Dynamics of ship collisions

The thesis studies ship collisions computationally and experimentally on large and model scales. On the basis of the experimental observations a 3D simulation model is proposed that couples the motions of the ships to the contact force, and considers all the major hydromechanical forces that act on colliding ships. Additionally, the effects of sloshing and the dynamic bending of the hull girder are investigated and implemented into the simulation model. Large-scale experiments were analysed in order to get a deeper insight into the collision dynamics. On the basis of the large-scale experiments a model-scale test setup is designed using the Froude's scaling law. There, the emphasis was laid on the external dynamics and the structural response, properly scaled from the large-scale test, was modelled using homogeneous foam in the side structure of the struck ship model. It is shown that the model-scale experiments illustrated the large-scale tests both qualitatively and quantitatively. A wide range of symmetric, both with and without sloshing, and non-symmetric collision scenarios are studied on a model scale. The experimental findings are exploited in the development of a coupled collision simulation model. The model is formulated in three-dimensional space, and the contact force between the colliding ships considers both the normal and frictional components. A discrete mechanical model for sloshing is implemented into this time-domain model. This linear sloshing model describes the fluid in partially filled tanks with a single rigid mass and with a number of oscillating mass elements that interact with the ship structure through springs and dampers. The dynamic bending of the ship hull girder is included by modelling it as an Euler-Bernoulli beam. Both the experiments and the simulations emphasised the importance of the coupling between the motions and the contact force. It was especially obvious in the case of non-symmetric collisions and in the experiments with sloshing. The penetration paths calculated with the developed time-domain simulation model agreed well with those from the experiments. The total deformation energy was predicted with a deviation of about 10%. The hydrodynamic radiation forces acting on colliding ships proved to have a strong influence on the energy distribution as at the end of the contact they accounted for up to 25% of the total available energy. However, if the interest is in the maximum deformation, the approach with the hydrodynamic damping ignored yields an error of about 5% in the deformation energy. The results of the large- and model-scale experiments with partially filled liquid tanks emphasised the importance of sloshing for collision dynamics. The structural deformation energy in the tests with sloshing was only about 70%-80% of that in similar collision tests without sloshing. The simulation method with the linear sloshing model overestimated the deformation energy by up to 10% for low filling levels of water, but in the case of medium filling levels the predictions agreed amazingly well.