Modeling combustion systems using a chemical reactor network methodology

Computational fluid dynamics (CFD) has become a powerful tool for analysis and design but significant computational time is required to simulate reacting flows with detailed chemistry. A chemical reactor network (CRN) can greatly reduce the computational burden yet provide critical relationships for the chemical mechanisms and flow paths in a complex reacting-flow system. By decomposing a combustor into several small reactors that are linked to species diffusion, it is possible to achieve more accurate predictions of NO_xusing detailed reaction mechanisms and less computational time. In this work, a novel method for connecting CFD and CRN is proposed to capture temperature, species as well as pollutant emission with a higher level of accuracy. A CRN is developed based on the CFD predictions of thermochemical characteristics, such as flame patterns, gas entrainment and flow in main recirculation zone of the flame. Species and pollutants such as NO_xhave been predicted in the CRN model using the detailed chemical kinetic reaction database of GRI-Mech 3.0. By calculating a few zones using a complex network rather than thousands or millions of nodes using CFD, temperatures and NO_xemissions can be predicted efficiently with relative errors as low as 5%. An updated flame structure and reaction pathway using CRN to relate sub-regions in the combustor will be presented to provide a better understanding of reacting flows.