Behaviour of riveted stringer-to-floorbeam connections in cyclic load tests to failure. (September 2019)

• Record Type: Journal Article
• Authors: Tulonen, Joonas; Siitonen, Tuomo; Laaksonen, Anssi
• Is Part Of: Journal of constructional steel research. Volume 160(2019)
• Page Start: 101
• Page End: 109
• Publication Date: 2019-09
• Subjects: Rivets -- Stringer-to-floor-beam connections -- Load tests -- Steel fatigue -- Railway bridges; Steel, Structural -- Periodicals; Building, Iron and steel -- Periodicals; Acier de construction -- Périodiques; Construction métallique -- Périodiques; 624.1821
• DOI: 10.1016/j.jcsr.2019.05.018
• Languages: English
• ISSNs: 0143-974X

Abstract

Abstract: Structural connections in bridges supporting directly the railway track have cumulated large amount of load cycles from passing trains. The connections were not originally designed for fatigue and some of them have found to be in poor condition by inspection. To study their structural behaviour under large number of repeated load cycles, five 85 year old stringer-to-floorbeam connections from a decommissioned railway bridge were loaded in laboratory. Each connection had four angle plates connecting stringer webs to floorbeams and two connection plates connecting stringer top flanges together. During the load tests, the connection plate maximum stress ranged from 60 MPa to 150 MPa. It was estimated that the peak stresses under traffic were in the range of 40–60 MPa. Despite large stresses in load tests and long active use under traffic, only three connections failed after 0.5–1.6 million load cycles. Two connection tests were run out after 2.4 and 2.6 million cycles. By measuring strains and displacements continuously during testing it was seen that failures occurred gradually as components fractured slowly during tens of thousands of cycles. Any of the connection failures was not complete and the connections showed substantial residual bending stiffness and capacity after failure. Load tests showed that even if connection is seen to be in poor condition and close to its theoretical fatigue life, it still might function safely for hundreds of thousands of stress cycles if stress redistribution between connection components is possible and connection is monitored for major fractures. Highlights: Three out of five 85 year old riveted connections failed in heavy cyclic load tests. Despite their poor initial condition, the connections were very robust. Though connections were similar in construction, their bending behaviour was not. Secondary bending stresses in tension members may lead to quicker failure. Health monitoring with strain and deflection measurements proved to be challenging.