Numerical Method Study Of Three-dimension Turbulent Flow Based On SIMPLE Algorithm

As the development of modern industry, high capability is needed to turbo machinery, the mastery of flow field law in turbo machinery is the key to optimize design. In compared with experimental method and theory method, numerical method has its unique advantages. In this paper, SIMPLE algorithm and relative techniques are discussed and the program for the numerically simulation of turbulent flow is developed.According to the trait of flow for turbo machinery, the flow governing equations and the pressure correction equation are presented in 3-dimensional general curvilinear coordinates. Furthermore, the equations are deducted in the finite control volume in this study. Pressure is adopted as a primary dependent variable and SIMPLE algorithm is used. In order to develop SIMPLE algorithm to compressible flow, retarded density is utilized. Non-stagger grid technique is used to make programming more clear and all-purpose. And the pressure-weighted interpolation method, proposed by Rhie and Chow, is conducted to eliminate the numerical concussion bringed by non-stagger grid. The standard k-s model is utilized to describe the turbulent flow process. The wall function method is adopted for dealing with the boundary condition at the standing wall. The disadvantages of the standard k-ε model have been discussed and the developing models of the standard k-ε model are discussed to correct the effect of curvature or eddy too. The difference scheme of convective transport terms has been discussed and high resolution scheme MUSCL scheme is implemented in the program. Block-correction technique and TDMA algorithm are used to solve the discrete equations. The techniques of grid generation are discussed. The grids for samples are generated through algebraic method.The program for the numerically simulation of turbulent flow is developed, the flow fields in 90° curved duct and contract-expand nozzle are numerically simulated use the program. The reliability of the algorithm and the program has been proved by the agreement between the calculated results and the experimental results available in literature.