Novel methodology of fail-safe reliability-based topology optimization for large-scale marine structures

In this paper, a novel reliability-based topology optimization (RBTO) framework integrating fail-safe is first presented to boost reliability levels and load path redundancy for complex marine structures. The sequential optimization and reliability assessment (SORA) approach using the conjugate gradient (CG) algorithm ( SORACG) is proposed to decouple the RBTO procedure into sequential deterministic topology optimization (DTO) loops and reliability assessment (RA) loops. The computational efficiency and solution accuracy are enhanced benefiting from the decoupling feature of SORA. A popular fail-safe model simulating the local material failure using damaged zones with prescribed shape and size is introduced into DTO. Non-differentiable fail-safe worst-case problem is transformed into an equivalent bound formulation via the beta-method. Combing the three-stage continuation technique (3SCT) which considers both iterative efficiency and global optimality, a multi-model optimization strategy is suggested to address the fail-safe model. In RA, the CG algorithm is developed to derive the most probable point (MPP) for the optimal fail-safe DTO design. Numerical cases concerning a cantilever beam and engineering applications for a long-span open deck and 10,000-ton container ship demonstrate the effectiveness of the framework.