Shock wave propagation and their damage to solid structures, which involve complex multiphase and multiphysics phenomena, are difficult problems to address. In this paper, a three-dimensional graphics processing unit (GPU)-accelerated finite difference-smoothed particle hydrodynamics (FD-SPH) method was developed for the prediction of strongly compressible fluid flow and strong fluid-structure interactions. The conventional mesh-based finite difference method was used to simulate shock wave propagation, while the smoothed particle hydrodynamics method was employed to predict the dynamic behavior of solid materials. In addition, the immersed boundary method (IBM) was implemented in the GPU-accelerated FD-SPH solver for the boundary treatment of the interface between solid and fluid materials. Four numerical cases, namely shock-cylinder obstacle interaction, elastic panel deformation induced by shock wave propagation, the response of stainless steel tubes under internal blast loading, and the damage of blast loading on reinforced concrete slab, were conducted for verification of the GPU-accelerated FD-SPH solver. The numerical results were compared against the available experimental data, which shows that the GPU-accelerated FD-SPH solver is capable of capturing the shock wave propagation and its damage to structures very well.

中文翻译：

---

用于冲击-结构相互作用和动态裂缝扩展建模的耦合 FD-SPH 方法


冲击波传播及其对固体结构的破坏涉及复杂的多相和多物理现象，是难以解决的问题。本文开发了一种三维图形处理单元（GPU）加速的有限差分平滑粒子流体动力学（FD-SPH）方法，用于预测强可压缩流体流动和强流固相互作用。传统的基于网格的有限差分方法用于模拟冲击波传播，而平滑粒子流体动力学方法用于预测固体材料的动态行为。此外，在GPU加速的FD-SPH求解器中实现了浸入边界法（IBM），用于固体和流体材料之间界面的边界处理。为了验证 GPU 的性能，我们进行了冲击筒与障碍物相互作用、冲击波传播引起的弹性面板变形、内部爆炸荷载作用下不锈钢管的响应以及爆炸荷载对钢筋混凝土板的破坏这四个数值算例进行验证。加速 FD-SPH 求解器。将数值结果与现有实验数据进行比较，表明 GPU 加速的 FD-SPH 求解器能够很好地捕捉冲击波传播及其对结构的损伤。