2020/11/02 by Olga A. Doronina, Doronina, Olga A., Scott M. Murman +3
Computer Science · Decision Sciences · Mathematics · Physics and Astronomy · #FOS: Physical sciences #Fluid Dynamics (physics.flu-dyn) #Gaussian Processes and Bayesian Inference #Markov Chains and Monte Carlo Methods #Model Reduction and Neural Networks #Probabilistic and Robust Engineering Design
paper · pdf · doi:10.48550/arxiv.2011.01231
openalex publication_date 2020/11/02 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
We use approximate Bayesian computation (ABC) to estimate unknown parameter\nvalues, as well as their uncertainties, in Reynolds-averaged Navier-Stokes\n(RANS) simulations of turbulent flows. The ABC method approximates posterior\ndistributions of model parameters, but does not require the direct computation,\nor estimation, of a likelihood function. Compared to full Bayesian analyses,\nABC thus provides a faster and more flexible parameter estimation for complex\nmodels and a wide range of reference data. In this paper, we describe the ABC\napproach, including the use of a calibration step, adaptive proposal, and\nMarkov chain Monte Carlo (MCMC) technique to accelerate the parameter\nestimation, resulting in an improved ABC approach, denoted ABC-IMCMC. As a test\nof the classic ABC rejection algorithm, we estimate parameters in a\nnonequilibrium RANS model using reference data from direct numerical\nsimulations of periodically sheared homogeneous turbulence. We then demonstrate\nthe use of ABC-IMCMC to estimate parameters in the Menter\nshear-stress-transport (SST) model using experimental reference data for an\naxisymmetric transonic bump. We show that the accuracy of the SST model for\nthis test case can be improved using ABC-IMCMC, indicating that ABC-IMCMC is a\npromising method for the calibration of RANS models using a wide range of\nreference data.\n