Developing a Model-Consistent Reduced-Dimensionality training approach to quantify and reduce epistemic uncertainty in separated flows

TL;DR

This paper presents a novel data-assimilation-assisted neural network training method that leverages physics-based models and autoencoders to reduce epistemic uncertainty in separated flow predictions.

Contribution

It introduces a model-consistent training approach combining autoencoders and ensemble Kalman methods for improved flow state estimation.

Findings

Effective reduction of epistemic uncertainty demonstrated

Enhanced accuracy in turbulent flow predictions

Robust framework integrating physics-based models and machine learning

Abstract

This proposed work introduces a data-assimilation-assisted approach to train neural networks, aimed at effectively reducing epistemic uncertainty in state estimates of separated flows. This method, referred to as model-consistent training, ensures that input features are derived directly from physics-based models, such as Reynolds Averaged Navier Stokes (RANS) turbulence models, to accurately represent the current state of the flow. Autoencoders have been selected for this task due to their capability to capture essential information from large datasets, making them particularly suitable for handling high-dimensional data with numerous discretization points in both spatial and temporal dimensions. This innovative approach integrates the ensemble Kalman method to enhance the training process, providing a robust framework for improving model accuracy and performance in turbulent flow predictions.