The Development of Design Tools for Optimized Next-Generation Propulsion Systems: High-Fidelity Direct Numerical Simulations of Rotating Detonation Engines

Abstract

To achieve more efficiency and sustainability in engines of the future, step changes in design leading to enhanced combustion efficiency, lower emissions, and greater power output are essential. Recent research within the combustion community has shown that detonative combustion offers a promising solution, providing numerous benefits over deflagrative combustion used in traditional gas turbine engines. A specific propulsion system of interest that leverages detonative combustion is the rotating detonation engine (RDE), which involves a detonation wave propagating around an annular chamber consuming and burning fuel/oxidizer mixture for energy release. The unique flow structure seen in RDEs leads to problems in designing them, as large fluctuations in pressure and temperature must be considered, as well as fuel/oxidizer mixing effects and other non-idealities present in the system. The goal of this dissertation is to provide a detailed examination of RDEs using high-fidelity simulations for the purpose of 1) identifying critical metrics of interest in RDE research that can garner significant benefit to the design of RDE systems, 2) furthering the fundamental understanding of flow structures within RDEs through the analysis of such metrics, and 3) demonstrating a range of tools and methodologies for use in extracting valuable information from RDE simulations.

The simulations employed in this work were analyzed to gain insight into the highly coupled flow mechanisms within RDEs. The first key metric explored in hydrogen-air RDEs was NOx emissions, with a strong connection found between NOx formation and wave strength. Specifically, average percentage increases in relative NOx across the wave for three pre-detonation temperature conditions (16,721%, 2,839%, and 417%) were proportional to percentage increases in pressure (648%, 261%, and 28%). Additionally, absolute NOx levels were correlated to pre-detonation temperature, as increasing the mass flow rate from the lowest to highest case reduced average residence time in the combustor (0.305 ms to 0.198 ms), which reduced the percentage of high-temperature particles before detonation wave contact (38.60% to 11.69%) and reduced NOx levels accordingly. Furthermore, Lagrangian Particle Tracking (LPT) was implemented into the full-scale RDE simulations to understand the link between physical and thermodynamic behavior through flow processes, and this methodology was validated against Eulerian velocity data with reasonable results in RMS errors (< 11%). Thermodynamic cycle plots were constructed from Lagrangian data for multiple injection locations relative to the detonation wave. However, even the most ideal location for realizable pressure gain, observed to be 26° ahead of the wave, failed to achieve ideal detonative behavior represented by a reduced-order model. Finally, injector response was evaluated within RDEs using the discharge coefficient, with significant increases seen in values for fuel and oxidizer (12% and 22%) when changing injector geometry from radial to axial airflow. CD values also successfully informed a reduced-order model, and output parameters between the model and high-fidelity simulation showed differences within 3%.

Overall, this work collectively advances the capabilities available for analyzing RDEs from a system-level perspective and applying that to the design of more stable, efficient RDEs with key performance metrics in mind. High-fidelity simulations can ultimately provide tools for the better design and optimization of RDEs and further bridge the gap between experiments and low-fidelity models. This research has significant implications for designing future detonation engine systems that seek to alleviate current bottlenecks in propulsive performance through radical changes in combustion technologies.