Search

Proper orthogonal decomposition and dynamic mode decomposition are evaluated for the study of self-excited longitudinal combustion instabilities in laboratory-scaled single-element gas turbine and rocket combustors. Since each proper orthogonal decomposition mode comprises multiple frequencies, specific modes of the pressure and heat release are not related, which makes the analysis more qualitative and less efficient for identifying physical mechanisms. On the other hand, dynamic mode decomposition analysis generates a global frequency spectrum in which each mode corresponds to a specific discrete frequency so that different dynamics can be correlated. In addition, proper orthogonal decomposition results are found to be inaccurate when only a limited amount of spatial information is provided in contrast with dynamic mode decomposition results, which provide more reliable results. Overall, dynamic mode decomposition analysis proves to be a robust and systematic method that can give consistent interpretati...

Combustion instability in rocket chambers is strongly influenced by acoustic interactions at the boundaries of the configuration. Many CFD simulations employ approximate boundary conditions in order to simplify the geometry but the impact that they have on the solution is not well understood. The present study focuses on the use of detailed (exact) boundary representations and an approximate boundary condition in a given longitudinal mode test chamber. The actual inlet boundary of the injector is comprised of a series of choked slots while the approximate boundary condition is a uniform constant mass flow. Both two and three-dimensional simulations are carried out. Differences in the flowfield are evident in the combustion region away from the inlet, including the size of the recirculation region and location of the peak heat release. The amplitudes of the acoustic modes are well predicted for the first two modes especially in three-dimensional simulation, while higher modes are poorly predicted. These results suggest that such boundary condition approximations must be judiciously used and having access to more detailed treatments is important to verify accuracy.

Three-dimensional simulations of combustion instability can provide extremely valuable insight into the problem physics, but they represent an enormous investment in time and, thus, may preclude parametric studies that are practical with two-dimensional simulations. Three numerical simulations of an unstable laboratory rocket combustor were performed to assess the effect of grid resolution and dimensionality on the ability to predict and model combustion instability. Computations show that two-dimensional axisymmetric simulations using both coarse and fine grids are able to capture the amplitude of the first longitudinal mode within an order of magnitude of what was measured in the experiment. Whereas the three-dimensional calculation based on the coarse grid resolution shows excellent agreement with the experimental level of instability two-dimensional simulations are valuable in parametric evaluations in developing diagnostic methods and in increasing the efficiency of expensive three-dimensional simula...

near the tip of the potential core and downstream. Axial and radial variation in radiation intensity fluctuations is similar to those reported for flames. Autocorrelation coefficients of the radiation intensity are approximated reasonably well by exponential curves. Integral time and length scales increase monotonically downstream of the core region and are consistent with Taylor’s hypothesis. The break frequency and slope of the normalized power spectral density function are comparable to those reported for turbulent jet flames. These findings suggest that reacting flows can be used to predict trends in turbulent radiation properties of exhaust plumes.

A comparison of a single step global reaction and the detailed GRI-Mech 1.2 for combustion instability modeling in a methane-fueled longitudinal-mode rocket combustor was performed. A single element shear-coaxial injector was operated under two different conditions corresponding to marginally stable and unstable operation in order to evaluate the performance of the chemical kinetics mechanisms on combustion stability. Results show improved prediction in the frequencies and amplitudes with the detailed kinetics but the underlying source of the instability phenomena remains the same. In contrast to previous two-dimensional results, these three-dimensional results demonstrate that the present non-premixed injector configuration is primarily mixing-controlled and that the global chemical kinetics are sufficient to capture the stability characteristics.

In this study a series of three-dimensional unsteady reacting flow simulations are used to investigate the effect of swirl on the instability amplitude of a single-element gas-gas rocket combustor. The baseline combustor of interest is unstable because of a fuel cut-off event caused by the high-pressure waves in the combustor. Previous two-dimensional simulations have shown that swirl reduces the amplitude of the pressure oscillations compared with that of the baseline configuration. The current three-dimensional simulations show that swirl is indeed able reduce the amplitude of the instabilities, albeit not to the same extent observed in the two-dimensional simulations. We further observe that the enhanced mixing due to the swirling flow leads to a reduction in the recovery time associated with the fuel cut-off event, thereby allowing the combustor to experience a more continuous heat release. Nevertheless, unlike the two-dimensional case, the three-dimensional simulations show that the flame does not stay anchored to the dump-plane, which explains the higher relative amplitudes in this case.

Computational fluid dynamics simulations are used to study spontaneous combustion instabilities in a rectangular chamber that contains seven coaxial injector elements. The simulation predicts self-excited transverse oscillations that have variable amplitudes, similar to what is seen in a companion experiment. Several processes are identified in the simulation that may promote the rise of instability by increased heat release during a local compression, including the interaction between vortices shed from adjacent injectors and interactions between flames and the combustor walls.

Combustion instability continues to be a challenge in the design of rocket engines. The use of computational fluid dynamics (CFD) simulations screen potential designs offers the ability to reduce the number of costly tests and improve understanding of the underlying instability mechanism. In this study a series of axisymmetric CFD simulations are used to investigate the instability sensitivity to four design changes. The design changes are selected in an attempt to reduce the level of instability. The parameters considered are the combustor wall temperature, the effect of adding swirl to the fuel injector and two geometric changes, namely, fuel injector area reduction and the introduction of a chamfer in the injector face. The simulations show that both the wall temperature and swirl are able to significantly lower the amplitude by 70%. The results of the geometric changes are mixed with both decreases and increases in the instability amplitude. The parametric study has enhanced the understanding of the instability mechanisms by demonstrating that the when a continuous fuel supply to the combustor is maintained the instability amplitude is decreased.

An investigation of the instability mechanism present in a laboratory rocket combustor is performed using computational fluid dynamics (CFD) simulations. Three cases are considered which show different levels of instability experimentally. Computations reveal three main aspects to the instability mechanism, the timing of the pressure pulses, increased mixing due to the baroclinic torque, and the presence of unsteady tribrachial flame. The stable configuration shows that fuel is able to flow into the combustor continuously allowing continuous heat release. The unstable configuration shows that a disruption in the fuel flow into the combustor allows the heat release to move downstream and new fuel to accumulate in the combustor without immediately burning. Once the large amounts of fuel in the combustor burn there is rapid rise in pressure which coincides with the timing of the acoustic wave in the combustor. The two unstable cases show different levels of instability and different reignition mechanism.

Combustion instability in rocket chambers is strongly influenced by acoustic interactions at the boundaries of the configuration. Many CFD simulations employ approximate boundary conditions in order to simplify the geometry but the impact that they have on the solution is not well understood. The present study focuses on the use of detailed (exact) boundary representations and an approximate boundary condition in a given longitudinal mode test chamber. The actual inlet boundary of the injector is comprised of a series of choked slots while the approximate boundary condition is a uniform constant mass flow. Both two and three-dimensional simulations are carried out. Differences in the flowfield are evident in the combustion region away from the inlet, including the size of the recirculation region and location of the peak heat release. The amplitudes of the acoustic modes are well predicted for the first two modes especially in three-dimensional simulation, while higher modes are poorly predicted. These results suggest that such boundary condition approximations must be judiciously used and having access to more detailed treatments is important to verify accuracy.

Concurrent experiments and computations are used to analyze combustion instabilities in a transverse mode combustion chamber. The experiments employ a shear-coaxial injector element, positioned within a rectangular chamber. The reacting flow portion of the study element is optically accessible and the chamber is extensively instrumented with high-frequency pressure transducers. High amplitude transverse acoustics modes are driven by unstable injector elements located near the chamber end-walls. Different levels of instability are obtained by varying the operation of these driving elements. High-fidelity computational fluid dynamics simulations are used to model this set-up, although only the study element is fully represented and the transverse acoustics modes are generated by vibrating the side walls at the appropriate frequencies. The computational results are compared quantitatively with the high frequency pressure measurements, and qualitatively by using the CH* chemiluminescence signal from the experiment. The combustion response of the first, second and third transverse modes obtained using a dynamic modal decomposition procedure show excellent agreement between the experiments and simulations. The overall approach shows significant promise for screening the combustion response of candidate injector configurations for rocket applications.

Combustion instability arises from the coupling between unsteady combustion and acoustic modes in a combustion chamber. A study comprising concurrent experiment and LES simulation of a single element rocket combustor was conducted. The goal was to evaluate the a priori predictive ability of the computational model with regards to self-excited combustion instability, and to use the detailed results from the simulation to interpret the mechanism of self-excitation. Pressure modes and chemiluminescence from CH* were measured at high sampling rates. Direct comparisons between the experimental and computational results were made on the basis of instability frequency, pressure mode shapes, and limit cycle amplitude. All indicated generally good agreement. The light emission representing heat release from the experiment was qualitatively similar to the simulation. The detailed results from the simulation provided much greater insight into the complex phenomena, and showed the likely mechanism that drove high amplitude instability was a periodic ignition in the recirculation zone just downstream of the dump plane. The ignition occurred nearly simultaneously with the arrival of a compression wave traveling from downstream.

Numerical simulations of gaseous methane and decomposed hydrogen peroxide in a laboratory rocket combustor were performed in order to determine how to best calculate the Rayleigh index. Calculations using four levels of spatial re nement show the level of re nement necessary to identify the regions of ampli cation and damping is signi cantly coarser than the mesh used for simulation. The Rayleigh index for the rst longitudinal mode is bimodal and shows an isolated area of damping. The two areas of ampli cation are the result of two di erent driving mechanisms. Mode shape, phase angle, and frequency analysis of pressure, velocity, and heat release are presented to validate the combustion modeling approach and interpret the results.

Turbulent radiation statistics are reported for the core region of three subsonic exhaust plumes. Spectrally integrated radiation inte nsity measurements were acquired using an infrared camera fitted with a narrowband filter (4.34+/-0.1 µm). Unsteady two-dimensional simulations were used to predict time-resolved and average temperature and partial pressure values of the exhaust species in the flow. The radiation intensity along diametric paths was calculated using the scalar values and a narrowband radiation intensity solver. Mean intensity values typically within agreed within 25% or better when using average and time-resolved scalar values. This indicates that tu rbulent fluctuations have little effect on the radiation intensity emitted from the cores of jets. Fluctuations in the measured radiation intensity increased markedly beyond the plume core. This corresponded with an increase in the skewness and kurtosis of the radiation intensit y. These findings indicate that the size of the core of a heated jet might be estimated using turbulent radiation statistics.