机理与数据驱动的物理仿真计算范式及引擎架构

doi:10.11996/JG.j.2095-302X.2024061207

图学学报 ›› 2024, Vol. 45 ›› Issue (6): 1207-1221.DOI: 10.11996/JG.j.2095-302X.2024061207

• “大模型与图学技术及应用”专题 • 上一篇下一篇

机理与数据驱动的物理仿真计算范式及引擎架构

何小伟¹(), 石剑², 刘树森¹, 任丽欣¹, 郭煜中¹, 蔡勇³, 王琥³, 朱飞⁴, 汪国平⁴()

1.中国科学院软件研究所，北京 100190
2.中国科学院自动化研究所，北京 100190
3.湖南大学汽车车身先进设计制造国家重点实验室，湖南长沙 410082
4.北京大学计算机学院，北京 100871

收稿日期:2024-07-31 接受日期:2024-10-05 出版日期:2024-12-31 发布日期:2024-12-24
通讯作者:汪国平(1964-)，男，教授，博士。主要研究方向为计算机图形学、虚拟现实与物理仿真。E-mail：wgp@pku.edu.cn
第一作者:何小伟(1985-)，男，研究员，博士。主要研究方向为计算机图形学与物理仿真。E-mail：xiaowei@iscas.ac.cn
基金资助:
国家重点研发青年科学家项目(2021YFB1715800);国家自然科学基金(62302490)

The computational paradigm and software framework for mechanism and data-driven physical simulation

HE Xiaowei¹(), SHI Jian², LIU Shusen¹, REN Lixin¹, GUO Yuzhong¹, CAI Yong³, WANG Hu³, ZHU Fei⁴, WANG Guoping⁴()

1. Institute of Software, Chinese Academy of Sciences, Beijing 100190, China
2. Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China
3. State Key Laboratory of Advanced Design and Manufacturing for Vehicle Body, Hunan University, Changsha Hunan 410082, China
4. School of Computer Science, Peking University, Beijing 100871, China

Received:2024-07-31 Accepted:2024-10-05 Published:2024-12-31 Online:2024-12-24
Contact: WANG Guoping (1964-), professor, Ph.D. His main research interests cover computer graphics, virtual reality and physical simulation. E-mail：wgp@pku.edu.cn
First author：HE Xiaowei (1985-), researcher, Ph.D. His main research interests cover computer graphics and physical simulation. E-mail：xiaowei@iscas.ac.cn
Supported by:
National Key Research and Development Program of China(2021YFB1715800);National Natural Science Foundation of China(62302490)

摘要/Abstract

摘要：

物理仿真作为现代工业软件的基石，其计算范式可分为机理驱动、数据驱动及混合驱动等模式。面对多样化物理仿真需求，构建一个既能灵活适应各类物理仿真计算范式，又能实现不同计算范式之间高效耦合的通用引擎架构，已成为软件设计与开发领域亟待解决的关键难题与挑战。针对该问题，提出面向多物理仿真计算范式的FNMS架构 Data Field-Node-Module-Scene Graph，其核心在于四层结构的设计：数据域(Data field)、节点(Node)、模块(Module)与场景图(Scene graph)。具体而言，数据域层为仿真过程提供统一的数据管理与访问接口，解决物理仿真计算数据共享的便捷性与高效性；模块层封装各类物理仿真算法，实现算法的模块化与可重用，解决仿真计算、渲染与交互的异步协同问题；节点层通过数据与算法模块的解耦实现算法在不同物理仿真计算范式之间的复用，同时便于实现多物理场耦合过程的交换与共享；而场景图层通过将节点组织成有向无环图，支撑多种物理仿真计算范式的高效耦合计算。通过该四层结构的结合，FNMS架构不仅能提升物理仿真的计算效率与灵活性，更为跨学科、跨领域的物理仿真研究提供了强有力的技术支持。

关键词: FNMS引擎架构, 机理驱动, 数据驱动, 物理仿真, 计算范式

Abstract:

As the cornerstone of modern industrial software, physical simulation encompasses various computational paradigms, including mechanism-driven, data-driven, and hybrid-driven models. Meeting the demands of diverse physical simulation requires the construction of a general framework capable of flexibly adapting to various physical simulation computational paradigms while achieving efficient coupling across various computational paradigms, presenting a critical challenge in software design and development. To address this, the Data field—Node—Module— Scene graph (FNMS) architecture was proposed, targeting multi-physics simulation computational paradigms. Its core lies in the design of a four-layer structure: Data field, Node, Module, and Scene graph. Specifically, the Data field layer provides a unified data management and access interface for the simulation process, enhancing the convenience and efficiency in data sharing for physical simulation computations. The Module layer encapsulated various physical simulation algorithms, realizing algorithm modularization and reusability while solving the asynchronous coordination of simulation computation, rendering, and interaction. Through data and algorithm decoupling, the Node layer enabled algorithm reuse across different physical simulation computational paradigms, and it facilitated the exchange and sharing within multi-physics coupling processes. The Scene graph layer supported efficient coupled computations of various physical simulation computational paradigms by organizing nodes into a directed acyclic graph. Through the combination of these four layers, the FNMS architecture not only enhanced the computational efficiency and flexibility in physical simulations but also provided strong technical support for interdisciplinary and cross-domain physical simulation research.

Key words: FNMS framework, mechanism driven, data driven, physical simulation, computing paradigm

中图分类号:

TP391

何小伟, 石剑, 刘树森, 任丽欣, 郭煜中, 蔡勇, 王琥, 朱飞, 汪国平. 机理与数据驱动的物理仿真计算范式及引擎架构[J]. 图学学报, 2024, 45(6): 1207-1221.

HE Xiaowei, SHI Jian, LIU Shusen, REN Lixin, GUO Yuzhong, CAI Yong, WANG Hu, ZHU Fei, WANG Guoping. The computational paradigm and software framework for mechanism and data-driven physical simulation[J]. Journal of Graphics, 2024, 45(6): 1207-1221.

图/表 18

图1 物理仿真计算范式流程图((a)理论建模；(b)数字建模；(c)数值求解；(d)可视分析)

Fig. 1 The computational paradigm of physical simulation mainly ((a) Theoretical modeling; (b) Digital modeling; (c) Numerical solving; (d) Visual analysis)

图2 边值问题示意图

Fig. 2 Diagram of a boundary value problem

图3 经典物理仿真计算范式分类

Fig. 3 Classification of computing paradigms of physical simulation

图4 数据驱动的物理仿真计算范式流程图

Fig. 4 Flowchart of the data-driven physical simulation

图5 FNMS架构示意图

Fig. 5 Diagram of the FNMS architecture

图6 数据域示意图

Fig. 6 Diagram of the data domain

图7 流固耦合场景((a)效果图；(b)节点图)

Fig. 7 Solid-fluid coupling scenario ((a) Visual effect;(b) Node graph)

图8 网格拓扑层次化结构表示

Fig. 8 Hierarchical representation of mesh topology

图9 吉普车多模态表示

Fig. 9 Multimodal representation of a jeep ((a) TextureMesh; (b) TriangleSet; (c) DiscreElements))

图10 吉普车刚体动力学仿真管线

Fig. 10 Simulation pipeline of rigid body dynamics for a jeep

图11 数据域同步机制示意图

Fig. 11 Diagram of the data field synchronization mechanism

图12 仿真与交互异步协同示意图

Fig. 12 Diagram of asynchronous collaboration between simulation and interaction

图13 基于扩散边界的血流仿真场景((a1)自适应网格二维截面图；(a2)速度场；(b)节点图)

Fig. 13 Diffuse interface model for blood flow simulation ((a1) A two-dimensional section of the adaptive grid; (a2) Velocity field; (b) Node graph)

图14 基于有限元法的泵体刚强度仿真分析((a)效果图；(b)节点图)

Fig. 14 FEM-based simulation and analysis for pump ((a) Visual effect; (b) Node graph)

图15 基于对偶粒子的流体仿真((a)效果图；(b)节点图)

Fig. 15 Fluid simulation based on a dual particle approach ((a) Visual effect; (b) Node graph)

图16 基于图神经网络的瞬态碰撞问题求解

Fig. 16 Solving transient collision problems based on graph neural networks

图17 面向二维动态问题的智能求解

Fig. 17 Intelligent solving for two-dimensional dynamic problems

图18 面向三维碰撞问题的智能快速仿真结果与实验对比

Fig. 18 Comparison of intelligent rapid simulation results and experiments for three-dimensional collision ((a) 0.00 ms; (b) 4.38 ms; (c) 7.29 ms; (d) 11.56 ms; (e) 18.00 ms)

参考文献 59

[1]	REN L X, WAN S, WEI Y, et al. Towards a non-invasive diagnosis of portal hypertension based on an Eulerian CFD model with diffuse boundary conditions[C]// The 24th International Conference on Medical Image Computing and Computer Assisted Intervention. Cham: Springer, 2021: 107-116.
[2]	KROLL N, GAUGER N R, BREZILLON J, et al. Flow simulation and shape optimization for aircraft design[J]. Journal of Computational and Applied Mathematics, 2007, 203(2): 397-411.
[3]	Geyer M. BMW group starts global rollout of NVIDIA omniverse[EB/OL]. (2023-03-21). [2024-05-31]. https://blogs.nvidia.com/blog/bmw-group-nvidia-omniverse/.
[4]	SZLUFARSKA I, RAMESH K T, WARNER D H. Simulating mechanical behavior of ceramics under extreme conditions[J]. Annual Review of Materials Research, 2013, 43: 131-156.
[5]	MARLOW F, JACOB J, SAGAUT P. A multidisciplinary model coupling Lattice-Boltzmann-based CFD and a social force model for the simulation of pollutant dispersion in evacuation situations[J]. Building and Environment, 2021, 205: 108212.
[6]	THOMÉE V. From finite differences to finite elements: a short history of numerical analysis of partial differential equations[J]. Journal of Computational and Applied Mathematics, 2001, 128(1/2): 1-54.
[7]	COURANT R, FRIEDRICHS K, LEWY H. Über die partiellen differenzengleichungen der mathematischen physik[J]. Mathematische Annalen, 1928, 100: 32-74.
[8]	CRANK J, NICOLSON P. A practical method for numerical evaluation of solutions of partial differential equations of the heat-conduction type[J]. Mathematical Proceedings of the Cambridge Philosophical Society, 1947, 43(1): 50-67.
[9]	COURANT R, ISAACSON E, REES M. On the solution of nonlinear hyperbolic differential equations by finite differences[J]. Communications on Pure and Applied Mathematics, 1952, 5(3): 243-255.
[10]	FOSTER N, FEDKIW R. Practical animation of liquids[C]// The 28th Annual Conference on Computer Graphics and Interactive Techniques. New York: ACM, 2001: 23-30.
[11]	RASMUSSEN N, NGUYEN D Q, GEIGER W, et al. Smoke simulation for large scale phenomena[J]. ACM Transactions on Graphics, 2003, 22(3): 703-707.
[12]	HONG J M, SHINAR T, FEDKIW R. Wrinkled flames and cellular patterns[C]// ACM SIGGRAPH 2007 Papers. New York: ACM, 2007: 47.
[13]	HRENNIKOFF A. Solution of problems of elasticity by the framework method[J]. Journal of Applied Mechanics, 1941, 8(4): A169-A175.
[14]	COURANT R. Variational methods for the solution of problems of equilibrium and vibrations[J]. Bulletin of the American Mathematical Society, 1943, 49(1): 1-23.
[15]	冯康. 基于变分原理的差分格式[J]. 应用数学与计算数学, 1965, 2(4): 238-262.
	FENG K. Difference schemes based On variational principle[J]. Journal of Applied and Computational Mathematics, 1965, 2(4): 238-262. (in Chinese)
[16]	YOSIBASH Z, WEISS D, HARTMANN S. High-order FEMs for thermo-hyperelasticity at finite strains[J]. Computers & Mathematics with Applications, 2014, 67(3): 477-496.
[17]	MARTIN A, ESNAULT J B, MASSIN P. About the use of standard integration schemes for X-FEM in solid mechanics plasticity[J]. Computer Methods in Applied Mechanics and Engineering, 2015, 283: 551-572.
[18]	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.
[19]	LUCY L B. A numerical approach to the testing of the fission hypothesis[J]. The Astronomical Journal, 1977, 82: 1013-1024.
[20]	SOLENTHALER B, PAJAROLA R. Predictive-corrective incompressible SPH[J]. ACM Transactions on Graphics, 2009, 28(3): 40.
[21]	HE X W, LIU N, LI S, et al. Local Poisson SPH for viscous incompressible fluids[J]. Computer Graphics Forum, 2012, 31(6): 1948-1958.
[22]	IHMSEN M, CORNELIS J, SOLENTHALER B, et al. Implicit incompressible SPH[J]. IEEE Transactions on Visualization and Computer Graphics, 2014, 20(3): 426-435.
[23]	BENDER J, KOSCHIER D. Divergence-free smoothed particle hydrodynamics[C]// The 14th ACM SIGGRAPH/Eurographics Symposium on Computer Animation. New York: ACM, 2015: 147-155.
[24]	HE X W, LIU N, WANG G P, et al. Staggered meshless solid-fluid coupling[J]. ACM Transactions on Graphics, 2012, 31(6): 149.
[25]	BAND S, GISSLER C, IHMSEN M, et al. Pressure boundaries for implicit incompressible SPH[J]. ACM Transactions on Graphics, 2018, 37(2): 14.
[26]	PEER A, TESCHNER M. Prescribed velocity gradients for highly viscous SPH fluids with vorticity diffusion[J]. IEEE Transactions on Visualization and Computer Graphics, 2017, 23(12): 2656-2662.
[27]	LIU S S, HE X W, WANG W C, et al. Adapted SIMPLE algorithm for incompressible SPH fluids with a broad range viscosity[J]. IEEE Transactions on Visualization and Computer Graphics, 2022, 28(9): 3168-3179.
[28]	OÑATE E, IDELSOHN S, ZIENKIEWICZ O C, et al. A stabilized finite point method for analysis of fluid mechanics problems[J]. Computer Methods in Applied Mechanics and Engineering, 1996, 139(1/4): 315-346.
[29]	ATLURI S N, ZHU T L. New concepts in meshless methods[J]. International Journal for Numerical Methods in Engineering, 2000, 47(1/3): 537-556.
[30]	SILLING S A. Reformulation of elasticity theory for discontinuities and long-range forces[J]. Journal of the Mechanics and Physics of Solids, 2000, 48(1): 175-209.
[31]	HE X W, WANG H M, WU E H. Projective peridynamics for modeling versatile elastoplastic materials[J]. IEEE Transactions on Visualization and Computer Graphics, 2018, 24(9): 2589-2599.
[32]	LU Z X, HE X W, GUO Y Z, et al. Projective peridynamic modeling of hyperelastic membranes with contact[J]. IEEE Transactions on Visualization and Computer Graphics, 2024, 30(8): 4601-4614.
[33]	YANG Y B, PERDIKARIS P. Adversarial uncertainty quantification in physics-informed neural networks[J]. Journal of Computational Physics, 2019, 394: 136-152.
[34]	HAGHIGHAT E, RAISSI M, MOURE A, et al. A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics[J]. Computer Methods in Applied Mechanics and Engineering, 2021, 379: 113741.
[35]	JIN X W, CAI S Z, LI H, et al. NSFnets (Navier-Stokes flow nets): physics-informed neural networks for the incompressible Navier-stokes equations[J]. Journal of Computational Physics, 2021, 426: 109951.
[36]	SAAD Y. Iterative methods for sparse linear systems[M]. 2nd ed. Philadelphia: Society for Industrial and Applied Mathematics, 2003: 105-132.
[37]	NOCEDAL J, WRIGHT S J. Numerical optimization[M]. New York: Springer, 1999: 101-134.
[38]	KIRCHDOERFER T, ORTIZ M. Data-driven computational mechanics[J]. Computer Methods in Applied Mechanics and Engineering, 2016, 304: 81-101.
[39]	XU L Y, HE X W, CHEN W, et al. Reformulating hyperelastic materials with peridynamic modeling[J]. Computer Graphics Forum, 2018, 37(7): 121-130.
[40]	LAI W M, RUBIN D, KREMPL E. Introduction to continuum mechanics[M]. 4th ed. Boston: Butterworth-Heinemann, 2010: 220-225.
[41]	MÜLLER M, CHARYPAR D, GROSS M. Particle-based fluid simulation for interactive applications[C]// The 2003 ACM SIGGRAPH/Eurographics Symposium on Computer Animation. Goslar: Eurographics Association, 2003: 154-159.
[42]	CAI S Z, MAO Z P, WANG Z C, et al. Physics-informed neural networks (PINNs) for fluid mechanics: a review[J]. Acta Mechanica Sinica, 2021, 37(12): 1727-1738.
[43]	CIFTCI K, HACKL K. A physics-informed GAN framework based on model-free data-driven computational mechanics[J]. Computer Methods in Applied Mechanics and Engineering, 2024, 424: 116907.
[44]	ABADI M, BARHAM P, CHEN J M, et al. TensorFlow: a system for large-scale machine learning[C]// The 12th USENIX symposium on Operating Systems Design and Implementation. Berkeley: USENIX Association, 2016: 265-283.
[45]	LU L, JIN P Z, PANG G F, et al. Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators[J]. Nature Machine Intelligence, 2021, 3(3): 218-229.
[46]	LIU M B, LIU G R. Smoothed particle hydrodynamics (SPH): an overview and recent developments[J]. Archives of Computational Methods in Engineering, 2010, 17(1): 25-76.
[47]	KEYES D E, MCINNES L C, WOODWARD C, et al. Multiphysics simulations: challenges and opportunities[J]. The International Journal of High Performance Computing Applications, 2013, 27(1): 4-83.
[48]	FAURE F, DURIEZ C, DELINGETTE H, et al. SOFA: a multi-model framework for interactive physical simulation[M]//Soft Tissue Biomechanical Modeling for Computer Assisted Surgery. Heidelberg:Springer, 2012:283-321.
[49]	HE X W, CHANG YUE, SHI J, et al. PeriDyno: an AI-Targeted physical simulation platform[EB/OL]. [2024-06-15]. https://github.com/peridyno/peridyno/.
[50]	WANG R, QIAN X L. Welcome to osg3-cookbook’s documentation[EB/OL]. [2024-06-15]. https://osg3cookbook.readthedocs.io/en/latest/index.html.
[51]	Unity Technologies. ECS for unity[EB/OL]. [2024-06-15]. https://unity.com/ecs#resources.
[52]	SCHECHTER H, BRIDSON R. Ghost SPH for animating water[J]. ACM Transactions on Graphics, 2012, 31(4): 61.
[53]	MACKLIN M, MÜLLER M. Position based fluids[J]. ACM Transactions on Graphics, 2013, 32(4): 104.
[54]	LIU S S, HE X W, GUO G Z, et al. A dual-particle approach for incompressible SPH fluids[J]. ACM Transactions on Graphics, 2024, 43(3): 28.
[55]	LI Y, WANG H, LI B Y, et al. An image-driven uncertainty inverse method for sheet metal forming problems[J]. Journal of Mechanical Design, 2022, 144(2): 022001.
[56]	YIN J C, WEN Z M, LI S H, et al. Dynamically configured physics-informed neural network in topology optimization applications[J]. Computer Methods in Applied Mechanics and Engineering, 2024, 426: 117004.
[57]	WEN Z M, LI Y, WANG H, et al. Data-driven spatiotemporal modeling for structural dynamics on irregular domains by stochastic dependency neural estimation[J]. Computer Methods in Applied Mechanics and Engineering, 2023, 404: 115831.
[58]	STOMAKHIN A, SCHROEDER C, CHAI L, et al. A material point method for snow simulation[J]. ACM Transactions on Graphics, 2013, 32(4): 102.
[59]	ZHU Z, WANG X F, ZHAO W B, et al. Is Sora a world simulator? A comprehensive survey on general world models and beyond[EB/OL]. [2024-05-31]. https://doi.org/10.48550/arXiv.2405.03520.

机理与数据驱动的物理仿真计算范式及引擎架构

The computational paradigm and software framework for mechanism and data-driven physical simulation

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献 59

相关文章 3

编辑推荐

Metrics

本文评价

[1]	何庆, 荆传玉, 孙华坤, 姚力, 徐井芒, 王平. 基于BIM和语义网的轨道智能运维管理方法[J]. 图学学报, 2024, 45(3): 601-612.
[2]	黄忠山 1,2，田凌 1,2 . 智能装备故障预测与健康管理系统研究[J]. 图学学报, 2018, 39(5): 855-861.
[3]	徐江斌，赵健，杨超，吴玲达. 云图数据驱动的三维云景可视化[J]. 图学学报, 2010, 31(3): 68-75.