Mechanical and Civil Engineering Seminar: PhD Thesis Defense

Abstract:

Hydrogen has gained popularity as an alternative source of clean energy generation due to its high specific energy density and lack of carbon and soot emissions. To reduce NOx emissions, hydrogen is typically burnt lean. However, lean hydrogen flames have a propensity to develop thermodiffusive instabilities due to the low Lewis number (ratio of thermal to mass diffusivity) of hydrogen, which introduces challenges for predictive design and modelling. In the presence of turbulence, lean premixed hydrogen/air flames have substantially increased flame speeds, commonly attributed to differential diffusion effects. In this thesis, the relationships between turbulence, chemistry, and modelling are studied through Direct Numerical Simulation (DNS) and Large Eddy Simulation (LES).

The effect of turbulence on lean hydrogen combustion is studied through DNS using detailed chemistry and detailed transport. Simulations are conducted at six Karlovitz numbers and four integral length scales. The evolution of local properties, such as the flame structure, and global properties, such as the flame area and flame speed, are quantified, and the results are summarized in a newly proposed expression for the burning efficiency. This expression depends on the conditional mean chemical source term and gradient of a progress variable, and the percentage of superadiabatic burning. Neglecting Soret diffusion (mass diffusion due to temperature gradient) is shown to reduce the flame speed, area, and burning efficiency. The changes in the flame structure with increasing Karlovitz number have previously been quantified via an effective Lewis number model. In this work, we also assess the impacts of Soret diffusion and integral length scale on this model.

To study the LES modelling of lean hydrogen flames, we simulate a Low-Swirl Burner. The LES modelling of these flows remains challenging because the transition of small-scale instabilities into large-scale turbulent structures cannot be modelled by conventional strategies. Traditional one-equation tabulated chemistry formulations require only a progress variable and cannot capture differential diffusion and curvature effects. In this work, we study the effects of tabulating different conditional mean source terms. It is shown that tabulating the appropriate conditional mean source term leads to improvements in the flow field prediction, however, key features such as the main recirculation region are not reproduced. Then, a two-equation tabulated chemistry model which accounts for differential diffusion and curvature effects is tested. This model provides the best agreement with experimental results. The work is a first effort in evaluating the performance of the two-equation model in the LES framework.