Optimization Design of Hull Compartment Structure Based on 3D Modeling

In this study, a finite element model was constructed and the optimized design variables were obtained through sensitivity analysis to optimize the structure of ship compartments. Sample data were obtained through orthogonal experiments, and back propagation neural network algorithm was used to optimize the structure of ship compartments. However, traditional back propagation neural network algorithms are prone to converge to local minima. A study was conducted to optimize the algorithms through particle swarm optimization. By using the staged mutation strategy, the algorithm performance was further improved. The obtained algorithm was applied to cabin structure optimization. The results showed that compared to back propagation neural network algorithms based on particle swarm optimization, the research method had smaller optimal solutions for equivalent stress and shear stress. The minimum value of the optimal solution for equivalent stress in the research method was 0.167, which was 1.7041 less than the comparison algorithm. The maximum value of the optimal solution for shear stress in the research method was 0.0640, which was 0.9761 less than the comparison algorithm. After optimizing the design, the stresses of the cabin components were increased, but all stresses were within the allowable values. The equivalent stresses before and after optimizing the inner bottom plate were 140N/mm2 and 160N/mm2, respectively. The research method effectively improves the stress of cabin components and achieves structural optimization.