首页|基于Kriging模型的船舶主机隔离系统抗冲击优化

基于Kriging模型的船舶主机隔离系统抗冲击优化

扫码查看
原文链接
  • 国家科技期刊平台
  • NETL
  • NSTL
  • 万方数据
为了提高系统的抗冲击性能和优化效率，本文以限位隔离系统的刚度阻尼等 7 个参数为设计变量，以系统冲击响应中的相对位移响应和绝对加速度响最大峰值为响应值建立Kriging代理模型。通过NSGA-II多目标优化算法对构建的代理模型进行多目标优化，优化目标为相对位移响应和绝对加速度响最大峰值两者和的最小值。与初始方案相比，优化后系统相对位移响应降低了 27。139%，绝对加速度响应降低了 37。846%。结果表明:建立的代理模型对隔离系统的冲击响应有较精确的预报作用，优化后船舶主机隔离系统的抗冲击得到明显提高。
Antiimpact optimization of the isolation system of a marine engine based on the Kriging model
The impact response analysis of the isolation system of marine engines based on the traditional finite ele-ment model is complicated and requires multiple calculations.To improve the impact resistance and optimization ef-ficiency of the system,this study takes seven parameters,such as stiffness and damping of the limit isolation sys-tem,as the design variables and the relative displacement response and maximum peak value of the absolute accel-eration response of the impact response of the isolation system as the response values.The Kriging surrogate model was established,and multiobjective optimization was conducted using the Kriging surrogate model constructed by NSGA-II.The optimization objective was the minimum sum of the relative displacement response and the maximum peak value of the absolute acceleration response.Compared with the initial scheme,the relative displacement re-sponse is reduced by 27.139%,and the absolute acceleration response is reduced by 37.846%.The complementa-ry results showed that the established surrogate model can accurately predict the impact response of the isolation system.After optimization,the impact resistance of the isolation system is significantly improved.

marine engineisolation systemmultiobjective optimizationKriging surrogate modelshock responseexperimental designoptimal Latin squareNSGA-II algorithm

王茀凡、赵华讯、王爽、郭君

展开 >

哈尔滨工程大学 船舶工程学院,黑龙江 哈尔滨 150001

中海石油(中国)有限公司海南分公司,海南 海口 570100

船舶主机 隔离系统 多目标优化 Kriging代理模型 冲击响应 试验设计 最优拉丁方 NSGA-Ⅱ算法

国家科技重大专项

J2019-I-0017-0016

2024

10.11990/jheu.202207038

哈尔滨工程大学学报
哈尔滨工程大学

哈尔滨工程大学学报

CSTPCD北大核心
影响因子：0.655
ISSN：1006-7043
年,卷(期)：2024.45(7)
  • 9