In modern gas turbines parts of combustion chamber and turbine section are under heavy heat load, for example, the rotor inlet temperature is far higher than the melting point of the rotor blade material. These high temperatures causes thermal stresses in the material, therefore it is very important to cool the components for safe operation and to achieve desired component life. But on the other hand the cooling reduces the turbine efficiency, for that reason it is vital to understand and optimize the cooling technique. In this project Thermochromic Liquid Crystals (TLCs) are used to measure distribution of heat transfer coefficient over a scaled up combustor liner section. TLCs change their color with the variation of temperature in a particular temperature range. The color-temperature change relation of a TLC is sharp and precise; therefore TLCs are used to measure surface temperature by painting the TLC over a test surface. This method is called Liquid Crystal Thermography (LCT). LCT is getting popular in industry due to its high-resolution results, repeatability and ease of use. Test model in present study consists of two plates, target plate and impingement plate. Cooling of the target plate is achieved by impingement of air coming through holes in the impingement plate. The downstream surface of the impingement plate is then cooled by cross flow and re-impingement of the coolant air. Heat transfer on the target plate is not uniform; areas under the jet which are called stagnation points have high heat transfer as compare to the areas away from the center of jet. It is almost the same situation for the impingement plate but the location of stagnation point is different. A transient technique is used to measure this non-uniform heat transfer distribution. It is assumed that the plates are semi-infinitely thick and there is no lateral heat transfer in the plates. To fulfill the assumptions a calculated time limit is followed and the test plates are made of Plexiglas which has very low thermal conductivity. The transient technique requires a step-change in the mainstream temperature of the test section. However, in practical a delayed increase in mainstream temperature is attained. This issue is dealt by applying Duhamel’s theorem on the step-change heat transfer equation. MATLAB is used to get the Hue data of the recorded video frames and calculate the time taken for each pixel to reach a predefined surface temperature. Having all temperatures and time values the heat transfer equation is iteratively solved to get the value of heat transfer coefficient of each and every pixel of the test surface. In total fifteen tests are conducted with different Reynolds number and different jet-to-target plate distances. It is concluded that for both the target and impingement plates, a high Reynolds number provides better overall heat transfer and increase in jet-to-target distance decreases the overall heat transfer.

Co-firing of biomass with coal in fluidized bed combustors is a promising alternative which leads to environmentally friendly use of coal by reducing emissions and provides utilization of biomass residues. Therefore, effect of biomass share on gaseous pollutant emissions from fluidized bed co-firing of various biomass fuels with high calorific value coals have extensively been investigated to date. However, effect of co-firing of olive residue, hazelnut shell and cotton residue with low calorific value lignites having high ash and sulfur contents has not been studied in bubbling fluidized bed combustors to date. In this thesis study, co-firing of typical Turkish lignite with olive residue, hazelnut shell and cotton residue in 0.3 MWt METU Atmospheric Bubbling Fluidized Bed Combustion (ABFBC) Test Rig was investigated in terms of combustion and emission performance and ash behavior of different fuel blends. The results reveal that co-firing of olive residue, hazelnut shell and cotton residue with lignite increases the combustion efficiency and freeboard temperatures compared to those of lignite firing with limestone addition only. O2 and CO2 emissions are not found sensitive to increase in olive residue, hazelnut shell and cotton residue share in fuel blend. Co-firing lowers SO2 emissions considerably while increasing CO emissions. Co-firing of olive residue and hazelnut shell has no significant influence on NO emissions, however, reduces N2O emissions. Co-firing cotton residue results in higher NO and N2O emissions. Regarding to major, minor and trace elements partitioning, co-firing lignite with biomasses under consideration shifts the partitioning of these elements from bottom ash to fly ash. No chlorine is detected in both EDX and XRD analyses of the ash deposits. In conclusion, olive residue, hazelnut shell and cotton residue can easily be co-fired with high ash and sulfur containing lignite without agglomeration and fouling problems.

A numerical analysis was conducted to study the effects of geometrically scaling scramjet inlet-combustor isolators. Three-dimensional fully viscous numerical simulation of the flow inside constant area rectangular ducts, with a downstream back pressure condition, was analyzed using the SolidWorks Flow Simulation software. The baseline, or 1X, isolator configuration has a 1” x 2.67” cross section and 20” length. This baseline configuration was scaled up based on the 1X configuration mass flow to 10X and 100X configurations, with ten and one hundred times the mass flow rate, respectively. The isolator aspect ratio of 2.67 was held constant for all configurations. To provide for code validation, the Flow Simulation program was first used to analyze a converging-diverging channel and a wind tunnel nozzle. The channel case was compared with analytical theory and showed good agreement. The nozzle case was compared with AFRL experimental data and showed good agreement with the entrance and exit conditions (Pi0= 40 psia, Ti0= 530ºR, Pe= 18.86 psia, Te= 456ºR, respectively). While the boundary layer thickness remained constant, the boundary layer thickness with respect to the isolator height decreased as the scale increased. For all the isolator simulations, a shock train was expected to form inside the duct. However, the flow simulation failed to generate this flow pattern, due to improper sizing of the isolator and combustor for a 3-D model or having a low pressure ratio of 2.38. Instead, a single normal shock wave was established at the same relative location within the length of each duct, approximately 80% of the duct length from the isolator entrance. The shape of the shock changed as the scale increased from a normal shock wave, to a bifurcated shock wave, and to a normal shock train, respectively for the 1X, 10X, and 100X models.

The goal of the present work was to identify and implement modifications to a density-based unstructured RANS CFD algorithm, as typically used in turbomachinery flows (represented here via the RoIIs-Royce 'Hydra' code), for application to Iow Mach number gas-turbine combustor flows. The basic algorithm was modified to make it suitable for combustor relevant problems. Fixed velocity and centreline boundary conditions were added using a characteristic based method. Conserved scalar mean and variance transport equations were introduced to predict scalar mixing in reacting flows. Finally, a flarnelet thermochemistry model for turbulent non-premixed combustion with an assumed shape pdf for turbulence-chemistry interaction was incorporated. A method was identified whereby the temperature/ density provided by the combustion model was coupled directly back into the momentum equations rather than from the energy equation. Three different test cases were used to validate the numerical capabilities of the modified code, for isothermal and reacting flows on different grid types. The first case was the jet in confined cross flow associated with combustor liner-dilution jetcore flow interaction. The second was the swirling flow through a multi-stream swirler. These cases represent the main aerodynamic features of combustor primary zones. The third case was a methane-fueled coaxial jet combustor to assess the combustion model implementation. This study revealed that, via appropriate modifications, an unstructured density-based approach can be utilised to simulate combustor flows. It also demonstrated that unstructured meshes employing nonhexahedral elements were inefficient at accurate capture of flow processes in regions combining rapid mixing and strong convection at angles to cell edges. The final version of the algorithm demonstrated that low Mach RANS reacting flow simulations, commonly performed using a pressure-based approach, can successfully be reproduced using a density-based approach.

The performance of dry, low NOx gas turbines, which employ lean premixed (or partially premixed) combustors, is often limited by combustor stability. To overcome this issue, a novel design, referred to as a Stagnation Point Reverse Flow (SPRF) combustor, has been recently demonstrated. The SPRF combustor has been shown to produce low NOx emissions with both gaseous and liquid fuels. The objective of this thesis is to elucidate the interactions between the flowfield and combustion processes in this combustor for gas- and liquid-fueled operation. This is achieved with experimental measurements employing various optical diagnostic techniques. These include Particle Image Velocimetry (PIV), chemiluminescence imaging, Planar Laser-Induced Fluorescence (PLIF) of OH radicals and laser scattering from liquid droplets. Velocity measurements in gas-fueled operation show that both nonreacting and reacting flows exhibit a stagnation region with low mean velocity and high turbulence intensities. The high shear between the forward and reverse flows causes significant recirculation resulting in enhanced entrainment and mixing of the returning product gases into the incoming reactant jet for the reacting flow cases, which enables stable operation of the combustor at very lean equivalence ratios. Nonpremixed operation produces a flowfield similar to premixed case except in the near-field region where high turbulence intensities result in significant fuel-air mixing before combustion occurs. Operation of the SPRF combustor with liquid Jet-A is also investigated experimentally. The results indicate that while the overall flow features are similar to the gas-fueled SPRF combustor, the combustion characteristics and NOx performance in liquid operation are strongly controlled by fuel dispersion and evaporation. Injecting the liquid at the exit of the air annulus results in a highly lifted flame, similar to nonpremixed gaseous operation. On the other hand, retracting the fuel injector well inside the air annulus produces a well-dispersed fuel pattern at the reactant inlet leading to a reduction of the equivalence ratio in the fuel consuming reaction zones. Since the effective Dahmkohler number increases with global equivalence ratio, the difference in NOx emissions is more pronounced at higher fuel-air ratios as the retracted injector lowers the relative mixing time compared to the flush case.

In this study, experimental data was collected in order to create a test case that can be used to validate computational fluid dynamics (CFD) simulations and the individual models used therein for gas turbine combustor applications. In many cases, the CFD results of gas turbine combustors do not correlate well with experimental results. For this reason, there is a requirement to test the simulation method used before CFD can successfully be used for combustor design. This test case encompasses all the features of a gas turbine combustor such as a swirler, primary, secondary and dilution holes as well as cooling rings. Experiments were performed on the same combustor geometry for both non-reacting and reacting flows. The non-reacting flow experiments consisted of stereoscopic particle image velocimetry (PIV) measurements performed at various planes in the three zones of the combustor. Data was collected on planes, both in line with the holes and in between the holes of each zone. For the reacting experiments, the temperatures on the outlet plane were measured using a thermocouple rake, thus a temperature contour plot on the outlet plane was produced. Further, the combustor can was modified with passive inserts, which were tested to determine their influence on the outlet temperature distribution during reacting runs. In this set-up, the outlet velocity profiles were also measured using a Pitot tube during both non-reacting and reacting flows. In addition to the outlet temperature distribution and velocity profiles, images of the flame patterns were captured, which showed the positions of flame tongues, fluctuating flames and steady flames. Carbon burn patterns on the walls of the combustor liner were also captured. From the data collected during the reacting runs, the pattern factor, profile factor, overall pressure loss and pressure loss factor were calculated. The non-reacting experiments performed using the PIV, produced three-dimensional velocity vector fields throughout the combustor. These experiments were performed at various flow rates, which gave an indication of which features of the combustor flow were affected by the flow rate. When comparing the individual PIV images alongside one another, the temporal nature of the combustor flow was also evident. The reacting experiments revealed a hot region of exhaust gas around the outer edge of the exhaust while there was a cooler region in the centre of the outlet flow. The PIV flowfield results revealed the reason for then hot outer ring-like region was due to the path the hot gasses would take. The hot combustor gas from the primary zone diverges outwards in the secondary zone then is further forced to the outside by the dilution recirculation zone. The hot flow then leaves the combustor along the wall while the cooler air from the jets leaves the combustor in the centre. The experiments performed produced a large variety of data that can be used to validate a number of aspects of combustor simulation using CFD. The non-reacting experimental data can be used to validate the turbulence models used and to evaluate how well the flow features were modelled or captured during the non-reacting stage of the combustor simulation process. The typical flow features such as jet penetration depths and the position and size of the recirculation regions are provided for effective comparison. The thermal results presented on the outlet plane of the combustor can be used for comparison with CFD results once combustion is modelled. Copyright / Dissertation (MEng)--University of Pretoria, 2010. / Mechanical and Aeronautical Engineering / unrestricted

A new Trapped Vortex Combustor (TVC) concept has been studied for applications such as those in Unmanned Aerial Vehicles (UAVs) as it offers potential for superior flame stability and low pressure loss. Flame stability is ensured by a strong vortex in a physical cavity attached to the combustor wall, and low pressure loss is due to the absence of swirl. Earlier studies on a compact combustor concept showed that there are issues with ensuring stable combustion over a range of operating conditions. The present work focuses on experimental studies and numerical simulations to study the stability issues and performance optimization in this compact single-cavity TVC configuration. For performing numerical simulations, an accurate and yet computationally affordable Modified Eddy Dissipation Concept combustion model is built upon the KIVA-3V platform to account for turbulence-chemistry interactions. Detailed validation with a turbulent non-premixed CH4/H2/N2 flame from literature showed that the model is sufficiently accurate and the effect of various simulation strategies is assessed. Transient flame simulation capabilities are assessed by comparison with experimental data from an acoustically excited oscillatory H2-air diffusion flame reported in literature. Subsequent to successful validation of the model, studies on basic TVC flow oscillations are performed. Frequencies of flow oscillations are found to be independent of flow velocities and cavity length, but dependent on the cavity depth. Cavity injection and combustion individually affect the magnitude of flow oscillations but do not significantly alter the resonant frequencies. Reacting flow experiments and flow visualization studies in an existing experimental TVC rig with optical access and variable cavity L/D ratio show that TVC flame stability depends strongly on the cavity air velocity. A detailed set of numerical simulations also confirms this and helps to identify three basic modes of TVC flame stabilization. A clockwise cavity vortex stabilized flame is formed at low cavity air velocities relative to the mainstream, while a strong anticlockwise cavity vortex is formed at high cavity air velocities and low L/Ds. At intermediate conditions, the cavity vortex structure is found to be in a transition state which leads to large scale flame instabilities and flame blow-out. For solving the flame instability problem, a novel strategy of incorporating a flow guide vane is proposed to establish the advantageous anticlockwise vortex without the use of cavity air. Experimental results with the modified configuration are quite encouraging for TVC flame stability at laboratory conditions, while numerical results show good stability even at extreme operating conditions. Further design optimization studies are performed in a multi-parameter space using detailed simulations. From the results, a strategy of using inclined struts in the main flow path along with the flow guide vane seems most promising. This configuration is tested experimentally and results pertaining to pressure drop, pattern factor and flame stability are found to be satisfactory.

The focus of the world is on the reduction of greenhouse gases, such as carbon dioxide, which contribute to the global warming currently experienced. Because most of the carbon dioxide emitted into the atmosphere is from fossil fuel combustion, alternative energy sources were developed and others are currently under study to see whether they will be good alternatives. One of these alternative sources of energy is the combustion of wood instead of coal. The advantages of wood are that it is a neutral carbon fuel source and that currently installed infrastructure used to combust coal can be retrofitted to combust wood or a mixture of wood and coal in an attempt to reduce the carbon dioxide emissions. Spent nuclear fuel has to be cooled so that the decay heat generated does not melt the containment system, which could lead to the unintentional release of radioactive material to the surroundings. The heat transfer mechanisms involved in the cooling have historically been analysed by assuming that the fluid and solid phases are at local thermal equilibrium (LTE) in order to simplify the analysis. The exergy destruction of the combustion of pine wood in an adiabatic combustor was investigated in this thesis using analytical and computational methods. The exergy destruction of the combustion process was analysed by means of the second law efficiency, which is the ratio of the maximum work that can be achieved by a Carnot engine extracting heat from the combustor, and the optimum work of the combustor. This was done for theoretical air combustion and various excess air combustions, with varied inlet temperatures of the incoming air. It was found that the second law efficiency reached an expected maximum for theoretical air combustion, and this held true for all varying air inlet temperatures. However, it was found that as the air inlet temperature was increased more and more, the maximum second law efficiency was the same for all excess air combustions, including the theoretical air combustion. It was also found that the results of the analytical and commercial computational fluid dynamics code compared well. Another analysis was conducted of irreversibilities generated due to combustion in an adiabatic combustor burning wood. This was done for a reactant mixture varying from a rich to a lean mixture. A non-adiabatic non-premixed combustion model of a numerical code was used to simulate the combustion process where the solid fuel was modelled by using the ultimate analysis data. The entropy generation rates due to the combustion and frictional pressure drop processes were computed to eventually arrive at the irreversibilities generated. It was found that the entropy generation rate due to frictional pressure drop was negligible when compared with that due to combustion. It was also found that a minimum in irreversibilities generated was achieved when the air-fuel mass ratio was 4.9, which corresponded to an equivalent ratio of 1.64, which was lower than the respective air-fuel mass ratio and equivalent ratio for complete combustion with theoretical amount of air of 8.02 and 1. Studieswere conducted to numerically analyse irreversibilities generated due to combustion in an adiabatic combustor burning wood. The first study analysed the effect of changing the incoming air temperature from 298 K to 400 K. The second study analysed the effect of changing the wall condition of the combustor from adiabatic to negative heat flux (that is heat leaving the system) for an incoming air temperature of 400 K. The irreversibilities generated in the combustor were calculated by computing the entropy generation rates due to the combustion, heat transfer and frictional pressure drop processes. For the first part of the study, it was found that for the minimum irreversibilities generated in the adiabatic combustor, the optimal air-fuel ratio (AF) corresponding to minimum irreversibilities slightly reduced from 4.9 to 4.8. In the second part of the study, it was found that by changing the wall condition from adiabatic to heat flux on the combustor, the AF corresponding to the minimum irreversibilities increased from 4.8 to 6. For the third part of the study, the combustor with a heat flux wall condition and a wall thickness simulated at an AF of 6, the sum of twice the wall thickness and the optimum diameter always added up to 0.32 m, resulting in the minimum irreversibilities. An analytical model was developed to minimise the thermal resistance of an air-cooled porous matrix made up of solid spheres with internal heat generation. This was done under the assumption of LTE. It was found that the predicted optimum sphere diameter and the minimum thermal resistance were both robust in that they were independent of the heat generation rate of the solid spheres. Results from the analytical model were compared with those from a commercial numerical porous model using liquid water and air for the fluid phase, and wood and silica for the solid phase. The magnitudes of the minima of both the temperature difference and the thermal resistance seemed to be due to equal contribution from the thermal conduction heat transfer inside the solid spheres and heat transfer in the porous medium. Because the commercial numerical porous model modelled only the heat transfer occurring in the porous medium, it expectedly predicted half of the magnitudes of the minima of the temperature difference and thermal resistance of those by the analytical model. / Thesis (PhD)--University of Pretoria, 2017. / Mechanical and Aeronautical Engineering / PhD / Unrestricted

No description available.

This thesis discusses the application, evaluation, and extension of dimensionally adaptive meshing to the numerical solution of velocity and pressure fields inside cylindrical reactors. Due to the high length to diameter ratios of many cylindrical reactor vessels the flow field can become axisymmetric, allowing for simplification of the governing equations and significant reduction in computational time required for solution. A fully 3D solver is developed from existing computational tools at BYU and validated against theoretical velocity profiles for pipe flow at various Reynolds numbers, as well as with experimental data for an axial-fired center jet with recirculating flow. Dimensionally adaptive meshing is then incorporated into the validated 3D solver. The boundary conditions and assumptions at the dimensional boundary are discussed. The flow information is passed across the boundary through spatial mass-weighted averaging. The 3D and axisymmetric computational domains are decoupled from one another so information can only be passed from the 3D domain downstream to the axisymmetric domain. The dimensional boundary placement must meet two main requirements, the flow must be one-way and axisymmetric. It is found that the flow becomes axisymmetric early on in the reactor (~0.3-0.4 m), but recirculation exists farther downstream (until ~0.61 m) and thus governs the placement of the dimensional boundary. The resulting computational tool capable of running simulations using dimensionally adaptive meshes is called CEDAR (Computationally Efficient Dimensionally Adaptive Recirculating flow solver). Several studies are then undertaken to examine CEDAR's ability to reproduce exit velocity profiles comparable to those produced by a fully 3D mesh, including variations in pressure, firing rate, and geometry. It is found that the flow structure inside the reactor is self-similar over a wide range of operating parameters as long as the burner jets are turbulent. This observation is supported by free and confined jet theory. These theories also provide a method for placing the dimensional boundary, which is a linear function of the confining geometry diameter only (assuming that the jet diameter is less than 1/10 the diameter of the confining geometry). All exit velocity profiles produced by CEDAR are on average within 5% of the fully 3D profiles. Timing studies reveal an average 5.16 times speedup in computational time over fully 3D computations.