Optimization design of icebreaker ice-belt side stringers considering nonlinear strength constraints

Objectives In order to meet the requirements for lightweight design of ice-strengthened structures in the ice belt region of icebreakers, this study investigates an optimization design method for side stringers considering nonlinear strength constraints. Methods A finite element model of the midship section of an icebreaker was established . The ultimate load criterion was adopted as the nonlinear strength assessment criterion, and a mathematical model for the side stringer optimization was formulated with the minimization of structural weight as the objective. An integrated Isight-Abaqus optimization framework was developed, in which three heuristic algorithms, including adaptive simulated annealing (ASA), multi-island genetic algorithm (MIGA), and particle swarm optimization (PSO), were employed for structural optimization. Meanwhile, Latin hypercube sampling was used to generate the sample database, and a Gaussian process regression (GPR) surrogate model was constructed. The trained GPR model was further combined with exhaustive search (ES) to perform optimization over the entire design space, and the optimization performance of different design methods was comparatively evaluated. Results The results indicate that the GPR-ES-based optimization method exhibits higher computational efficiency and better solution stability. The optimal design obtained by this method achieves a structural weight reduction of 30.26% compared with the initial design. Conclusions The proposed side stringer optimization design method based on nonlinear strength constraints can effectively utilize the plastic load-carrying capacity of the structure, improve design efficiency while maintaining structural safety, and provide a methodological reference for the optimization design of ice belt structures in icebreakers.