Investigation of Deep Learning-Based Filtered Density Function for Large Eddy Simulation of Turbulent Scalar Mixing

Shubhangi Bansude,Reza Sheikhi

2023-09-29

Abstract:A filtered density function (FDF) model based on deep neural network (DNN), termed DNN-FDF, is introduced for large eddy simulation (LES) of turbulent flows involving conserved scalar transport. The primary objectives of this study are to develop the DNN-FDF models and evaluate their predictive capability in accounting for various filtered moments, including that of non-linear source terms. A systematic approach is proposed to select DNN training sample size and architecture via learning curves to minimize bias and variance. Two DNN-FDF models are developed, one utilizing FDF data from Direct Numerical Simulations (DNS) of constant-density temporal mixing layer, and the other from zero-dimensional pairwise mixing stirred reactor simulations. The latter is particularly intended for cases where generating DNS data is computationally infeasible. DNN-FDF models are applied for LES of a variable-density temporal mixing layer. The accuracy and consistency of both DNN-FDF models are established by comparing their predicted filtered scalar moments with those of conventional LES, where moment transport equations are directly solved. The DNN-FDF models are shown to outperform a widely used presumed-FDF model, especially for multi-modal FDFs and higher variance values. Results are further assessed against DNS and the transported FDF method. The latter couples LES with Monte Carlo for mixture fraction FDF computation. Most importantly, the study shows that DNN-FDF models can accurately filter highly non-linear functions within variable-density flows, highlighting their potential for turbulent reacting flow simulations. Overall, the DNN-FDF approach is shown to offer an accurate yet computationally economical approach for describing turbulent scalar transport.

Fluid Dynamics,Computational Engineering, Finance, and Science

What problem does this paper attempt to address?

The paper primarily focuses on addressing the issue of filtered density function (FDF) modeling involved in turbulent scalar mixing in large eddy simulations (LES). Specifically, the research objectives include:

1. **Developing a deep neural network (DNN)-based FDF model (DNN-FDF)**: Constructing a DNN to predict the FDF used to describe turbulent mixing in LES, aiming to improve simulation accuracy.

2. **Evaluating the predictive capability of the DNN-FDF model**: Verifying whether the developed DNN-FDF model can accurately represent various filtered statistical moments, especially those related to nonlinear source terms.

3. **Selecting training sample size and network architecture**: Proposing a method based on learning curves to select the DNN training sample size and network structure to minimize bias and variance.

4. **Developing two DNN-FDF models**: One model is trained using direct numerical simulation (DNS) data; the other model is trained using data generated from zero-dimensional paired mixing stirred reactor (PMSR) simulations, aiming to provide an alternative for situations where DNS data cannot be generated.

5. **Application and validation**: Applying the DNN-FDF model to LES of variable density turbulent mixing layers and comparing its predictions with traditional LES methods, widely used presumed FDF models, and methods that directly solve the FDF transport equation.

6. **Performance evaluation**: Demonstrating the superior performance of the DNN-FDF model in cases of multimodal FDFs and high variance values, and proving its capability in filtering highly nonlinear functions.

7. **Potential applications**: Highlighting the potential of the DNN-FDF model in simulating turbulent reactive flows.

In summary, this research introduces a deep learning-based FDF model to provide a new solution for improving the simulation of turbulent scalar mixing in LES, and demonstrates that this approach can offer accurate predictions while maintaining computational efficiency.