Development of a CFD code using TVD schemes and advanced turbulence models for incompressible flow simulations

A computational fluid dynamic (CFD) code is developed for simulating two-dimensional steady incompressible turbulent flow. The time-averaged incompressible Navier-Stokes equations are modified by using the pseudo-compressible formulation. Therefore the time-marching schemes which are developed for compressible flow can be used for solving incompressible flow problems. There are three critical elements that this CFD code consists of. First, it can generate good quality grid to discretize the domain of the problem that is under study. Second, turbulence models are used so that the turbulence phenomenon can be simulated. And third, an efficient numerical scheme is used so that the solutions for the governing equations can be obtained efficiently and accurately.; The grid is generated by using grid generation techniques. A preliminary grid first is generated by using an algebraic grid generator, and then is modified by using a elliptic grid generator. The grid generators can control the quality of the grid so that the desired arrangement of the grid points can be obtained. In order to properly simulate flow fields with turbulence, turbulence models need to be used. This code provides three choices: the Baldwin-Lomax algebraic model, Chien's low-Reynolds number k-{dollar}varepsilon{dollar} model, and Wilcox's modified k-{dollar}omega{dollar} model. The solutions for the time-averaged incompressible Navier-Stokes equation and the turbulence model equation are obtained consecutively by using the TVD scheme incorporated with the approximate factorization method.; The validity of the code is demonstrated by applying it to a fully developed channel flow. Comparing the results with the DNS data shows that the code can produce competitive velocity profile. The code is also applied to the NREL (National Renewable Energy Laboratory) S809 airfoil at various angles of attack. The flow properties that the code predicts show good agreement compared with the Delft two-dimensional wind tunnel test data (Somers, 1989) and the Eppler code data.