Accuracy improvement of second-order central scheme using smart post-process filter for hybrid RANS/LES method

The present research aims to propose a new approach to improve the performance of computational codes employing the RANS/LES technique with 2nd order central scheme. To accomplish this, an unsteady compressible 3D coupled RANS/LES solver using two-equation turbulence model is developed. A smart post-processing filter is also embedded as the dissipation term in the code to maintain both the stability and the accuracy. A predesign sensor that detects the amount and position of non-physical fluctuations in numerical solution, controls the smart post-processing filter. Detailed study is performed to investigate the performance of this sensor and filter. The methodology is validated against several experimental observations and numerical findings for compressible high-Reynolds flow regimes. The unsteady 3D flow over a wing with NACA 0015 section at high angle of attack is computed to evaluate the model performance. The results indicate that the proposed methodology shows promising accuracy for both steady and unsteady flow simulations, especially in high angles of attack, in which the accuracy of the other schemes is normally decreased considering the flow separation phenomena.