WMLES Simulations for the 4th High-Lift Prediction Workshop using BCFD

Abstract

Wall Modeled Large Eddy Simulation (WMLES) results of the NASA High-Lift Common Research Model are presented at Rec = 5.49 million. Free air simulations are carried out using an ODE-based WMLES capability implemented in the BCFD solver at angles of attack corresponding to Case 2a of AIAA High-Lift Prediction Workshop-4. Overall, WMLES results show reasonably good agreement of aerodynamics coefficients with experimental data demonstrating its improved predictive capabilities compared to RANS at higher angles of attack. The unphysical outboard wing separations seen in RANS solutions are not seen in WMLES solutions. The present free air WMLES simulations under-predict lift coefficient at higher angles of attack and misses the pitch break observed in experimental data post-stall. Flow visualization shows differences in the flow features leading up to stall compared to experimental observation. No significant changes are seen in the aerodynamic coefficients and flow features with a refined grid, potentially indicating the need to include the tunnel wall in the simulation to mimic the experiments. An initial assessment of non-equilibrium wall model effects shows no significant difference compared to an equilibrium wall model solution at a post-stall angle of attack. For an implicit solver, time step and number of iteration per time step play a major role in overall throughput. In this study, it is seen that a time step much larger than the convection rate across a typical cell has an adverse effect on accuracy and with an appropriate time step, 4-orders of magnitude residual drop per time step provides sufficient convergence for an efficient balance of throughput and accuracy for the WMLES simulation.