Completion at the end of last year of an elegant new railway bridge over the Hinterrhein River in eastern Switzerland has enabled work to start on a programme of renovation and improvement work on its historic neighbour. The new bridge, near Reichnau, has been dubbed Sora Giuvna — ‘Little Sister’ — in deference to the 19th century steel truss bridge that it stands alongside.
Both the new 200m-long bridge and its historic neighbour carry single-track railway lines, and they are in a highly significant and sensitive location. Not only is this the meeting point of the two principle tributaries of the upper Rhine, but the surrounding area includes notable bridges by Christian Menn, Max Bill and Mirko Roš, and used to be the site of innovative timber bridges by the Grubenmann brothers. Hence the location has long had a historical significance for Swiss bridge engineering.
Client for the new bridge is the Swiss regional rail company Rhätische Bahn; the construction of the new structure allows the last single-track stretch of line between Chur and Bonaduz to be eliminated, enabling two-way working on the Chur to St Moritz and Chur to Disentis-Muster routes.
As well as increasing capacity and robustness of these lines, the new layout with two independent single-track bridges will make maintenance easier. The bridges form part of the feeder line to Rhätische Bahn’s historical Albula route, which was added to the list of Unesco World Heritage Sites in 2008.
Consultants Cowi and Walt Galmarini developed the design for an international design competition in 2015, working with Danish architect Dissing & Weitling and local landscape architect Hager Partner. The steel superstructure construction was built by Schneider Stahlbau, Joerimann Stahl and Toscano Stahlbau, with the foundations and abutments by Erni Bauunternehmung.
All bridge designs must relate to their context, responding to the specific constraints presented by the location; in this case there were very particular challenges. The mountainous topography and beautiful landscape, the confluence of the two rivers and the many historical and cultural factors make this a very special place. The challenge was to respect these factors and above all to devise an elegant, modern design capable of standing directly alongside the historic 1896 truss bridge, complementing the original, while still retaining its own notable character.
The existing truss has an unusual bracing pattern, and a majestic presence, so the design team quickly concluded that the new bridge should be as transparent and slender as possible — not easy in a railway bridge — so as not to detract visually from the historic structure. Refurbishment of the existing bridge is part of the overall project and was planned to start after the new bridge was completed.
The landscape works that were necessary for the new bridge included cutting back the mountain slope between bridge and railway station to make room for the additional track; this solution was preferred to the alternative, which was to build large additional retaining walls, and it turned the focus towards the bridges instead. The design needed to find a synthesis between the environmental and technical constraints, between architecture, landscape and the art of engineering, enabling both bridges to shine.
In addition to crossing the river, the new railway line, which follows a gentle plan curved alignment, had to cross the busy A13 highway on the Reichenau side. The old concrete bridge which carried the existing railway over the A13 also needed replacement. The alignment meant that this crossing would be highly skewed, and with limited headroom clearance over the road, both of the new railway bridges required a shallow construction depth.
The ‘quadropods’ on the piers enabled a reduced main span of 63m (Roman Sidler)
As part of the competition, participants were able to choose whether to use a single structure to cross both the river and the highway, or whether to do so with separate bridges. Several factors supported a clear preference for combining them into a single bridge. Spatial constraints alongside the road, the highly skewed alignment, access difficulties for construction and the desire to respect the elevations of the existing bridge and its stone abutment all contributed to the decision. For visual consistency, the separate, single-span replacement bridge for the existing rail track had to have the same cross-section and depth.
One of the early decisions taken by the team was to build the bridge in steel. This was a decision driven as much by landscape and context considerations as technical and engineering expedience. The other notable road bridges in the area are generally concrete, and the existing railway bridge is steel; concrete for roads, and steel for railways. This seemed to fit with the spirit of the place and the historical context, in keeping with the characteristic steel structures of so much railway infrastructure. But also, as slenderness and transparency were going to be key design considerations, and as light weight was going to be an essential factor for constructability, steel became the natural choice. The bridge design needed to minimise the deck construction depth in order to provide sufficient headroom over the road. To achieve this, the structure is a trough girder, with a U-shaped cross-section formed by trapezoidal steel boxes on each side, with the tracks supported on ballast on top of a steel plate stiffened by shallow transverse cross-beams underneath. The cross-beams are closely spaced at 1m centres to keep stresses low, particularly for fatigue, and to keep the bridge soffit as clean as possible there are no visible longitudinal stiffeners.
In order to reduce the effective spans and minimise overall depth, the trough girder is supported on inclined integral props standing on concrete piers aligned with the stone columns of the existing bridge. This results in a main span of 63m and a girder depth of only 1.7m, achieving the desired slenderness of the concept. The inclined props form a four-point support to the girder, creating so-called ‘quadropods’ which became a major feature of the design. Careful shaping and detailing of these quadropods was required, particularly at the top where they taper to a welded connection to the girder soffit at a point where fatigue is a primary design consideration. The quadropods are integral with the girder, which acts as a tie between them, and are supported on the slender concrete piers by large-diameter pin bearings concealed behind removable steel cover plates.
Between the river and the road there is a small triangle of land within which a supporting pier was required alongside the stone abutment of the old bridge. After testing various options, the team converged on a solution involving a skewed V-shaped pier placed parallel to the road. This avoided any headroom clearance issues over the edge of the highway and created a coherent language for the support arrangement over the entire bridge length by echoing the shape of the quadropods. When viewed along the bridge axis, the V-pier and the quadropods present a consistent form and a clear expression of the structural system. But the skewed arrangement also meant that the two longitudinal girders have different spans, introducing some interesting asymmetry in the bridge behaviour.
View along the centreline showing pier-top bearings, which will be concealed behind removable cover plates (Andreas Galmarini)
The design went through several iterations before settling on the optimum articulation scheme, which involves a combination of bearings that allow movement and continuous structures flexible enough to accommodate movement. The bridge is fixed at the west abutment, and all longitudinal traction and braking forces are accommodated here. In keeping with the bridge owner’s request, there is just one expansion joint to accommodate all thermal and other longitudinal movements, at the extreme east end. In between, the bridge girder is fully continuous and integral with the quadropods.
This means that the concrete piers have to be flexible enough to accommodate some longitudinal deflection at the top, where the pin bearings are, and the quadropods experience rotations with one pair of arms moving up and one pair moving down under the passage of a train.
At the V-pier, sliding bearings support the girder to allow longitudinal movements, and the legs are tied together at the top with post-tensioned high strength steel bars in order to stop them spreading apart under load. There are three bars to permit them to be replaced one at a time in future, should that be necessary.
The slender concrete piers stand on piled foundations, and the design brief required that pile caps were placed at a low level, approximately 7.5m below ground level on the east side, to reduce the risk of undercutting in the event of extreme scour. Thus the piers extend some distance below ground level, increasing their flexibility and enabling them to accommodate the required movements.
There is a fine line between allowing sufficient flexibility for thermal effects while at the same time retaining enough strength and stiffness to support the loads. This was the subject of extensive investigation.
The initial concept was to fix the bridge longitudinally at the east abutment, and terminate the bridge at a skew angle, parallel to the road. However, with this arrangement, the girder would have experienced a large twist under railway loading, and more importantly a very large rate of change of twist as trains passed over the abutment. This would not only have infringed the serviceability limits of the code, but no doubt would have also caused discomfort for the passengers.
The sharply-skewed end means that the vertical deflection of the two edge girders under a passing train is very different, causing a transverse crossfall or twist of the section, so the train leans over sideways. As it passes from the bridge onto firm ground beyond the abutment, the train returns upright again, and it is this dynamic rocking motion that needs to be controlled. The problem was compounded when the decision was made to move the longitudinal fixed point to the west abutment, because a skewed expansion joint is definitely undesirable.
So the skewed end was replaced with a square end, involving an extra 11.5m of steel girder on one side. This extra length of span would have required a significantly deeper and heavier girder, at least on that side, but this was highly undesirable. So the solution was to introduce an additional bearing at the point where the skewed end bearing would have been. With this arrangement the axial rotation still occurs at the intermediate bearing position, but there is no sudden change in rotation as the train passes over the abutment.
In order to avoid uplift at the end bearing when train loads are applied in the span over the road, it was necessary to control the amount of permanent load in this intermediate bearing. Thus the bridge was first installed without the intermediate bearing so that all the permanent load at that end of the girder was initially carried by the end bearing. Then the intermediate bearing was jacked in to a pre-determined load, reducing the permanent load on the end bearing but not so much that uplift would occur under transient live load.
Lifting the eastern section onto the western section ready for the launching operation
Building a structure so close to a live railway line with overhead electrification and limited space for temporary works, coupled with the proximity of the busy highway, forced the team to carefully consider possible construction methods from very early stages. It was clear that the steel trough girder, quadropods and V-pier — some 960t in total — would need to be fabricated off site and delivered in short sections by road for site assembly. This is common practice and achieves best quality with minimum disruption, reducing risk and construction time.
An accessible area of level ground beside the river on the west bank was available for lay-down and assembly, and a temporary river bank extension was permitted here so as to reduce the reach of the crane. A smaller assembly area could also be made available behind the eastern abutment alongside the railway. So the questions became: how large could the assembled pieces be? Heavier pieces meant a larger crane but fewer operations — how could they be erected? By crane, by launching, or both? And could it all be done without additional temporary supports?
The eastern spans were launched over the western spans in a single operation
The team established that a large crawler crane standing on the extension of the river bank in the assembly area would be able to reach far enough to install the quadropods and the river spans as individual lifts, and could just manage to place the V-pier. The bridge is fully welded throughout, so this meant welding the girder sections together and to the quadropod legs out over the river. But the real challenge was how to place the girder sections over the V-pier and the highway.
The busy road could only be closed to traffic for a very few nights and there was no possibility of access off the road at that point or room to place a crane there. The initial idea was to assemble the spans on the east side and launch from there towards the west. In the end, the solution chosen was to use the crane to install the four western sections from the west abutment to the eastern quadropod, and then to assemble the eastern sections as a single piece, place it on top of the already erected spans, and launch it across the road in a single operation, finally lowering it into position. This all went remarkably smoothly.
Finally, once all the sections had been fully welded and the final paint system applied, with the ballast added and tracks and railway equipment installed, it was time for the load test. All went well with the bridge performing satisfactorily, and the new track was successfully opened to rail traffic in November 2018. After completion, trains were diverted onto the new bridge to enable refurbishment and improvement works to be carried out on the original structure and for the existing concrete bridge over the road to be removed. Then the 51m-long replacement span for the original rail track, which has the same form as the new bridge, will be launched over the road, with completion expected in late 2019.
Ian Firth is a consultant at Cowi; Andreas Galmarini is director and Matthias Ludin bridge engineer at Walt Galmarini; and Steen Savery Trojaborg is managing director at Dissing & Weitling.