Utilization of artificial neural networks to resolve chemical kinetics in turbulent fine structures of an advanced CFD combustion model
Dr JE Hoffmann
The research entailed the development and testing of a novel chemical kinetics integrator which uses artificial neural networks to predict incremental species mass fraction changes occurring in the turbulent fine structures of a flame. The goal was to reduce the computational resources required to solve advance gas phase turbulent combustion numerical flow simulations. The novel neural network predictive model is developed, implemented and tested on a well published experimental setup, of a fully turbulent piloted jet diffusion flame of methane and air. The novel technique generated results with comparable accuracy to the experimental measurements at a significant computational cost reduction of 250 to 600%.