Uncertainty analysis of a structural-acoustic problem using imprecise probabilities based on p-box representations
Uncertainty analysis of a structural-acoustic problem using imprecise probabilities based on p-box representations
复制标题
使用基于 p 盒表示的不精确概率对结构声学问题进行不确定性分析
DOI：
10.1016/j.ymssp.2016.04.009
复制
发表时间：
2016
影响因子：
8.4
通讯作者：
Beer Michael
中科院分区：
文献类型：
--
作者：
Chen Ning;Yu Dejie;Xia Baizhan;Beer Michael
不精确的概率可以捕获知识的不确定性，这反映了有限的可用知识，因此无法建立精确的概率模型。本文利用p-box来表示结构声问题的参数，以捕获模型中的认知不确定性。为了对p-box结构声问题进行必要的分析，提出了一阶矩阵分解摄动法（FMDPM）进行区间分析，并推导了一种基于FMDPM的高效区间蒙特卡罗方法。在基于FMDPM的有效区间蒙特卡罗方法的实现中，首先在矩阵分解技术的基础上，通过不确定参数提取，得到常数矩阵。然后用一阶摄动法对这些常数矩阵进行多区间分析。算例说明了该方法的可行性和有效性。
Imprecise probabilities can capture epistemic uncertainty, which reflects limited available knowledge so that a precise probabilistic model cannot be established. In this paper, the parameters of a structural–acoustic problem are represented with the aid of p-boxes to capture epistemic uncertainty in the model. To perform the necessary analysis of the structural–acoustic problem with p-boxes, a first-order matrix decomposition perturbation method (FMDPM) for interval analysis is proposed, and an efficient interval Monte Carlo method based on FMDPM is derived. In the implementation of the efficient interval Monte Carlo method based on FMDPM, constant matrices are obtained, first, through an uncertain parameter extraction on the basis of the matrix decomposition technique. Then, these constant matrices are employed to perform multiple interval analyses by using the first-order perturbation method. A numerical example is provided to illustrate the feasibility and effectiveness of the presented approach.