Study of Fuel-Oxygen Mixing in a Cold Single Shot Linear Rotating Detonation Engine

Abstract

An experimental and numerical study was undertaken to examine the detonation behaviour and the effect of mixing in a Cold Single Shot Linear Rotating Detonation Engine. Observations of the experimental work were performed using schlieren photography as well as stereo soot impressions. A planar detonation wave was developed by spark ignition of a premixed hydrogen oxygen mixture upstream of a horizontal sliding door which separated an inert gas from the premixed combustible mixture. The detonation was then sustained through an inert gas (Ar) by way of a hydrogen and oxygen layer injected in the optical section. This layer was created via separate feed lines which fed each gas into a manifold which consisted of a linear array of injection holes to inject the gas into the channel. The height of the injected layer was controllable through a delay generator which applied a preset delay to the solenoid valves controlling the hydrogen and oxygen injection into the manifold. The width of the optical section was variable between 12, 10 and 8 mm to evaluate the effect the width of the channel had on the resulting detonation. Reducing the channel width was observed to improve the detonation indicated by increased average velocity, and improved steadiness which was observed through the cellular structure left by the detonation. Tests performed in the narrower channel were also able to successfully propagate through the optical section using shorter injection times, meaning decreased height and reduced mixing time of the injected layer. A numerical mixing study of the injection process was also performed using ANSYS Fluent 2020 R2. This computational fluid dynamic (CFD) simulation modelled the gas injection process from the moment the valves opened to the arrival of the detonation into the optical section. This calculation was performed to characterize the injection process which takes place in the experimental tests and to verify and evaluate proposed modifications to the original experimental apparatus and more specifically the injection scheme. The results from these simulations corroborated the results that were observed in the experimental testing that the current injection geometry produces poor mixing in the injected layer and leads to significant argon dilution. Numerical results with the narrower channel showed significantly improved mixing which confirmed the observations that were made during the experimental testing of the narrower channel.