Large Eddy Simulation Techniques for Predicting Turbulent Combustion of Hydrogen Fuel

Abstract

The concept of a hydrogen economy has reified in recent times and many envision hydrogen to become the fuel of choice in the not-too-distant future. Whether it will be used for power generation, domestic heating, or vehicular propulsion, the benefits of hydrogen are well documented and highly desirable. This thesis investigates numerical methods currently used to study the turbulent combustion of conventional hydrocarbon- based fuels and how a similar framework can be applied to the special case of hydrogen combustion. A discussion on the importance of seeking alternatives to current fuel sources is given. The case for adopting hydrogen as the preferred alternative is outlined, along with the challenges that must be overcome if we are to do so. The theoretical foundations of combustion theory presented, includes the relevant fundamental concepts from the fields of thermodynamics, chemical kinetics and turbulent flow. The procedure for constructing large eddy simulations for turbulent combustion systems is described. This procedure is used to build a model for propane combustion which is validated against experimental data. In the absence of empirical validation data for hydrogen combustion, techniques are instead devised to adapt hydrogen into the working model for propane combustion. Certain quantities related to the flame speed and heat release of the propane case are used to estimate a specific setup configuration to be used for the hydrogen case. The goal is to configure the hydrogen case in such a way that the system will exhibit flame speed and heat release characteristics that are comparable the corresponding flame speed and heat release characteristics of the propane case. By designing the new hydrogen model in such a way that the flame speed and heat release characteristics are equivalent to those of the trusted propane model, a point of reference is available for which the two different systems can be compared and contrasted. The results obtained from each combustion model are presented followed by an evaluation of modelling strategy.