Training convolutional neural networks to estimate turbulent sub-grid scale reaction rates
In the combustion community, the determination of the sub-grid scale contribution to the filtered reaction rate in reacting flows Large Eddy Simulation is an example of closure problem that has been daunting for a long time. We propose a new approach for premixed turbulent combustion modeling based on convolutional neural networks by reformulating the problem of subgrid flame surface density estimation as a machine learning task.
In order to train a neural network for this task, a Direct Numerical Simulation and the equivalent LES obtained by a spatial filtering of this DNS (more details can be found here) is needed.
- In a first step, two DNS of a methane-air slot burner are run and then filtered to create the training dataset. Models are trained on this data in a supervised manner.
- In a second step, a new, unseen and more difficult case was used to ensure network capabilities. This third DNS is a short-term transient started from the last field of the second DNS, where inlet velocity is doubled, going from 10 to 20 m/s for 1 ms, and then set back to its original value for 2 more ms.
Description of the dataset
The dataset can be downloaded here. Each of these files corresponds to a time step of a simulation and contains 3 fields :