北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2024, Vol. 47 ›› Issue (4): 44-49.

• 体系化人工智能专题 • 上一篇    下一篇

基于空间极坐标卷积的流体动力学预测

杜飞龙1,2,班晓娟1,2,3,4,5,张雅斓1,2,3,董子瑞2,王笑琨1,2,3,4   

  1. 1. 北京科技大学 北京材料基因工程高精尖创新中心
    2. 北京科技大学 智能科学与技术学院
    3. 北京科技大学 顺德创新学院 4.北京科技大学 智能仿生无人系统教育部重点实验室 5.辽宁材料实验室 材料智能技术研究所
  • 收稿日期:2023-12-29 修回日期:2024-03-11 出版日期:2024-08-28 发布日期:2024-08-26
  • 通讯作者: 班晓娟 E-mail:banxj@ustb.edu.cn
  • 基金资助:
    科技创新2030-重大项目; 国家自然科学基金项目; 广东省基础与应用基础研究基金项目; 海南省重点研发项目

Fluid Dynamics Prediction Based on Spatial Polar Coordinate Convolution

DU Feilong1,2, BAN Xiaojuan1,2,3,4,5, ZHANG Yalan1,2,3, DONG Zirui2, WANG Xiaokun1,2,3,4   

  • Received:2023-12-29 Revised:2024-03-11 Online:2024-08-28 Published:2024-08-26
  • Contact: Xiao-juan Ban E-mail:banxj@ustb.edu.cn

摘要: 流体作为自然界中最基本的物质之一,对其模拟往往需要在精度和效率之间进行取舍。为此,提出了一种端到端体系化的新型卷积网络流体模拟器———PolarNet。首先,将粒子数据转换为三维极坐标表示,设计了 PolarConv空间卷积结构。其次,结合物理流体模拟器,构建了四层网络结构,设计了网络流体模拟器 PolarNet,实现流体的端到端预测。此外,精心设计了基于物理的约束,以强化流体的不可压缩性。实验结果表明,与传统基于建模模拟器相比,PolarNet 在提高流体边界的精确性和保持不可压缩性方面有显著提升,同时保证了效率。相较于其他基于学习的流体模拟器,得益于极坐标的空间紧凑表示,PolarNet 在具有更少训练参数的情况下,保持了最高的预测稳定性。PolarNet 为多模态信息处理提供了新的方法和视角。

关键词: 流体动力学模拟, 空间卷积网络, 数据驱动, 体系化模拟器

Abstract: Fluid, as one of the most fundamental substances in nature, often requires a trade-off between accuracy and efficiency in its simulation. Thus, a novel end-to-end systematic convolutional network fluid simulator called PolarNet is proposed. Firstly, particle data is converted to a 3D polar coordinate representation, and the PolarConv spatial convolution structure is designed. Secondly, combined with a physical fluid simulator, a four-layer network structure is constructed to design the network fluid simulator PolarNet, achieving end-to-end fluid prediction. Additionally, physically-based constraints are carefully designed to enhance fluid incompressibility. Experimental results show that compared with traditional modeling-based simulators, PolarNet significantly improves the accuracy of fluid boundaries and maintains incompressibility while ensuring efficiency. Compared with the other learning-based fluid simulators, PolarNet maintains the highest prediction stability with fewer training parameters, benefiting from the compact representation of polar coordinates. PolarNet provides new methods and perspectives for multimodal information processing.

Key words:

fluid dynamics simulation , spatial convolutional network ,  data-driven ,  systematic simulator

中图分类号: