基于传统加筋形式的船舶舷侧结构，提出一种新式折叠V型夹层板舷侧结构，针对其复杂的耐撞性优化问题，分别利用GA-BP-GA方法和直观分析法对其开展耐撞性能优化设计，并验证了GA-BP-GA方法的可行性与准确性。结果表明：经遗传算法优化后的BP神经网络具有较优的训练精度和泛化能力；与原设计相比，GA-BP-GA最优设计的耐撞性能提高了21.0%，高于正交最优设计的16.5%和直观优化的6.3%；GA-BP-GA最优设计关于耐撞性指标的预测值与有限元仿真值之间的相对误差均小于3.5%，具有较高的可信度。
Based on the traditional stiffened ship's side structure, a new type of side structure with V-shaped folding sandwich panel was proposed. Aiming at its complex crashworthiness optimization, the crashworthiness optimization method GA-BP-GA and direct analysis method were used, and the feasibility and accuracy of the GA-BP-GA method were verified. The results showed that the BP neural network optimized by genetic algorithm has greater prediction accuracy and generalization ability. Compared with the original design, the crashworthiness of the GA-BP-GA optimal design is improved by 21.0%, which is higher than 16.5% of the orthogonal optimal design and 6.3% of the direct optimization. The relative error of crashworthiness indexes between the predicted value and finite element results for GA-BP-GA optimal design is less than 3.5%, which has high reliability.
2022,44(2): 29-33 收稿日期：2021-02-22
DOI：10.3404/j.issn.1672-7649.2022.02.006
分类号：U661.43
作者简介：高明星(1993-)，男，硕士，助理工程师，研究方向为海洋结构物设计与制造
参考文献：
[1] 张延昌, 葛珅玮, 刘昆, 等. 基于正交设计与BP-GA算法的船体结构耐撞性能优化设计[J]. 江苏科技大学学报(自然科学版), 2013, 27(6): 511–517
[2] 王哲, 赵辰水, 刘昆. V型折叠式夹层板改进设计及抗冲击性能数值分析[J]. 船海工程, 2017, 46(4): 98–101
[3] 张延昌, 王自力, 张世联, 等. 基于折叠式夹层板船体结构耐撞性设计[J]. 船舶工程, 2009, 31(6): 1–5
[4] 高明星, 刘刚, 黄一, 等. 基于改进BP-GA方法的FPSO舷侧结构耐撞性能优化设计[J]. 船舶工程, 41(1): 38−43.
[5] 刘刚, 高明星, 陈志颖, 等. 基于SGA-BP-GA方法的FPSO舷侧结构耐撞性能优化设计[J]. 振动与冲击, 2019, 38(21): 62–70
[6] HOU S, ZHAO S, REN L, et al. Crashworthiness optimization of corrugated sandwich panels[J]. Materials & Design, 2013, 51: 1071–1084
[7] 曹金凤. Python语言在Abaqus中的应用[M]. 北京: 机械工业出版社, 2011.
