Generalization of Machine-Learned Turbulent Heat Flux Models Applied to Film Cooling Flows

Abstract

The design of film cooling systems relies heavily on Reynolds-averaged Navier–Stokes (RANS) simulations, which solve for mean quantities and model all turbulent scales. Most turbulent heat flux models, which are based on isotropic diffusion with a fixed turbulent Prandtl number (Prt), fail to accurately predict heat transfer in film cooling flows. In the present work, machine learning models are trained to predict a non-uniform Prt field using various datasets as training sets. The ability of these models to generalize beyond the flows on which they were trained is explored. Furthermore, visualization techniques are employed to compare distinct datasets and to help explain the cross-validation results.

Turbulent scalar flux in inclined jets in crossflow: counter gradient transport and deep learning modelling

“…(2018) and Milani et al. (2020). Then, , and in (4.4) and (4.6) are non-dimensionalized using local values of and before being fed as inputs to the network.…”

“…So, given the mean LES velocity and pressure fields, the transport equations for k and are solved, which also generates ν t as a by-product. This approach to non-dimensionalization has been used by others such as Sandberg et al (2018) and Milani et al (2020). Then, S, R and ∇c in (4.4) and (4.6) are non-dimensionalized using local values of k and before being fed as inputs to the network.…”

“…Sandberg et al (2018) used a genetic optimization algorithm to generate data-driven analytical expressions for turbulent anisotropy and turbulent diffusivity, and Maulik et al (2019) explored the use of a neural network towards subgrid scale modelling in large eddy simulations. The bulk of the aforementioned literature focuses on Reynolds stress modelling, but some authors have had success in applying machine learning techniques for closure of the passive scalar equation Sandberg et al 2018;Sotgiu, Weigand & Semmler 2018;Milani, Ling & Eaton 2020;Weatheritt et al 2020). In the present paper a different scalar closure model is proposed, leveraging the tensor basis neural network concept of Ling et al (2016a) to generate a more general model form.…”

“…2018; Sandberg et al. 2018; Sotgiu, Weigand & Semmler 2018; Milani, Ling & Eaton 2020; Weatheritt et al. 2020).…”

Challenges and opportunities for artificial intelligence and high-fidelity simulations in turbomachinery applications: A perspective

“…Because turbomachinery design often involves many simulations of related flows, it is important to use visualization and interpretability approaches such as t-SNE to understand when data-driven models are extrapolating and when they can be used with confidence. Milani et al (2020) adapted the t-SNE approach to film cooling flows, using this visualization approach to assess the generalization of his scalar flux closure model to new film cooling geometries and blowing ratios. They also investigated the use of spatially-varying feature importance metrics to shed light on which closure terms were important in which regions of the flow.…”