Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Le développement de nouveaux outils de diagnostic est primordial pour assurer la maintenance des grands alternateurs. La tension présente entre les deux extrémités de l’arbre est considérée comme une quantité indésirable car elle peut induire des dégradations dans les paliers et, est ainsi mesurée de manière préventive. Néanmoins, son analyse fine pourrait s’avérer utile pour la détection de défauts. Comme la tension d’arbre est présente même en fonctionnement sain de l’alternateur, il est nécessaire de découpler toutes les causes qui peuvent l’engendrer. L’objectif de ce travail consiste donc à mener une investigation sur les multiples sources à l’origine de cette tension d’arbre afin de pouvoir quantifier leurs impacts respectifs, mais également de déterminer la signature d’un défaut de court-circuit entre spires ou d’excentricité rotorique. Une première étude a porté sur l’analyse de machines simplifiées dans le but d’isoler l’impact d’un défaut donné. Deux approches ont été développées, la première à partir d’un modèle analytique et la seconde à l’aide de la modélisation par éléments finis. Les conclusions émises par ces deux approches ont été analysées et comparées montrant des concordances qualitatives intéressantes. Ensuite, deux turboalternateurs de grande puissance (900 MW et 1300 MW) ont été modélisés par une approche numérique en 2D. Une démarche d’investigation systématique a été menée pour caractériser l’influence de plusieurs causes afin de cerner et quantifier leurs impacts sur la tension d’arbre. Enfin, une modélisation tridimensionnelle d’un des 2 alternateurs a permis de montrer l’effet des têtes de bobines et ce, en régime sain comme en défaut. / The development of new diagnostic tools is compulsory for the maintenance of large generators. The voltage which appears between the two ends of the shaft of the machine is an undesirable quantity because it can induce degradations in the bearings and thus, it is measured in a preventive manner. However, its analysis could be useful for the detection of some given defects. Since this voltage is non-nil even in healthy operation, it is necessary to be able to dissociate all the causes which can generate it. The objective of this work is to conduct an investigation on the multiple sources which can induce this shaft voltage in order to quantify their respective impacts, but also to highlight more specifically the signature of an interturn short circuit or rotor eccentricity. A first study was focused on the analysis of simplified machines in order to isolate the impact of the inherent effect of a given defect without interaction with other parameters. Two approaches were then developed, the first from an analytical approach and the second one using a finite element modeling. Conclusions of these approaches have been analyzed and compared thus showing interesting qualitative concordances. Then, two high power turbogenerators (900 MW and 1300 MW) were modeled using a numerical approach in 2D. An enquiry process was conducted to characterize the influence of several causes in order to identify and quantify their impacts on the shaft voltage. Finally, a three-dimensional modeling of one of these alternators showed a significant effect of the rotor and stator end windings on this shaft voltage either in healthy and faulty operating.

It is possible to drain slender containers filled with wetting liquids via capillary flows along the interior corners of the container. Usually the well established equations governing such flows demand numerical techniques. In the case of container draining unique boundary conditions resulting from local section geometry allow for a quasi-steady assumption and in turn permit analytical solutions. The quasi-steady assumption may also be employed for certain problems in which the corner flows cause passive capillary migration of the fluid within the container. The analytic solutions are useful because of the ease in which geometric effects may be observed. Container draining and capillary migration by means of corner flows are studied in a variety of container geometries. It is shown that careful selection of cross sectional shape can be used to maximize drain rates and minimize capillary migration times. Three-dimensional effects for these flows are investigated in tapering containers. Some simple micro-scale experiments are reported that provide confidence in the assumptions and application of the important boundary conditions that enable the solutions.

Graduation date: 1998

The study uses the Phase-field method to simulate the bubble behavior of liquid-solid interface in the solidification.The flow use the mass equation,momentum equation,and energy equation for simulating the variation of bubble. This pattern considers that three-phase of solid, liquid,and gas coexist with the different density and viscosity coefficient,and the external force considers surface tension and the gravity force. In addition,the mass transfer also can¡¦t neglect around interface. The result discuss the production of pore when the velocity of solidification is fast,but bubble leaves when the velocity of solidification is slow.

Thesis (Ph. D.)--University of Michigan. / Cover title. "Reprinted from the Botanical gazette, 50:257-284, no. 4, October, 1910."

No description available.

The surface tension of an interface separating two bulk phases is one of the most widely studied properties in surface science research. The importance of surface or interfacial tension is reflected in the diverse number of applications which are influenced by surface tension related effects. This thesis represents a comprehensive experimental and theoretical investigation on molecular adsorption and surface tension from a class of organic compounds in aqueous solutions. The research illustrates the effect of both liquid and vapor phase adsorption on the interfacial properties. Adsorption from both sides of the vapor/liquid interface is considered simultaneously rather than exclusive of one another, which has been the conventional practice. In the experimental study, the surface tension of a number of different volatile organic compounds is measured using the Axisymmetric Drop Shape Analysis-Profile (ADSA-P) method. The experiments were performed in a controlled environment under conditions where the surface tension can be affected by both vapor and liquid phase adsorption. The vapor phase was exerted by the presence of an environment solution containing the same organic component as in the drop solution. The results show that initially the surface tension is influenced by the organic concentration in both the liquid and the vapor phase. At the final steady-state the liquid phase becomes less important and the primary factor influencing the surface tension is the vapor phase concentration. The ADSA-P technique is verified by reproducing a select number of cases using the Wilhelmy plate method. A possible consequence of the surface tension phenomenon is illustrated through time-dependent contact angle experiments. The behavior of the interface at steady-state conditions is investigated by measuring the surface tension response to a change in drop volume. It is concluded that the organic compounds considered in the current study may represent a rather general group of molecules whose surface behavior is unique to that of many conventional surfactants. In the theoretical study an empirical model is proposed to describe the relation between the steady-state surface tension and the concentration of the environment and drop solutions. The results confirm the experimental observation that the final steady-state surface tension is determined primarily by the organic concentration in the vapor phase. In addition, a modified adsorption isotherm is developed to account for simultaneous adsorption from both sides of the vapor/liquid interface at steady-state conditions. The derivation is based upon the classic Langmuir analysis, and the new equation is consistent with the Langmuir isotherm under traditional conditions where adsorption occurs from one side of the interface. The modified isotherm is shown to be consistent with the experimental data and is used to generate the equilibrium parameters for three of the systems studied in this research. The adsorption isotherm is then extended to model the dynamic adsorption process through the creation of a new kinetic transfer equation. As with the adsorption isotherm, the transfer equation is based on Langmuir kinetics and is capable of simulating adsorption from both sides of the interface during surface equilibration. The kinetic transfer equation is validated against experimental data from two systems which exhibit a transfer-controlled adsorption mechanism. The theoretical predictions from the transfer equation fit well with the experimental data for both systems. However, significant variability is observed in the least squares estimates of the kinetic rate constants. The variability is attributed to the limitations of empirical models that utilize adjustable fitting parameters to optimize the model predictions, and the wide range of surfactant concentrations studied. Specific concentration regions are identified where the variability in the rate constants is minimal and thus, where the model is most appropriate.

The DC casting process used for the production of cast aluminium alloy products intended for processing by rolling, extrusion or forging is an economically important process with approximately 10 million tonnes of DC cast product being produced annually world wide [1]. Process productivity, particularly with respect to elimination of casting defects and hence process scrap is an important factor to DC cast product producers. The literature reporting the DC casting process, particularly with respect to the formation of defects on the cast surface, is reviewed and the mechanisms for the formation of such defects examined. A universally understood and accepted explanation was found for only one of the normal surface defects encountered in practice. A number of samples of commercially cast DC products were subject to detailed cast surface examination, particularly surface microstructures. The results of this examination and the literature survey identified molten metal surface tension as a possible contributing factor affecting the molten metal meniscus stability within the DC casting mould. Meniscus instability is linked with the formation of surface defects. The literature on surface tension of aluminium alloys provided only limited information hence the surface tension of a range of binary and ternary alloys, including commercial alloys was determined, producing data hitherto not available. Of the common alloying elements used in commercial aluminium alloys, iron and magnesium were found to significantly reduce the surface tension. Surface fracture also resulted in a reduction in surface tension for the majority of alloys tested. The surface tension data is combined with mould thermal and physical model calculations to propose a mechanism for the formation of the cast surface defects. The model proposes a maximum stable size for the meniscus according to the alloy and mould conditions. Conditions outside these limits result in meniscus instability and the formation of cast surface defects. The model suggests possible operating changes that may reduce the incidence of surface defect formation. The work also identifies a number of areas requiring further investigation before major practical process changes aimed at cast surface defect elimination, may be formulated.

The DC casting process used for the production of cast aluminium alloy products intended for processing by rolling, extrusion or forging is an economically important process with approximately 10 million tonnes of DC cast product being produced annually world wide [1]. Process productivity, particularly with respect to elimination of casting defects and hence process scrap is an important factor to DC cast product producers. The literature reporting the DC casting process, particularly with respect to the formation of defects on the cast surface, is reviewed and the mechanisms for the formation of such defects examined. A universally understood and accepted explanation was found for only one of the normal surface defects encountered in practice. A number of samples of commercially cast DC products were subject to detailed cast surface examination, particularly surface microstructures. The results of this examination and the literature survey identified molten metal surface tension as a possible contributing factor affecting the molten metal meniscus stability within the DC casting mould. Meniscus instability is linked with the formation of surface defects. The literature on surface tension of aluminium alloys provided only limited information hence the surface tension of a range of binary and ternary alloys, including commercial alloys was determined, producing data hitherto not available. Of the common alloying elements used in commercial aluminium alloys, iron and magnesium were found to significantly reduce the surface tension. Surface fracture also resulted in a reduction in surface tension for the majority of alloys tested. The surface tension data is combined with mould thermal and physical model calculations to propose a mechanism for the formation of the cast surface defects. The model proposes a maximum stable size for the meniscus according to the alloy and mould conditions. Conditions outside these limits result in meniscus instability and the formation of cast surface defects. The model suggests possible operating changes that may reduce the incidence of surface defect formation. The work also identifies a number of areas requiring further investigation before major practical process changes aimed at cast surface defect elimination, may be formulated.