摘要
为提高抽壳滑板的疲劳寿命,进而满足射速在1500发/min左右的某10连发高射速自动机寿命达1000发的最低寿命要求,提出一种基于Kriging回归的代理模型用于抽壳滑板的疲劳优化。在与试验结果相符的有限元模型基础上,通过拉丁超立方采样设置初始样本点,构建与样本点相对应的结构模型,并计算每一组样本的理论寿命。根据初始样本点构建Kriging代理模型的同时,采用以改善期望准则为加点准则、遗传算法为子优化求解算法的代理优化算法来对目标函数进行寻优。优化后抽壳滑板的疲劳寿命提高到1193发,且经试验验证,优化结果满足自动机战术技术指标。研究结果表明,通过基于Kriging和遗传算法的代理优化算法能够快速有效地寻优得到全局最优解,适用于高射速自动机内破断零部件的疲劳优化,对工程应用具有一定的借鉴意义。
In order to improve the fatigue life of shell extractor skateboard and meet the minimum life requirement of a certain 10-shot-high-firing-speed automatic gun with a firing rate of around 1500 shots per minute up to 1000 rounds,a surrogate model based on Kriging regression is proposed for the fatigue optimization of shell extractor skateboard.On the basis of the finite element model consistent with the experimental results,the initial sample points are set by Latin hypercube sampling,a structural model corresponding to the sample points is constructed,and the theoretical life of each group of samples is calculated.A Kriging surrogate model is constructed according to the initial sample points,and the surrogate optimization algorithm,which takes the expected improvement(EI)criterion as the addition point criterion and the genetic algorithm as the sub-optimization algorithm,is used to optimize the objective function.The fatigue life of shell extractor skateboard is increased to 1193 rounds after optimization,and the optimized result meets the tactical technical index of the automatic gun after experimental verification.The research results show that the surrogate optimization algorithm based on Kriging and genetic algorithm can quickly and effectively find the global optimal solution,which is applicable to the fatigue optimization of broken parts in the high-firing-speed automatic gun,and has certain reference significance for engineering applications.
作者
田恒旭
林圣业
李浩
巫英豪
王茂森
戴劲松
TIAN Hengxu;LIN Shengye;LI Hao;WU Yinghao;WANG Maosen;DAI Jinsong(School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,Jiangsu,China;State-owned No.152 Factory,Chongqing 400071,China)
出处
《兵工学报》
EI
CAS
CSCD
北大核心
2024年第10期3585-3595,共11页
Acta Armamentarii
关键词
有限元模拟
疲劳分析
Kriging回归
改善期望准则
遗传算法
finite element simulation
fatigue analysis
Kriging regression
expected improvement criterion
genetic algorithm
