基于多源协方差数据的快中子能谱调整方法研究

Multi-Source Covariance Data Adjustment Method for Fast Neutron Spectrum in Reactor Pressure Vessels

摘要: 快中子注量是核反应堆压力容器(RPV)结构完整性评估的核心参数，现有计算模型的近似处理与测量系统的固有误差影响结果的可信度，能谱调整技术能够有效地提高注量计算精度并降低不确定度。开展先验知识与专家经验多源融合的协方差矩阵构造方法研究，初步表征计算–测量的不确定性传播特性。建立目标函数方差最小化的约束型最小二乘算法，形成能谱最佳估算模型，实现能谱精度与注量率不确定度的协同优化。以H.B.Robinson-2屏蔽基准实验为测试算例，数值结果表明：探测器比活度计算精度提高5%，计算值与实验值比值平均在0.98左右，快中子注量率不确定度相对降幅达15%以上。能谱调整技术能够提高快中子注量的计算精度，同时降低其不确定性，对RPV高精度寿命预测具有重要应用价值。

Abstract: The Fast neutron fluence is a core parameter for assessing the structural integrity of nuclear reactor pressure vessels (RPVs). Carry out research on the covariance matrix construction method that integrates prior knowledge and expert experience from multiple sources, and preliminarily characterize the uncertainty propagation characteristics of the calculation-measurement process. Spectrum adjustment techniques offer an effective means to enhance fluence calculation accuracy and reduce uncertainty. This study investigates covariance matrix construction methods that integrate multi-source prior knowledge and expert experience, systematically characterizing the propagation of uncertainty throughout the entire calculation-measurement process. A constrained least-squares algorithm minimizing the variance of the objective function was established, forming a best-estimate neutron spectrum model to achieve the co-optimization of spectrum accuracy and fluence rate uncertainty. Using the H.B. Robinson-2 shielding benchmark experiment as a test case, numerical results demonstrate: a 5% improvement in calculated reaction rate accuracy, with calculated-to-experimental (C/E) ratios averaging approximately 0.98, and a relative reduction exceeding 15% in fast neutron fluence rate uncertainty. Spectrum adjustment techniques thus improve the calculation accuracy of the fast neutron fluence while simultaneously reducing its uncertainty, demonstrating significant application value for high-accuracy RPV lifetime prediction.