Statistical approach for determining parameters of a turbulence model

In this study, a data assimilation technique was applied to determine the values of parameters for a turbulence model, and its effectiveness was investigated. The turbulence model is a system model required in fluid simulation in the engineering field. Almost all turbulence models include parameters to represent turbulence phenomena. To date, however, the values of the parameters were “calibrated” by the model proposers themselves considering some simple assumptions or to correspond to representative turbulent flows, such as the flat plate boundary layer, wake, and mixing layer. This study showed that the values could be determined objectively or statistically by data assimilation techniques and not by subjective “calibration” as performed by the model proposers themselves.