The Qualitative Influence of Model Parameters on the Dynamic Response of a Viaduct

Abstract

The qualitative influence of several model parameters on the dynamic response of a railways viaduct, modelled after an actual structure located in Santana do Cartaxo, Portugal, is analysed through a parametric statistical analysis. Two-level factorial experimentation was carried out with the help of a parametric model of the viaduct developed with ANSYS APDL. The range of values admitted for the key parameters and the corresponding deviations from mean values were evaluated from data made accessible by the road authorities and experience. For the sake of simplicity only one axle load was applied on a single module of the viaduct. It was concluded that the analysis of individual effects is satisfactory to obtain key results based on maximum displacements. Nevertheless, regarding peak accelerations, especially at the viaduct deck, interaction effects are important and the influence of key parameters cannot be analysed individually. Therefore, simultaneous consideration of several key parameters provides a better representation of reality associated with the application of statistical analysis.

A statistical analysis of the dynamic response of a railway viaduct

“…The Santana do Cartaxo segment, where a new railway was included over a viaduct built at km 59 + 000 to km 60 + 000 ( Figure 1) is an exception in the rehabilitation, which otherwise follows closely the original railway route design. A more detailed description of the structure can be found in previous work of the authors [13]. Figure 2) the viaduct develops linearly and at the end starts a left transition curve of 1750 m radius.…”

“…In the present work the moving force model can be safely used since the ratio of the moving mass load over the mass of the bridge does not exceed 30 % and the load velocity will not reach 20 % of the crit-ical one, as shown by parametric analysis [12]. This simplification was already used in the authors' previous work [13] and it is also mentioned in the monographs [1,2].…”

“…It is preferable to run several parametric studies first and gradually select the most relevant factors. In a previous work [13] ballast stiffness, concrete stiffness, soil stiffness, train speed, ballast damping and railpad damping were selected as key factors. In this context the concrete stiffness is represented by Young's modulus.…”

“…The ballast behaviour model is considered in this paper as the only qualitative factor. Only one single force and one value for the reference train speed 180 km/h was used in [13]. Hence, special attention is placed here on the train speed and on a more realistic train model.…”

“…Springs represent the rigidity of the surrounding layers and dampers ensure smooth wave propagation into surrounding layers without reflections from the artificial boundary. The method of coefficients estimation of these representative springs and dampers and their validation is presented in the previous work [13]. Some remarks are also included in Appendix C. Springs and dampers are also used to model the embankments.…”