Abstract This paper discusses one of the most important issues facing the regulatory body while performing the uncertainty analysis of the nuclear reactor parameter during accident conditions. This problem is the long computational time required by the statistical sampling methods to compute the uncertainty. We overcome this problem by introducing the Unscented Transform (UT) algorithm and singular value decomposition (SVD). Where both algorithms are combined (SVD/UT) to generate a set of sigma points, these sigma points are the representatives of whole probability distribution. The uncertainty quantification is performed during Loss of coolant accident in Pressurized Water Reactor (PWR), where the input variable of uncertainty is the coolant density reactivity. The SCALE 6.2 code is used for calculating the reactivity coefficients and the covariance matrix. The response variables are the peak cladding temperatures during the accident which are computed by ATHLET thermal-hydraulic code. The results obviously confirm the efficiency of the SVD/UT sampling in predicting the new mean values, and assure its ability to reduce the sampling size leading to a dramatic reduction of computational cost.

中文翻译：

---

在 LOCA 期间使用无迹变换和奇异值分解对 PWR 进行有效的不确定性量化

摘要 本文讨论了监管机构在事故工况下对核反应堆参数进行不确定性分析时面临的最重要问题之一。这个问题是统计抽样方法计算不确定性所需的长计算时间。我们通过引入无迹变换 (UT) 算法和奇异值分解 (SVD) 来克服这个问题。将两种算法结合起来（SVD/UT）生成一组 sigma 点，这些 sigma 点是整个概率分布的代表。不确定性量化是在压水反应堆 (PWR) 冷却剂损失事故期间进行的，其中不确定性的输入变量是冷却剂密度反应性。规模 6. 2 代码用于计算反应系数和协方差矩阵。响应变量是事故期间的峰值包层温度，由 ATHLET 热工水力代码计算。结果显然证实了 SVD/UT 采样在预测新平均值方面的效率，并确保其能够减少采样大小，从而显着降低计算成本。