[열유체부문] 유한요소를 이용한 MIRA 모델주위의 유동해석

최형권(Hyounggwon Choi),강성우(Sungwoo Kang),유정열(Jung Yul Yoo) 한국자동차공학회 2001 한국자동차공학회 춘 추계 학술대회 논문집 Vol.2001 No.11_1

A finite element code based on P2PI or PIPI tetra element has been developed for the large eddy simulation (LES) of turbulent flows around a complex geometry. Fractional 4-step algorithm is employed to obtain time accurate solution since it is less expensive than the integrated formulation, in which the velocity and pressure fields are solved at the same time. Crank-Nicolson method is used for second order temporal discretizarion and Galerkin method is adopted for spatial discretization. For very high Reynolds number flows, which would require a fom1idable number of nodes to resolve the flow field, SUPG (Streamline Upwind Petrow-Galerkin) method is applied. As a benchmark problem for code validation. turbulent flows around a sphere and turbulent channel flows<br/> have been studied at various Reynolds numbers. Turbulent flow around MIRA model has been solved using P2PI element and Smagorinsky model at Re= 2.6 x 106. Furthermore, a parallel code based on domain-decomposition method using PIPI element has been developed for LES of turbulent flow and tested with turbulent channel flows. Turbulent flow analysis around MIRA model using the present LES-parallel code is ongoing and will be presented in the near future<br/>

병렬화된 Dynamic LES 를 이용한 자동차 모델 주위의 유동해석

강성우(Sungwoo Kang),최형권(Hyounggwon Choi),유정열(Jung Yul Yoo) 대한기계학회 2002 대한기계학회 춘추학술대회 Vol.2002 No.5

For large scale computation of turbulent flows around an arbitrarily shaped body, a parallelized LES (large<br/> eddy simulation) code has been developed in which domain decomposition method is adopted. 4-step splitting<br/> finite element algorithm is adopted for unsteady computation of the incompressible Navier-Stokes equation<br/> along with Smagorinsky or dynamic LES model which is chosen for the modeling of small eddies in turbulent<br/> flows. For the validation and performance-estimation of the parallel code, a three-dimensional laminar flow<br/> generated by natural convection inside a cube has been solved. Regarding the speed-up of the code, the<br/> present parallel code with parallel block-Jacobi preconditioner is about 50 times faster than the corresponding<br/> serial code with 64 processors when approximately one million grid points are used. For the validation of LES<br/> models, turbulent channel flows are simulated at Re = 180. Lastly, we consider the turbulent flow around a<br/> MIRA (Motor Industry Research Association) model at Re = 2.6×10^6. Both Smagorinsky and dynamic models<br/> are tested, comparing estimated drag coefficients and pressure distribution along the model surface with the<br/> existing experimental data. The calculated drag coefficient agrees better with the experimental result than<br/> those using two equation turbulence models.

자유 낙하하는 실린더 주위의 층류 유동에서 St-Re 관계

남궁각(Kak NamKoong),최형권(HyoungGwon Choi),유정열(JungYul Yoo) 대한기계학회 2001 대한기계학회 춘추학술대회 Vol.2001 No.9

In the FSI (Fluid-Structure Interaction) problems, two different governing equations are to be solved together. One is for fluid and the other for structure. Furthermore, a kinematic constraint should be imposed along the boundary between the fluid and structure. We use the combined formulation (Hesla, 1991) which incorporates both the fluid and structure equations of motion into a single coupled variational equation so that it is not necessary to calculate the fluid force on the surface of structure explicitly when solving the equations of motion of the structure. Before treating complex FSI problems, laminar flow around a freely-falling cylinder is considered. The Navier-Stokes equations are solved using a P2P1 Galerkin finite element formulation with ALE (Arbitrary Lagrangian-Eulerian) algorithm and Newton’s equations of motion for cylinder are solved. The adaptive mesh refinement technique is also adopted which uses stress error as a posteriori error estimator together with an efficient variable-reordering and element-reordering method for unstructured finite element meshes. The numerical simulation results reveal the effect of vortex shedding on the motion of a falling cylinder. The cylinder falls oscillating in the spanwise direction and rotating about the center. A correlation of St-Re relationship for a freely-falling cylinder is to be proposed from the numerical data of the present study.

비압축성 2 상유동의 모사를 위한 Level Set 방법의 Reinitialization 방정식의 해법에 관한 연구

조명환(Myung Hwan Cho),최형권(HyoungGwon Choi),유정열(Jung Yul Yoo) 대한기계학회 2008 大韓機械學會論文集B Vol.32 No.10

Computation of moving interface by the level set method typically requires the reinitialization of level set function. An inaccurate estimation of level set function φ results in incorrect free-surface capturing and thus errors such as mass gain/loss. Therefore, an accurate and robust reinitialization process is essential to the simulation of free-surface flows. In the present paper, we pursue further development of the reinitialization process, which evaluates level set function directly using a normal vector on the interface without solving the re-distancing equation of hyperbolic type. The Taylor-Galerkin approximation and P1P1 splitting/SUPG (Streamline Upwind Petrov-Galerkin) FEM are adopted to discretize advection equation of the level set function and the incompressible Navier-Stokes equation, respectively. Advection equation and re-initialization process of free surface capturing are validated with benchmark problems, i.e., a broken dam flow and timereversed single vortex flow. The simulation results are in good agreement with the existing results.

영역분할 Navier-Stokes 방정식의 병렬 예조건화

강성우(Sungwoo Kang),최형권(Hyounggwon Choi),유정열(Jung Yul Yoo) 대한기계학회 2001 대한기계학회 춘추학술대회 Vol.2001 No.9

Convergence of iterative solvers like CG(Conjugate Gradient method) are influenced by stiffness of matrix. Preconditioner matrix is used to decrease the condition number of a matrix accelerating the convergence of CG. Parallel Block-Jacobi, Iterative Block-Jacobi, and distributed ILU(Incomplete LU factorization) methods are tested to precondition a SPD(symmetric positive definite) matrix. In this paper, we investigate various parallel preconditioners to measure the effectiveness of the algorithms when P1P1 finite element discretization is used for the parallel solution of the Navier-Stokes equations. To parallelize a finite element code for the simulation of Navier-Stokes equations, the calculation domain is divided into subdomains and each subdomain is allocated to each processor for calculation. One and two-dimensional domain decomposition based on node are adapted and their performances are compared. MPI(Message Passing Inteface) is used for communication between processors and structured grid is used for the ease of domain decomposition in this benchmark study. Two-dimensional and three-dimensional Laplace equations are calculated on structured uniform and nonuniform grids to compare the speed-up of various parallel preconditioning methods. For the domain-decomposed-parallel computation of the Navier-Stokes equations we have solved three dimensional lid-driven cavity problem and natural convection problem as benchmark problems using parallelized fractional 4 step finite element method. The present results are in good agreement with previous results. The speed-up for each preconditioning methods are to be compared upto 64 processors.

비정렬격자를 이용한 자동차 주위의 유동해석

김범준(Beomjun Kim),강성우(Sungwoo Kang),최형권(Hyounggwon Choi),유정열(Jung Yul Yoo) 대한기계학회 2002 대한기계학회 기타 간행물 Vol.2002 No.11

다양한 LES 모델을 고려한 3차원 비대칭 스테노시스 유동에 대한 수치해석

전병진(ByoungJin Jeon),김정우(Jungwoo Kim),최형권(Hyounggwon Choi) 대한기계학회 2014 대한기계학회 춘추학술대회 Vol.2014 No.11

에어프라이어 바닥면 형상이 상승대류 속도에 미치는 영향의 고찰

임세환(Lim Sehwan),장윤호(Jang Yoonho),최형권(Choi Hyounggwon),한상조(Sangjo Han) 한국가시화정보학회 2020 한국가시화정보학회지 Vol.18 No.2

Airfryer is used to heat a food up by convecting hot air upward around the food. In this study, we investigated the effect of the bottom-shape of the food container in airfryer on the upward convection velocity of hot air to find an optimal bottom-shape by computational fluid dynamics. Numerical experiments were performed by solving the incompressible Navier-Stokes equations with turbulence model. We found that the maximum upward velocity with concave flow-passage on the bottom was bigger than that with the flat bottom and that the maximum upward convection velocity was achieved when the number of concave flow-passage with fan-shape is around six. The pressure drop by the internal flow was found to increase as the number of the concave flow-passage on the bottom increased probably due to increase of the surface area of the bottom. Therefore, it can be said that the optimal number of the concave flow-passage is around six for the flow rate considered in this study.