We carried out comprehensive statical analysis for prestressed concrete railway bridges in the DB German Rail network, including refurbishment with new cross-sections and added reinforcement.
Germany’s rail network carries a large stock of prestressed concrete bridges built from the 1950s onward. Most are in sound condition. The reason they come up for assessment is usually not deterioration but change: load models have been revised, traffic has grown heavier, clearance requirements have tightened, or the line is being upgraded and the structure has to accommodate something it was never designed for.
An assessment answers a narrow question. Not “is this bridge adequate” in general, but “does this bridge satisfy current requirements for the loading it will now carry”. A structure can pass comfortably under the code it was designed to and still fail against a current load model.
The analysis has to be built on what the structure is, not what the drawings say it should be. For prestressed concrete that means resolving several unknowns.
Prestress losses. Creep, shrinkage and relaxation reduce the effective prestress over decades. The remaining force is what governs behaviour under service loads, and it has to be estimated from the age and history of the structure rather than assumed at the design value.
Tendon condition and position. Duct grouting quality and tendon layout affect both durability and capacity. Where records are incomplete, this needs investigation rather than assumption.
Concrete and reinforcement properties. Actual material strengths often exceed the specified values, which can work in the assessment’s favour, but only if they are established by testing rather than assumed.
Getting these right matters more than refining the analysis model. A precise calculation built on an assumed prestress value is a precise calculation of the wrong structure.
Detailed finite element simulation was used to assess structural integrity and performance under a range of load conditions. For a prestressed deck the model needs to represent the prestress as an applied action rather than a material property, capture the construction sequence where it affects the locked-in stress state, and cover the load cases that govern at different points along the span.
Rail loading adds requirements that road bridge analysis does not have. Dynamic amplification has to be applied to the static load model, and deflection and end rotation have to be checked against track geometry limits, not just structural ones. On refurbishment these serviceability checks frequently govern, because the existing structure was designed against limits that have since been tightened.
The refurbishment included upgrading the bridge’s cross-section and adding new reinforcement to improve durability and safety.
Widening a deck changes more than the deck. Additional width adds permanent load across the whole span, shifts the load path into the existing bearings and substructure, and alters how the structure distributes load transversely. The supporting elements have to be checked against the revised loading even where no work is planned on them, and it is common for the substructure rather than the deck to become the limiting element.
Added reinforcement has to be made to work with a structure that is already carrying load. New reinforcement only takes up force from additional load applied after it is installed, so it does not relieve the existing stress state. Where strengthening is needed against permanent load, the structure has to be unloaded first, or the analysis has to account for the fact that the new material starts from an unstressed position.
None of this happens on a closed site. Access is limited to possession windows, often at night and measured in hours. That constraint shapes the engineering rather than just the programme. Details that can be installed quickly, in sequence, and checked immediately are worth more than details that are marginally more efficient but need extended access.
Analysis and design were carried out to the Eurocodes as adopted in Germany, DIN EN 1990 through 1998, with DB guideline drawings governing standard details such as cap widening and parapet adaptation. Finite element modelling used SOFiSTiK, Dlubal and RFEM. Drawings were produced in Nemetschek and AutoCAD.
The output is a static analysis document suitable for submission and approval, containing the analysis, cross-sectional drawings, plan overviews and the FEM model results.
]]>We have carried out design and analysis of oblique-angled concrete frame structures for railway bridges in the DB German Rail network, on more than one occasion.
Skew is rarely chosen. It is imposed by whatever the structure crosses, since roads, watercourses and other railways run where they run. Once the angle is below roughly seventy degrees, a set of effects appear that a square-structure analysis does not capture.
Load does not travel the way the span suggests. In a skew slab or frame, load takes the shortest stiff path to the supports rather than running parallel to the free edges. The effective span is shorter than the measured one, and the load concentrates along the line connecting the obtuse corners.
The obtuse corners attract reaction. Support reactions are not distributed evenly along the abutment. They concentrate sharply at the obtuse corners, and the higher the skew, the sharper the concentration.
The acute corners can lift. As reaction concentrates at the obtuse corners, the acute corners shed load and can go into uplift. A bearing designed only for compression is then in the wrong condition, and the structure has to be checked for whether hold-down is required.
Twisting moments become significant. Skew slabs carry a substantial part of the load in torsion. Principal moments rotate away from the span direction, which means reinforcement laid out on the span axis is no longer aligned with the way the structure actually wants to carry load.
Skew geometry rules out simplified strip methods. The structure has to be modelled as a plate or shell so the twisting moments and the corner effects appear in the results rather than being smoothed away.
The reinforcement then has to be resolved from the principal moment field. Bars are practical to place either parallel to the free edge or parallel to the support line, and neither direction matches the principal moments across the whole deck. The design has to convert the principal moments into resistance in the directions the bars actually run, which is what governs the layout in the corner regions.
Corner detailing carries a disproportionate share of the design effort. The obtuse corners need reinforcement against the concentrated reaction and the local hogging that develops there, and this is the region where skew structures show distress if the effect has been underestimated.
For a frame rather than a simply supported deck, the walls and the deck act together. The skew angle carries through the whole frame, so the connection between deck and walls has to transfer moment across a joint that is itself oblique. Continuity between the elements changes both how load distributes and how the structure responds to settlement and temperature.
Earth pressure on the walls has to be considered alongside the structural loading, and for a rail structure the surcharge from the track and its loading is part of that. Geotechnical input and structural analysis are not separable on this kind of frame.
Design and analysis followed the Eurocodes as adopted in Germany, DIN EN 1990 through 1998, including the sections covering concrete design and actions on structures from rail traffic. Finite element modelling used SOFiSTiK, Dlubal and RFEM, with GGU used for geotechnical checks on the frame walls and foundations.
The deliverable is a static analysis document covering the model, the load cases, the reinforcement design and the corner detailing, in a form suitable for approval.
]]>This project covered design and analysis of a concrete frame structure with a special anchor foundation for a railway bridge in the DB German Rail network.
The section through this structure records what the analysis had to work with: distinct soil layers each with their own unit weight and friction angle, a groundwater level sitting above founding level, and a bored pile solution taken down past the weaker material.
Those numbers set the problem. Layers with low friction angles will not carry the frame on spread footings at a sensible depth, so load has to be taken to competent material below. Groundwater above founding level introduces uplift on the structure and requires the excavation to be supported and dewatered rather than simply cut.
A buried frame in high groundwater has a problem that surface structures do not. The structure displaces water, and the water pushes back. When the buoyant force exceeds the weight of the structure and the material above it, the frame wants to float.
There are three ways to deal with it. Add enough weight to hold the structure down, which is expensive and adds load to ground that was weak to begin with. Rely on friction from surrounding material, which is unreliable and cannot be counted on in saturated ground. Or anchor the structure into competent material below, which is what a special anchor foundation does.
Anchored piles work in tension rather than compression, which reverses the usual design assumption. The pile has to develop uplift resistance through shaft friction along its embedded length, the connection between pile and base slab has to transfer tension rather than bearing, and the reinforcement has to be continuous through a joint that would otherwise just be a construction detail.
The governing case is often not the finished structure in service. It is the structure complete but not yet backfilled, when the weight holding it down is at its lowest and the water table has recovered after dewatering stops. That temporary condition has to be checked explicitly, and it frequently controls the anchor design.
A frame founded on piles cannot be analysed as a frame sitting on fixed supports. Pile stiffness affects how moment distributes through the frame, and differential settlement between piles introduces forces that a rigid-support model never produces.
The practical approach is to model the piles as springs with stiffness derived from the geotechnical assessment, then check the frame against a range of stiffness values rather than a single figure. Ground stiffness is never known precisely, and a design that only works for one assumed value is not a design.
Earth pressure on the frame walls is part of the same system. Active and at-rest pressure give different answers, and which applies depends on whether the wall can move enough to mobilise active conditions. For a stiff closed frame it usually cannot, so at-rest pressure governs and the walls carry more than a retaining wall of the same height would.
Design followed the Eurocodes as adopted in Germany, DIN EN 1990 through 1998, with DIN EN 1997 governing the geotechnical design of the piles and the verification against uplift. Structural analysis used SOFiSTiK, Dlubal and RFEM. Geotechnical checks, including pile capacity and earth pressure, were carried out in GGU.
The static analysis document covers the frame model, the pile design in both compression and tension, the uplift verification for the temporary and permanent cases, and the reinforcement drawings for the pile-to-slab connection.
]]>This project covered design and construction of a composite steel and concrete railway bridge with three spans of thirty-three metres for the DB German Rail network.
A composite deck puts steel girders below and a reinforced concrete slab above, connected so the two act as a single section rather than two independent elements.
Under sagging moment, which governs most of the span, the concrete slab is in compression and the steel girders are in tension. Concrete is strong and cheap in compression. Steel is efficient in tension. The neutral axis sits high in the section, close to the slab, which means most of the steel is working at or near its full capacity rather than sitting near the neutral axis contributing little.
The result is a shallower and lighter deck than reinforced or prestressed concrete would give at the same span. On a railway that matters twice over: less permanent load into the piers and foundations, and a shallower construction depth, which often decides whether the required clearance underneath can be achieved at all.
Composite action exists only if the slab and the girders are prevented from sliding relative to each other. The shear connectors that do this are what make the section behave as one, and they are designed for the longitudinal shear flow at the steel and concrete interface rather than for any load applied directly to them.
That shear flow is highest near the supports, where the rate of change of moment is greatest, and lowest at midspan. Connector spacing follows this, which is why it varies along the girder rather than being uniform.
For a railway bridge, fatigue governs the connection rather than static strength. The structure sees a large number of significant load cycles over its life, and the connector detail has to be assessed against fatigue as a primary check.
Three continuous spans behave differently from three separate simply supported ones. Continuity reduces the sagging moment at midspan and lets the deck be shallower, but it introduces hogging moment over the intermediate piers, and there the advantage of composite action reverses. In hogging the slab is in tension, where concrete contributes little and cracks, and the lower steel flange is in compression, where it may buckle without restraint.
The pier regions therefore need heavier reinforcement in the slab to control cracking, and attention to restraint of the compression flange. It is normal for the section to change along the length of a continuous composite deck, with more steel over the supports than at midspan.
Continuity also means the structure is sensitive to support settlement in a way a simply supported deck is not. Differential settlement between piers redistributes moment through the whole deck, which is why the substructure was modelled together with the deck rather than treated as fixed support points.
The FEM work covered the deck slabs and the piled piers as one system. Modelling the substructure alongside the deck is what allows settlement sensitivity and foundation stiffness to feed back into the deck design.
Time-dependent effects need explicit treatment on a composite structure. Concrete creeps and shrinks, steel does not. Shrinkage of the slab is restrained by the girders, which puts the slab into tension and the steel into compression before any traffic load arrives. Creep gradually transfers stress from the concrete into the steel over the life of the structure. Short-term and long-term conditions give different answers, and both have to be checked.
Design followed the Eurocodes as adopted in Germany, DIN EN 1990 through 1998, with DIN EN 1994 governing composite design and the fatigue provisions applying to the shear connection and the steelwork. Analysis used SOFiSTiK, Dlubal and RFEM. Steel detailing was produced in Tekla, with drawings in Nemetschek and AutoCAD.
Construction supervision followed the design, checking that the shear connection, the concreting sequence for the slab and the erection of the girders matched the assumptions the analysis was built on.
]]>