As in many thermal processing technologies, there is a delicate balance between productivity and quality during ingot cooling process. Higher cooling velocities increase productivity but also create higher temperature gradients inside the ingot. Such a fast cooling does not leave sufficient time to establish the equilibrium within the solid, thus in the most affected surface layer the composition and crystalline structure are different from those in the bulk metal. The heat flux plays a particular role for the production of alloys where different melting points and complex structure formation -depending on the temperature- are present. To prevent the two worst cases - cracking and remelting - during cooling a balance has to be found between good productivity and quality on the one side but also a high security on the other side. To avoid the negative effects of cracking and remelting, it is necessary to determine the heat flux as a function of the influencing parameters and to control the cooling in order to obtain a maximal productivity with the required quality. There is no widely accepted method for the quantitative characterization of the cooling capacity of the water. The cooling may be characterized by the heat transfer coefficients measured in different boiling regimes on the surface or directly by the heat flux. As the fluid flow and heat transfer phenomena are very complicate in the falling liquid-vapour film, using the heat transfer coefficients does not necessarily help the understanding of the underlying mechanisms. However, the correlation between the surface heat flux and temperature for a given surface roughness and water quality includes all the relevant information about the cooling process. Thus in the present project our main objective was to determine the surface heat flux for a water cooled ingot as a function of the water quality in order to provide a tool for assuring a uniform quality in the cast-shop. A further objective was to improve the understanding of the flow of boiling liquid film along a solid surface. The first challenge in the project was the development of a surface temperature and heat flux measurement method which does not disturb the ingot surface. A heat flux sensor attached onto the surface would have negative effects on the film-nucleate boiling process and increase the surface roughness that affects the nucleation of bubbles. Furthermore it was important that the measurement method be fast responding and sensitive enough to detect very rapid and weak variations in the surface temperature. Thus an innovative surface temperature sensing method - using an open-tip thermocouple - was developed. This sensor was inserted into a null-point cavity from the backside of the aluminium ingot. The open-tip sensor combined with the null-point cavity forms a null-point calorimeter. Using the inverse solution of the general heat conduction equation, it is possible to determine the surface heat flux from the measured temperature history. For the analytical solution of the heat conduction equation in one dimension we used the so called Cook-Felderman equation. It was also investigated how well the assumptions of the Cook-Felderman equation are applicable to the actual conditions in our experimental setup. For this purpose a computer program was developed. The algorithm uses the finite volume approach; the program was written in C++ and calculates the transient 2D temperature distribution in the solid sample. The two-dimensional model uses the measured temperature histories as boundary conditions. The computer predictions were used to determine the horizontal and the vertical heat flux components during the cooling process. Furthermore it was analyzed how the heat flux results obtained analytically by the Cook-Felderman equation compare to the more detailed numerical solutions. The cooling process is strongly influenced by the different water types and different ingot surfaces. Thus the investigations were focused on the difference between several water types such as tap water (conductivity: 90fiS), plant water (conductivity: 1900/uS') and water with higher conductivity (conductivity: 3200/uS). These water types were tested on different ingot surfaces to eliminate the influence of the surface structure. In addition, the effects of some water additives and a water-oil mixture on the surface heat flux were investigated. The surface temperature measurements were accompanied by a visualization of the surface boiling effects. Furthermore the difference between different ingot surfaces was analyzed. The tests made on machined and a rough surface ingots provided information about the differences in the temperature cooling curves as well as information about the structure of the cooling water film. All collected measurement and visualization data were analyzed to explain the surface flow and boiling effects during the cooling process. The dynamic structure of the falling water film was investigated. Certain characteristics in the temperature cooling curves can be explained with the structure of the water film. We achieved a very good experimental repeatability. We found that the repeatability of the boiling phenomena itself depends on the measurement position (height along the ingot). It was observed that there are more fluctuations on a machined ingot than on a rough surface ingot. Even two different rough plates have different heat flux results. The tests using different water types have shown that there is almost no difference between tap and plant water, neither on a machined nor on a rough ingot. A difference could only be found between tap water and high conductivity water.
| Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:QCU.750 |
| Date | January 2003 |
| Creators | Meenken, Thomas |
| Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
| Detected Language | English |
| Type | Thèse ou mémoire de l'UQAC, NonPeerReviewed |
| Format | application/pdf |
| Relation | http://constellation.uqac.ca/750/ |
Page generated in 0.0026 seconds