Thesis (MScEng (Mechanical and Mechatronic Engineering))--University of Stellenbosch, 2006. / The project aimed at numerically modeling the flow and thermal processes occurring in a Volkswagen Citi Golf Chico underhood using computational fluid dynamics (CFD). The motivation for this investigation was to develop and demonstrate the capability of CFD as an automotive analysis tool. This would allow local automobile analysts and designers enhanced analyses of the thermal and flow conditions occurring in this com-pact environment, leading to improved local vehicles. A review of relevant literature indicated that the CFD community in South Africa is small with comparison to the international sector. The application of CFD to analyse automo-biles in South Africa is limited and practised by few. This experience requires develop-ment and refinement, such that South Africa may improve vehicles manufacture in the country. The review also indicated that CFD used in the international communities pro-vides good results, promoting simulation-based engineering. The experimental investigation involved parking a vehicle in the subsonic wind tunnel intake at the Mechanical Engineering Department in Stellenbosch. This tunnel is 3.7 m wide, 4 m long and 2.8 m tall, capable of wind speeds up to 90 m/s. Various equipment including thermocouples, a thermal imager and a hand held hot-wire anemometer pro-vided temperature and velocity measurements within the underhood. A pitot-static probe connected to a pressure transducer measured the wind tunnel velocities. The numerical investigation started with the creation of a three-dimensional geometry of the underhood from measurements taken of the vehicle. This geometry, created with Solid Edge version 14, formed the domain for automatically generating discretised grids using STAR-Design version 3.2. Subsequently, boundary conditions and numerical models were applied to the grids, which included simplified fan and radiator models. The analysis concluded with results obtained from the numerical CFD simulations, per-formed with STAR-CD version 3.24. The validity and accuracy of the numerical solutions was verified and quantified with the numerical results. The evaluation consisted of two test cases (wind tunnel speeds of 0 m/s and 5 m/s), each simulated at three different grid resolutions. Each simulation con-tinued until they fully converged to a single solution. The comparison of the three simu-lations from each case indicated that the results were grid independent. The final in-spection of the results in terms of y+ values and boundary conditions indicated that the models implemented were valid. The comparison of the numerical results for temperatures and fan inlet velocities with the experimentally measured data served as a measure to quantify the applicability of CFD for underhood investigations. The comparison between the two sets of results proved acceptable, with a maximum difference of 10%, indicating that CFD is capable of predicting temperatures and flow fields with reasonable accuracy. The numerical results indicated that while the vehicle travels at higher velocities, the underhood remains well ventilated. The underhood tends to trap the hot air from the radiator and other heat sources when the vehicle remains stationary, causing the air to heat further. This can be addressed by the installation of vents in the side panels near the top of the underhood environment. This should allow the hot air to escape, possibly resulting in a significant reduction of the underhood temperatures. Momentum and energy source terms modelled the effects from the fan and radiator. These models worked well for both cases, but improvement is necessary. Special at-tention should be given to the condition where the radiator fan obstructs the flow through the radiator. A further result of the project was the establishment of a flexible foundation for conduct-ing numerical simulations on automobiles. It allows for the inclusion of additional com-ponents and the implementation of more advanced models for representing effects from various engine components.

This thesis is concerned with the development and application of laser induced thermal grating spectroscopy (LITGS) as a tool for thermometry in reacting and non-reacting flows. LITGS signals, which require resonant excitation of an absorbing species in the measurement region to produce a thermal grating, are acquired for systematic measurements of temperature in high pressure flames using OH and NO as target absorbing species in the burned gas. The signal obtained in LITGS measurements appears in the form of a time-based signal with a characteristic frequency proportional to the value or the sound speed of the local medium. With knowledge of the gas composition, the temperature can be derived from the speed of sound measurement. LITGS thermometry using resonant excitation of OH in the burned gas region of in oxygen enriched CH4/O2/N2 and CH4/air laminar flames was performed at elevated pressure (0.5 MPa) for a range of conditions. Measurements were acquired in oxygen enriched flames to provide an environment in which to demonstrate LITGS thermometry under high temperature conditions (up to 2900 K). The primary parameters that influence the quality of LITGS signal were also investigated. The signal contrast, which acts as a marker for the strength of the frequency oscillations, is shown to increase with an increase in the burnt gas density at the measurement point. LITGS employing resonant excitation of NO is also demonstrated for quantitative measurements of temperature in three environments – a static pressure cell at ambient temperature, a non-reacting heated jet at ambient pressure and a laminar premixed CH4/NH3/air flame operating at 0.5 MPa. Flame temperature measurements were acquired at various locations in the burned gas close to a water-cooled stagnation plate, demonstrating the capability of NO-LITGS thermometry for measuring the spatial distribution of temperature in combustion environments. In addition, the parameters that in influence the local temperature rise due to LITGS were also investigated in continuous vapour flows of acetone/air and toluene/air mixtures at atmospheric conditions. Acetone and toluene are commonly targeted species in previous LITGS measurements due to their favourable absorption characteristics. Results indicate that LITGS has the potential to produce accurate and precise measurements of temperature in non-reacting flows, but that the product of the pump intensity at the probe volume and the absorber concentration must remain relatively low to avoid significant localised heating of the measurement region.

No description available.