随着水下无人装备的发展，对精细化仿真的需求日益增加，尤其固（刚体）-液边界处的模拟效果，对于提升整体模拟的真实性具有重要意义。传统物理模拟方法虽稳定，尤其在流固边界上表现优异（如反弹边界法），但计算成本大，模拟耗时长。相比之下，基于数据驱动方法的计算效率高、模拟耗时短，但因缺乏对流固边界的特殊处理，常出现粒子穿透等失真现象。结合二者方法优势，本文提出在基于数据驱动的基础框架上，融入专门提取流-固细节特征的网络，使框架具备处理流固边界交互的能力，提升局部模拟精度。仿真实验结果表明，本文方法比传统基于数据驱动的方法模拟效果更佳，误差降低10%左右。
With the development of underwater unmanned vehicles (UUVs), the demand for high-fidelity fluid simulation has increased significantly, particularly in the simulations effects at the solid (rigid body)-liquid interface, which is important for enhancing the authenticity of the overall simulation. Traditonal physics-based simulation methods, while stable and particularly effective in handing fluid-solid boundaries (e.g., reboudboundary method), are associated with high computational costs and long simulation times. In contrast, data-driven methods offer efficient computation and shorter simulation durations but often suffer from distortions such as particle penetration due to the lack of specialized treatment for fluid-solid boundaries. Combining the advantages of both approaches, this paper proposes incorporating a network dedicated to extracting fluid-solid detail features within a data-driven framework, enabling the framework to handle fluid-solid boundary interactions and significantly enhancing local simulation accuracy. The simulation results show that this method outperforms traditional data-driven methods, with a reduction in error by approximately 10%.
2025,47(24): 185-190 收稿日期：2025-3-28
DOI：10.3404/j.issn.1672-7649.2025.24.030
分类号：U662.9
基金项目：河南省重点研发专项资助项目（231111212100）；河南省重点研发与推广专项资助项目（232102210053）；河南省重点研发与推广专项（24102210026）；河南省重点研发专项（231111221400）
作者简介：刘子豪(1996-),男,硕士,助理工程师,研究方向为水下机器人智能控制
参考文献：
[1] GOSWAMI P, SCHLEGEL P, SOLENTHALER B, et al. Interactive SPH simulation and rendering on the GPU[J]. Eurographics/ ACM SIGGRAPH Symposium on Computer Animation, 2010, 2(1): 1–10.
[2] FOSTER N, METAXAS D. Realistic animation of liquids[J]. Graphical Models and Image Processing, 1996, 58(5): 471-483.
[3] GINGOLD R A, MONAGHAN J J. Smoothed particle hydrodynamics: theory and application to non-spherical stars[J]. Monthly notices of the Royal Astronomical Society, 1977, 181(3): 375-389.
[4] MONAGHAN J J. Smoothed particle hydrodynamics[J]. Reports on Progress in Physics, 2005, 68(8): 17-20.
[5] FOSTER N, METAXAS D. Controlling fluid animation[C]//Proceedings Computer Graphics International. IEEE, 1997.
[6] STAM J. Stable fluids[C]//Proceedings of the 26th annual conference on Computer Graphics and Interactive Techniques, 1999.
[7] TREUILLE A, LEWIS A, POPOVIĆ Z. Model reduction for real-time fluids[J]. ACM Transactions on Graphics (TOG), 2006, 25(3): 826-834.
[8] UMMENHOFER B, PRANTL L, THUEREY N, et al. Lagrangian fluid simulation with continuous convolutions[C]//International Conference on Learning Representations, 2020.
[9] QI C R, SU H, MO K, et al. Pointnet: Deep learning on point sets for 3d classification and segmentation[C]//Proceedings of the IEEE conference on computer vision and pattern recognition, 2017.
[10] QI C R, YI L, SU H, et al. Pointnet++: deep hierarchical feature learning on point sets in a metric space[J]. Advances in Neural Information Processing Systems, 2017.
