有限元法
光滑粒子流体力学
汽车工业
流固耦合
晃动动力学
材料点法
计算机科学
机械工程
工程类
机械
结构工程
物理
航空航天工程
作者
Essam Al-Bahkali,Hisham Elkenani,Mhamed Souli,Fouad Erchiqui,Mojtaba Moatammedi
出处
期刊:Elsevier eBooks
[Elsevier]
日期:2021-01-01
卷期号:: 1-20
标识
DOI:10.1016/b978-0-12-817899-7.00002-2
摘要
Simulation of airbag and membrane deployment under pressurized gas problems becomes more and more the focus of computational engineering, where finite element methods (FEMs) for structural mechanics and finite volume for computational fluid dynamics are dominant. New formulations have been developed for fluid structure interaction (FSI) applications using mesh free methods as smooth particle hydrodynamic (SPH) method. Up to these days very little has been done to compare different methods and assess which one would be more suitable. For small deformation, FEM Lagrangian formulation can solve structure interface and material boundary accurately, the main limitation of the formulation is high mesh distortion for large deformation and moving structure. One of the commonly used approaches to solve these problems is the arbitrary Lagrangian Eulerian (ALE) formulation which has been used with success in the simulation of FSI with large structure motion such as sloshing fuel tank in automotive industry and bird impact in aeronautic industry. For some applications, including bird impact and high velocity impact problems, engineers have switched from ALE to SPH method to reduce central processing unit (CPU) time and save memory allocation. Both ALE and SPH methods are described and compared here using similar mesh size, each ALE element is replaced by an SPH particle at the element center. From different simulation, it has been observed that for the SPH method to provide similar results as ALE or Lagrangian formulations, the SPH meshing needs to be finer than the ALE mesh. A contact algorithm is performed at the FSI for both SPH and ALE formulations. A simulation of airbag membrane deployment generated by high pressurized gas is performed.
科研通智能强力驱动
Strongly Powered by AbleSci AI