Thesis (Ph.D.)--University of Delaware, 2006. / Principal faculty advisor: Stanley Sandler, Dept. of Chemical Engineering. Includes bibliographical references.

Thesis (Ph. D.)--University of Chicago, 1916. / "A Private edition distributed by the University of Chicago libraries." "Reprinted from the Botanical gazette, vol. LXIII, no. 5, May 1917." "Literature cited": p. 395-397. Also issued online.

This thesis discusses the effect of an obstruction on the evolution of the Rayleigh-Taylor instability in a confined geometry at low Atwood numbers. Laboratory experiments are the principal method of investigation, though these data are supplemented with implicit large eddy simulations (ILES). The laboratory data are captured using an innovative laser scanning system which is able to simultaneously record density and velocity data in 3D. A new approach for calculating density data from laser induced fluorescence measurements is developed and demonstrated. The technique is used to improve the accuracy of the density measurement from laser induced fluorescence, by correcting for the damage to dye caused by the laser. The introduction of an obstacle at the height of the initial interface results in dramatic changes to the dynamics of mixing, even when this obstacle is only a few percent of the domain width. Two obstructed scenarios are considered. In both of these an obstruction is placed on the interface between an upper heavy layer and lower light layer. In the first case, a single horizontal opening connects the upper and lower layers. A bidirectional flow exchanges fluid through the opening, establishing a circulation cell in each layer. These cells exist quasi-steadily for long periods, constantly recirculating and mixing the fluid in each layer. This acts to increase the time required for mixing compared with the classical unobstructed case, but results in a more uniformly mixed final stratification. The second case has two horizontal openings, one either side of the obstruction. This results in markedly different dynamics. The flow through each of the openings switches back and forth between being bidirectional (as with the single opening case) and unidirectional, with unidirectional exchange reversing direction with a constant period. These results are consistent with the ILES data. For both of these cases a wide range of analytical techniques are used to connect the new obstructed dynamics with previously conducted research, such as calculating the molecular mixing fraction, energetics and mixing efficiency. A multistage mixing process is identified, unique for cases with an obstruction. For the single opening case a hierarchy of models are developed that accurately capture the density change of each layer for both the experimental and numerical data. The effect of changing the aspect ratio of the domain is investigated using ILES, from which different dynamical regimes are observed, discussed and analysed.

Made available in DSpace on 2014-06-11T19:31:37Z (GMT). No. of bitstreams: 0 Previous issue date: 1998-08-17Bitstream added on 2014-06-13T19:20:52Z : No. of bitstreams: 1 battaglini_nmp_dr_botfca.pdf: 581238 bytes, checksum: 6976e40352726ea0079d8119e8b4c482 (MD5) / Fluidos de viscosidade entre 0,874 10-3 e 787 10-3 N m s-2 foram agitados pelos impelidores tipos âncora convencional, âncora modificada, fita helicoidal e parafuso em tanques cilíndricos de fundos reto e toroesférico para determinação da potência requerida e do tempo de mistura. Na agitação efetuada com os impelidores centrados houve o aparecimento de vórtices em algumas rotações. As medidas de potência foram efetuadas na região em que o número de Reynolds (NRe ) variou de 16 a 70.000 com a âncora convencional e modificada. Os diagramas do número de potência( ) P N em função do número de Reynolds, construídos com os dados experimentais, mostraram que, para todos os fluidos usados, ocorreu um desvio na curva quando o vórtice, formado na superfície do fluido, alcançou os impelidores e teve início a dispersão de ar no interior da massa fluida. A presença de bolhas de ar desenvolveu, no tanque, um sistema de duas fases (líquido–gás), alterando as características dos fluidos agitados. Nos cálculos do número de potência e do número de Reynolds foram utilizados os valores de viscosidade e densidade do fluido, e não os correspondentes ao sistema líquido-gás o que provocou, provavelmente, o desvio nas curvas. Na região em que o número de Reynolds fica entre 16 e 300 não houve formação de vórtices acentuados com aeração e a curva foi definida pela correlação ( )a NP A NRe = , onde A é função do tipo de impelidor e das dimensões do vaso e o expoente a, um valor empírico que varia com o número de Reynolds obtendo a relação ( )1,7 Re N = 19.953 N − P . O desempenho dos impelidores na homogeneização foi analisado pela energia requerida para se obter uma mistura de açúcar e água com concentrações correspondentes a 40 e 60o Brix. O método desenvolvido foi baseado na verificação... / Fluids with viscosity between (0,874 and 787 10-3 Nm s-2 were agitated by conventional anchor, modified anchor, helicoidal ribbon and screw impellers in cilindric tanks straight and concave (torospheric) bottoms for the measurement of power and time of mixture. The arisal of vortex occurred when agitation was accomplished with centred impellers and in some rotations. The measures of power were obtained in the interval where the number of Reynolds (NRe) varied from 16 to 70.000 with conventional and modified anchor. The diagrams of number of Reynolds in function of number of Froude constructed with experimental data showed that, to all fluids employed, a deflection in the curve occurred when the vortex formed at the surface of the fluid reached the impellers and the dispersion of air inside the fluid mass started. The presence of air bubbles develops in the tank a system of two phases altering the characteristics of the agitated fluids. In the calculations of the number of Reynolds were used the values of the fluid viscosity and density and not ones corresponding to the liquid-gas system, provoking a deflection in the curves. In the interval where the number of Reynolds remains between 16 and 300 the formation of a well developed vortex with aeration did not occur and the curve was defined by the correlation A N = ANRE Re , where A is function of the type of the impeller and of the dimensions of the tank, and the exponent a, an empirical value, varies with the number of Reynolds. The performance of the impellers in the homogenization was analysed in terms of the energy required in order to obtain a mixture...(Complete abstract click electronic access below)

Thesis (MTech (Chemical Engineering))--Cape Peninsula University of Technology, 2005 / The prediction of head losses in a pipe system is very important because head losses affect the performance of fluid machinery such as pumps. In a pipe system, two kinds of losses are observed: major losses and minor losses. In Newtonian and non-Newtonian flow, major losses are those that are due to friction in straight pipes and minor losses are those that are due to pipe fittings such as contractions, expansions, bends and valves. Minor losses must be accurately predicted in a pipe system because they are not negligible and can sometimes outweigh major losses (Edwards et al., 1985). There is presently little data for the prediction of non-Newtonian head losses in pipe fittings in the literature and little consensus amongst researchers (Pienaar et al., 2004). In the case of diaphragm valves, usually, only one loss coefficient value is given in turbulent flow or in laminar flow with no reference to a specific size of the valve, assuming geometrical similarity that would lead to dynamic similarity. However, no one has done a systematic study of various sizes of diaphragm valves from the same manufacturer to establish if this is true. This could be the main reason for discrepancies found in the literature (Hooper, 1981; Perry & Chilton, 1973; Miller, 1978 and Pienaar et al., 2004). This work addresses this issue. A literature revIew on the flow of Newtonian and non-Newtonian fluids has been presented. The work of Hooper (1981) on diaphragm valves and the works of Edwards et al., (1985), BaneIjee et aI., (1994) and Turian et al., (1997) for non-Newtonian fluids in globe and gate valves were found to be relevant to this work.

Thesis (DTech (Chemical Engineering))--Cape Peninsula University of Technology, 2005. / The subject of this study is highly concentrated emulsion explosive (HCEE). These emulsions are dispersions of an aqueous phase (up to 90 v-%) in an oil phase. The dispersed droplets consist of an aqueous solution of nitrate salts, which is supersaturated at room temperature, comprising less than 20% of water by mass. Compounds of this kind are thermodynamically unstable and their instability is related to the coarsening of emulsion (coalescence) and phase transition (crystallization) of dispersed phase. However it is demonstrated that the dominating mechanism is slow crystallization inside the super-cooled droplets. The main goal of this thesis is a phenomenological study of the dependence of structural parameters, such as droplet size and volume fraction, as well as the ageing processes, on the rheological properties of these emulsions. The results of the measurements include the flow and viscoelastic properties of the materials. The rheological parameters are correlated with the kinetics of structural changes during ageing, as a function of emulsion formulation. The emulsions under study are non-Newtonian liquids. Experiments in shear rate sweep mode demonstrate that the upward and downward branches of the flow curves coincide above some specific shear rate value. The upward experiments show the existence of a low shear Newtonian asymptote, while the effect of yielding is observed in the downward curve. Wall slip is investigated and shown to be negligible.

No description available.

No description available.

No description available.

No description available.