摘要
精确评估炸药冲击Hugoniot参数的可信区间,有效量化冲击起爆、状态方程等唯象模型的不确定度,提高模型的鲁棒性和可靠性,会极大地降低爆压标定的成本。利用线性回归拟合方法得到冲击Hugoniot参数,采用参数估计法推导出Hugoniot参数的置信区间,使用国内试验数据确认方法的有效性。在合理假设的基础上,结合Chapman-Jouguet理论和冲击Hugoniot关系,导出爆压与装药厚度、冲击波走时、初始密度、Hugoniot斜率和0-压声速之间的函数关系式。分别用对数正态分布和Beta分布表征输入不确定度,通过Rosenblatt变换将输入不确定度转化成相互独立的标准正态分布,使用自适应基函数多项式混沌不确定度传播量化方法,给出了爆压的概率密度函数和置信区间。研究结果表明,爆压与样品厚度、爆轰走时、爆速具有拟单调性,在特殊前提下证实了前人的判断。应用于PBX9502炸药,发现给出的置信区间较宽,与试验专家的先验论断吻合。该研究发展了一套有效量化爆轰不确定度的传播量化方法,为发展可信、强预测能力爆轰模拟软件提供了技术支撑。
The confidential interval of explosive shock Hugoniot parameters is exactly assessed and the uncertainties associated with shock ignition and equation of state are efficiently quantified to improve the robustness and reliability of the model and enormously reduce the calibration cost of detonation pressure.The linear regression method is utilized to calibrate the shock Hugoniot parameters,and the confidential interval of Hugoniot parameters are also deduced from parameter estimation.Moreover,the availability of Hugoniot model is validated through domestic experiment data.On the basis of reasonable assumptions,Chapman-Jouguet theory and shock Hugoniot relationship are combined to the functional relationships among detonation pressure.sample thickness,shock passing time,initial density,Hugoniot slope and 0-presure sound speed are deduced.The input uncertainty is characterized by log-normal distribution and Beta distribution,and it is transformed into independent standard normal random variables by Rosenblatt transformation.Polynomial chaos with basis adaptation is applied to implement uncertainty propagation,and the probability density function and confidential interval of detonation pressure is obtained.The results show that the detonation pressure satisfies the quasi-monotonicity with respect to sample thickness,passing time and detonation speed.The assertion of previous studies is confirmed in specific conditions.The confidence interval is much wider when it is used in PBX9502,which coincides with the prior judgment given by experimental expert.This study develops an efficient uncertainty quantification and propagation method.The result can provide technique support for developing highly predictable and confidential software.
作者
梁霄
王瑞利
LIANG Xiao;WANG Ruili(School of Mathematics and System Science,Shandong University of Science and Technology,Qingdao 266590,Shandong,China;Institute of Applied Physics and Computational Mathematics,Beijing 100094,China)
出处
《兵工学报》
EI
CAS
CSCD
北大核心
2024年第5期1673-1680,共8页
Acta Armamentarii
基金
山东省自然科学基金面上项目(ZR2021MA056)
国家自然科学基金委员会-中国工程物理研究院联合基金资助项目(U2230208)
国家数值风洞工程项目(NNW2017ZT7-A13)。
