【摘要】 发展了基于代理模型的不确定度量化方法,以少量确定性模拟结果为样本,通过代理模型来构建目标变量和不确定性输入参数的响应关系,以此来分析目标变量的不确定性信息,包括均值、标准差、概率密度分布等统计信息以及各个输入参数对目标变量的敏感性大小。本文以烧蚀热响应计算为例,通过分析材料物性参数的不确定性对烧蚀热响应预测的影响,表明该方法具有准确度高、效率高的特点,为工程中的高维不确定度量化问题提供了很好的解决思路。更多还原

【Abstract】 In this paper, the uncertainty quantification method based on the surrogate model is developed. The surrogate model is used to construct the response relationship between the target variable and the uncertainty input parameters with samples of some deterministic numerical simulation results. The uncertainty information of the target variable, including the statistical information such as mean, standard deviation, probability density distributions, and sensitivity of each input parameter, is analyzed. Taking the ablative thermal response calculation as an example, the influence of the uncertainty of the physical property parameters on the ablative response prediction shows that the method has high accuracy and efficiency. This paper provides a good approach to solve the multidimensional uncertainty quantification in engineering.更多还原