Thesis (MScEng (Mechanical and Mechatronic Engineering))--University of Stellenbosch, 2010. / AFRIKAANSE OPSOMMING:Die navorsing wat hier uiteengesit word ondersoek die hipotese dat kondensasie en die gevolglike verdamping van water wat teenwoordig is in die uitlaatgas van ‘n binnebrandenjin, plaasvind in die gedeeltes van die uitlaatstelsel tussen die uitlaatklep en die katalitiese-omsetter se uitlaat. Daar word verder veronderstel dat hierdie tweefasevloeieffekte die tydafhanklike temperatuurprofiele in die uitlaatstelsel beïnvloed, wat moontlik kan lei tot ‘n vertraging in die tydsduur vir die katalitiese omsetter om temperature van 200-300 °C te bereik, wat nodig is om noemenswaardige omsetting te bewerkstellig. Om hierdie veronderstelling te evalueer is ‘n tydafhanklike, eendimensionele wiskundige model van die termo-vloei gedrag in die uitlaatstelsel gedurende ‘n koue inwerktreeding, insluitende vogtigheidseffekte, opgestel en opgelos deur van ‘n rekenaaralgoritme gebruik te maak. Warmte- en massaoordragsteorie was gebruik om die ongestadigde massa-, energie- en momentumbehoudsvergeleikings te formuleer. Die tweefasige vogeffekte was gemodelleer deur gebruik te maak van die verhouding tussen warmte- en massaoordrag, wat verdamping en heterogeniese kondensasie (die kondensasie van damp teen die pypwand) voorspel as gevolg van die dampdrukgradient tussen die grootmaat damp en die versadigde damp by die oppervlak van die vloeistoffilm. Homogene kondensasie (die kondensasie van vloeistof in die vorm van druppels in die dampstroom) was ook in aanmerking geneem indien die grootmaatgas temperatuur onder die versadigingstemperatuur van die grootmaatdamp gedaal het. ’n Eksperimentele ondersoek was gedoen deur van twee enjins gebruik te maak, ’n 1.6 L Volkswagen Bora en ’n 1.6 L Ford RoCam, in die toetsselle van Cape Advanced Engineering Pty (Ltd). Om die gastemperature so akkuraat moontlik te meet, was spesiale radiasiegeskermde sensore met vinnige reaksietyd ontwerp en installer in die pypseksies van die uitlaatstelsels van beide enjins. Die geskermde sensore het temperature van tot 50 °C hoër as konvensionele termokoppels in dieselfde areas gemeet. Dit is in koers is met resultate wat deur die foutbeperkingsteorie, geassosieer met die meet van temperature in vloeïende gas in uitlaatstelsels, voorspel word. Vergelyking van die numeriese simulasie met die eksperimenteel gemete temperature het aangedui dat in dele van die uitlaatstelsel voor die katalitieseomsetter, die vog min uitwerking het op die termiese gedrag van die stelsel. In hierdie gedeeltes is die konveksie warmte-oordrag dominant. In die katalitieseomsetter was die vogeffekte invloedryk. Die eksperimentele resultate toon ‘n duidelike vroeë toename in die gastemperature, gevolg deur ‘n tydperk van konstante temperature by nagenoeg die versadigingstemperatuur van die grootmaatdamp (verwys na as die temperatuurplato) by die katalitiese-omsetter se kern en uitlaat. Die numeries gesimuleerde gastemperature het ook hierdie gedrag getoon, maar ‘n baie hoë en skerp piek by die begin van die plato het voorgekom. Hierdie piek was nie te sien by die eksperimentele resultate nie en is toegeskryf aan nie-ewewigstoestande in die verdampingsproses, wat aandui dat die tempo van verdamping wat deur die massa-oordragmodel voorspel word te hoog is vir die model en dat dit verfyn moet word. Verdere ondersoek van die invloed van die individuele massa-oordragprosesse het getoon dat die homogene kondensasie die dominante proses is in die vorming van vloeistof in die katalitiese-omsetter. Heterogeniese kondensasie het plaasgevind, maar ‘n kleiner massa vloeistof is produseer. Die maksimum hoeveelheid vloeistof wat voorspel is om in die katalitiese-omsetter te vorm was 12 g/cm wat gelykstaande is aan ‘n film van 0.05.mm dik indien eweredig versprei oor die binneoppervlak van die monoliet. Daar was in die simulasie gevind dat beide verdamping en kondensasie benodig word om die temperatuurplato te simuleer, vanwaar die gevolgtrekking gemaak kan word dat beide prosesse wel plaasvind en dat die eerste stelling in die oorspronklike hipotese wel geldig is. Daar was egter teen die einde van die toetsperiode gevind dat beide temperature wat met en sonder vogeffekte simuleer was, die eksperimentele temperature nagevolg het, wat aandui dat die invloed van vog beperk is tot die vroeë stadiums van die katalitiese-omsetter se opwarmingstydperk. Die tweede gedeelte van die hipotese wat veronderstel dat die voggedrag ‘n vertraging in die tydsduur om omsetting te bewerkstellig veroorsaak, is dus bevind om ongeldig te wees. Die wiskundige model wat opgestel is tydens die ondersoek is weens noodsaaklikheid ‘n vereenvoudigde simulasie van komplekse termo-vloei prosesse. Dit dien as nuttige grondwerk vir verdere in-diepte ondersoeke en afronding van die teorie met betrekking tot voggedrag en die uitwerking daarvan op die tydsafhanklike temperature in ‘n uitlaatstelsel. / ENGLISH ABSTRACT: The research presented here investigates the hypothesis that condensation and subsequent evaporation of water vapour present in the exhaust gas of an internal combustion engine occur in the sections of the exhaust system between the exhaust port and the catalytic converter exit. It is further hypothesised that these two-phase moisture effects influence the transient temperature profiles in the exhaust system, and potentially cause a delay in the time it takes for the catalytic converter to reach temperatures of 200-300 °C, which are required for light-off to occur. In order to evaluate this hypothesis a transient, one-dimensional mathematical model of the thermo-fluid behaviour in the exhaust system during a cold start, including moisture effects, was created and solved by means of a computer algorithm. Heat and mass transfer theory was used to formulate the unsteady conservation equations for mass, energy and momentum. The two phase moisture effects were modelled using the analogy between heat and mass transfer, which predicts evaporation and heterogeneous condensation (the condensation of vapour against the pipe wall) due to a vapour pressure gradient between the bulk vapour and a saturated vapour at the surface of the liquid film. Homogeneous condensation (the condensation of liquid in the form of droplets in the gas stream) was also accounted for if the bulk gas temperature dropped below the bulk vapour saturation temperature. An experimental investigation was performed using two engines, a 1.6.L Volkswagen Bora and a 1.6.L Ford RoCam, in the test cells of Cape Advanced Engineering Pty (Ltd). In order to measure the gas temperatures as accurately as possible specialised radiation shielded sensors with fast time response were designed and installed in the pipe sections of the exhaust systems of both engines. The shielded sensors measured temperatures up 50 °C higher than the conventional thermocouples installed at the same positions, which is in keeping with the results predicted by the theory governing errors associated with temperature measurement in the flowing gas in the exhaust system. Comparison of the numerically simulated and experimentally measured temperatures indicated that in the sections of the exhaust system leading up to the catalytic converter the moisture has little influence on the temperature behaviour of the exhaust system. In these sections the convective heat transfer is dominant. In the catalytic converter the moisture effects were found to be influential. The experimental results clearly show an early rise in the gas temperatures, followed by a period of constant temperature at approximately the saturation temperature of the bulk vapour (referred to as the temperature plateau) at the catalytic converter mid-bed and exit. The numerically simulated gas temperatures also exhibited this plateau, but an initial very high and sharp peak in the simulated gas temperatures occurred at the start of the plateau. This was not seen in the experimental results and is attributed to non-equilibrium in the evaporation process, indicating that the rate of evaporation predicted by the mass transfer model used is too high for this application and that the model needs to be refined. Further investigation of the influence of the individual mass transfer processes indicated that the homogeneous condensation is the dominant process in the formation of liquid in the catalytic converter. Heterogeneous condensation was found to occur, but produced a smaller mass of liquid. The maximum amount of liquid predicted to form in the catalytic converter was 12 g/cm, which translates to a film 0.05 mm thick if evenly distributed over the inner surface of the monolith. In the simulation it was found that both evaporation and condensation are needed in order to simulate the temperature plateau, from which it was concluded that both these processes do occur and the first statement in the original hypothesis is valid. However, by the end of the test period temperatures simulated both with or without the moisture effects closely approached the final temperatures of the experimental investigation, indicating that the influence of the moisture is limited to the early stages of the catalytic converter warm-up. The second part of the hypothesis, postulating that the moisture behaviour caused a delay in the time taken to reach light-off temperature, is therefore concluded to be invalid. The mathematical model constructed in this research is by necessity a simplified solution to complex thermo-fluid processes. It serves as useful groundwork for further elaboration and refinement of the theory related the moisture behaviour and its influence on the transient temperatures in the exhaust system.