基于贝叶斯推断的堆芯功率分布重构

Reactor Core Power Mapping Based on Bayesian Inference

基于贝叶斯推断理论,实现了一种有效融合堆内中子探测器实际测量值与中子学理论计算值两类信息的堆芯功率分布重构斱法。应用大亚湾核电站1号机组的测量数据对贝叶斯推断斱法的功率分布重构精度进行了验证,幵将贝叶斯推断斱法与卡尔曼滤波斱法以及耦合系数法进行了精度对比。验证结果显示,贝叶斯推断斱法在整个循环寿期内的均斱根误差、最大相对误差、功率峰重构误差分别不大于0.31%、1.64%和0.07%,且重构精度优于卡尔曼滤波斱法以及耦合系数法。重构精度以及计算速度表明贝叶斯推断斱法有潜力被应用于功率分布在线监测系统。

The reactor core power mapping method based on Bayesian inference has been implemented,and it provides an effective way to combine the information from the measurements of in-core neutron detectors and the numerical neutronics simulation results.Measurements from Unit 1 of Daya Bay Nuclear Power Plant are used to verify the accuracy of the Bayesian inference method,and comparisons are made among the Bayesian inference method,the Kalman filter method and the coupling coefficients method.The root mean square errors,the maximum relative errors,and the power peak reconstruction error of the Bayesian inference method are less than 0.31%,1.64%and 0.07%for the entire operating cycle,respectively,and the Bayesian inference method outperforms the Kalman filter method and the coupling coefficients method in terms of accuracy.The reconstructed assembly power distribution results and the calculation speed show that the Bayesian inference method is a promising candidate for the on-line core power distribution monitoring system.