As the need for high-powered, high endurance devices continues to rise, new technologies that can provide the required energy densities are being investigated. One such technology is micro-combustion. Micro-combustors are small devices that make use of hydrocarbon fuels to produce thermal energy. The extremely high energy densities of hydrocarbon fuels, when compared to modern battery systems, is a motivating factor in their study. Research has shown that the heat recirculation in the wall structure of a micro-combustor is one of the governing parameters of the ow within.
The primary goal of this thesis was to develop a boundary condition that models the heat recirculation within the wall structure of a single-channel micro-combustor. Along with this, this thesis sought to evaluate the possibility of modelling more complex micro- combustor designs through the use of user-dened boundary conditions. Prior to the development of the boundary condition script, the transient, two-dimensional heat con- dution problem that exists within the micro-combustor wall was solved. A number of dierent boundary conditions were applied to the dierent edges of the micro-combustor wall; these were convection on the southern boundary, convection and radiation on the northern boundary and insulated boundaries at the east and west. The development of the boundary condition was completed using the programming language Lua.
Once the boundary condition, and associated les were completed, the heat conduction problem was solved externally, as well as part of a combined simulations with Eilmer3. These simulations allowed for the assessment of the boundary condition in regards to its computational performance. A convergence analysis of boundary was conducted for both the time and spatial discretisations used. This analysis demonstrated that the transverse dimension required a greater discretisation than the axial dimension. An analysis of the computational time for the heat conduction problem and coupled1 simulations was also undertaken. This analysis showed that as the discretisation in the wall structure increase, the percentage of the total computation time required for the wall calculation increased dramatically.
With these investigations conducted, it can be recommended that the use of user- dened boundaries conditions in Eilmer3 be avoided for the analysis of more complex ows and geometries. At present, the Lua scripts used to solve the heat conduction problem are creating a bottleneck within the simulations. As such, future work in this area would be to begin development of a C++ boundary condition, within the Eilmer3 source code.