Ivan Lucian, Hummel David, Lebel Luke
Canadian Nuclear Laboratories, 286 Plant Road, Chalk River, Ontario, K0J 1J0, Canada.
Canadian Nuclear Laboratories, 286 Plant Road, Chalk River, Ontario, K0J 1J0, Canada.
J Environ Radioact. 2018 Dec;192:551-564. doi: 10.1016/j.jenvrad.2018.07.014. Epub 2018 Aug 22.
Assessing the risks of radioactive dose in a radiological dispersal device (RDD) attack requires knowledge of how the radiological materials will be spread through the air surrounding the site of the detonation. Two essential parts of the accurate prediction of the behaviour of this dispersion are a characterization of the initial cloud size, directly after the blast, and detailed modelling of the behaviour of different size particulates. Capturing the transport of contaminants from the initial blast wave is integral to achieving accurate predictions, especially for regions where the blast dynamics dominates, but performing such calculations over a wide range of particle sizes and spatial scales is computationally challenging. Formulation of efficient computational techniques for such advanced models is required to provide predictive tools useful to first responders and emergency planners. In this work, a Multi-Cloud Radiological EXplosive Source (MCREXS) modelling approach for RDD is investigated. This approach combines a stochastic, particle-based, mechanistic model with a standard atmospheric dispersion model. The former is used to characterize the distribution of radioactive material near the source of the explosion, where the blast wind effects are important, while the latter is used to model the transport of the contaminant in the environment over large areas. The particle transport in the near-field of the explosion site is computed based on a Lagrangian description of the particle phase and a reconstructed-Eulerian field for the carrier phase. The information inferred from this physics-based model is then used as a starting point for a subsequent standard Gaussian puff model to calculate the dispersion of the radioactive contaminant. The predictive capabilities of the MCREXS model are assessed against the 2012 DRDC Suffield full-scale RDD experiments. The results demonstrate improved predictions relative to those performed using only a Gaussian puff calculation from an empirical initial cloud distribution.
评估放射性散布装置（RDD）袭击中的辐射剂量风险，需要了解放射性物质将如何在爆炸地点周围的空气中扩散。准确预测这种扩散行为的两个关键部分是爆炸刚发生后对初始云团大小的表征，以及对不同大小颗粒行为的详细建模。捕捉污染物从初始爆炸波中的传输对于实现准确预测至关重要，特别是在爆炸动力学起主导作用的区域，但在广泛的粒径和空间尺度上进行此类计算在计算上具有挑战性。需要为这种先进模型制定高效的计算技术，以提供对第一响应者和应急规划者有用的预测工具。在这项工作中，研究了一种用于RDD的多云放射性爆炸源（MCREXS）建模方法。这种方法将基于粒子的随机机理模型与标准大气扩散模型相结合。前者用于表征爆炸源附近放射性物质的分布，在那里爆炸风效应很重要，而后者用于模拟污染物在大面积环境中的传输。爆炸现场近场中的粒子传输基于粒子相的拉格朗日描述和载体相的重构欧拉场进行计算。然后，从这个基于物理的模型中推断出的信息被用作后续标准高斯烟羽模型的起点，以计算放射性污染物的扩散。MCREXS模型的预测能力是根据2012年国防研究与发展司令部萨菲尔德全尺寸RDD实验进行评估的。结果表明，相对于仅使用基于经验初始云团分布的高斯烟羽计算所得到的预测，该模型的预测有所改进。
