一种基于正交试验设计的土中爆炸数值模型确认方法

A METHOD FOR MODEL VALIDATION OF UNDERGROUND EXPLOSION BASED ON THE ORTHOGONAL EXPERIMENTAL DESIGN

提出了一种基于正交试验设计的土中爆炸数值模型确认方法。结合相应物理实验数据,该方法将数值模型确认问题转化为确定影响数值计算结果的各因素及其水平最佳搭配的优化问题。采用正交试验设计作为优化手段,达到了用较少的试验次数获得各因素及其水平最佳搭配的目的。通过极差分析,研究了在不同确认准则条件下,各因素及其水平对土中爆炸数值模拟结果的影响。计算结果表明:分别以最小计算耗时和最高计算精度为确认准则时,网格尺寸是影响数值结果的主要因素;以两者综合评估为确认准则时,计算方法是影响数值结果的主要因素。该文提出的这种方法为复杂数值模型确认问题的研究提供了新的思路。

A method based on orthogonal experimental design is presented to deal with model validation of underground explosion. In this method, the model validation problem is formulated as an optimization problem in which the best match of the factors and their levels can be determined. In order to save computational time, the orthogonal experimental design method is used as an optimization solver. The effect of the factors and their levels under different validation criteria is also investigated by means of range analysis. The results show that mesh size has a major influence on simulation results when the minimum computational time or maximum computational accuracy is adopted as a validation criterion. And computational method has a major influence on simulation results when both the minimum computational time and maximum computational accuracy are adopted as validation criteria. It is demonstrated that the proposed method offers a new viewpoint and strategy for solving the model validation problems.