基于蒙卡方法的高温气冷堆随机堆积球床模拟研究

Simulation Study of Randomly Packed Pebble Bed in High-Temperature Gas-Cooled Reactor Based on Monte Carlo Method

摘要: 对于采用球形元件的高温气冷堆，燃料球与石墨球经堆芯顶部加料管装入后，在堆芯中随机堆积成球床。为实现对球床堆物理特性的准确掌握，开展精确的物理仿真十分必要。本文以球床高温气冷堆为研究对象，首先采用离散元方法(DEM)和程序获取球床堆芯随机堆积下所有燃料球和石墨球的空间坐标，随后采用蒙卡程序构建出球床堆芯真实的随机堆积模型以进行有效增殖因子k_eff计算，此外，本文还基于“蒙卡重复结构规则排布 + 燃料球和石墨球随机分配”的方法构建了球床堆芯近似的规则排布模型。对于10兆瓦高温气冷实验堆(HTR-10)初次临界问题的计算结果表明：球床随机堆积模型和规则排布模型的k_eff计算结果都与实验结果较为接近(偏差小于0.4%)，说明本文建立的两种球床模型都具有较高的计算精度；当燃料球占比增加和球床高度增加时，两种球床模型的k_eff计算结果更为接近；相较于球床随机堆积模型，规则排布模型的蒙卡计算时间更短，并且易于满足蒙卡程序对于球形元件模拟规模要求，更推荐其作为工程计算模型。

Abstract: In high-temperature gas-cooled reactors using spherical fuel elements, fuel pebbles and graphite pebbles are loaded from the top of the core through the charging tube and randomly packed into a pebble bed. To accurately understand the physical characteristics of the pebble bed, conducting precise physical simulations is essential. This paper takes a pebble-bed high-temperature gas-cooled reactor as the research subject. First, the spatial coordinates of all fuel and graphite pebbles in the randomly packed core are obtained using the Discrete Element Method (DEM) and corresponding programs. Then, a Monte Carlo program is employed to construct an actual random packing model of the pebble bed core for calculating the effective multiplication factor (k_eff). Additionally, this paper develops an approximate regular arrangement model of the pebble bed core based on the method of “Monte Carlo repeated structures with regular arrangement + random assignment of fuel and graphite pebbles”. The calculation results for the initial criticality problem of the 10 MW high-temperature gas-cooled reactor (HTR-10) indicate that the k_eff results from both the random packing model and the regular arrangement model are relatively close to the experimental results (with deviations less than 0.4%), demonstrating that both pebble bed models established in this study have high computational accuracy. When the proportion of fuel pebbles increases or the height of the pebble bed rises, the k_eff results from the two models become even closer. Compared to the random packing model, the regular arrangement model requires shorter Monte Carlo computation time and is easier to meet the simulation scale requirements of Monte Carlo programs for spherical elements, making it more recommended as an engineering computational model.