Structural optimisation of single span steel-concrete high-speed railway bridges
Time: Fri 2025-06-13 13.00
Location: M108, Brinellvägen 23, Stockholm
Video link: https://kth-se.zoom.us/j/63902445951
Language: English
Subject area: Civil and Architectural Engineering, Structural Engineering and Bridges
Doctoral student: Marcel Hofstetter , Bro- och stålbyggnad
Opponent: Associate Professor Gunnstein Thomas Frøseth, NTNU – Norwegian University of Science and Technology, Trondheim, Norway
Supervisor: Professor John Leander, Bro- och stålbyggnad, Byggteknik och design
QC 20250522
Abstract
This thesis focuses on optimisation of single span steel-concrete composite high-speedrailway (HSR) bridges. The research aims to investigate how different parameters, suchas span length, steel grade, cross-section type, and damper usage, influence the designof optimised bridges. A case-study bridge with an installed long-term monitoring systemis used in a model updating procedure using error domain model falsification (EDMF)and then optimised by a Genetic Algorithm (GA). Then, a parametric study on theaforementioned parameters is performed, using the bridge as a reference.The thesis is a compilation of three papers, in which the first, Paper I, a simplified 2D FEmodel of a simply supported bridge was calibrated using EDMF. A sensitivity analysiswas performed to investigate the parameters that govern the dynamic response. Thedecisive parameters were then updated using falsification to obtain accurate parameterranges. It is shown that the most accurate dynamic behaviour is achieved using updatedmaterial parameters instead of Eurocode parameters, and that falsification is an efficientapproach to perform model calibration.In Paper II, the calibrated model was subjected to an optimisation procedure, wherethe mass of the steel girders of the bridge was minimised. The constraints were Eurocodedesign limits in the ultimate limit state (ULS), serviceability limit state (SLS), and fatiguelimit state (FLS). Dynamic evaluation was performed using a single degree of freedom(SDOF) system and Duhamel’s integral. Two optimisations were performed, one thatincluded retrofitted viscous dampers on the bridge, and one that did not. Using accuratematerial parameters obtained in Paper I, results show that the ULS in conjunction withaccelerations limits the design of the bridge.In Paper III, the case-studied bridge from the previous two papers was subjected to aparametric study. Instead of using calibrated material parameter values, Eurocode valueswere used to provide generality. In the parametric study, simply supported steel-concretecomposite railway bridges were optimised to investigate how different parameters affectedthe optimised design. Five different steel grades (S235, S275, S355, S420, and S460), 51different span lengths (from 20-70 m using 1 m increments) and four different cross-sectiontypes (double I-girders, semi-box girder with joined lower flanges, double box girdersand box girder with inclined webs) were analysed. Results indicate that accelerationsand stresses in the SLS limit the design. Results also indicate that the double I-girderprovide the lowest mass, and that mass of the steel girders (expressed in kg/m) is virtuallyunchanged for span lengths from 20-50 m.