Ultimate Strength and Reliability based Design of Ship Hulls with Emphasis on Combined Global and Local Loads

During the last few decades, the emphasis in structural design of ships has moved from the allowable stress design to the ultimate limit state design. The ultimate limit state represents the collapse of the structure due to loss of structural stiffness and strength. Recently IACS CSR projects for tankers and bulk carriers introduced the first real implementation of calibrated ULS rule requirements. As a result, there is a need to compare and harmonize structural requirements for different ship types, not only the tankers and bulk carriers separately. This thesis deals with the ultimate strength and reliability-based design of ship hull girders with particular emphasis on combined global and local loads. The ultimate strength of both stiffened panels and hull girder under combined loadings was investigated numerically using FE analysis. The stiffened panel is assumed to be under pitting corrosion. This is considered to be relevant for conventional ships, e.g. tankers and bulk carriers. The ultimate strength of the hull girder under combined global and double bottom bending was also investigated. The possibility of applying simplified approaches, such as linear elastic approach or simplified progressive collapse approach, for a hull girder under combined loads was discussed in this thesis. The development of a semi-probabilistic design format for ultimate strength design of ship hulls under combined global and local loads was discussed. The thesis consists of four papers presenting the main findings as well as a summary part giving the context of the work. Here it is shown that thickness” model does not represent the true structural b plate under pitting corrosion. The main parameters of influence were found to be the level of DoP (Degree of Pitting), the smallest sectional area and the location of concentration of pits. It is also shown that the ultimate hull girder strength of a bulk carrier under AHL condition is significantly reduced as a result of combined global and double bottom bending. The FE analyses showed that an interaction exists between the initial imperfections and combined global and local loads which depend upon the geometry of the hull girder and loading conditions. The difference in the stress distribution in both longitudinal and transverse directions may cause first failure to occur at a different location. Based on the results obtained by FE analyses, it was found that the linear approach yields very conservative results. It is also shown that a linear elastic analysis of a multi span “strip-beam” model of the double bottom could be applied to include the rotation of the double bottom into a “Smith-type” progressive collapse approach. The most viable simplified method suitable to deal with combined global and local loads in design was found to be an interaction formula based on an “effective weighted average” pressure that depends upon the average outside pressure and average inside pressure. A linear elastic approach with first failure criterion was found to be practical for the time being. Regarding reliability-based design of ship hulls, it was found that the most relevant definition of characteristic still-water load for tankers and bulk carriers is μ+1.5σ (with μ = mean and σ = standard deviation), while μ+2.24σ is more relevant for container vessels. For bulk carriers and other vessels with significant double bottom bending, it was shown that the interaction formula results in a consistent safety level. The effect of high correlations between global and local loads for both still-water and wave loads was found to be relatively important. It was shown that for container vessels, the effect of nonlinearity on the failure probability is less for hogging than the sagging, because wave loads in hogging contrary to sagging wave loads are less sensitive to nonlinearity. The container vessels had approximately the same failure probability as the other main class vessels, given that the flexibility of the ship hull is properly accounted for. The safety level for different ships was found to be in reasonable harmony.