Vereinfachte dynamische Bemessung von WiB-Eisenbahnverbundbrücken für den Hochgeschwindigkeitsverkehr

Bigelow, Hetty; Hoffmeister, Benno (Thesis advisor); Feldmann, Markus (Thesis advisor); de Roeck, Guido (Thesis advisor)

Aachen : Shaker (2018, 2019)
Book, Dissertation / PhD Thesis

In: Schriftenreihe Stahlbau - RWTH Aachen 82
Page(s)/Article-Nr.: 1 Online-Ressource (iii, 186, XXXVI Seiten) : Illustrationen, Diagramme

Dissertation, RWTH Aachen University, 2018


Filler beam railway bridges have many advantages. They are durable, can feature high slenderness and require only a small amount of framework material during construction. They are usually constructed as simply supported beams, i.e. their high slenderness can also be in some respects disadvantageous. Calculated eigenfrequenciesare often relatively small, thus leading to anticipation of resonance effects already at low crossing velocities induced by equally spaced axle loads. Measurements have repeatedly shown though, that filler beam bridges behave much better in reality than predicted by calculations. The anticipated resonance effects occur eventually at much higher crossing velocities than predicted. The sometimes significantly high differences between measurements and calculations have been generally noticed before and are affiliated with additional contributions of non-structural elements to system stiffness and damping, e.g. the contributions of ballast, tracks, sleepers or edge caps. The thesis at hand presents a detailed examination of the contributions of individual parameters influencing the dynamic behavior of filler beam bridges. The examination aims at separating parameters, which can be clearly identified and isolated from other effects and thus could be considered in design, from those parameters, that require further extensive research. Experimental tests regarding interaction effects between bridge decks, which are separated by longitudinal gaps but share a ballast bed, are performed with a newly developed test set up. This set up eliminates effects, which occur simultaneously with the interaction effects on real bridges. The contribution of the interaction effects to stiffness and damping are tested. Based on German codes, modelling simplifications are derived. The derivation of a horizontal equivalent spring stiffness, representing restraining effects of railway tracks exceeding bridges, enables a significant reduction of modelling effort and computational time. The equivalent horizontal spring, which is effective in the centroidal axis of the track, is afterwards converted into an equivalent rotationalspring, which can be applied at the hinges of simple beam models, thus simplifying the design of simple span bridges. The developed beam system with rotational springs is then converted into an equivalent single degree of freedom (SDOF) system. With the derived analytical formulas, the fundamental frequency n0 can directly be calculated considering there straining effects of tracks. Users can now include the restraining effects of tracks into a simplified estimation of resonance risks, where n0 is compared to limiting values given by the codes. If resonance can be ruled out and crossing velocities are 200 km/h (56 m/s) at most, the dynamic design of single span bridges with spans up to 40 m (thus the scope of application of filler beam bridges) can be performed with equivalent static loads only. If resonance cannot be ruled out using the simplified approach, or, if crossing velocities are above 200 km/h (56 m/s), dynamic simulations have to be performed considering all train types designated for the bridge in question. The results of all simulations then have to be interpreted. In scope of the presented thesis, computation tools using SDOFs are programmed, which enable automated simulations of large numbers of train crossings in a short time, including the restraining effects of tracks. The tool requires only the input parameters bending stiffness, mass, span length, the rotational spring stiffness just derived in this thesis and a damping ratio provided by the codes. This again leads to a significant reduction of modelling effort and computation time compared to conventional calculation software.