基于确定性采样的不确定性分析

Uncertainty Analysis Based on Deterministic Sampling

采用确定性采样方法研究输入参数不确定性传播问题。使用计算流体力学(CFD)与基于模块化特征线法的三维中子物理计算程序(MOC)的耦合计算程序对5×5光棒区冷却剂通道进行计算分析,关注燃料包壳外表面最高温度的不确定性量化结果,并分析了边界条件(冷却剂入口流速、冷却剂入口温度、系统压力、燃料区域功率、功率峰值因子)的不确定度对模拟结果的影响。通过理论分析和算例研究对比了确定性采样方法和广泛使用的拉丁超立方抽样方法的不确定性量化和敏感性分析结果。结果表明:在相同精度要求下,确定性采样能够大幅提高CFD/MOC耦合计算不确定性分析的效率。

An uncertainty analysis for the propagation of the input uncertainties based on deterministic sampling was presented.Uncertainty analysis was performed on a 5×5 rod bundle model using the coupling code of computational fluid dynamics(CFD)and neutronics calculation based on method of characteristic(MOC),focusing on the qualitative results of the uncertainty of the maximum outer surface temperature of the fuel clad.The effect of the boundary conditions(coolant inlet velocity,coolant inlet temperature,system pressure,fuel power)on the simulation results is analyzed.The uncertainty quantification and the sensitivity analysis from the deterministic sampling and Latin hypercube sampling are compared.Results show that the deterministic sampling is an efficient tool for the uncertainty analysis of coupled CFD/MOC code.