Received: 2026-05-03
Accepted: 2026-07-23
Published: 2026-10-02

Comparative assessment of FE modelling assumptions for railway filler-beam bridge decks based on proof load deflections

Jakub Jarosz

Abstract

Differences between computed and measured bridge deflections are frequently observed during proof load tests and can lead to conservative or inconsistent decisions in serviceability assessment and load rating. The article focuses on railway filler-beam decks and uses their proof load response as the basis for experimental validation of alternative FE modelling assumptions. Proof load tests performed on five single-span filler-beam railway bridges on the E30/C-E30 corridor are used to compare measured midspan deflections with finite element predictions obtained from three composite-action assumptions: A) full composite action of the concrete slab in compression and tension, B) partial composite action consistent with UIC recommendations and compatible with the general provisions of Eurocode 4, and C) compression-only composite action. The results demonstrate that models adopting full composite action provide the closest agreement with measured deflections, while conservative cracked-section assumptions significantly overestimate deflection. The study also quantifies the influence of secondary stiffness contributors, in particular service walkway slab, on both the magnitude and transverse distribution of deflections, and confirms that the observed trends are consistent across the analysed bridge set.

Keywords:

bridge assesment, composite action, deflection, finite element modelling, load testing, railway bridges, system behaviour

Download files

PDF

Citation rules

Jarosz, J. (2026). Comparative assessment of FE modelling assumptions for railway filler-beam bridge decks based on proof load deflections. Roads and Bridges – Drogi I Mosty, 25(3), 171–192. https://doi.org/10.7409/rabdim.026.012

Cited by / Share

Licence


This website uses cookies for proper operation, in order to use the portal fully you must accept cookies.