A Novel Framework for Automated Grid Generation and CFD Simulation of Complex Geometries

A novel technical framework for the automatic generation of grids and subsequent flow field computation in complex geometries was introduced. This framework integrated body-fitted grids that accommodate intersecting grid lines, the flexible-node finite difference method (FN-FDM) designed for unordered point clouds, and a universal algorithm for freestream preservation (discrete equivalence equation and its discrete rule, DEER). It achieved full automation in the process from importing discrete surface points to generating flow field computation results. The grid generation algorithm, starting from a vector of discrete points on the object surface, generated body-fitted grides in a single pass based on parameters such as the height of the grid layer, growth rate, and number of layers. It removed overlapping grides and reconstructed the computational stencil required for FN-FDM. FN-FDM utilized the reconstructed computational stencil to perform the differential solution of flow field parameters. DEER was used to eliminate the geometrically induced errors introduced during the differential computation from the physical plane to the computational plane, achieving freestream preservation and improving computational accuracy. Verification results demonstrate that the new framework can automatically generate grids under complex geometries, significantly improving the efficiency of grid generation, while maintaining computational accuracy comparable to that of conventional methods. This framework has practical value in engineering applications.