The gravel granular filter has been widely used to eliminate the strong shock waves induced by deliberated and occasional explosions. The modelling of shock wave propagation and attenuation in the gravel granular filter is still not well resolved and is a topic of current research. In this study, a new three-dimensional mesoscopic model of gravel granular filter is presented. This involves establishing the geometric model of three-dimensional gravel particles with the consideration of random shape based on the Voronoi algorithm, then generating the corresponding FE mesh based on the mapping grid algorithm and finally precisely controlling the porosity by applying a given displacement load. The three-dimensional mesoscopic model is then validated by numerically simulating two field tests on shock wave propagation and attenuation in a gravel granular filter. Detailed analysis of numerical simulations demonstrates that the drag force between gaseous phase and solid filter and the multiple complex reflection and diffraction between the particles are the main causes of the attenuation of shock wave with low amplitude. Finally, parametric study on the five dominant characteristics of filter, i.e., the particle shape, filter porosity, filter length, particle size and particle wave impedance are conducted. It is found that the particle shape, porosity, and wave impedance of filter have a strong influence on the capability of shock wave attenuation, and the filter length has limited influence beyond a critical optimal length, while the particle size has negligible effect.

砾石颗粒过滤器已被广泛用于消除故意和偶然爆炸引起的强烈冲击波。砾石颗粒过滤器中冲击波传播和衰减的建模仍未得到很好的解决，是当前研究的课题。在这项研究中，提出了一种新的砾石颗粒过滤器的三维细观模型。这涉及基于Voronoi算法建立考虑随机形状的三维砾石颗粒几何模型，然后基于映射网格算法生成相应的有限元网格，最后通过施加给定的位移载荷精确控制孔隙度。然后通过数值模拟两个关于砾石颗粒过滤器中冲击波传播和衰减的现场测试来验证三维细观模型。数值模拟的详细分析表明，气相和固体过滤器之间的拖曳力以及颗粒之间的多次复杂反射和衍射是导致低振幅冲击波衰减的主要原因。最后，对过滤器的五个主要特性，即颗粒形状、过滤器孔隙率、过滤器长度、颗粒尺寸和颗粒波阻抗进行了参数研究。发现过滤器的颗粒形状、孔隙率和波阻抗对冲击波衰减能力有很大影响，超过临界最佳长度的过滤器长度影响有限，