Generalization Capability of Convolutional Neural Networks for Progress Variable Variance and Reaction Rate Subgrid-Scale Modeling

Abstract

Deep learning has recently emerged as a successful approach to produce accurate subgrid-scale (SGS) models for Large Eddy Simulations (LES) in combustion. However, the ability of these models to generalize to configurations far from their training distribution is still mainly unexplored, thus impeding their application to practical configurations. In this work, a convolutional neural network (CNN) model for the progress-variable SGS variance field is trained on a canonical premixed turbulent flame and evaluated a priori on a significantly more complex slot burner jet flame. Despite the extensive differences between the two configurations, the CNN generalizes well and outperforms existing algebraic models. Conditions for this successful generalization are discussed, including the effect of the filter size and flame–turbulence interaction parameters. The CNN is then integrated into an analytical reaction rate closure relying on a single-step chemical source term formulation and a presumed beta PDF (probability density function) approach. The proposed closure is able to accurately recover filtered reaction rate values on both training and generalization flames.

Investigation of the generalization capability of a generative adversarial network for large eddy simulation of turbulent premixed reacting flows

“…Among the four configurations, two different ratios are essentially conserved: the ∆{η « 8 and ∆{δL « 1.45. Those might then represent ratios fundamental to achieve extrapolation capabilities, as it was previously found when predicting the progress variable variance [19]. The relative importance of these two parameters on the predictive capability is then analyzed on the following sections.…”

“…This might cause convergence issues in a-posteriori applications, as suggested by Lapeyre et al [17]. Attili et al [18] and Xing et al [19], using similar architectures, performed some preliminary a-priori generalization studies. Their work demonstrated that the CNNs can be trained on canonical simple cases and applied to practical configurations indicating good extrapolation capabilities.…”