基于三维变分和神经网络算法的压水堆堆芯燃耗分布数据同化方法研究  

Method Research on Data Assimilation for Burnup Distribution of PWR Based on Three-dimensional Variation and Neutral Network Algorithm

作　　者：郭林 万承辉 李云召[1] 吴宏春[1] GUO Lin;WAN Chenghui;LI Yunzhao;WU Hongchun(School of Nuclear Science and Technology,Xi’an Jiaotong University,Xi’an 710049,China)

机构地区：[1]西安交通大学核科学与技术学院,陕西西安710049

出　　处：《原子能科学技术》2022年第11期2408-2414,共7页Atomic Energy Science and Technology

基　　金：国家自然科学基金(12005164)。

摘　　要：由于数值计算建立的物理模型与真实堆芯之间不可避免的差异性,堆芯功率分布计算值与实测值之间将存在一定误差,最终使得堆芯燃耗分布的计算值与真实值之间存在误差,从而影响堆芯的经济性和安全性。为了减小堆芯燃耗分布计算值与真实值之间的差异,本文采用三维变分算法和神经网络算法,利用堆芯功率分布的实测值,建立了堆芯燃耗分布的数据同化模型。其中,神经网络算法用于构造堆芯燃耗分布和功率分布之间的复杂函数关系,三维变分算法用于对燃耗分布进行同化计算,最终得到更高精度的堆芯燃耗分布。将本文提出的燃耗分布同化模型用于我国某商用压水堆,数值结果表明:通过堆芯燃耗分布的数据同化,可以显著地降低堆芯燃耗分布计算值与实测值之间的误差,表征为功率分布相对误差的最大值从9.53%降到5.11%。因此,本研究为获得更加精确的堆芯燃耗分布提供了重要的技术手段,能够进一步保障压水堆的经济性和安全性。In this paper,a data-assimilation method was proposed and applied for the burnup distribution of PWR.The burnup distribution is significant to the safety and economy of the reactor,as it is essential for the fuel-reloading design and optimization.Due to the burnup distribution can’t be measured directly during the reactor operation,the numerical simulation is widely applied in reactor engineering.However,there are definitely exist differences between the physical model by the numerical simulation and the actual core,which would induce the errors to the simulation values of power distributions and hence to the burnup distributions.Therefore,a data-assimilation method has been proposed based on the three-dimensional variation(3DVAR)algorithm and the neutral network algorithm,aiming at reducing the errors of the burnup-distribution simulations with application of the power-distribution measurements.In our proposed method for the burnup distribution,the 3DVAR algorithm was applied to get the more-accurate burnup distribution,the neutral network algorithm was applied to obtain the complex relations between the burnup distributions and the power distributions which were essential for the 3DVAR.As method verification,the proposed method has been applied to a commercial PWR operated in China.It can be observed that through the data assimilation,the burnup-distribution errors can be reduced notably,as the maximum value of relative errors for power distribution can be notably reduced from 9.53%to 5.11%.Moreover,the key parameters including power-peak factors(F_(q) and F_(ΔH))can also be improved.As concluded,the data-assimilation method based on the power-distribution measurements is efficient to improve the simulation accuracy of burnup distributions,and could further guarantee the economy and safety of PWR.

关 键 词：三维变分算法 神经网络算法 燃耗分布数据同化 

分 类 号：TL329[核科学技术—核技术及应用]