Mechanical Engineering - Theses
Permanent URI for this collection
Search Results
1 - 5 of 5
-
ItemLow-Reynolds-number turbulent boundary layers( 1988-12)This thesis documents an extensive experimental investigation into low-Reynolds-number turbulent boundary layers flowing over a smooth flat surface in nominally zero pressure gradients. The way in which these layers are affected by variations in R(theta), i.e. the Reynolds number based on the boundary-layer momentum thickness, type of tripping device used and variations in freestream velocity, each considered independently, are investigated.
-
ItemSome aspects of turbulent boundary layers( 1969)A detailed experimental programme of two dimensional rough wall turbulent boundary layers developing in zero and arbitrary adverse pressure gradients is used to investigate aspects of turbulent boundary layer development and surface roughness. It is shown that two types of roughness can be clearly distinguished on the basis of the flow variables involved. The more common ‘k’ or ‘sand grain’ type follows the well documented ‘Nikuradse-Clauser’ correlation scheme where the effect of roughness on the flow depends on the size or scale of the roughness elements. The second type of roughness, typified by a smooth wall containing a pattern of narrow cavities, is independent of the scale of the roughness and does not follow• the ‘Nikuradse-Clauser’ correlation scheme. It is shown that previous pipe flow experiments have involved this second type of roughness and these results are used to show that the dependent variable is pipe diameter. This roughness has therefore been named ‘d’ type in this thesis. No length scale associated with the boundary layer could be found to replace pipe diameter except for zero pressure gradient layers. However, it is found that the distance below the crests of the roughness from where the logarithmic distribution of velocity is measured will correlate both types of roughness action. It is shown that a zero pressure gradient turbulent boundary layer developing on a ‘d’ type rough wall conforms to Rotta's condition of precise self preserving flow. The results are used to illustrate several theoretical consequences of this type of flow. Wall shear stresses are determined by measuring in detail, the pressures on the faces of the roughness elements and thereby calculating their form drag. Similarity laws for these pressure patterns are developed for the ‘d’ type results and explicit expressions for the functions are proposed for the zero pressure gradient case. Pressure patterns around 'k' type roughness elements cannot be described by the similarity laws developed here. Theories proposed by several authors to describe the velocity profiles in regions above the logarithmic distribution are compared in detail and critically examined. Some new work related to these theories is introduced. The predictions of mean velocity distribution are tested against an extensive range of experimental data including the results of this thesis. It is shown that all the theories have important shortcomings in their present form and a recommendation for a basis for future work is offered. The problem of the transition of a turbulent boundary layer from a rough (‘d' type) to smooth wall in an adverse pressure gradient is investigated experimentally for two boundary layers. It is found that the outer regions of the boundary layer appear to be unaffected by this change in wall condition whereas the inner flow makes a rapid adjustment to it. This result is at variance to the published work on flow in conduits and for zero pressure gradient boundary layers. An explanation of this is offered. Literature surveys introduce the work in each topic.
-
ItemExperiments in smooth wall turbulent channel and pipe flows( 2011)An experimental investigation of smooth wall, fully developed turbulence was undertaken in a high aspect ratio channel flow facility and a circular pipe flow facility. High fidelity hot-wire measurements of streamwise velocity with carefully matched experimental and anemometer parameters were made in order to facilitate the direct comparison of the two internal flows over a range of Reynolds numbers. Measurements where the inner scaled hot-wire sensor length was held constant at 22 with length-to-diameter ratio greater than 200 revealed that channels and pipes have similar streamwise mean velocity distributions in the near wall and logarithmic regions, with pipe flow mean velocity profiles showing a larger deviation from the log law in the outer region compared to the channel flow. The turbulence intensity profiles were found to match well throughout the flow, and importantly, the near wall peak in turbulence intensity was shown to increase with friction Reynolds numbers up to 3000. Further, the centreline turbulence intensity in both flows was also shown to increase with increasing Reynolds number in a manner consistent with data presented in the literature. The one-dimensional pre-multiplied energy spectra revealed that the near wall peak in turbulence intensity grows due to the increasing energy contribution of the large scales of turbulent motion, however, whilst the spectral energy distribution is similar in channels and pipe for the near wall and overlap regions, the channel flow maintains energy at longer wavelengths than the pipe flow at the centreline, indicating that there are some subtle yet important difference between the core regions of these two flows. Measurements where hot-wire sensor length and aspect ratio were systematically varied reveal that increasing sensor lengths lead to attenuation of an increasingly large range of small scales, whereas an insufficient aspect ratio causes the measured energy to be attenuated over a very broad range of scales. It was demonstrated that the competing effects of spatial filtering and increasing Reynolds number may be the cause for the conflicting reports regarding the Reynolds number dependency of the near wall peak in turbulence intensity. Spatial filtering effects were found to be no different in channels and pipes for the range of sensor lengths and Reynolds numbers investigated. While the effect of spatial filtering diminished when moving away from the wall such that little to no attenuation was observed in the outer region, an insufficient sensor aspect ratio appeared to attenuate energy over a broad range of scales in both the inner and outer regions of the flow. It was determined that a sensor aspect ratio greater than 200 is sufficient for the pure platinum hot-wires used in this study which is consistent with the findings of Ligrani & Bradshaw (1987). In a further set of experiments, an array of multiple wall skin friction sensors was simultaneously employed with a wall normal traversing hot-wire to capture the ‘footprint’ of large scale motions in a turbulent channel flow. These experiments were carried out at a friction Reynolds number of 1000. Conditional averaging revealed that in the presence of low skin friction, there exists an elongated low speed region flanked by high speed regions consistent with `Superstructure' type events. The turbulence intensity within these low speed regions was modulated, with increased activity in the presence of local relative streamwise deceleration and decreased activity in the presence of local relative streamwise acceleration. These observations are consistent with the large scale counter rotating roll modes proposed by Hutchins & Marusic (2007b) and fit both the hairpin packet model of Adrian et al. (2000) as well as the cluster-wake model of del Alamo et al. (2006).
-
ItemNumerical study of internal wall-bounded turbulent flows( 2011)Direct numerical simulation (DNS) of turbulent pipe flow has been performed at Reynolds numbers ranging from Reτ ≈ 170 to 2000. A literature review highlights a need for higher Reynolds number pipe flow DNS data. There have been many numerical studies for internal geometry (pipe and channel) wall-bounded turbulent flows. Many of the numerical data for both pipe and channel flows, which are now readily accessible are at lower Reynolds numbers. At higher Reynolds numbers, there is a lack of pipe flow DNS data as compared to channel flow DNS data. As the highest Reynolds numbers in numerical simulations are starting to overlap the lower region of experiments, validation of both experimental and numerical results is now possible. Moreover, numerical simulations are extremely useful in complementing experimental results in the near-wall region where accurate experimental data are often difficult to obtain. However, available DNS data of internal wall-bounded turbulent flows are performed with different grid resolutions and computational domain sizes, making it difficult to directly compare between them. An undertaking of this thesis involves a systematic study (using constant grid resolutions) of the domain length effect on the convergence of turbulence statistics. Investigations carried out using numerical data from fully developed pipe flow simulations indicate a recommended computational length of 8π pipe radius or half channel height for turbulence statistics to converge. It is hoped that this will serve as a benchmark computational domain length for future numerical simulations performed. A study is also carried out to better understand the similarities and differences of the flow physics between turbulent channel and pipe flows. This is performed using the newly obtained pipe flow DNS data and channel flow DNS data of del ´ Alamo et al. (2004) at a comparable Reynolds number of Reτ ≈ 1000. Different turbulence statistics investigated including mean flow, turbulence intensities, correlations and energy spectra. Comparison of both wall-bounded channel and pipe flows shows little discrepancies in the near-wall region but differences are observed in the outer-region. Although there is abundant literature for both experimental and numerical wall bounded turbulent flows, further analysis reveals discrepancies in the open literature. One of the primary contributing factors that plagues reported results are spatial resolution issues. In this thesis, the numerical data is used to investigate the effects of insufficient spatial resolution in wall-bounded turbulence by averaging the streamwise velocity component in the spanwise direction. A correction scheme is proposed to correct experimental results suffering from insufficient spatial resolution. The correction scheme is applied to attenuated experimental results such as streamwise turbulence intensity and one-dimensional energy spectra and is shown to be effective. The method of using DNS data to analysis and correct experimental results can be extended to other experimental techniques such as particle image velocimetry and laser doppler velocimetry.
-
ItemRecovery of fluid mechanical modes in unsteady separated flows( 2010)This study is concerned with the recovery of fluid mechanical modes that can be used to describe the physical properties of unsteady separated flows. The flow configuration of interest is a spanwise homogeneous NACA 0015 airfoil with leading edge laminar separation and turbulent recirculation. An in-depth understanding of the unsteady flow dynamics and fluid mechanical stability properties, can assist in the future development of more efficient separation control strategies. In order to provide a richer understanding of the physics, the flow fields are numerically generated, and characterised at various key Reynolds numbers leading up to the target turbulent case. Proper Orthogonal Decomposition modes are recovered to most efficiently represent the unsteady scales of motion, and linear stability modes are sought to identify how a perturbation will evolve in this unsteady environment. The generation of the Proper Orthogonal Decomposition modes can require very large amounts of data, and the current study presents a means of recovering these modes using parallel computation. To enable the stability analysis, a means of performing the calculation in steady two-dimensional flows of semi-complex geometry has been developed. The corrections required to perform the stability analysis in unsteady turbulent flows has also been identified by using a non-linear eddy viscosity model to close the triple decomposition stability equations. It is intended that the means of recovering these fluid mechanical modes can assist in the future development of reduced order models necessary for the control of unsteady separated flows.