DualFluidNet: an Attention-based Dual-pipeline Network for FLuid Simulation (2312.16867v2)
Abstract: Fluid motion can be considered as a point cloud transformation when using the SPH method. Compared to traditional numerical analysis methods, using machine learning techniques to learn physics simulations can achieve near-accurate results, while significantly increasing efficiency. In this paper, we propose an innovative approach for 3D fluid simulations utilizing an Attention-based Dual-pipeline Network, which employs a dual-pipeline architecture, seamlessly integrated with an Attention-based Feature Fusion Module. Unlike previous methods, which often make difficult trade-offs between global fluid control and physical law constraints, we find a way to achieve a better balance between these two crucial aspects with a well-designed dual-pipeline approach. Additionally, we design a Type-aware Input Module to adaptively recognize particles of different types and perform feature fusion afterward, such that fluid-solid coupling issues can be better dealt with. Furthermore, we propose a new dataset, Tank3D, to further explore the network's ability to handle more complicated scenes. The experiments demonstrate that our approach not only attains a quantitative enhancement in various metrics, surpassing the state-of-the-art methods but also signifies a qualitative leap in neural network-based simulation by faithfully adhering to the physical laws. Code and video demonstrations are available at https://github.com/chenyu-xjtu/DualFluidNet.
- Interaction networks for learning about objects, relations and physics. Advances in neural information processing systems, 29, 2016.
- Divergence-free smoothed particle hydrodynamics. In Proceedings of the 14th ACM SIGGRAPH/Eurographics symposium on computer animation, pages 147–155, 2015.
- Physics-informed neural networks (pinns) for fluid mechanics: A review. Acta Mechanica Sinica, 37(12):1727–1738, 2021.
- Rotation-invariant completion network. arXiv preprint arXiv:2308.11979, 2023.
- Attentional feature fusion. In Proceedings of the IEEE/CVF winter conference on applications of computer vision, pages 3560–3569, 2021.
- Smoothed particle hydrodynamics: theory and application to non-spherical stars. Monthly notices of the royal astronomical society, 181(3):375–389, 1977.
- A bi-lipschitz continuous, volume preserving map from the unit ball onto a cube. Note di Matematica, 28(1):177–193, 2008.
- Neurofluid: Fluid dynamics grounding with particle-driven neural radiance fields. In International Conference on Machine Learning, pages 7919–7929. PMLR, 2022.
- Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016.
- Monte carlo convolution for learning on non-uniformly sampled point clouds. ACM Transactions on Graphics (TOG), 37(6):1–12, 2018.
- Smoothed particle hydrodynamics techniques for the physics based simulation of fluids and solids. arXiv preprint arXiv:2009.06944, 2020.
- Multidisciplinary topology optimization for reduction of sloshing in aircraft fuel tanks based on sph simulation. Structural and multidisciplinary optimization, 58:1719–1736, 2018a.
- Learning particle dynamics for manipulating rigid bodies, deformable objects, and fluids. arXiv preprint arXiv:1810.01566, 2018b.
- Reynolds averaged turbulence modelling using deep neural networks with embedded invariance. Journal of Fluid Mechanics, 807:155–166, 2016.
- Sfusion: Self-attention based n-to-one multimodal fusion block. In International Conference on Medical Image Computing and Computer-Assisted Intervention, pages 159–169. Springer, 2023.
- Position based fluids. ACM Transactions on Graphics (TOG), 32(4):1–12, 2013.
- Deep dynamical modeling and control of unsteady fluid flows. Advances in Neural Information Processing Systems, 31, 2018.
- Flexible neural representation for physics prediction. Advances in neural information processing systems, 31, 2018.
- Guaranteed conservation of momentum for learning particle-based fluid dynamics. Advances in Neural Information Processing Systems, 35:6901–6913, 2022.
- Pointnet: Deep learning on point sets for 3d classification and segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 652–660, 2017a.
- Pointnet++: Deep hierarchical feature learning on point sets in a metric space. Advances in neural information processing systems, 30, 2017b.
- Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational physics, 378:686–707, 2019.
- The earth mover’s distance as a metric for image retrieval. International journal of computer vision, 40:99–121, 2000.
- Learning to simulate complex physics with graph networks. In International conference on machine learning, pages 8459–8468. PMLR, 2020.
- Predictive-corrective incompressible sph. In ACM SIGGRAPH 2009 papers, pages 1–6. 2009.
- Accelerating eulerian fluid simulation with convolutional networks. In International Conference on Machine Learning, pages 3424–3433. PMLR, 2017.
- Lagrangian fluid simulation with continuous convolutions. In International Conference on Learning Representations, 2019.
- Dynamic graph cnn for learning on point clouds. ACM Transactions on Graphics (tog), 38(5):1–12, 2019.
- Physics-informed machine learning with smoothed particle hydrodynamics: Hierarchy of reduced lagrangian models of turbulence. Physical Review Fluids, 8(5):054602, 2023.
- Smoothed particle hydrodynamics (sph) for complex fluid flows: Recent developments in methodology and applications. Physics of Fluids, 31(1), 2019.
- Cddfuse: Correlation-driven dual-branch feature decomposition for multi-modality image fusion. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 5906–5916, 2023.
- Topology optimization on fuel tank rib structures for fuel sloshing suppression based on hybrid fluid–solid sph simulation. Thin-Walled Structures, 165:107938, 2021.