The uncertainty of design variables is incorporated to design optimization of ship structure. Since conventional double loop strategy for probabilistic design optimization involves iterative optimization process and probabilistic assessment, a strategy of reliability-based design optimization is proposed by synthesizing kriging model, sequential optimization and reliability assessment (SORA) method and particle swarm optimization (PSO). And the strategy is applied to the problem of reliability-based design optimization of a multipurpose ship. SORA method decouples the process of reliability assessment and optimization, which improves the efficiency of reliability-based optimization. PSO method guarantees the efficiency of calculating global optimal solution in the optimization part of SORA. Kriging model substitutes for time-consuming finite element model calculation in the process of optimization and reliability assessment. And inverse First Order Reliability Method (FORM) is used to calculate reliability index and locate the most probable points (MPP) which is used to shift the boundaries of violated constraints. The optimization results and computing time of the multipurpose ship are verified by the results of finite element (FEM) model calculation and Monte Carlo simulation, which shows that the proposed strategy guarantees the accuracy of solutions and significantly alleviates the computational burden.
The deterministic optimization of ship structure is widely studied in recent years (Kong et al. 2006; Ehlers 2012), while few researchers concentrate on the reliability-based optimization considering the impact of uncertainty (Hannapel and Vlahopoulos 2010). In real life, design variables contain uncertainties due to the uncontrollability of manufacturing process, material properties, environmental impact and other factors. Consequently, the results of deterministic optimization may be too sensitive to the variation of design variables. And if the results are close to the constraint boundaries, it may cause structure failure. For this reason, it is necessary to use reliability-based design optimization considering uncertainties for ship structure design optimization.
