基于遗传优化小波网络的随机载况下裂纹扩展预报  

Crack growth prediction under random loading conditions based on genetic algorithm wavelet neural network

作　　者：张明宇 孙力 黄小平[1] ZHANG Ming-yu;SUN Li;HUANG Xiao-ping(State Key Laboratory of Ocean Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;Marine Design and Research Institute of China,Shanghai 200010,China)

机构地区：[1]上海交通大学海洋工程国家重点实验室,上海200240 [2]中国船舶及海洋工程研究院,上海200010

出　　处：《船舶力学》2024年第9期1430-1440,共11页Journal of Ship Mechanics

摘　　要：疲劳问题作为工程领域常见的破坏形式受到了广泛关注。基于断裂力学的疲劳分析方法可以获取可无损检测的疲劳损伤-裂纹尺寸,但计算较为复杂。针对海洋工程结构物疲劳分析中的谱分析法,本文通过遗传优化的小波神经网络建立一种同一热点下各随机载况的应力强度因子谱,结合有限元分析获取的应力强度因子进行网络训练。结果表明,该模型可对各随机载况下的SIF谱进行较好的预测。本文所提出的方法可大幅减少重复性有限元计算,为裂纹扩展方法应用于随机载况下工程结构的疲劳寿命预报提供一种思路。最后,结合裂纹扩展单一曲线模型实现随机载况下裂纹扩展量的快速预报。Fatigue problem as a common failure form in the engineering field has been widely concerned.The fatigue damage-crack size can be obtained by the fatigue analysis method based on fracture mechanics,but the calculation is relatively complicated.In this paper,aiming at the spectrum analysis based fatigue analysis of ocean engineering structures,the stress intensity factor(SIF)spectrum under random loading conditions of the same hot spot through genetic algorithm wavelet neural network(GAWNN)was established,and the network training with the SIF obtained from finite element analysis was conducted.The results show that the model can predict the SIF spectra under random loading conditions well.The method proposed in this paper can considerably reduce the repetitive finite element calculation and provide a reference for the fatigue life prediction of engineering structures under random load conditions by applying crack propagation method.Finally,combined with the unique crack growth rate curve model,the rapid prediction of crack growth under random loading conditions was realized.

关 键 词：疲劳裂纹扩展 随机载况 应力强度因子谱 小波神经网络 有限元分析 

分 类 号：O346.1[理学—固体力学] P751[理学—力学]