A ship collision accident can result in severe environmental damage and loss of life. Therefore the non-linear finite element method with shell elements is used to assess the crashworthiness of ship steel structures through collision simulations. However, a non-linear finite element-based benchmark revealed inconsistencies and inaccuracies in the results of collision analysis using current material relations and failure criteria. To overcome these problems in this thesis, the steel material's true strain and stress relation is derived in a novel way from tensile experiments until failure on the basis of optical measurements. The novel material relation is obtained until failure with respect to the strain reference length. Furthermore, this material relation, including failure, can be varied to accommodate different finite element sizes. By this means good correspondence in numerical results for the simulation of tensile and plate specimens and complex topologies under indentation loading is achieved for different mesh sizes ranging from 0.88 mm to 140 mm. It is shown that the choice of a constant strain failure criterion suffices for thin steel ship structures. Furthermore, a procedure to optimise a conventional ship side structure for crashworthiness in the conceptual design stage is presented. This procedure extends the assessment procedure for structural arrangements from Germanischer Lloyd. The energy absorbed until inner plate rupture during a right-angle ship collision is used as an optimisation objective. This procedure exploits the novel element length-dependent strain and stress relation, including failure. A particle swarm algorithm is used to identify the crashworthy conceptual design. By this means a crashworthy conceptual ship side structure is obtained, which can absorb significantly more energy than the initial rules-based concept with a reasonable weight increase.
展开▼
