Numerical optimization procedures have been employed for the design of airfoils in transonic flow based on the transonic small-disturbance (TSD) and Euler equations. A sequential approximation optimization technique was implemented for solving the design problem of lift maximization with wave drag and area constraints. A simple linear approximation was utilized for the approximation of the lift. Accurate approximations for sensitivity derivatives of the wave drag were obtained through the utilization of Nixon's coordinate straining approach. A modification of the Euler surface boundary conditions was implemented in order to efficiently compute design sensitivities without recreating the grid. Our design procedures experienced convergence problems for some TSD solutions, where the wave drag was found not to vary smoothly with the design parameters and consequently create local optimum problems. A procedure interchanging the role of the objective function and constraint, initially minimizing drag with a constraint on the lift was found to be effective in producing converged designs, usually in approximately 10 global iterations. This procedure was also shown to be robust and efficient for cases where the drag varied smoothly, such as with the Euler solutions. The direct lift maximization with move limits which were fixed absolute values rather than fractions of the design variables, was also found to be a reliable and efficient procedure for designs based upon the Euler equations. / Ph. D.

An experimental study of the transverse injection of a particle laden liquid jet into a supersonic air stream was conducted. Five micron diameter silicon dioxide particles with a specific gravity of 2.35 were suspended in water and injected across a Mach 3.0 air stream. A stagnation pressure of 4.2 atm. and a stagnation temperature of 15° C were maintained throughout the tests. Particle loadings of up to 68% by mass were injected. The penetration and behavior of the jets were examined through a light extinction survey and through a series of streak arid nanoflash photographs. Some of the particles were found to agglomerate into clumps ranging up to 40 microns in diameter. The clumps separated from the liquid plume and penetrated up to 45% further into the air stream than the liquid. For mass loadings greater than 55%, the jet column appeared more rigid than those of lower or zero particle loadings. / Master of Science

Mean flow measurements and short-duration Schlieren and Shadowgraph photographs of heated and unheated supersonic slot injection of air into a supersonic air stream are presented for the purpose of observing the mixing which occurs in the resulting shear layer. The heated injected jet (M_j, P_tj = 10.7 psia, and T_tj = 760°R) passes through a slot of height 0.475 inches (1.2 cm) tangent to a free stream with M_∞ = 3.0, P_∞= 95 psia and T_t_∞ = 540°R. The resulting density ratio is P_j / P_∞ = 0.3. The free stream Re/ft = 6.1 x 10⁷ (Re/cm = 2 x 10⁶). The thickness of the plate which separates the primary and secondary flows is 0.021 inches (0.052 cm). Pitot pressure, cone static pressure and stagnation temperature profiles are obtained at four axial stations downstream of the slot (x/H = 0.25, 4, 10, 20). An additional set of measurements is obtained at Station 4 for the case of a weak shock (P₂/P₁ = 1.8) interacting with the shear layer just upstream of Station 4. From the pressure and temperature measurements, Mach number, velocity, density, mass flux and static pressure profiles are generated. An unheated injection study is also performed at the same flow conditions for comparison to the heated case. The heated and unheated cases are very similar except in the slot where the temperature difference creates changes inp, U, and pU. This study is thus concerned with the effects of changes in P_j / P_∞ and U_j / U_∞, especially (U_j - U_∞), which are created by heating the injected flow. The heated slot flow did not create a marked difference in the location of the merging of the free stream boundary layer with the slot flow when compared to the unheated slot flow. Indeed, the appearance of the two flows on Schlieren photographs is similar even though the injected mass flow in the heated case is about 15% less than that in the unheated case. The pressure adjustments in the slot lip region are different for the two different cases. The flowfields are documented for both the heated and unheated cases with the added measurements and photographs for the shock impingement case. / Master of Science

A combination of tangential and normal air injection into a Mach 3 airflow was experimentally studied. A rearward facing slot producing tangential injection at a nominal Mach number 1.7 was operated at several different total pressures. An array of transverse tubes of height equal to the slot height and placed just downstream of the slot was operated at two dynamic pressure ratios as well as at Mach 1 and 2.2. Mean flow measurements of static and total pressures were taken up to 20 slot heights downstream from which Mach number, density, velocity and entrainment rates were calculated. Various dimensions and spreading angles of the mixing regions were measured directly from Nanoshadowgraphs and Schlieren photographs. Large eddy structures were produced in several cases, leading to increased entrainment of the free stream. For some cases, heated air was injected through the normal tubes, and the jet total temperature decay was measured downstream. It can be seen from the data as a whole that the mixing rate can be significantly increased by the combined tangential-normal injection design over tangential slot injection alone, with up to as much as 92% more entrained mass. / Master of Science

Air-cooled condensers are routinely designed for a variety of applications, including residential air-conditioning systems. Recent attempts at improving the performance of these heat exchangers have included the consideration of microchannel tube, multilouver fin heat exchangers instead of the more conventional round tube-plate fin designs. In most packaged air-conditioning systems, however, the condenser surrounds the compressor and other auxiliary parts in an outdoor unit, with an induced draft fan at the top of this enclosure. Such a configuration results in significant mal-distribution of the air flow arriving at the condenser, and leads to a decrease in performance. This work addresses the issue of mal-distribution by adapting the air-side geometry to the expected air flow distribution. A microchannel tube, multilouver fin condenser is first designed to transfer the desired heat rejection load for an air-conditioning system under uniform air flow conditions. Tube-side pass arrangements, tube dimensions, and fin and louver geometry are varied to arrive at a minimum mass, 2.54 kg condenser that delivers the desired heat load of 14.5 kW. The design model is then used to predict the performance of the condenser for a variety of air flow distributions across the heat exchanger. It is found that for a 50% air flow mal-distribution, the required condenser mass increases to 2.73 kg. The air-side geometry (fin density and height) of the condenser is then systematically changed to optimally distribute the air-side surface area across the condenser to best address the mal-distributed air flow. It is found that linear fin density and height variations from the mean value of 40% and 20%, respectively, keeping the mean fin density and height the same, reduce the required condenser mass to 2.65 kg even for this mal-distributed air-flow case. The influence of geometry variations on heat transfer coefficients, fan power and other performance measures is discussed in detail to guide the judicious choice of surface area and tube-side flow area allocations for any potential air flow mal-distribution. The results from this study can be used for the design of air-cooled condensers under realistic flow conditions.

Baigiamajame magistro darbe aptariama oro srauto greičio netolygumo aerodinaminiame vamzdyje problema ir aktualumas. Išanalizuota būdinga mokslinė literatūra ir aprašyti būdingiausi oro srauto greičio darbinės dalies skerspjūvyje nustatymo metodai. Naudojant CAD aplinkos programą sumodeliuotas trimatis aerodinaminio vamzdžio modelis ir pritaikius analizės baigtiniais elementais metodą, atliktas oro srauto greičio pasiskirstymo aerodinaminio vamzdžio darbinėje dalyje ir jos viduriniame skespjūvyje tyrimas. Tyrimo rezultatai išanalizuoti ir pateiktos išvados ir siūlymai. / In this master's thesis deals discusses the inadequacy of air flow speed in wind tunnel and the relevance of the problem. Performed analysis of the scientific literature and describe the most typical technique of air flow mesurements in wind tunnel. Using a CAD program modeled three-dimensional wind tunnel model and performed air flow velocity testing in wind tunenel test section and their midle cros area section usig CFD program. The results is evaluated and stated the conclusions and suggestions.

Typescript (photocopy). / Digitized by Kansas Correctional Industries

Thesis (MScIng)--University of Stellenbosch, 2004. / ENGLISH ABSTRACT: This study investigated the flow field above a proposed solar chimney power plant, without a cross wind, using a commercial numerical solver in the form of CFX 4-4 by AEA Technology plc. The governing equations solved are for an incompressible steady state solution. Variation in density due to buoyant effects is modelled with the Boussinesq approximation, and turbulence is approximated by the k-ε model with modifications due to buoyancy. The effect of different turbulence conditions at atmospheric inflows is also investigated. Modifications to the k-ε turbulence model in the context of atmospheric turbulence are investigated and applied to the standard k-ε model. These modifications include the addition of source terms in the turbulence kinetic energy and the turbulence dissipation rate equations to allow for the production/destruction of turbulence due to buoyancy. Other modifications include an additional source term in the turbulence dissipation rate equation accounting for atmospheric stability and the specification of model constants relevant to atmospheric flows. Initial results for the flow field using the Boussinesq approximation show reasonable correlation between the current study and the study by Thiart (2002) whereby the flow field exhibits characteristics of the axially symmetric turbulent jet. One of the primary and most noticeable differences between the current study and that of Thiart (2002) is the difference in height at which air is drawn into the collector. In order to account for variation in density with height, a modification on the Boussinesq approximation, the Deep Boussinesq model, Montavon (1998), is applied to a simplified 100 m by 100 m rectangular geometry. The results obtained are compared to a similar model using the Boussinesq approximation and it is seen that the average velocities reached in the deep Boussinesq model are significantly larger than those obtained in the Boussinesq approximation. / AFRIKAANSE OPSOMMING: Hierdie studie ondersoek die vloeiveld bo 'n aangewese son-skoorsteen kragstasie, deur die gebruik van 'n kommersiële numeriese oplosser in die vorm van CFX 4-4; van AEA Technology plc. Die heersende vergelykings wat opgelos moet word, is vir 'n niesaampersbare, tyd-onafhanklike oplossing. Die wisseling in digtheid as gevolg van saamdrukbaarheids effekte, word gemodelleer met die Boussinesq benadering en turbulensie-skatting deur die k-ε model met aanpassings as gevolg van dryfkrag. Die effek van die verskillende turbulensie-toestande teen atmosferiese invloeing word ook ondersoek. Aanpassings op die k-ε turbulensie-model in die konteks van atmosferies turbulensie word ondersoek en toegepas op die standaard k-ε model. Hierdie veranderinge sluit die byvoeging van bron terme in die turbulente kinetiese energie en die turbulensieverspillings vergelykings om ruimte te laat vir die produksie/afbreking van turbulensie as gevolg van dryfkrag in. Ander aanpassings sluit in 'n bykomende bron term in die turbulensie-verspillings vergelyking wat rekenskap gee vir atmosferiese stabiliteit en die spesifikasie van model onveranderlikes met betrekking tot atmosferiese vloei. Aanvanklike resultate vir die vloeiveld met die gebruik van die Boussinesq benadering dui op geredelike ooreenstemming tussen de huidige studie en die studie van Thiart (2002), waar die vloeiveld eienskappe toon van 'n simmetriese turbulente spilpunt. Een van die mees primêre en duidelikste verskille tussen die huidige studie en diè van Thiart (2002) is die verskil in die hoogte waarteen die lug in die samesteller ingesuig word. Om rekenskep te kan gee vir die verskil in digtheid met betrekking tot hoogte is 'n aanpassing van die Boussinesq benadering, die Diep Bussinesq model, Montavon (1998), aangebring. Hierdie aanpassing word toegepas op ‘n vereenvoudigde 100 m by 100 m reghoek. Die resultate wat verkry word, word vergelyk met 'n soortgelyke model waar die algemene snelheid wat met die diep Boussinesq model bereik word merkbaar groter is as die in die Boussinesq benadering.

Thesis (MScIng)--University of Stellenbosch, 2004. / ENGLISH ABSTRACT: One of the areas of the fluid dynamic design of solar chimney power plants that has not been investigated sufficiently is the collector-to-chimney transition section of a single turbine layout. The transition section contains the turbine inlet guide vanes (IGVs) that support the whole chimney and guide the flow entering the turbine. The primary objective of the study was to determine the dependence of the loss coefficient of the section on inlet guide vane stagger angle and collector roof height. Experiments were done on a nominal 900 mm chimney diameter rig, with four combinations of two collector roof heights and two IGV stagger angles. Velocity components and pressures in the transition section were measured in three conical planes, respectively at the IGV exit and midway to, and just below the turbine position, using a five-hole pneumatic pressure probe. Very good agreement was found between experimental values and commercial CFD code predictions of flow angles, velocity components and internal and wall static pressures. The agreement between measured and predicted total pressure loss coefficient was reasonable when considering that most of the loss occurred in the weak wakes of the IGVs and in the very thin transition section wall boundary layers. The CFD code served to extend the predictions to a proposed full scale geometry. The losses are less than previously assumed. The study led to correlations between respectively loss and turning angle as dependent variables, and collector roof height and IGV stagger angle as independent variables. / AFRIKAANSE OPSOMMING: Een van die gebiede van vloeidinamiese ontwerp van sonskoorsteen kragstasies wat nog nie voldoende navorsing geniet het nie is die kollektor-tot-skoorsteen oorgangs gedeelte van `n enkel turbine opstelling. Die oorgangs gedeelte bevat die turbine se inlaat lei lemme (ILL) wat die hele skoorsteen dra en die vloei lei wat by die turbine ingaan. Die hoof doel van hierdie studie was om die verlies koëffisiënt van hierdie oorgangs gedeelte te bepaal as afhanklike van die ILL stel hoek en die hoogte van die kollektor se dak hoogte. Die eksperimente het op `n nominale 900mm deursnee skoorsteen skaal model geskied, met vier kombinasies van twee dak hoogtes en twee ILL stel hoeke. Snelheidskomponente en drukke is met `n pneumatiese 5-punt buis in drie koniese vlakke in die oorgangs gedeelte gemeet. Die vlakke was by die ILL se stert, halfpad deur die oorgangs gedeelte en by die turbine se inlaat. Baie goeie ooreenstemming is gevind met die eksperimentele waardes en `n kommersiële CFD kode se voorspellings van vloei hoeke, snelheidskomponente en interne- en wand statiese drukke. Die ooreenstemming tussen die gemete waardes en die berekende waardes vir die totale druk verlies koëffisiënt was redelik siende dat die meeste verliese van die klein versteurings van die ILL en die oorgangs gedeelte se dun-wand grenslae kom. Die CFD kode is toe ingespan om verdere voorspellings te maak vir `n voorgestelde volskaal geometrie. Die verliese is minder as wat daar van te vore voorspel is. Hierdie studie het gelei tot korrelasies tussen onderskeidelik verlies en draai hoeke as afhanklike veranderlikes, en kollektor dak hoogte en ILL plasings hoek as onafhanklike veranderlikes.

Thesis (MScEng)--University of Stellenbosch, 2002. / ENGLISH ABSTRACT: Mixed flow fans find application in an operating region between that of axial flow and centrifugal fans. The candidate investigated the working of these mixed flow fans and formulated a design algorithm. The algorithm was based on work done by previous authors, most of which have tried to modify existing axial flow fan design methods to suit the mixed flow case. A fan was designed, built and tested. Tests showed that the fan did not perform as designed, producing a lower pressure rise at a lower flow rate. A five hole probe was used to measure the exit flow vector and the results showed that most of the work was done by the lower half of the blade. Based on this discrepancy between the designed and actual fan performance, a number of changes to the design procedure were recommended. Amongst other things: the use of numerical blade modelling procedures instead of modified cascade correlations and the disregard of the Coriolis work in the design of the blade shape. / AFRIKAANSE OPSOMMING: Gemengde vloei waaiers bevredig die behoefte aan 'n waaier in 'n spesifieke werksgebied waar aksiaal en sentrifugale waaiers minder effektief is. Die projek ondersoek die werking van gemengde vloei waaiers en formuleer 'n ontwerp-algoritme. Die algoritme is gebaseer op bestaande navorsing, waarvan die meeste 'n poging is om aksiaalvloeiwaaiermetodes aan te pas vir die gemengde vloei geval. 'n Waaier is ontwerp, gebou en getoets. Toetsresultate toon dat die waaier 'n laer drukstyging by 'n laer volume vloei gee as wat die ontwerp voorspel. 'n Vyfgat-anemometer is gebruik om die uitlaatvloei-vektor te meet en die resulate het bewys dat die gedeelte van die rotorlem naaste aan die naaf meeste van die werk doen. Hierdie afwyking van die ontwerptoestand is ondersoek en 'n aantal veranderings aan die ontwerp algoritme is voorgestel, naamlik: die gebruik van numeriese lemprofiel simulasie, instede van die glip- en deviasie- modelle, en die ontwerp van die lemprofiel deur die Coriolis werk te Ignoreer.