압축성 유동에 대한 GEKO 모델 계수의 불확실성 정량화 연구

Abstract

In the present work, supersonic flows over an axisymmetric base and 24-deg compression ramp are investigated using Generalized k-ω (GEKO) model included in commercial package of ANSYS FLUENT. GEKO is a two-equation model, based on the k - ω formulation, and able to be tuned for a variety of flows. Compressibility correction is applied to a turbulence model to improve a pressure level along the base surface. Uncertainty Quantification analysis (UQ) is incorporated to quantify the uncertainty of the model coefficients and to calibrate the coefficients for the base and 24-deg compression ramp flow. Latin Hypercube Sampling (LHS) method is used for sampling input parameters which are independent as a uniform distribution. Metamodel is constructed by using ordinary least-squares (OLS) and least angle regression (LARS) and both algorithms are compared to assess fidelity of models. Affine Invariant Ensemble Algorithm (AIES) is selected to characterize the posterior via Markov Chain Monte Carlo sampling. Through Forward problem, the most influential coefficient among the coefficients of GEKO model is known. Calibrated model coefficients are obtained through Backward problem. The results obtained using the calibrated coefficients by UQ corresponding to each flow show better agreement against available experimental measurements than obtained using default coefficients.