Parallelized Cartesian Grid Methodology for Non-Equilibrium Hypersonic Flow Analysis of Ballutes

Hypersonic;Ballute;Cartesian grid

Lee, Jin Wook ; Aerospace Engineering

University:Georgia Institute of Technology

Department:Aerospace Engineering

关键词: Hypersonic;    Ballute;    Cartesian grid;   

Others  :  https://smartech.gatech.edu/bitstream/1853/16150/1/Lee_JinWook_200708_phd.pdf

美国|英语

来源: SMARTech Repository

【 摘 要 】

Hypersonic flow analysis is performed on an inflatable aerocapture device called a "Ballute" for Titan's Mission.An existing unstructured Cartesian grid methodology is used as a starting point by taking advantage of its ability to automatically generate gridsover any deformed shape of the flexible ballute.The major effort for this thesis work is focused on advancing the existing unstructured Cartesian grid methodology.This includes implementing thermochemical nonequilibrium capability and porting it to a parallel computing environment using a Space-Filling-Curve (SFC) based domain decomposition technique.The implemented two temperature thermochemical nonequilibrium solver governs the finite rate chemical reactions and vibrational relaxation in the high temperature regimes of hypersonic flow.In order to avoid the stiffness problem in the explicit chemical solver, a point implicit method is adopted to calculate the chemical reaction source term.The AUSMPW+ scheme with MUSCL data reconstruction is adopted as the numerical scheme to avoid non-physical oscillations and the carbuncle phenomenon.The results for five species air model and for thirteen species N2-CH4-Ar model to simulate Titan entry are included for verification against DPLR (NASA Ames' structured grid hypersonic flow solver).The efficient parallel computation of any unstructured grid flow solver requires an adequate grid decomposition strategy because of its complex spatial data structure.The difficulties of even and block-contiguous partitioning in frequently adapting unstructured Cartesian grids are overcome by implementing the 3D Hilbert SFC.Grids constructed by the SFC for parallel environment promise short inter-CPU communication time while maintaining perfect load balancing between CPUs.The load imbalance due to the local solution adaption is simply apportioned by re-segmenting the curve into even pieces.The detailed structure of the 3D Hilbert SFC and parallel computing efficiency results based on this grid partition method are also presented.Finally, a structural dynamics tool (LS-DYNA) is loosely coupled with the present parallel thermochemical nonequilibrium flow solver to obtain the deformed surface definition of the ballute.