Modelling and Analysis of a Floating Bridge

Abstract

The Norwegian National Public Road Administration is working on a project to improve the coastal road E39, connecting the cities along the west coast of Norway. Today, this road has seven ferry crossings which are to be replaced by permanent connections, to a total expected cost of 340 billion NOK.

Several of these fjords are wider and deeper than what existing designs can manage, such that new technology needs to be developed. For some of the fjord crossings, floating bridge concepts have been developed and concluded to be feasible solutions. For the about 4500 meters wide Bjørnafjorden, south of Bergen, there are two floating bridge concepts which are in the process of being further assessed.

One of these concepts is a straight bridge, laterally supported by pre-tensioned mooring lines. This concept was modelled in the software SIMO-RIFLEX, where a static, eigenvalue and dynamic analysis were performed in order to illustrate modelling aspects and calculation procedures. Panel models of the pontoon were modelled in GeniE and imported into HydroD where hydrodynamic analyses were carried out in Wadam.

From the eigenvalue analysis, a significant limitation was identified in the eigenvalue calculation codes in SIMO-RIFLEX, as the catenary mooring lines were not properly accounted for. Therefore, a second model was made where the mooring system was replaced by a linearised implementation. For this model, the eigen periods and mode shapes were coinciding well with those obtained by the NPRA. The first 30 eigen periods were differing with a mean value of 3.9 % when only the infinite-frequency added mass matrix was considered. By updating the added mass for a selected set of modes, differences of less than about 2 % were found.

The eigenvalue calculations revealed several modes that can be triggered by environmental loads. Laterally dominated modes at high periods with negligible damping, which can be important for the response in slowly varying wind, and laterally dominated modes close to the peak period for the 100-year wind waves, were identified. Additionally, modes dominated by pontoon motions along the bridge girder close to this period were found, possibly important for the dynamic weak axis bending moments in the high bridge.

From the dynamic analyses in regular waves, response patterns related to the identified modes were present when the bridge was subjected to waves from different directions, respectively. The maximum weak axis bending moment in the bridge girder for the conditions examined was found in the high bridge for a response pattern related to the mentioned modes dominated by pontoon motions along the bridge girder. This moment had a magnitude of 9.1E + 05 kNm, where the dynamic moment only contributed with 15 %. The results from the analyses performed were therefore seen to give indications on possible room for girder length to girder height ratio optimisation and should be further assessed based on analyses performed for the actual environmental conditions in the fjord.