基于时间演化响应面模型的JWL状态方程参数标定方法  

Parameters calibration method of JWL equation of state based on time varying response surface methodology

作　　者：梁霄 王朔 王瑞利[2] 陈江涛[3] LIANG Xiao;WANG Shuo;WANG Ruili;CHEN Jiangtao(School of Mathematics and System Sciences,Shandong University of Science and Technology,Qingdao 266590,Shandong,China;Institute of Applied Physics and Computational Mathematics,Beijing 100094,China;China Aerodynamics Research and Development Center,Mianyang 621000,Sichuan,China)

机构地区：[1]山东科技大学数学与系统科学学院,山东青岛266590 [2]北京应用物理与计算数学研究所,北京100094 [3]中国空气动力研究与发展中心,四川绵阳621000

出　　处：《山东大学学报（理学版）》2024年第4期117-126,共10页Journal of Shandong University(Natural Science)

基　　金：山东省自然科学基金资助项目(ZR2021BA056);国家数值风洞工程项目(NNW2019ZT7-A13);国家自然科学基金-中国工程物理研究院联合基金项目(U2230208);国家自然科学基金资助项目(12171047)。

摘　　要：综合试验数据和建模与模拟结果,标定爆轰产物JWL(Jones-Wilkins-Lee)状态方程中的唯象参数,对理解爆轰膨胀驱动过程和评价做功能力至关重要。时间演化代理模型能构建重要参数与系统响应量(system response quantity,SRQ)之间较为简洁的函数关系,为描述爆轰动力行为和参数标定提供一种快速高精度算法。本文首先利用拉丁超立方抽样(Latin hypercube sampling,LHS)技术对爆轰产物JWL状态方程中的不确定参数进行抽样,选取有效数据并分成两组:第一组进行高分辨率仿真,使用具有自主知识产权爆轰数值软件CFD2D计算得到径向壁位置和径向壁速度,输出结果用于构建不确定JWL参数和SRQ之间的时间演化响应面模型;第二组用于评估时间演化代理模型的精确度。其次,定义代理模型与试验数据之间的残差,进而搜索最小残差获得最优参数。结果表明:将最优参数代入高分辨率模型后导出的SRQ的输出结果和试验数据吻合较好,从而完成了参数标定和模型确认。结果为武器型号设计和装备性能评估提供依据。Calibration of phenomenological parameters of Jones-Wilkins-Lee(JWL)equation of state based on experimental data and modeling and simulation result plays a vital role in comprehension of detonation-driven expansion process and evaluating the work performance of detonation product.Time varying response surface methodology(RSM)can establish a simpler functional relation between prominent parameters and system response quantity(SRQ).It provides a rapid and high-accurate computational method of describing the dynamical behavior and calibrating parameters.Firstly,Latin hypercube sampling(LHS)technique is used to sample the uncertainty in the uncertain parameters of JWL equation of state in the detonation production,and then effective data is selected and divided into two groups.High-resolution simulation is implemented based on group I,and the radial position and radial velocity is obtained through full intellectual software—CFD2D.Time varying RSM between uncertain JWL parameters and SRQ is constructed through output of the high fidelity simulation.Group II is used to evaluate the accuracy of the time varying RSM.Then Residual error between the simulation results of surrogate model and experiment data is constructed.Furthermore,the minimum residual error is searched among all the residual error and the corresponding parameters are obtained as the optimum parameters.The result shows that output of SRQ coincides with and experimental data perfectly,when these optimum parameters are substituted into the high fidelity computational methods.That means,the parameters are calibrated and model validated.The result will provide guidance for weapon design and munitions assessment.

关 键 词：参数标定 拉丁超立方抽样 时间演化响应面模型法 圆筒试验 JWL状态方程 

分 类 号：O385[理学—流体力学]