RotatingDetonationEngineCFD

Overview

This dissertation investigated the numerical modelling of rocket-mode rotating detonation engines for space propulsion applications. Rotating detonation engines are pressure-gain combustion systems where one or more detonation waves continuously travel around an annular combustor.

The aim was to determine whether a computationally constrained ANSYS Fluent model could reproduce the key qualitative features of detonation behaviour in a simplified rocket-mode RDE. The study used a stoichiometric hydrogen-oxygen mixture, NASA CEA Chapman-Jouguet benchmarking, a one-dimensional shock tube validation model and a three-dimensional annular combustor geometry based on the Air Force Research Laboratory RDE configuration.

The final model successfully generated an initial detonation wave in the annular combustor, with pressure levels comparable to theoretical CJ values. Sustained rotating detonation was not achieved because backflow across the inlet boundary prevented continued reactant injection.

Background

Rotating detonation engines use detonation combustion rather than conventional deflagration combustion. In a conventional rocket engine, propellants burn through a process that approximates constant-pressure heat addition. In an RDE, the detonation wave produces a pressure rise inside the chamber, creating a pressure-gain combustion process.

This makes RDEs attractive for space propulsion because higher combustion efficiency and reduced upstream pressure requirements could reduce engine mass, turbomachinery complexity and overall propulsion system size.

Methodology

The methodology was structured around progressively validating the numerical model before attempting the full 3D combustor simulation.

NASA CEA Benchmarking

NASA CEA was used to calculate theoretical Chapman-Jouguet detonation properties for a stoichiometric hydrogen-oxygen mixture. These values provided a benchmark for the Fluent simulations.

• CJ pressure: 20.4 MPa
• CJ temperature: 4137.8 K
• Detonation velocity: 2958.8 m/s
• Pressure ratio: 20.159

One-Dimensional Shock Tube Validation

A one-dimensional shock tube tested whether the solver, chemistry model and timestep settings could generate and propagate a detonation wave before moving to the full annular combustor.

Mesh Sensitivity Study

Mesh sizes of 0.5 mm, 1.0 mm, 1.5 mm and 2.0 mm were tested. A 2 mm mesh was selected because it produced similar wave-speed accuracy to finer meshes while significantly reducing the number of required timesteps.

CFL Study

A CFL value of 0.4 was selected because it remained within the literature-recommended range and gave consistent wave-speed behaviour without relying on pressure results that appeared to be numerical artefacts.

Three-Dimensional RDE Simulation

The final 3D model used an AFRL-based annular combustor geometry with outer diameter 154 mm, inner diameter 139 mm, combustor length 100 mm, a premixed hydrogen-oxygen inlet, pressure outlet and ignition patch of high-pressure, high-temperature combustion products.

Numerical Modelling Workflow

Numerical Setup

The final solver setup used a density-based transient formulation in ANSYS Fluent. A density-based approach was chosen because it is more suitable for compressible shock-dominated flows.

• Density-based solver
• First-order implicit transient formulation
• Inviscid flow model
• Species transport with finite-rate chemistry
• CHEMKIN hydrogen oxidation mechanism
• Roe Flux Difference Splitting scheme
• Second-order upwind spatial discretisation
• CFL number: 0.4
• Premixed mass-flow inlet controlled by a pressure-dependent UDF
• Pressure outlet at 1 atm
• Initial nitrogen-filled domain
• Ignition patch at 50 atm and 3442 K

Results

The one-dimensional shock tube simulations confirmed that the numerical setup was capable of producing and propagating a detonation wave. With the selected 2 mm mesh and CFL value of 0.4, the shock tube produced a peak pressure of approximately 16.1 MPa and a wave speed of 2469 m/s.

The 3D annular combustor simulation successfully generated an initial high-pressure detonation wave after ignition. The peak pressure reached approximately 24.3 MPa, which compared favourably with the NASA CEA Chapman-Jouguet pressure of 20.4 MPa.

Sustained rotating detonation was not achieved. The wave decayed after the initial reactant supply was depleted, and inlet analysis showed negative mass flux confirming backflow across the inlet boundary.

Limitations

The model was intentionally computationally constrained. The ANSYS Student license limited computational resources, and the three-dimensional simulation required approximately 128 hours of runtime.

The selected 2 mm mesh reduced computational cost, but detonation modelling literature suggests that sub-millimetre resolution is typically required to resolve detonation reaction zones and shock gradients accurately. The model should therefore be interpreted as a qualitative demonstration of detonation behaviour rather than a quantitatively exact prediction of RDE performance.

Future Work

The next stage would be to improve the inlet UDF and injector boundary conditions so reactant injection can recover after the detonation wave passes. Additional simulations should investigate ignition patch size, ignition location and thermodynamic state to determine their effect on detonation formation and stability.

A more complete model would include realistic injector and nozzle geometries rather than a simplified premixed inlet. Higher computational resources would also allow finer mesh resolution, longer runtimes and more detailed wave-structure analysis.

Key Metrics