined. Four different damage limit states have been determined considering the plastic rotations for DGBMÇ bridge elements. For the steel truss bridge elements, axial comprression and tension capacities are defined as collapse damage state. EN 1990 Annex 2 horizontal displacement limits are used for displacement limits obtained for the middle of the bridge span, and the serviceability limit states are obtained for three different speeds of use. Four different damage conditions for low-type steel bearing are determined according to bearing displacements. Using result of the nonlinear analysis, probabilistic seismic demand analyzes are carried out in order to determine the relationship between the seismic demand of the bridge elements and the intensity measure. The probability seismic demand models for each bridge components are obtained separately and the regression analysis constants are shown in tabular form. Correlations between seismic demands of bridge components are also shown in tabular form. The convenience of the intensity measure used in the fragility analyses are determined as a result of the examination of 11 different paremeters in the literature. With the help of the probabilistic seismic demand models, the practicality, suitability, efficiency and sufficiency of the intensity measure parameters are compared and the intensity measure parameters that best represent the seismic demands of the steel railway bridges are determined. Three different analytical fragility analysis methods in the literature are investigated in order to analyze the fragility of bridge elements. Fragility curves are obtained with the help of truncated incremental dynamic analysis (IDA) in the Baker method. Fragility analysis is carried out using seismic demands obtained by proportioning earthquake records up to a certain seicmic intensity measure, so that the run of large analysis is alleviated. Separate fragility curves are obtained for each bridge components so that the most fragile ones are identified, and the performance of the bridge components is evaluated for possible retrofitting projects. For each bridge component, the median and dispertion values obtained as a result of the brittleness analysis are shown with the help of tables. The median values of DGBMÇ bridge components fragility curves obtained for the slight, medium, large and collapse damage state conditions are close to each other. It is determined that the median values obtained for the damage limits are close to each other since the load redistribution on the bridge members can not be realized after the formation of the plastic hinge in one of the bridge members. Longitudinal direction of the fixed support and the horizontal direction of the roller support are determined as the most brittle elements for the DGBMÇ bridge. Only the upper wind brace of the TABMÇM bridge components is reached the collapse limits. The collapse of the upper wind brace affects the buckling lengths of upper chord of the main truss and reduces the compression capacity of the bridge so it may cause the bridge to collapse. For this reason, the collapse of the upper wind brace is considered as a collaspe damage state. Analysis for ÇASÇM bridge components show that the longitudinal direction of the fixed support and the horizontal direction of the roller support are the most fragile components. Analyzes made for the ÇABMÇM bridge members show that the transverse and truss column elements are the most fragile elements. Due to the ballast load on the bridge, ÇABMCM bridge has different nonlinear behaviour from other truss bridge. In order to obtain the system fragility curves of bridges, existing methods in the literature are examined. From the largest probability of damage determined for each bridge element, is used to obtain the system fragility curves. As a result of the system fragility analysis, median and dispersion values are obtained for four different damage limit states and these values are compared with the fragility analysis median and dispersion values obtained for steel road bridges in the literature. The system fragility curves obtained for the DGBMÇK, TABMÇMK and TASÇMK yield similar results with fragility curves in the literature. The fragility curves obtained for the ÇABMÇM bridge are more fragile than the fragility curves in the literature. With the help of the Turkey earthquake hazard map web interface PGA 475 year and PGA 2475 year period seismic intensity values are optained for the analysed bridge and with using derived fragility curve damage probability of the bridges are obtained. ÇASÇMK is determined as the bridge having the highest probability of damage and TABMÇMK is determined as the bridge with the lowest probability of damage.