Hydrodynamic and structural aspects of ship collisions

Abstract

Ships and offshore installations operating at sea are exposed to the risk of collision accidents, which may cause severe consequences. It is therefore crucial to understand the physics of the collision process and to design structures so that they have sufficient strength to avoid progressive collapses. A traditional way of dealing with ship collisions is to decouple the process into the external dynamics and the internal mechanics. The external dynamics deals with ship motions and energy absorption while the internal mechanics handles the deformation and resistance of structures. The decoupled method is simple to implement and gives fast predictions of the dissipated energy and structural damage. However, the method fails to capture the influence of ship motions and the consideration of hydrodynamic forces is crude, which may lead to poor prediction accuracy in skew collisions with small collision angles and collisions with long durations. To better understand the fluid structure interactions in ship collisions, one main purpose of the thesis is to couple the external dynamics and the internal mechanics and to discuss the influence of hydrodynamic forces and six degrees of freedom (6DOF) ship motions on the damage prediction of the decoupled method.

By taking advantage of the user defined load subroutine and the user common subroutine, two coupled approaches for ship collision simulation were developed by implementing two different hydrodynamic models into the nonlinear finite element code LS-DYNA. The first approach uses a traditional ship maneuvering model for the in-plane surge, sway and yaw motions and three single-degree-of-freedom (SDOF) spring-damper vibration systems for the out-of-plane heave, roll and pitch motions. This method provides improved accuracy of hydrodynamic representation, and more importantly, enables a full 6DOF coupled dynamic simulation of ship collisions for the first time without simplifying the collision resistance. The second approach improves the representation of hydrodynamic forces by using linear potential flow theory. Models both with and without considering the forward speed effect were implemented.

Various collision scenarios were simulated with the proposed coupled models, including colliding with oblique plates, grounding on a sloping sea floor, crushing into rigid plates with normal vectors misaligned with coordinate axes, and collision with a submersible platform. Note that the struck objects are assumed to be fixed in the studies, but there should be no limitation for the coupled models to account for the 6DOF motions of the struck objects.

A comparative study was carried out, where the accuracy of the decoupled method to predict the demand for strain energy dissipation and damage extent was checked. A new phenomenon of ‘secondary impacts’ was observed when the periodic motions of heave, roll and pitch were introduced in the coupled method. This is not accounted for in the de-coupled method. The influence of hydrodynamic forces, 6DOF ship motions and the forward speed effect on the energy dissipation was investigated. The assumptions and simplifications of the external dynamic models were reviewed and the validity was discussed by comparison with the coupled simulation results. Potential limitations of the external dynamic models were pointed out. In the second part of the thesis, the internal deformation mechanics of two of the most commonly used structural components in offshore industries, i.e. tubular members and stiffened panels, were studied by the use of numerical simulation and simplified analytical methods.

The responses of offshore tubular members subjected to vessel bow and stern impacts were investigated using LS-DYNA. Extensive collision simulations with a total energy of 30-50 MJ were carried out with varying tube diameters, lengths and thicknesses. The effect of ship-platform interaction was considered by modeling both the ship and the tubular braces/legs with nonlinear shell finite elements. An existing analytical denting model was extended to account for distributed loads and was verified against simulation results. The requirements for tubular members to keep compactness were reviewed and discussed. A new concept, ‘transition indentation ratio’ from local denting to global bending, was proposed for tube deformation. The intention of the concept is to help judge the governing deformation patterns with given tube dimensions and material properties, and to unify existing compactness requirements, providing theoretical support to the Rc (characteristic denting resistance) criterion in the new version DNV-GL RP C204. Design considerations of braces/legs subjected to ship impacts were discussed with emphasis on the ship-platform interactions.

A simplified formulation was proposed for the resistance of large inelastic deformation of stiffened plates subjected to lateral loading. The method is based on rigid plastic material assumptions and the use of yield functions formulated in terms of stress resultants. The method considers the flexibility of the panel ends with respect to inward motion, while the rotational boundary conditions are free or clamped. Concentrated and distributed loads are considered, as well as patch loading. The resistance-deformation curves predicted by the proposed method were compared with results using LS-DYNA. The formulation may be used for quick estimates of the resistance of stiffened panels subjected to abnormal or accidental static and transient transverse loads such as explosions, slamming, hydrostatic pressure, and ice actions.