سال انتشار: ۱۳۹۰
محل انتشار: نوزدهمین همایش سالانه مهندسی مکانیک
تعداد صفحات: ۴
S. Sadeghi – Ph.D. Student, University of Shahid Bahonar
M. Ameri – Associate Professor, University of Shahid Bahonar
This work describes the combination of adaptive grid-embedding, equation adaptation and local time-step methods on semi-structured grids for compressible turbulent flows. Semi-structured grids involve unstructured agglomerations of macro-cells, each containing a structured inner grid. Equation adaptation can automatically sense viscous macrocells. Navier- stokes equations are solved within appreciably viscous macro-cells, whereas for the remaining macro-cells the equations are reduced to the Euler equations. In the grid-embedding methods, the algorithm senses high gradient regions and automatically divides grid cells in such regions. In order to accelerate convergence, a local time-step method is included. The strategy of local time-step method is to use different time-step for different cells because CFL condition depends on cell size and turbulent viscous value. This combination is used to solve three internal and external flow model problems. The combination of adaptive methods iscomplex, but it is shown in this paper that it is remarkably reduces the number of necessary nodes, memory usage and computational time. Also the results show that this combination is more effective for turbulent flows than for laminar flows.