To prevent significant structural damage and associated losses during the ice-breaking process when submarines surface in polar regions, finite element analysis is commonly employed in the structural safety design. In finite element simulations, it is computationally impractical to model the infinite sea ice in actual ice-breaking processes. Thus, a predefined domain is employed to approximate the infinite. Whether this modeling choice significantly affects the accuracy of finite element analysis remains an open question. This study conducts finite element simulations of the ice-breaking process involving the sail structure of submarine under various operational conditions to investigate the influence of sea ice model dimension on the resulting ice-breaking loads. Cosine similarity and relative error metrics are jointly applied to evaluate the similarity of ice load time-history curves in terms of both magnitude and shape across different sea ice modeling dimensions. The results indicate that when the modeled sea ice dimension exceeds three times the structural dimension, the similarity of the ice load curves is relatively high, suggesting that it can serve as a reasonable approximation for infinite sea ice. These findings provide practical guidance for determining appropriate sea ice modeling dimensions in finite element simulations of submarine-ice collisions.
Zhang et al. (Sun,) studied this question.