Mpi Platform Two Dimensional Euler Equations Numerical Solution

As we know, much more grid points are required in order to differentiate the structure of flow field, which is extremely complex. In addition, unsteady flow simulation needs to gain flow movement along time. As a result, the computing task of the numerical simulation is very great. In order to make the long computing can be expediently applied to engineering, the computing time have to be decreased and the efficiency must be increased. There are two main ways that will be effective to solve the question currently: one is to improving numerical algorithm by all kinds of skills to accelerate convergence; the other is to introducing parallel computing.In order to work out the numerical simulation of flow in NACA0012, Euler equations are selected as the control equations of the flow field in this paper. The main numerical method of this code is coming from scheme (Jameson, Schimit and Turkel): using cell-centered finite volume method as spatial discretization tools, and a system of ordinary differential equations for time variable is obtained, which is solved by utilizing five-step Runge-Kutta scheme as time marching method , introducing artificial dissipation to damp high frequency oscillations near the shock and stagnation point.The thesis is focused on the following aspects:1 .As the base of comparison with parallel computing, an example which is completed on single processor is given.2.Cluster of workstation (COW) has been established as a parallel computing platform.3.A parallel computing program of two-dimension flow numerical calculation has been completed. Some important techniques such as communication among nodes and load balancing are analyzed.