Modeling the Surface Pressure Spectrum Beneath Turbulent Boundary Layers in Pressure Gradients

A wide variety of models and methods for the prediction of the surface pressure spectrum beneath turbulent boundary layers is presented and assessed. A thorough review is made of the current state of the art in empirical and analytical pressure spectrum models; and predictions of zero, adverse, favorable, and nonequilibrium pressure gradient boundary layers are examined using a steady Reynolds-averaged Navier-Stokes (RANS) prediction of a subset of a pressure gradient boundary-layer benchmark flow case. The existing empirical models show either an inability to adapt to pressure gradient conditions or an oversensitivity to model inputs, producing nonphysical results under certain flow conditions. New empirical models are created using a gene expression programming machine-learning algorithm based on both experimental and RANS inputs. The various input options for analytical Toegepast Natuurwetenschappelijk Onderzoek (TNO) modeling are presented and assessed, and recommendations for best practices are made. The developed models show improvement in both accuracy and robustness over existing models.

Automated summary

This paper presents and evaluates a wide range of models and methods for predicting the surface pressure spectrum beneath turbulent boundary layers. The current state of empirical and analytical pressure spectrum models is thoroughly reviewed, and predictions of different types of pressure gradient boundary layers are examined using a steady Reynolds-averaged Navier-Stokes (RANS) prediction. Existing empirical models show limitations in adapting to pressure gradient conditions or being overly sensitive to inputs. New empirical models are developed using a gene expression programming machine-learning algorithm based on experimental and RANS inputs. The accuracy and robustness of the developed models are improved compared to existing models.