Parameter uncertainty quantification for a four-equation transition model using a data assimilation approach

Abstract

A majority of the simulation discrepancies for RANS models are due to parameter uncertainties. In this paper, a data assimilation approach called the ensemble Kalman filtering (EnKF), was used to investigate the parameter uncertainties of a four-equation k-ω-γ-A r transition/turbulence integrated model. The transitional flows past two wind-turbine airfoils were analyzed, while the transition onset and end locations, as well as the skin friction coefficients, were taken as the observation variables. The results show that the posterior distributions of the parameters can be efficiently obtained through the filtering process, while the estimated parameter variations are observed among different angles of attack (AoAs), as well as different sides at the same AoA. After the EnKF analysis, the most sensitive parameter, C 2 , is changed into an adaptive parameter, which demonstrates the pressure gradients effects through a modified shape factor. More accurate transition locations are calculated by the adaptive C 2 at different AoAs for NLF(1)-0416, and on both sides for NACA-0012 than the original model having constant C 2 . • The EnKF is used to investigate the parameter uncertainties of a transition model. • The estimated parameter variations are observed among different cases. • The most sensitive parameter is modified based on the boundary layer shape factor. • The modified model demonstrates more accurate transition prediction.