| In modern aircraft design, there is an urgent requirement for accurate, efficient and easy-use aerodynamic data and computational analysis tools. However, the Euler codes are not satisfied all the demand for simulating the complex flow-fields. More powerful tools solving Navier-Stokes (N-S) equations are longed for viscous effect, more accurate and more detail information of flow-fields. It is possible to solve N-S equations with the improvement of the hardware and software in computer field, and the development of numeric methods in computational fluid dynamics (CFD). To numerically solve N-S equations, the present work is mainly focus on the following aspects:1. Solving the elliptic grid generation together with an algebraic method marching along the normal-to-wall direction, viscous grids around complex geometries are generated. The inner-layer grids with the algebraic method is othogonality and easy to control the distance to the wall. According to the Hilgenstock, the source items are calculated to control the othogonality and spacing of grid lines on boundaries.2. The strategy of multi-block grid system makes it possible to proceed for structured grid when dealing with great complex geometries, like several components, et al. In order to reduce the computational complexity and computational costs, the fields are solved with "ghost" cells. The implementations with "ghost" cells largely improve the computational efficiency at computers with vetorizing capabilities.3. Three spatially discrete schemes about the convection terms of the N-S equations: the centered difference with artificial viscosity by Jameson, the van Leer scheme of flux vector splitting, and the Roe scheme of flux difference splitting are studied respectively. In order to apply the central scheme to higher Mach number flows, the sensor about pressure or temperature aremodified with some TVD-like properties. Modification to van Leer schemeabout the energy component can preserve a constant total enthalpy in the steady flows. In order to achieve a computer code with a maximum degree of vectorization, the numerical flux function must be written uniformly with sign function. To get higher order accuracy, the MUSCL interpolation functions are applied for van Leer and Roe schemes. At the same time, the continuous differential van Albada limiter is used to overcome the spurious non-monotonicity (wiggles or over and undershoots).4. Some time stepping schemes, explicit four stages modified Runge-Kutta scheme, implicit LU-SGS and 2-Order accuracy LU-SGS-TS schemes, are used to solve 3-dimensional compressible N-S equations. The explicit method is widely used for its simpleness and little memory consumed with local time step and variable coefficients implicit residual smooth to accelerate the convergence procedure. According to Yoon and Jameson's ideas, an efficient implicit LU-SGS algorithm is carefully constructed by combing the advantages of LU factorization and Symmetric-Gauss-Seidel technique in such a way to make use the L and U operators scalar diagonal matrices, thus the numeric algorithm requires only scalar inversion. The computational efficiency is greatly improved with this scheme. To simulate unsteady flows accurately, a Newton-like sub-iteration, which is pseudo-time sub-iteration (TS) method, has been applied to the original LU-SGS scheme.5. Algebraic Baldwin-Lomax model, semi-equation Johnson-King model (there are two versions as J-K90A and J-K92 models), and two-equation k-g model derived from k- model are applied to simulate turbulent flows. All models can predict the attached and little separation flows. For the high angle-of-attack separated flows with distinct history effects, the J-K and k-g models perform better than B-L model. For multi-body and multi-block mesh system, model k-g, which does not need the distance normal-to-wall, performs much better adaptability for complex geometries.6. The delta wing and double delta win...
|
PDF Full Text Request