REDUCED LARGE EDDY FLAME SIMULATION, DIMENSIONAL SURROGATE APPROACH

Our long-term aim is to enable detailed simulations of turbulent reactive flows using reduced models, making these simulations feasible on personal computers. This will provide deeper insights into the structure of complex turbulent flames, with the goal of improving combustion efficiency and reducing pollution. In this work, we seek to leverage AI and dimensionality reduction techniques to develop a fast and accurate method for describing reactive flows in a highly efficient manner. Our central hypothesis is that thermochemical dynamics rapidly relax onto a low-dimensional attractive manifold. This hypothesis is grounded in the success of our prior low-dimensional simulations of one-dimensional flame simulations. The rationale behind our approach is that these simulations demonstrated the effectiveness of a reduced manifold-based method for capturing reactive flows, suggesting its potential as a promising alternative to traditional moment-based techniques. This study aims to achieve the following objectives: i) Efficiently map thermochemical scalars into PC manifolds, ii) Construct PC transport equation’s transport terms for turbulent reacting flows, and iii) Model transport of PCs in the context of turbulent reacting flows.