TURBULENT COMBUSTION MODELING OF SWIRL STABILIZED BLENDED CH4/H2 FLAMES BY USING FLAMELET GENERATED MANIFOLD

Hydrogen has been identified as one of the key elements of the decarbonization initiatives. The level of maturity with different original equipment manufacturers (OEMs) varies significantly for a 100% H-2 gas turbine combustor. The typical standard short-term goal is to blend hydrogen with existing fuel as a promising alternative to meet regulatory standards for emission. A typical Dry Low NOx (DLN) combustion system can handle a certain level of hydrogen blending. However, due to fundamental differences between the properties of hydrogen and methane, existing designs of combustion systems are not capable of handling moderate to high levels of hydrogen blending. Therefore, prior knowledge of blend ratios that a given combustion system can handle is essential for the system's stable operation. Computational Fluid Dynamic (CFD) simulation can help study the effect of different blend ratios on flame stability, peak temperature, pollutants, etc., without affecting the hardware. Thus, helping in reducing the overall cost and time spent deciding the allowable blend ratios. In this work, the accuracy and consistency of Flamelet Generated Manifold (FGM) with Large Eddy Simulation (LES) have been assessed to model swirling turbulent combustion of CH4/H-2 blends for gas turbine engine combustors. FGM characterizes the extent of reaction using a reaction progress variable typically defined as a weighted sum of some representative product species of hydrocarbon combustion like CO and CO2. With H-2 blending, the mixture now has multiple heat release time scales, and the prevailing choices of reaction progress definition are not optimal. Therefore, the first and foremost task is to correctly describe the reaction rate by choosing a reaction progress variable with validity over a range of H-2 blending ratios and equivalence ratios. Additionally, the variation in the laminar properties of the blended mixture, e.g., thermal conductivity and viscosity, is enhanced when H-2 is added to the fuel. In this work, we have used kinetic theory to compute these properties accurately as a function of temperature and composition. The flame configurations used to validate FGM in this work are CH4/H-2 swirl flame (SMH1) and HM3e. The burner designs belong to a detailed and widely simulated database from Sydney Swirl Burner, with a CH4/H-2 blend ratio of 1: 1 (by volume). The FGM generates flamelets from opposed flow diffusion flames and freely propagates premix flame configuration. The solution of both the FGM approaches is compared with Finite Rate detailed chemistry solution, and definitive advantages/disadvantages of each approach are identified based on computational speed and accuracy. The results are then compared with experimental data for velocity, temperature, major and minor species distribution to establish the computational accuracy of each approach. Together with the inclusion of modifications in the modeling framework and usage of detailed chemistry with FGM-LES, these results provide important insights into the simulation of hydrogen-blended methane flames.