CARTESIAN GRID FLUID SIMULATION METHOD BASED ON ADAPTIVE NORMAL RAY REFINEMENT

In viscous flow simulations, the velocity gradient in the normal direction of the wall is much larger than that in the tangential direction, presenting a significant anisotropic feature. Traditional isotropic Cartesian grid methods face the challenges of a sharp increase in the number of grids and a decline in computational efficiency when capturing the details of boundary layer flows. To address this issue, this paper proposes a Cartesian grid viscous fluid simulation method based on Adaptive Normal Ray Refinement (ANRR). The core of this method lies in adaptively generating normal ray seed points according to the degree of angle change in the tangential direction of the surface, and performing grid refinement near the normal rays to accurately capture the characteristics of boundary layer flows. Meanwhile, coarser grids are used for transition between rays, effectively reducing the overall number of grids while maintaining computational accuracy. Then, an efficient information transfer technology between rays is constructed through interpolation to ensure the accuracy of the flow field solution process. Finally, typical cases such as laminar flow over a flat plate, low Reynolds number flow around a circular cylinder, and flow around an NACA0012 airfoil are used for model verification. The results show that compared with the traditional geometric adaptive grid refinement method, the ANRR grid method significantly reduces the grid scale in the boundary layer flow region, improves computational efficiency while maintaining high accuracy, and provides a new solution for the efficient solution of viscous flow problems using adaptive Cartesian grids.