Malik Hassanaly

Turbulent Mixing and Combustion of Supercritical Jets

Supercritical flows are becoming increasingly relevant to aircraft engines, and have always been integral to rocket motors. More recently, supercritical combustion is being considered for stationary gas turbines in CO2 based cycles. The purpose of this study is to understand the turbulent mixing as the flame structure of fuel/air jets issuing at supercritical conditions. Direct numerical simulations (DNS) of a coflowing CH4/O2/CO2 jet with two different inflow configurations are studied at 200 bar pressure (1) jet with coflow, and (2) jet and annular with coflow. Further, a steady laminar flamelet model is adapted for supercritical conditions and the results of the DNS compared against the flame structure predicted by the flamelet model. It is seen that DNS results are roughly similar to the flamelet results, but behave as a more strained flame as compared to the 1-D results. This suggests that the weak heat release associated with strong dilution broadens the reaction zone, which partially invalidates the 1-D flamelet assumption. In comparing the two inlet configurations, the jet case is shown to have a lower maximum temperature at ∼ 1500K while the annular case has a much higher flame temperature at ∼ 1900K. The jet case is characterized by an attached flame while the annular case has a highly lifted flame with high strain rate mixing downstream that enhances mixing but forms high temperature, locally fuel rich region that produces an order of magnitude higher CO mass fraction than the jet case. These configurations demonstrate the extreme sensitivity of supercritical flames to inflow conditions. In particular, local hot spots that occur due to inadequate dilution present a design issue.

Perturbation Dynamics in Turbulent Flames

A new approach based on dynamical systems theory is introduced in order to study turbulent combustion. Here, instead of the conventional statistical approach to chaotic systems, a perturbation-based dynamical systems approach is used. The main focus is in the characterization of the attractor, which represents the loci of solutions given a set of initial, boundary and operating conditions. In this paper, we discuss the basic formulation, the definition of so-called Lyapunov vectors and exponents, and the numerical approach for computing these quantities. The method is applied to a canonical premixed turbulent flame in a periodic box. The algorithms are developed in the context of low-Mach number direct numerical simulation solvers, which introduces certain unique complications. Tests of numerical convergence and behavior of the Lyapunov quantities are discussed.