基于开源求解器CFL3D的主动流动控制数值模拟

Numerical Simulation of Active Flow Control Based on the Open-Source Solver CFL3D

摘要: 主动流动控制方法因其显著的增升效果备受关注，但其伴随的强剪切，大曲率分离等复杂流动对数值模拟提出了挑战。本文基于NASA开源求解器CFL3D，采用标模CC020-010EJ和CFJ6415翼型，系统研究了环量控制与协同射流两种典型主动流动控制方法的数值模拟策略，重点考察湍流模型与边界条件对精度的影响。研究发现：对于环量控制，标准SA模型难以准确捕捉翼型后缘Coanda效应所导致的流线偏转，分离预测精度不足；引入曲率修正的SA + RC模型则能有效改善后缘分离模拟。此外，主动射流对边界条件高度敏感。在不可压缩与等熵假设下，研究对比CFL3D中两种亚声速流动边界条件(2008/2018)，发现编号为2018的结果与实验吻合度更高，结合SA + RC模型可将升力预测误差降至3%。进一步针对协同射流的模拟显示，RANS方法因在大分离流动预测上固有局限导致阻力存在偏差，但整体气动趋势预测保持一致。

Abstract: Active flow control (AFC) have garnered significant attention due to their substantial lift enhancement capabilities. However, the complex flow phenomena induced by these methods, such as intense shear layers, high-curvature separation regions, and strong adverse pressure gradients, present substantial challenges for computational fluid dynamics (CFD) simulations. Based on the NASA’s open-source solver CFL3D, this study systematically evaluates simulation strategies for two representative AFC methods (circulation control and co-flow jet) using benchmark airfoils CC020-010EJ and CFJ6415, with particular emphasis on turbulence modeling and boundary condition formulations. For the circulation control: The standard SA model is difficult to accurately capture the streamline deflection caused by the Coanda effect at the trailing edge of the airfoil, resulting in insufficient accuracy for separation prediction. The SA + RC model with curvature correction can effectively improve the simulation of trailing edge separation. In addition, active jets are highly sensitive to boundary conditions. Under the assumptions of incompressibility and isentropic, a comparative study was conducted on two types of subsonic flow boundary conditions in CFL3D (2008/2018). It was found that the result of 2018 had a higher degree of agreement with the experimental data, and combined with the SA + RC model, the lift coefficient only differs by 3%. The simulation of co-flow jet shows that although the RANS method has inherent limitations in large separation flow, resulting the deviations in drag coefficient, the overall aerodynamic trend prediction remains consistent.