基于不确定度量化分析的炸药驱动飞片参数标定与预测方法

A UNCERTAINTY PARAMETER CALIBRATION AND PREDICTION METHOD WITH AN APPLICATION TO FLYER PLATE EXPERIMENTS

炸药驱动飞片实验作为检验炸药爆轰能力的基准实验,在爆轰领域应用广泛,但由于炸药反应迅速、剧烈、不易测量,能够根据现有信息预测炸药驱动飞片实验效果极为重要.然而由于爆轰系统中不确定参数众多,尤其对于精确描述爆轰驱动过程与做功能力都至关重要的爆轰产物状态方程不确定参数,需要量化其不确定性,及不确定性对最终计算结果的影响,才能做出有效预测.本文聚焦JWL状态方程参数,从不确定度量化角度出发,综合利用实验数据与现有数值模拟数据,基于贝叶斯思想量化给出其后验分布.并针对炸药LX-17,评估参数不确定度对整体计算结果的影响,利用参数后验分布预测不同长度飞片与炸药的驱动效果,并评估预测结果不确定度.本文不仅描述了不确定性从状态参数到最终爆轰驱动能力的传播规律,还预测了不同条件下炸药驱动飞片的能力,及其不确定度,为分析不确定度在爆轰系统中的传播特征,提出模拟爆轰驱动的高置信度计算方法,降低系统整体不确定度奠定了基础,也为通过数值实验预测未知爆轰实验效果提出新的方法思路.

Detonation driven flyer plate experiment is one of the fundamental experiments to ex- amine the detonation capability of the explosive. However, the reaction of the explosive is quick, severe and hard to measure. Therefore, it is of great importance that we can pre- dict the effect of the detonation driven flyer experiments based on the known information. Since there are many uncertainty parameters in the detonation system, especially those in equation of state （EOS） and the detonation product equation of state （EOS） is important to correctly describe the detonation driven process and the work capacity of the explosive detonation. Reasonable parameters in the EOS must be determined in the first place. In this paper, we are focusing on the uncertain parameters in the Jones-Wikins-Lee （JWL） EOS, and the Bayesian approach, in which both the parameter uncertainty and model discrepancy are considered, is employed to calibrate the uncertain parameters in the detonation prod- uct. Explosive LX-17 is chosen to test this method, and the posterior distribution of the uncertain parameters are adopted to predict the detonation effect of flyers and detonation with different length. This paper not only describes the propagation law of the uncertainty from the parameters to the final numerical results, but also predict the work capacity of the explosive under different conditions. It provides a new perspective to analyze the uncertainty propagation characteristics in the detonation systems and promotes a new method to forecast the experimental results based on the numerical results.