基于自适应和投影Wiener混沌的圆筒实验不确定度量化

由于炸药爆轰现象的复杂性和人们对它的认知缺陷,其表征爆轰流体力学过程的物理数学模型具有较强的不确定性,要降低基于爆轰建模与模拟的数值结果做出决策的风险,量化和评估不确定输入对爆轰系统输出结果的影响尤为重要。本文中针对具有高维随机变量的爆轰问题的不确定度量化,使用自适应基函数的Wiener混沌方法、耦合旋转变换和投影方法,减少截断空间的长度。针对输入变量相关性,使用Rosenblatt变换使其相互独立。针对不符合标准正态分布的变量使用等概率原则,将它化为标准正态分布。最后,使用自主研发的具有完全知识产权的爆轰数值模拟软件LAD2D,研究了具有高维不确定参数的圆筒实验的不确定度量化,给出期望、标准差、置信区间等统计信息,所得问题与实验数据比对,从而确认了模型的有效性。

Uncertainty quantification of cylindrical test through Wiener chaos with basis adaptation and projection

The mathematical-physical model used to describe the detonation  dynamics has many  uncertain factors due to the complexity and lack of knowledge for detonation phenomenon. Quantifying and assessing the impact of input uncertainties on output of detonation systems has a direct influence on reducing the risk based on the numerical model  and simulation results for detonation. The Wiener chaos based on adapted basis is used to deal with the uncertainty quantification of high-dimensional random variables for detonation simulation. The rotation transformation and projection method is used  to reduce the length of truncation number.  Rosenblatt transformation is used to transform the set of dependent random variables into independent random variables. The equality of probability principle is used to change the non-Gaussian random variables into standard random variables. Uncertainty quantifications of the cylinder test with high dimensional input uncertainties are studied. The statistical informations such as mean, standard deviations, and confidence intervals are presented. The simulation  results coincide with the experimental data, and the accuracy of the model is validated.