基于正交试验方法的重叠网格技术不确定度研究  

Research on uncertainty of overset mesh technology based on orthogonal experiment method

作　　者：洪智超 宗智[2] 王轶赓[2] 赵玫佳 刘蕾 梁冠军 HONG Zhi-chao;ZONG Zhi;WANG Yi-geng;ZHAO Mei-jiang;LIU Lei;LIANG Guan-jun(Jiangsu University of Science and Technology,Zhenjiang 212003,China;Dalian University of Technology,Dalian 116026,China;Chinese Society of Naval Architects and Marine Engineers,Beijing 100861,China)

机构地区：[1]江苏科技大学,江苏镇江212003 [2]大连理工大学,辽宁大连116026 [3]中国造船工程学会,北京100861

出　　处：《船舶力学》2020年第7期865-873,共9页Journal of Ship Mechanics

基　　金：国家重点基础研究发展规划(973计划)(2013CB036101);国家自然科学基金创新群体项目(51221961);国家自然科学基金面上项目(51279030;51379033)。

摘　　要：为研究重叠网格方法在背景网格和局部网格区域插值所引入的额外数值误差,本文结合正交试验方法和ITTC不确定度分析方法,对该数值误差、网格尺度、时间步长和湍流模型进行了多因子不确定度分析。数值计算工况采用L9(3^4)正交表进行设计,考虑网格形式、网格尺度、时间步长和湍流模型四个因子,每个因子采用3个水平。不确定度分析结果表明,因子时间步长和网格尺度得到了验证和确认。正交试验方法中的方差分析表明,因子湍流模型和网格形式对数值模拟误差的影响相比于时间步长和网格尺度并不显著,即量级相当。In order to study the numerical error induced by interpolation between background grids and local grids,a multi-factor uncertainty analysis of numerical error,scale of grids,time step and turbulence model was made in this paper by combining the orthogonal experiment method and the uncertainty analysis method suggested by ITTC.An L9(3^4)orthogonal table considers four factors of grid form,scale of grids,time step and turbulence model,each of which consists of three levels and is designed for numerical simulations.The results of uncertainty analysis show that the verification and validation process for time step and grid scale are both conducted completely while the results of orthogonal analysis show that the influence caused by the turbulence model and grid types is almost at the same level with that caused by time steps and grid scales.

关 键 词：重叠网格 不确定度 正交试验 

分 类 号：U661.1[交通运输工程—船舶及航道工程]