Research On Anisotropic Adaptive Cartesian Grid Method And Its Application To High-speed Trains

Cartesian grids have the characteristics of good grid quality,high degree of automation,and strong adaptive ability to complex shapes/flow fields,making them suitable for numerical simulation research on complex shapes/flow fields.However,when simulating the anisotropic flow characteristics of the boundary layer in detail,the required number of Cartesian grids shows an exponential growth with the Reynolds number,especially for high Reynolds number flow,facing the problem of“grid catastrophe”.Anisotropic adaptive Cartesian grids can refinement only in a specific direction according to the actual needs of the shape/flow field,in order to reduce the grid size while ensuring the accuracy of the calculation results.Therefore,this paper is based on anisotropic adaptive Cartesian grids,conducting research on grid generation,solver construction,and exploring their feasibility and reliability in the aerodynamic performance evaluation of trains.A robust,efficient,and low-memory-consumption anisotropic adaptive three-dimensional Cartesian grid generation method for complex shapes has been proposed.A data structure for surface triangle grids with a nested bounding box concept and an anisotropic spatial Cartesian grid data structure with member-encapsulated traces pointing to them have been constructed.An efficient neighbor query method is established,and three criteria for neighbor relationships are proposed to ensure the correctness,rationality,and uniqueness of neighbor relationships.A Cartesian grid-triangle intersection algorithm based on the Separating Axis Theorem and combined with the data structure of the surface triangle grid is developed to determine grid types,improving the robustness and efficiency of grid generation.A five-step geometric adaptive refinement strategy including shape adaptive,curvature adaptive,anisotropic adaptive,buffer layer addition,and grid quality check is constructed to ensure high-fidelity geometric information,reasonable grid size,and optimal grid quality.The anisotropic adaptive Cartesian grid generation method in this paper is one order of magnitude more efficient than previous publications,reduces memory consumption by more than 20%compared to traditional 2~N data structure,and reduces grid numbers by over 50%compared to isotropic adaptive methods.An efficient and high-precision surface boundary handling method and a suspension mesh flow field reconstruction method have been established.A wall treatment method has been developed with reasonable reference point location selection and accurate flow field reconstruction of the gohst cell.It has achieved the adaptive correction and application of the Spalding wall function under the non-body-fitted Cartesian grid framework.A method for selecting interpolation template points at suspended meshes and reconstructing flow field values at template points has been constructed.Combined with the surface triangle grid data structure,the backtracking process of data points far from the geometry surface has been optimized in the nearest neighbor search algorithm,and an accurate and efficient calculation method for the wall distance under the Cartesian grid framework has been designed.A flow field solution adaptive criterion based on velocity divergence,velocity rotation,and flow field variable gradient has been developed to solve the problem of ineffective flow field solution adaptive criteria under the anisotropic adaptive Cartesian grid framework.A numerical solver applicable to the anisotropic adaptive Cartesian grid framework has been constructed.A fully automatic anisotropic adaptive Cartesian grid program integrating grid generation and flow field solution has been developed,achieving accurate and efficient simulation of three-dimensional flow problems.Typical cases including cylinders,wedges,spheres,asymmetric airfoils,symmetric wing sections,and high-speed trains have been used to verify the correctness,reliability,and efficiency of the program by comparing with experimental results.For example,in the case of a wedge,the anisotropic adaptive grid reduces the number of grids by about 20%and the calculation time by about 13%compared to isotropic adaptive grids.Meanwhile,the errors of isotropic and anisotropic adaptive methods compared with analytical solutions are 7.2% and 6.9%,respectively.