M.S.

No description available.

Thesis (MScEng)--Stellenbosch University, 2013. / ENGLISH ABSTRACT: The majority of power stations in South Africa are located in coal rich, but arid regions where wet-cooled condenser systems are not feasible from an environmental and economic perspective. Consequently the focus on power generation cooling has shifted towards dry-cooling systems using air-cooled steam condensers (ACSC). The steam passing through the ACSC units is cooled by an air-draught, mechanically induced by large diameter axial flow fans. Consequently the effectiveness of the cooling is impacted by the performance of these fans, which ultimately affects the overall efficiency of the power plant. However, due to the large diameters (> 10 m) of these fans, their performance is predicted based on small scale test results using the fan scaling laws. The objective of this project was to develop a methodology which accurately predicts the fan performance of more than one fan configuration using computational fluid dynamics (CFD) software and validating the results with experimental tests. Four fans were considered in this study of which three were scaled fan models of large air-cooled axial fans. The performance of the scaled fan models (L1-, L2- and N-fan) were measured in a type A, BS 848 standard fan test facility. The geometries of the fans were scanned three-dimensionally to obtain the models for simulation purposes. The other fan considered was an 8- bladed axial fan designed by Bruneau (1994) and referred to as the B-fan. Simulations were carried out for the L2-, N- and B-fan for different computational domains while implementing the multiple reference frames (MRF) and steady RANS approach. Three variations of the k-ε turbulence model were also investigated. Noticeable differences were found between the experimental and numerical results of the B-fan. Good correlations between the numerical and experimental fan static pressure, fan power and fan static efficiency were found for the two scaled model fans over a large operating range. The performance of the full scale fans, however, did not correlate well with the performance of the scaled models. It is concluded that accurate simulations of axial fans are possible although these domains require a large number of mesh elements. It is recommended that further research is carried out to investigate the relationship between full scale and small scale fan models. / AFRIKAANSE OPSOMMING: Die meerderheid steenkool kragstasies in Suid-Afrika is geleë in droë, maar steenkool ryke streke waar natverkoelde kondensor stelsels uit ʼn omgewings en ekonomiese perspektief nie geskik is nie. Die fokus in kragopwekking verkoeling het dus verskuif na droë-verkoelings stelsels en spesifiek die gebruik van lugverkoelde stoomkondensors (LVSKs). Die stoom in LVSK eenhede word verkoel deur atmosferiese lug wat meganies geïnduseer word deur groot aksiaalvloeiwaaiers. Die effektiwiteit van die verkoelingsproses word gevolglik beïnvloed deur die werksverrigting van hierdie waaiers wat uiteindelik die algehele effektiwiteit van die kragstasie beïnvloed. As gevolg van hierdie waaiers se grootte word hulle werksverrigting egter bepaal op grond van kleinskaal toetsresultate en deur gebruik te maak van die waaierskaleringswette. Die hoofdoelwit van hierdie projek was om ‘n metodiek te ontwikkel wat die werksverrigting van ʼn aksiaalwaaier akkuraat kan voorspel vir ʼn verskeidenheid opstellings, deur gebruik te maak van berekenings vloei meganika (BVM) sagteware en die resultate eksperimenteel te verifieer. Die projek het vier waaiers ondersoek waarvan drie van hierdie waaiers geskaleerde modelle van groot lugverkoelde aksiaalwaaiers was. Die werksverrigting van die geskaleerde waaiers (L1-, L2- en N-waaier) was met ‘n tipe A, BS 848 standaard waaier toetsfasiliteit gemeet. Die geometrie van dié waaiers was ook drie-dimensioneel opgemeet vir simulasie doeleindes. Die B-waaier, ‘n 8 lem aksiaalwaaier, wat ontwerp is deur Bruneau (1994) was slegs numeries ondersoek. Die L2-, N- en Bwaaier was gesimuleer in verskillende berekeningsdomeine deur gebruik te maak van die multi verwysingsraamwerk en gestadigde vloeiberekenings benaderings. Drie k-ε turbulensie modelle was ook ondersoek. Merkbare verskille tussen die eksperimentele en numeriese resultate van die Bwaaier was waargeneem. Goeie korrelasie tussen die eksperimentele en numeriese resultate van die geskaleerde waaiers vir ‘n wye bedryfsbestek was gevind. ‘n Vergelyking tussen die volskaal en kleinskaal waaiers se werksverrigting het egter beduidende afwykings aangetoon. Deur gebruik te maak van ‘n groot aantal selle in die berekeningsdomein was dit moontlik om ʼn verskeidenheid aksiaalvloeiwaaiers akkuraat te simuleer. Verdere navorsing wat die verhouding tussen volskaal en kleinskaal waaiers ondersoek woord aanbeveel.

Thesis (MScEng)--Stellenbosch University, 2014. / ENGLISH ABSTRACT: Carbon dioxide is a natural gas that has been used as a refrigerant as far back as the 1850s. It was then primarily used because it was easily obtainable and non-toxic. Due to the high operating pressure and relatively low critical temperature (and high critical pressure) of CO2, it was gradually phased out and replaced with hydrochlorofluorocarbons and chlorofluorocarbons. With the discovery of the greenhouse effect, greenhouse gasses and ozone depleting gasses, the Montreal Protocol and the Kyoto Protocol were placed into effect. These two protocols call for the reduction in use of certain greenhouse gasses and the complete exclusion of others. The focus turned to natural gasses that are more environmentally friendly and easier to come by. Carbon dioxide is one such gas. However, CO2 has a low critical temperature and high pressure, 33.98 °C and 73.77 bar respectively. CO2 refrigeration systems are more effective, have a greater coefficient of performance and have a greater operating temperature difference (over the gas-cooler) when used under transcritical conditions. It is preferred to have transcritical CO2 refriger-ation systems. The drawback of such a system is the extreme operating conditions. A special system must be designed, built and commissioned at the University of Stellenbosch. The system will demand the use or manufacture of unique, special-ised components. Most of the components utilised are extremely expensive and/or difficult to come by, or designed and manufactured specifically for this system. The CO2 system at the University of Stellenbosch is exceptional and stands out from conventional systems in that: - it can operate under both the subcritical and transcritical conditions; - it has multiple expansion units, which can be alternated, giving a total of four different operating configurations; - the system can run with fully automated controllers or as a static system; - the internal heat exchanger can be included or excluded from the system; - all the heat exchangers can run in counter flow or parallel flow; and - The system has multiple testing points, for both temperature and pressure, to give the operator accurate measurements to be used in comparison with design software, etc. Three simulation programs are given that describe the physical system. The first simulation program is a steady-state simulator used to aid in the design of the heat exchangers and the capillary tube. The second simulator is a steady-state program that determines the mass flow rate in the capillary tube. The third simulation is a transient program, programmed to determine the steady-state conditions of a sys-tem, given set initial conditions and a transient start-up. / AFRIKAANSE OPSOMMING: Koolstofdioksied is 'n natuurlike gas wat gebruik word as 'n verkoelingsmiddel so ver terug as die 1850's. Dit is toe hoofsaaklik gebruik omdat dit maklik verkrygbaar en nie giftig is nie. As gevolg van die hoë werksdruk en relatief lae kritiese temperatuur (en hoë kritiese druk) van CO2, is dit geleidelik uitgefaseer en vervang met hidrochloorfluoorkoolstof en chloorfluoorkoolstowwe. Met die ontdekking van die kweekhuiseffek, kweekhuisgasse en osoon-afbrekende gasse, is die Montreal-protokol en die Kyoto-protokol in werking gestel. Hierdie twee protokolle vereis die afname in die gebruik van sekere kweekhuisgasse en die algehele uitsluiting van ander. Die fokus het verskuif na natuurlike gasse wat omgewingsvriendeliker en makliker is om te bekom. Koolstofdioksied is so 'n gas. Maar CO2 het 'n lae kritiese temperatuur en hoë kritiese druk, 33.98 °C en 73.77 bar onderskeidelik. CO2 verkoelingstelsels is meer effektief, het 'n groter koëffisiënt van werksverrigting en het 'n groter bedryfstemperatuur-verskil (oor die gas-verkoeler) wanneer dit gebruik word onder transkritiese toestande. Dit is dus verkieslik om transkritiese CO2 verkoelingstelsels te hê. Die nadeel van so 'n stelsel is die relatief uiterste bedryfstoestande. 'n Spesiale stelsel moet ontwerp word, dikwels met die gebruik of vervaardiging van unieke, hoogs gespesialiseerde komponente. So 'n stelsel is ontwerp vir die Universiteit van Stellenbosch. Die meeste van die komponente wat gebruik is, is baie duur en/of moeilik om te bekom, of is spesifiek ontwerp en vervaardig vir hierdie stelsel. Die CO2-stelsel by die Universiteit van Stellenbosch is uitsonderlik en staan uit bo konvensionele stelsels deurdat: - dit kan funksioneer onder beide subkritiese en transkritiese toestande; - dit verskeie gasuitsettings-eenhede het, wat afgewissel kan word, wat 'n totaal van vier verskillende bedryfskonfigurasies gee; - die stelsel bedryf kan word met volle outomatiese beheerders of as 'n statiese stelsel; - die interne hitte-uitruiler óf in-, óf uitgesluit kan word van die stelsel; - al die hitte-uitruilers bedryf kan word in kontra-vloei of parallelle vloei; en - die stelsel verskeie toetspunte het, vir beide temperatuur en druk, wat die operateur akkurate metings gee om te vergelyk met die ontwerp sagteware, ens., wat gebruik word. Drie gegewe simulasieprogramme beskryf die fisiese stelsel. In een simulasie word 'n bestendige toestand-simulator gebruik om te help met die ontwerp van die hitteruilers en die kapillêre buis. Die tweede simulasie program is 'n bestendige toestands-program wat die algehele vloeitempo in die kapillêre buis bepaal. Die laaste simulasie is 'n vlugtige program, geprogrammeer om die tydelike toestande van 'n stelsel te bepaal, gegewe vasgestelde aanvanklike toestande en 'n kortstondige aanskakeling.

Thesis (PhD)--Stellenbosch University, 2004. / ENGLISH ABSTRACT: During the last 30 years I have been involved in the theory and practice of thermal engineering and in particular, in the areas of air-cooled heat exchangers and cooling towers for the power, refrigeration, process and petrochemical industries in South Africa and internationally. During this period, I have authored and co-authored more than 120 papers that were published in technical journals or presented at conferences nationally or internationally. Most of these papers are included in a manuscript entitled "Air-cooled Heat Exchangers and Cooling Towers", in which Ipresent a systematic approach to the thermal performance evaluation and design of industrial air-cooled heat exchangers and cooling towers. This original publication also includes the relevant practice applicable to the design of cooling systems, based on my experience as a consultant to industry. Design offices throughout the world presently follow our design methods, or at least employ many of our research results. Our work has furthermore contributed to the development of improved cooling system designs (e.g. new dephlegmator header designs), components (e.g. single-row flattened finned tubes) and product improvement and quality control (e.g. performance testing and measurement of thermal contact resistance between fin and tube during production). Many of our research findings have found application in the modification of existing cooling systems. The manuscript has also been used as reference work during the presentation of short courses to practising engineers and consultants in industry and to engineering graduates at the University of Stellenbosch. A two-volume edition of this manuscript was published by PennWell Corp., Tulsa, Oklahoma, USA in 2004. / AFRIKAANSE OPSOMMING: Gedurende die laaste 30 jaar was ek betrokke by die teorie en praktyk van lugverkoelde warmteoordraers en koeltorings vir die kragopwekkings-, verkoelings-, proses- en petro-chemiesenywerhede in Suid-Afrika sowel as in die buiteland. Gedurende hierdie periode was ek outeur en mede-outeur van meer as 120 publikasies wat in tegniese tydskrifte, of by plaaslike of oorsese konferensies aangebied is. Die meeste van hierdie publikasies vorm deel van 'n manuskrip getiteld "Air-cooled Heat Exchangers and Cooling Towers" waarin ek 'n sistematiese benadering tot die bepaling van die termiese vermoë en ontwerp van industriële lugverkoelde warmteoordraers en koeltorings aanbied. Hierdie oorspronklike publikasie bevat ook die relevante praktyk wat van toepassing is op verkoelingsaanlegte. Ontwerpkantore wêreldwyd volg tans hierdie ontwerpsmetodes, of gebruik ten minste baie van ons navorsingsresultate. Ons werk het verder bygedra tot die ontwikkeling van verbeterde verkoelingsaanlegte (bv. nuwe deflegmatore), komponente (bv. enkelbuisry platvinbuise ) en verbeterde produkte en kwaliteitskontrole (bv. toetsing van verkoelingsvermoë oftermiese kontakweerstand tussen vin en buis gedurende produksie). Baie van ons bevindinge het toepassing gevind in die modifikasie van verkoelingsaanlegte. Die manuskrip is ook as verwysing gebruik gedurende die aanbieding van kort kursusse aan ingenieurs in die praktyk en aan nagraadse studente aan die Universiteit van Stellenbosch. 'n Twee-volume uitgawe van die manuskrip is deur PennWell Corp., Tulsa, Oklahome, VSA in 2004 gepubliseer.

Increased consumer demand for fresh fruit throughout the year has created a need for long term storage. Long term storage of fruit uses more energy than fresh market products, thus increasing production cost. Pacific Northwest energy costs and more competition for markets has made energy conservation an important factor to be considered by the fruit industry. A BASICA computer program, RLSIM, was developed to predict the transient refrigeration load throughout the storage season in apple and pear cold storage warehouses. RLSIM accurately predicted the seasonal and component refrigeration system energy demand curves during the 1985-1986 cold storage season. The results also indicated that the largest single energy use component is the continuous operation of evaporator fans. Simulation of a six hours on and six hours off fan cycling technique indicated a reduction of 23.75 percent could be achieved in overall refrigeration system energy use in the cold storage warehouse. Cold storage warehouse management can be improved by using the results of RLSIM. Fan cycling schemes could be properly employed without risk of increasing fruit temperature. Recommendations were made to update research in areas of cooling and respiration rates of various fruits in both controlled atmosphere and common storage. / Graduation date: 1988

Pulse tube cryocoolers (PTCs) are among the most attractive choices of refrigerators for applications requiring up to 1 kW of cooling in the temperature range of 4-123 K as a result of the high relative efficiency of the Stirling cycle, the reliability of linear compressors, and the lack of cryogenic moving parts resulting in long life and low vibration signature. Recently, PTCs have been successfully used in applications in the 150 K range, extending the useful range of the device beyond the traditional cryogenic regime. A carefully designed cylindrical cavity referred to as the pulse tube replaces the mechanical expander piston found in a Stirling machine. A network consisting of the pulse tube, inertance tube, and surge volume invoke out-of-phase pressure and mass flow oscillations while eliminating all moving parts in the cold region of the device, significantly improving reliability over Stirling cryocoolers. Terrestrial applications of PTCs expose a fundamental flaw. Many PTCs only function properly in a narrow range of orientations, with the cold end of the pulse tube pointed downward with respect to gravity. Unfavorable orientation of the cold head often leads to a catastrophic loss of cooling, rendering the entire cryocooler system inoperable. Previous research indicates that cooling loss is most likely attributed to secondary flow patterns in the pulse tube caused by free convection. Convective instability is initiated as a result of non-uniform density gradients within the pulse tube. The ensuing secondary flow mixes the cryogen and causes enhanced thermal transport between the warm and cold heat exchangers of the cryocooler. This study investigates the nonlinear stabilizing effect of fluid oscillation on Rayleigh-Bénard instability in a cryogenic gas subject to misalignment between gravitational body force and the primary flow direction. The results are directly applicable to the flow conditions frequently experienced in PTCs. Research has shown that the convective component can be minimized by parametrically driven fluid oscillation as a result of sinusoidal pressure excitation; however, a reliable method of predicting the influence of operating parameters has not been reported. In this dissertation, the entire PTC domain is first fully simulated in three dimensions at various angles of inclination using a hybrid method of finite volume and finite element techniques in order to incorporate conjugate heat transfer between fluid domains and their solid containment structures. The results of this method identify the pulse tube as the sole contributor to convective instability, and also illustrate the importance of pulse tube design by incorporating a comparison between two pulse tubes with constant volume but varying aspect ratio. A reduced domain that isolates the pulse tube and its adjacent components is then developed and simulated to improve computational efficiency, facilitating the model’s use for parametric study of the driving variables. A parametric computational study is then carried out and analyzed for pulse tubes with cold end temperatures ranging from 4 K to 80 K, frequencies between 25-60 Hz, mass flow - pressure phase relationships of -30◦ and +30◦, and Stokes thickness-based Reynolds numbers in the range of 43-350, where the turbulent transition occurs at 500. In order to validate the computational models reported and therefore justify their suitability to perform parametric exploration, the CFD codes are applied to a commercially developed single stage PTR design. The results of the CFD model are compared to laboratory-measured values of refrigeration power at temperatures ranging from 60 K to 120 K at inclination angles of 0◦ and 91◦. The modeled results are shown to agree with experimental values with less than 8.5% error for simulation times of approximately six days using high performance computing (HPC) resources through Georgia Tech’s Partnership for Advanced Computing (PACE) cluster resource, and 10 days on a common quad-core desktop computer. The results of the computational parametric study as well as the commercial cryocooler data sets are compiled in a common analysis of the body of data as a whole. The results are compared to the current leading pulse tube convective stability model to improve the reliability of the predictions and bracket the range of losses expected as a function of pulse tube convection number. Results can be used to bracket the normalized cooling loss as a function of the pulse tube convection number NPTC. Experimental data and simulated results indicate that a value of NPTC greater than 10 will yield a loss no greater than 10% of the net pulse tube energy flow at any angle. A value of NPTC greater than 40 is shown to yield a loss no greater than 1% of the net pulse tube energy flow at all angles investigated. The computational and experimental study completed in this dissertation addresses static angles of inclination. Recent interest in the application of PTCs to mobile terrestrial platforms such as ships, aircraft, and military vehicles introduces a separate regime wherein the angle of inclination is dynamically varying. To address this research need, the development of a single axis rotating cryogenic vacuum facility is documented. A separate effects apparatus with interchangeable pulse tube components has also been built in a modular fashion to accommodate future research needs.

No description available.

Atualmente, a análise de desempenho de sistemas de refrigeração domésticos é realizada através de ensaios experimentais normalizados. Durante esses experimentos, diversas variáveis como pressões de trabalho, temperaturas em diversos pontos do sistema, corrente elétrica e potência consumida, são monitoradas. Porém, em muitos casos são necessárias mais de 24 horas para execução de um teste experimental (e.g., teste abaixamento de temperatura). Tendo em vista o tempo despendido nestes testes, propõe-se no presente trabalho um modelo matemático semi-empírico capaz de predizer o comportamento das variáveis do sistema testado e, com isso, antecipar o final do ensaio. O modelo, desenvolvido através das leis de conservação da massa e da energia, apresenta parâmetros que são ajustados a partir de informações experimentais obtidas durante a execução do próprio teste. Após a inicialização do ensaio, a cada período de tempo prédeterminado, os dados medidos são utilizados para determinar os parâmetros empíricos do modelo. Obtidas as constantes, simula-se o comportamento das principais variáveis do sistema de refrigeração até a condição de regime permanente. Com isso, o teste experimental pode ser finalizado com antecedência. O modelo desenvolvido é capaz de prever com boa precisão, a partir de duas horas de teste, a variação da vazão mássica e da pressão de sucção (com diferenças da ordem de 10% em regime permanente quando comparadas às variáveis experimentais), da pressão de condensação (com diferença da ordem de 5%) e da temperatura da parede do condensador (diferença da ordem de 2°C). / The performance of household refrigeration systems are usually evaluated through experimental tests carried at in temperature and humidity controlled chambers. During the tests, the discharge and suction pressures, the temperature in several system positions, and the compressor power are measured. These tests are expensive and time-demanding, e.g., a single pull-down test can take more than 24 hours to be performed. Although the mathematical models have been proposed for decades as an alternative to the experiments, they are not sufficiently reliable to substitute completely the tests. Therefore, the current work proposes a semiempirical mathematical model to predict the system performance with the purpose of reducing the test time instead of replacing it. The model is based on the mass and energy conservation equations in which the constant parameters, such as conductance and capacitances, are calibrated from previous measured values of temperature and pressure. As soon as the parameters are obtained, a simulation is performed to forecast future values of temperature, pressure and compressor power and therefore, to anticipate the end of the test. Calibrations and simulations can be continuously performed as the test evolves. Preliminary results show that steadystate values of discharge and suction pressures can be predicted within error bands of 5 and 10%, respectively, after only two hours of a pull-down test being performed.

Dynamic finite element modeling and analysis of a refrigeration compressor was investigated as part of a noise emission study. Natural frequencies and normal mode shapes were calculated for the major structural components of the compressor. The components were later combined to form a model of the compressor assembly which was subsequently solved for its dynamic properties. Model development included coordination with test data for verification and revision to improve model prediction accuracy. Considerable efforts were made to accurately represent the hermetic shell which presents several inherent modeling difficulties due to its geometry and other characteristics which result from a deep drawn manufacturing process. The importance of physical simplifications such as geometry representation, thickness variation, attachments, the welded seam, and residual stresses were established. In addition, theoretical limitations of the finite element method were addressed as a cause for analysis-test discrepancies. Housing models developed were found to agree within 12% of experimental natural frequencies up to 1100 Hz. Compatibility of analytical normal modes with resonant dwell experimental deflection shapes was considered. Analytical forced vibration response showed situations when the deflected shapes can be a superposition of modes rather than the pure mode shape. Analytical simulation of the test setup improved the agreement of analysis and test data. Additional components modeled include the internal compressor mechanism and its supports. Analysis showed that interactions with the internal components, particularly resonances within the suspension springs, are important for a valid representation of the compressor assembly. Resonances within the internal suspension components more than double or nearly triple the number of resonance frequencies in the compressor assembly. / Master of Science