Feature selection and processing of turbulence modeling based on an artificial neural network

TL;DR

This paper improves turbulence modeling by selecting and processing features for neural networks, enhancing prediction accuracy and smoothness in complex flow scenarios.

Contribution

It introduces a novel feature selection and processing method, including a modified spatial orientation feature decomposition, to improve neural network-based turbulence predictions.

Findings

Enhanced prediction accuracy of turbulence models.

Improved smoothness and detail in flow separation predictions.

Validated on complex geometries with significant results.

Abstract

Data-driven turbulence modeling has been considered an effective method for improving the prediction accuracy of Reynolds-averaged Navier-Stokes equations. Related studies aimed to solve the discrepancy of traditional turbulence modeling by acquiring specific patterns from high-fidelity data through machine learning methods, such as artificial neural networks. The present study focuses on the unsmoothness and prediction error problems from the aspect of feature selection and processing. The selection criteria for the input features are summarized, and an effective input set is constructed. The effect of the computation grid on the smoothness is studied. A modified feature decomposition method for the spatial orientation feature of the Reynolds stress is proposed. The improved machine learning framework is then applied to the periodic hill database with notably varying geometries. The results of the modified method show significant enhancement in the prediction accuracy and smoothness, including the shape and size of separation areas and the friction and pressure distributions on the wall, which confirms the validity of the approach.