Case Study

Cost-Effective Steel Girder Rail Bridge Assessment

• steel girder rail bridge
• buckling assessment for 40 tonne lorries
• strength of structure proved by LUSAS

Many steel bridges have been shown to fail assessment to design code BS5400 Part 3. Two consulting firms, both using LUSAS Bridge have proved a 40 tonne capacity on a structure that, according to the code, had insufficient lateral restraint and would have required a severe weight restriction.

All bridges in the UK need assessing for a 40 tonne weight limit in order to meet new European legislation. As part of its bridge assessment program Railtrack plc required a 3 span skewed structure at London Road, Hackbridge, London to be assessed. The bridge carries the A237 across 2 railway lines south of Hackbridge station and consists of 2 brick arch approach spans and a central 12m span comprising steel edge girders. These girders support brick parapets and main steel troughing spanning between the masonry piers and onto their bottom flanges. An initial simple assessment identified a possible deficiency and, following discussion with Railtrack reviewing engineer Pell Frischmann, a more rigorous analysis was commissioned.

By using LUSAS Bridge, Robert West Consulting built a model of an edge girder. In a straightforward modelling approach, QSL8 shell elements, which take account of both membrane and flexural deformations and allow for eigenvalue extraction as required for the buckling analysis, were used to model the web and flanges of the girder. BSL3 beam elements modelled the flange to web angle connections and transverse web stiffeners. A 40 tonne deck assessment analysis provided the vertical loading for the girder and was applied as a series of increasing point loads to simulate the triangular loading of the deck troughing. A UDL represented the self weight of the girder together with the parapet load supported by the top flange.

In the eigenvalue buckling analysis the first 7 fundamental buckling modes (or eigenmodes) were investigated. The eigenvalues or load factors obtained were then applied to the initial analysis model to obtain the stresses in the top flange just prior to buckling. These stresses were then used to determine the limiting compressive stress at the critical buckling stage in the top flange in accordance with BS 5400 Part 3 Cl. 9.7.5. Additional calculations to confirm the stability of the girder under a lateral overturning moment and sliding force were carried out in accordance with Cl. 9.12.4 of the code.

In performing the category 3 check on the bridge Design and Building Services of Sheffield City Council, who perform numerous bridge assessments of this type, adopted a slightly more detailed modelling approach. QTS4 thick shell elements were used for all parts of the edge beam girder including the web stiffeners. By using rigid supports at abutments, spring supports at the position of transverse edge beams and a slightly finer mesh it was shown that the edge beams had adequate strength, confirming the original bridge assessment. A minimum load factor of 2.55 was obtained indicating that using spring supports to model the restraint provided by the deck and stiffeners can be significant.

Bridge assessment using the BS5400 code permits making use of finite element analysis to calculate effective lengths but does not make it an obvious alternative to using standard code provisions. The draft version of Eurocode 3 Part 2 - Bridges and Plated Structures explicitly states that eigenvalue analysis using finite element methods can be performed to ascertain buckling characteristics of a structure. This makes analysis software such as LUSAS Bridge an obvious alternative when carrying out an assessment.

LUSAS Bridge provides a very practical and cost-effective way of proving the integrity of all types of bridge structures that fail initial assessment to BS 5400 or indeed any other design code. It helps eliminate the need for otherwise potentially costly strengthening, remedial works and traffic management.