10.25394/PGS.11866974.v1 Kota Mikoshiba Kota Mikoshiba NUMERICAL STUDY OF HIGH-PRESSURE ROTATING DETONATION ENGINES Purdue University Graduate School 2020 Rotating Detonation Engine CFD analysis Pressure Gain Combustion Reacting flows injector dynamics Aerospace Engineering 2020-03-03 19:36:50 Thesis https://hammer.purdue.edu/articles/thesis/NUMERICAL_STUDY_OF_HIGH-PRESSURE_ROTATING_DETONATION_ENGINES/11866974 <div>The potential higher performance of rotating detonation engines (RDEs), due to the constant volume combustion process, has attracted researchers and engineers from around the world. However, additional research is necessary to achieve higher performance because the dynamic nature of RDEs. </div><div><br></div><div>While a few experiments have been conducted for the injector dynamics for RDEs, this parameter plays a key role relative to the development of high performance RDEs. A parametric study of three dimensional computational fluid dynamics (CFD) simulations is performed to investigate dynamics of an oxidizer injector utilized in an RDE. Input parameters of this parametric study are the simulated detonation strength at oxidizer injector throat (i.e. inlet of a combustion chamber), the injector geometry and number of wave. </div><div><br></div><div>The injector dynamics potentially change local equivalence ratio before detonation wave arrival in an RDE chamber. Also, the injector spacing plays a large role in determining local mixing efficiency. From these point of views, a two dimensional non-premixed detonation parametric study is performed in order to investigate effects of injector mixing and mixing efficiency. An unwrapped domain in the azimuthal direction of an RDE chamber is considered and the mixing efficiencies and injector spacings are modeled as initial conditions.</div><div><br></div><div>In general, from the injector dynamics study, all the results show a strong attenuation of the detonation overpressures in the near field of the injector exit and as a result, the wave traversing the annular passage is largely acoustic in nature. With the mass flow inlet boundary condition employed, reflections are present at this boundary and reflected waves have a non-negligible contribution to the transient mass flow of the injector. These mass flow pulsations could conceivably contribute to the formation of additional waves depending on their energy content and subsequent detailed mixing and combustion. Increasing the number of waves or shortening the length of the injector created reflections as one would predict from acoustic behavior. In the presence of higher amplitude waves these effects will likely be more pronounced. Together with the non-linearities in the heat release, the small fluctuations in the mass flow can significantly alter the detonation behavior. </div><div><br></div><div>The two dimensional non-premixed detonation parametric study further examines the effects of the potential non-uniform mixture due to the injector dynamics exposed to a single planer detonation wave. In all non-premixed cases, the detonation wave is decoupled with the pressure wave (the shock) and the combustion wave once the detonation wave arrives at the non-premixed target region. However, in the cases with the 45 mm and smaller injector spacings, the shock and combustion wave are recoupled. This decouple-explosion-recouple sequence becomes smoother and happens earlier with finer injector spacing. There are some higher pressure pockets than the CJ and the premixed case. The poor mixing efficiency cases show the similar decouple-explosion-recouple sequence.</div><div><br></div><div>The local high pressure in the detonated non-premixed and poor mixing cases is caused by the compression by the shock and other compression from the upstream high pressure pockets. Time scale separation of exothermic and endothermic reactions due to the non and poor mixing efficiencies allows the shock compression closer to the the classical Zel'dovich, Neumann, and D{\"o}ring (ZND) model : the shock is the only pressurization mechanism the poor and non-mixing cases where as the combustion starts at the same time as the shock arrives in the premixed and good mixing efficiencies cases such as the baseline and Gaillard mixing efficiencies. </div><div><br></div><div>The non-premixed and poor mixing efficiency cases with the 45 mm or finer injector spacings provide the higher time and space averaged pressure at the injector surface than it of the premixed case. The poor mixing profile with 11.3 mm injector spacing is the best performer among the all cases in this part of parametric study with respect to pressure thrust. It may be possible to attain a desired axial pressure profile through injector design for specific mixing profile. </div>