Surface-mounted finite-height circular cylinders and square prisms can be found in many industrial and engineering applications. The local flow fields around these bluff bodies are not yet well understood due to lack of experimental and numerical data close to the cylinder and prism. The aim of this thesis was therefore to gain an improved physical description of the flow field above the free end surface and around the cylinders and prisms. In the present experimental study, the particle image velocimetry (PIV) technique was used to measure the flow field very close to these bluff bodies in the test section of a low-speed wind tunnel. Four finite circular cylinders and square prisms of aspect ratios AR = 9, 7, 5 and 3 were tested at a Reynolds number of ReD = 4.2×104. At the location of the cylinder or prism, the boundary layer thickness relative to the cylinder diameter or prism width (D) was δ/D = 1.6. PIV velocity field measurements in the near-wake region were made in a vertical plane parallel to the mean flow direction on the flow centreline (the symmetry plane), within 2D upstream and 5D downstream of the cylinder or prism. Additional PIV measurements were carried out in three orthogonal x-z, x-y, and y-z planes above the free end surface of the models. In the near-wake region of the finite circular cylinders, the large recirculation zone contained a vortex immediately behind and below the free end; this vortex was found for all four aspect ratios. A second vortex was found behind the cylinder near the cylinder-wall junction; this vortex was not observed for the cylinder of AR = 3, indicating a distinct wake structure for this cylinder. Similar to the circular cylinder case, in the near-wake region of the square prisms, a vortex was observed immediately behind and below the free end in the recirculation zone. The size and strength of this vortex increased as the aspect ratio of the prism decreased. Also, a second vortex was found near the prism-wall junction downstream of the prisms of AR = 9 and 7, while this vortex was not observed for the prisms of AR = 5 and 3. The PIV results in the near-wake regions of the circular cylinders and square prisms show that the effect of the bluff body shape (circular or square cross-section) is evident in the maximum length of the mean recirculation zone. A considerable difference was seen between the maximum length of the mean recirculation zones of the circular cylinder and square prism of AR = 9, while the shape of the bluff body does not considerably affect the length of the recirculation zones for the bodies of AR = 7, 5, and 3. The present PIV results also provided insight into the separated flow above the free ends, including the effects of AR and body shape. Above the free end of the cylinders, flow separation from the leading edge led to the formation of a mean recirculation zone on the free-end surface. The point of reattachment of the flow onto the free-end surface moved towards the trailing edge as the cylinder aspect ratio was decreased. Large regions of elevated turbulence intensity and Reynolds shear stress were found above the free end. For the finite circular cylinders, the flow pattern above the free end was similar in all three x-z planes for all aspect ratios, consisting of a cross-stream vortex at approximately x/D = 0. According to the PIV results in the x-y planes, one of the main characteristics of the flow over the free end surface of the circular cylinders was a pair of focal points at x/D ≈ 0 and near the edge of the free end. As the cylinder aspect ratio increased, the size and strength of these vortices decreased. Also, the centers of the vortices moved downstream as the aspect ratio increased. For the finite square prism, the large, separated, recirculating flow region extended into the near wake. For the square prism of AR = 3, considerable difference was seen in the free-end flow pattern compared to the more slender prisms of AR = 9, 7 and 5. In particular, a cross-stream vortex formed due to interaction between the separated flow from the leading edge of the prism and the reverse flow over the trailing edge of the free end. This vortex was seen in all three planes at different cross-stream locations for AR = 3 but only in the symmetry plane for AR = 9. Hence, the present PIV results in the x-z planes revealed the effect of the near-wake flow on the flow above the prism free end. The results also showed a considerable effect of the aspect ratio on the mean velocity field as well as the Reynolds stress fields. The results in the x-y planes showed different flow patterns for the prism of AR = 3 including wall-normal vortices close to the free end at the sides of the prism as well as two saddle points close to the corners of the trailing edge and one node downstream of the trailing edge, while for AR = 9, no vortices and node were observed. Two streamwise vortices with opposite sign of rotation were seen in the y-z plane at x/D = 0.2 for all aspect ratios. The present results illustrate in-plane vorticities originating from the vertices of the leading edge of the prism for all aspect ratios.

Nesta pesquisa é apresentado um estudo do escoamento turbulento formado pela interação de “bluff-body” com o fluxo principal (ar). Os “bluff-bodies” estudados possuem a forma de um disco, cone e cilindro. A zona de recirculação formada por esta interação possibilita a aplicação de “bluff-body” como estabilizador mecânico de chama. Aplicando vários modelos de turbulência foram feitas simulações computacionais para analisar a formação da zona de recirculação do fluxo de ar em relação à forma do estabilizador de chama (disco, cone e cilindro), sua dimensão (grau de bloqueamento do canal) e velocidade de escoamento. Para escolher o modelo de turbulência que descreve melhor o comportamento do fluxo em escoamento turbulento com “bluff-body” foi feita uma comparação entre os dados experimentais e os obtidos computacionalmente. O modelo de turbulência K-Epsilon mostrou maior conformidade com os dados experimentais em comparação com os outros modelos aplicados. / 92 f.

This thesis involves the experimental testing of a bluff body with a blunt base to evaluate the effectiveness of end-plate tabs in reducing drag. The bluff body is fitted with interchangeable end plates; one plate is flush with the rest of the exterior and the other plate has small tabs protruding perpendicularly into the flow. The body is tested in the Cal Poly 3ft x 4ft low speed wind tunnel. Testing is conducted in three phases. The first phase was the hot-wire measurement of streamwise velocity of the near wake behind the bluff body. An IFA300 thermal anemometry system with a hot-wire probe placed behind the model measures the wake velocity fluctuations. The power spectral density on the model without tabs shows large spikes at Strouhal numbers of 0.266, 0.300, and 0.287 at corresponding Re = 41,400, 82,800, 124,200 where vortex shedding occurs. The model with tabs shows no such peaks in power and therefore has attenuated vortex generation in the wake flow at that location. The second phase of testing was pressure testing the model through the use of pressure ports on the exterior of the bluff body. A Scanivalve pressure transducer measured multiple ports almost simultaneously through tubing that was connected to the model internally and routed through the model’s strut mount and outside of the wind tunnel. This pressure testing shows that the model with tabs is able to achieve up to 36% increase in Cp at Reh = 41,400 on the base region of the bluff body and no negative pressure spikes that occur as a result of vortex shedding. The last phase of testing is the measurement of total drag on the model through a sting balance mount. This testing shows that the drag on the model is reduced by 14% at Re = 41,400. However it also shows that as velocity increased, the drag reduction is reduced and ultimately negated at Re = 124,200 with no drag loss at all. The addition of tabs as a passive flow control device did eliminate vortex shedding and alter the base pressure of the bluff body. This particular model however showed no reduction in total drag on the model at high Reynolds numbers higher than 124,000. Further study is necessary to isolate the exact geometry and flow velocities that should be able to produce more favorable drag results for a bluff body with this type of passive flow control device.

No description available.

Bluff body stabilization of flames is a commonly employed technique for combustion applications, such as thrust augmentors. These combustors are usually required to operate at lean conditions governed by a lower stability limit on combustion denoted by lean blow off. Lean blow off is believed to be a dynamically unstable phenomenon that leads to flame extinction or convection from a stable, usually desired, point in space. Current theories predict lean blow off based on models that were developed over specific domain of inflow parameters. This thesis sought to compile, re-evaulate, and analyze past blowoff data presented in literature using time scale correlations, Damkohler numbers, by employing modern chemical kinetic solvers to approximate characteristic chemical times. The research has conclusively shown that it is possible to express blowoff data for multiple flow conditions using a power law relationship between Damkohler number and Reynolds numbers. From the analysis of this power law relations, trends are validated using past empirical observations, and some new information regarding flame stability is also conveyed.

Two algorithms commonly used for solving low-speed flow fields are evaluated using an unsteady turbulent flow formulation. The first algorithm is the method of artificial compressibility which solves the incompressible Navier-Stokes equations. The second is a preconditioned system for solving the compressible Navier-Stokes equations. Both algorithms have been implemented into GASP Version 4, which is the flow solver used in this investigation. Unsteady numerical simulations of unsteady, 2-D flow over square cylinders are performed with comparisons made to experimental data. Cases studied include both a single-cylinder and a three-cylinder configuration. Two turbulence models are also used in the computations, namely the Spalart-Allmaras model and the Wilcox k-ω (1998) model. The following output data was used for comparison: aerodynamic forces, mean pressure coefficient, Strouhal number, mean velocity magnitude and turbulence intensity. The main results can be summarized as follows. First, the predictions are more sensitive to the turbulence model choice than to the choice of algorithm. The Spalart-Allmaras model overall produced better results with both algorithms than the Wilcox k-ω model. Second, the artificial compressibility algorithm produced slightly more consistent results compared with experiment. / Master of Science

The purpose of this research was to operate a bluff body flame holder with the objective of stabilizing a flame at elevated pressures over a range of equivalence ratios. The ability to have a ground-based test rig capable of maintaining stable flames at high pressures and temperatures is critical in understanding flames present in modern jet engines and gas turbine technologies. The facility was reconfigured multiple times and the resultant flame was imaged within the optical test section. A converging nozzle was utilized to choke the flow and vary the operating pressures up to 5 atm. By regulating mass flow rates of both the fuel and air, the target range of equivalence ratios was achieved. Jet fuel was successfully ignited on the bluff body and a flame was maintained in the recirculation zone. Visualization of flames during the flights of any aircraft is limited due to material and weight requirements, therefore, performing these studies in ground- based facilities is required. Further analysis was performed to characterize C2* and CH* radicals in fuel lean and rich flames.

No description available.

No description available.

The effects of shape and relative submergence (the ratio of flow depth to obstacle height, d/H) were investigated on the wakes around four different low-aspect-ratio wall-mounted obstacles: semi-ellipsoids with the major axes of the base ellipses aligned in the streamwise and transverse directions, two cylinders with aspect ratios matching the ellipsoids. Wake structure of a fully submerged, spherical obstacle was also investigated in the same flow conditions to provide insight into the flow obstacle interaction with ramification to sediment transport. A low-aspect-ratio semi-ellipsoid was chosen as broad representative of a freshwater mussel projecting from a river bed, and a sphere was employed as representative of a boulder. Two cylinders were used due to their similarity to geometries investigated in other studies. Digital Particle Image Velocimetry and thermal anemometry were used to interrogate the flow. For ellipsoids and cylinders, streamwise features observed in the mean wake included counter-rotating distributions of vorticity inducing downwash (tip structures), upwash (base structures), and horseshoe vortices. In particular, the relatively subtle change in geometry produced by the rotation of the ellipsoid from the streamwise to the transverse orientation resulted in a striking modification of the mean streamwise vorticity distribution in the wake. Tip structures were dominant in the former case while base structures were dominant in the latter. A vortex skeleton model of the wake is proposed in which arch vortex structures, shed from the obstacle, are deformed by the competing mechanisms of Biot-Savart self-induction and the external shear flow. An inverse relationship was observed between the relative submergence and the strength of the base structures for the ellipsoids, with a dominant base structure observed for d/H = 1 in both cases. The wake of the sphere is more complex than ellipsoidal geometries. Streamwise features observed in the mean wake including tip, horseshoe structures, and weak upwash. The shedding characteristics and dynamics of the wake were examined. Weak symmetric shedding was observed in the wakes of streamwise and transverse ellipsoids at d/H = 3.9 while cross-spectral measurements confirmed downstream and upstream tilting of arch structures shed by the transverse and streamwise ellipsoids, respectively. Much weaker peaks in the power spectrum were observed for low- and high-aspect-ratio cylinders. While the dominant Strouhal number remained constant as the relative submergence was reduced to d/H = 2.5 for the ellipsoids, it increased abruptly at d/H = 1 and transitioned to an antisymmetric mode. For sphere geometry at d/H = 3.9, a weak dominant frequency was observed close to obstacle junction and the cross-correlation function for symmetric measurements in the wake indicates symmetric shedding. These results demonstrate a means by which to achieve significant modifications to flow structure and transport mechanisms in the flow.