Modeling subgrid-scale scalar dissipation rate in turbulent premixed flames using gene expression programming and deep artificial neural networks

In this present study, gene expression programing (GEP) has been used for training a model the subgrid scale (SGS) scalar dissipation rate (SDR) large range of filter widths, using database statistically planar turbulent premixed flames, featuring different turbulence intensities and heat release parameters. GEP is based on idea to iteratively improve population candidates survival-of-the-fittest concept. The resulting mathematical that can be easily implemented, shared with community, analyzed physical consistency, as illustrated in work. Efficient evaluation cost function smart choice basis functions have found essential successful optimization process. outperform an existing algebraic from literature. However, process was quite intricate, SGS SDR closure turned out difficult. Some these problems explained model-agnostic interpretation method, which requires existence trained artificial neural network (ANN). ANNs are known their ability represent complex functional relationships serve additional benchmark solution model.