BEAVRS基准模型热零功率状态的JMCT分析

JMCT Monte Carlo analysis of BEAVRS benchmark:hot zero power results

使用JMCT(J Monte Carlo Transport Code)对来自MIT的全堆芯pin-by-pin精细建模的国际基准模型BEAVRS的热零功率(HZP)状态进行了模拟计算,并与测试数据进行了对比和分析.比较的物理量包括临界本征值、控制棒价值、反应性温度系数、轴向积分的全堆探测器测量值和不同位置四个组件轴向相对功率密度分布.HZP状态下不同控制棒位置插入和硼浓度的临界本征值计算,JMCT结果与理论值1.000的误差小于0.2%,控制棒价值计算结果与测量值符合.JMCT对轴向积分的探测器径向相对功率分布和四个组件的轴向归一化的探测器的计算结果与测量值进行了比较和分析,计算结果与测量值一致,同时清晰地展示了模型增加格架后,轴向功率曲线在相应位置出现下凹的现象.此外,JMCT给出了轴向积分的组件径向相对功率密度分布和轴向相对功率最大处(Z轴位置)的pin径向相对功率密度分布,并与国际知名程序MC21结果进行了对比,两个图像都符合得非常好.随着计算机与并行计算的高速发展,蒙特卡罗程序开始从传统的反应堆校验工具向反应堆设计工具转变.

J Monte Carlo transport code（JMCT）, a new three-dimentional（3D） Monte Carlo transport code, is introduced in this paper. The code is developed on the basis of 3D geometry infrastructure JCOGIN and composed of multilayer modules. JMCT is capable of simulating the collision of particles with multi-group energy or providing energy data libraries. Two forms of parallelism supported in JMCT are domain decomposition and domain replication. The code has very good expansibility. JMCT Monte Carlo results have been compared with hot zero power（HZP） measurements of BEAVRS benchmark model from the MIT Computational Reactor Physics Group. Included in the comparisons are the eigenvalues, control rod bank worths, isothermal temperature coefficients, axially integrated full core detector measurements, axial detector profiles, etc. The eigenvalues for the HZP condition with different control rods positions and boron concentrations are calculated and the error is less than 0.2% compared with the theoretical error 1.000. The results of JMCT for isothermal temperature coefficients are also listed together with MC21 results and measured data.Each calculation for the eigenvalue is run by 1000 cycles in total, discarding 600 cycles, tracking 4 million neutrons each cycle. It takes 5.3–5.7 hours to run on 200 CPU cores. The JMCT results of axially integrated radial detector relative power distribution（RPD） and axial normalized detector signal are compared with the measured data. Power depressions from grid spacers are clearly seen in the JMCT results and accord with the measured data. The JMCT results of axially integrated assembly RPD power distribution are in good agreement with MC21 results, the maximum difference being3.173% for 193 assemblies. So is the result of pin power RPD relative power at the axial elevation of peak power; the minimum relative power RPD 0.278 of JMCT is comparable to 0.283 of MC21, and the max relative power RPD 2.422 of JMCT is comparable to 2.452 of MC21. The calculation for RPD is run with 3000 inactive cycles and 5000 active cycles, tracking 4 million particles each cycle. It takes about 4.8 days to run on 200 CPU cores. Shannon entropy is used to demonstrate that the fission source distribution is converged after 3000 inactive cycles. With the development of computers and parallel computing, the Monte Carlo method can be used in reactor design instead of benchmarking other calculated results.