Measurement and analysis of combustion response to transverse combustion instability

Brian R Pomeroy, Purdue University

Abstract

This research aimed to gain a better understanding of the response of a gas-centered swirl coaxial injector to transverse combustion instability. The goals of the research were to develop a combustion chamber that would be able to spontaneously produce transverse combustion instability at elevated pressures and temperatures. Methods were also developed to analyze high-speed video images to understand the response of the injector. A combustion chamber was designed that produced high levels of instabilities. The chamber was capable of pressures as high as 1034 kPa (150 psi) and operated using decomposed 90% hydrogen peroxide and JP-8. The chamber used an array of seven gas-centered swirl coaxial injectors that exhibited linear instability to drive the transverse oscillations. The injector elements would operate in a monopropellant configuration flowing only decomposed hydrogen peroxide or in a bipropellant configuration. The location of the bipropellant injectors could be varied to change the level of the instability in the chamber from 10% of the chamber pressure up to 70% of the chamber pressure. A study element was placed in the center of the chamber where it was observed simultaneously by two high-speed video cameras which recorded a backlit video to show the location of the fuel spray and the location of the emitted CH* chemiluminescence. The videos were synchronized with high frequency pressure measurements to gain a full understanding of the physics in the combustion chamber. Results showed that the study element was coupled with the first mode velocity wave. This was expected due to the first mode velocity anti-node being located in the center of the chamber. The velocity is an absolute maximum twice during each cycle so the coupling with the second mode pressure was also investigated showing a possible coupling with both the velocity and pressure. The results of the first mode velocity showed that, as the velocity wave traveled through the chamber, the fuel spray was first displaced into an oxidizer rich region and secondly followed by a reaction in the direction of travel of the velocity wave as the peak velocity traveled through the region. The deflection into the oxidizer rich region was especially apparent in high-level instabilities. In low-level instabilities, the velocity wave was not strong enough to fully displace the fuel, and instead the oxidizer core was deflected into the fuel annulus causing a reaction in the direction of travel of the velocity wave. Neighboring oxidizer only injectors caused a lower reaction upstream as the neighboring oxidizer was deflected into the fuel annulus. The region of the fluctuating emitted light agreed well in size, shape and location with a correlation between the first mode velocity and combustion leading to the conclusion that the first mode is highly coupled with velocity. The second mode variance did not agree well with either the velocity or pressure correlation leading to a conclusion that it is coupled with both velocity and pressure. When comparing the variance to the pressure or velocity correlation, parts of the variance compared in shape and location to the pressure or velocity correlation, however, this was not true for all regions of response. This leads to a conclusion that both the pressure and velocity can be affecting the second mode. The second mode chemiluminescence emission occurs when the velocity is nearly zero in the chamber leading to the reaction to not be deflected and occurring downstream of the injector. At the same time, the second mode pressure is a minimum so an increase in mass flow could be responsible for the increased reaction. The methods and combustion chamber used to study the response of an injector can be used in the future to study any injector or combination of injectors placed at various locations in the chamber to study pressure or velocity coupling. The chemiluminescence data can be used to develop transfer functions for use in low fidelity computational models and can be used to validate high fidelity CFD.

Degree

Ph.D.

Advisors

Anderson, Purdue University.

Subject Area

Aerospace engineering

COinS