Suppr超能文献

一种用于不可压缩流中运动边界的高效离散浸入边界方法。

A high-efficiency discretized immersed boundary method for moving boundaries in incompressible flows.

机构信息

State Key Laboratory of Hydraulic Engineering Simulation and Safety, Tianjin University, Tianjin, 300072, China.

College of Water Conservancy & Hydropower Engineering, Hohai University, Nanjing, China.

出版信息

Sci Rep. 2023 Jan 30;13(1):1699. doi: 10.1038/s41598-023-28878-5.

Abstract

The Immersed Boundary Method (IBM) has an advantage in simulating fluid-structure interaction, owning to its simplicity, intuitiveness, and ease of handling complex object boundaries. The interpolation function plays a vital role in IBM and it is usually computationally intensive. For moving or deforming solids, the interpolation weights of all the immersed boundary points ought to be updated every time step, which takes quite a lot CPU time. Since the interpolation procedure within all uniform structured grids is highly repetitive and very similar, we propose a simple and generalized Discretized Immersed Boundary Method (DIBM), which significantly improves efficiency by discretizing the interpolation functions onto subgrid points within each control volume and reusing a predefined universal interpolation stencil. The accuracy and performance of DIBM are analyzed using both theoretical estimation and simulation tests. The results show speedup ratios of 30-40 or even higher using DIBM when compared with conventional IBM for typical moving boundary simulations like particle-laden flows, while the error is estimated to be under 1% and can be further decreased by using finer subgrid stencils. By balancing the performance and accuracy demands, DIBM provides an efficient alternative framework for handling moving boundaries in incompressible viscous flows.

摘要

浸没边界法 (IBM) 在模拟流固耦合方面具有优势,因为它具有简单、直观和易于处理复杂物体边界的特点。插值函数在 IBM 中起着至关重要的作用,但通常计算量很大。对于移动或变形的固体,每次时间步长都需要更新所有浸没边界点的插值权重,这需要大量的 CPU 时间。由于所有均匀结构网格内的插值过程高度重复且非常相似,因此我们提出了一种简单而通用的离散化浸没边界法 (DIBM),通过将插值函数离散到每个控制体积内的子网格点,并重新使用预定义的通用插值模板,显著提高了效率。我们通过理论估计和模拟测试分析了 DIBM 的准确性和性能。结果表明,与传统的 IBM 相比,DIBM 在处理类似颗粒负载流的典型移动边界模拟时,可以将速度提高 30-40 倍甚至更高,而误差估计在 1%以下,并且可以通过使用更精细的子网格模板进一步降低误差。通过平衡性能和准确性需求,DIBM 为处理不可压缩粘性流中的移动边界提供了一种高效的替代框架。

https://cdn.ncbi.nlm.nih.gov/pmc/blobs/eb12/9887058/14fa64389e64/41598_2023_28878_Fig1_HTML.jpg

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验