Heat transfer through the sidewall accounts for a significant proportion of the energy loss from an aluminium reduction cell. At the same time, the ledge formed on the sidewall has important consequences with regards to the cell service life and the dynamic heat balance during various disturbances associated with the smelting operation such as alumina feeding, anode changing, metal tapping and anode effect. Several studies have been carried out to investigate the ledge heat transfer. However, the reported heat transfer coefficients in the literature not only vary over a wide range of values but provide insufficient information on the ledge heat transfer. A heat transfer probe and measurement techniques were developed for studying the ledge heat transfer in a full-scale 3-D air-water model. Quantitative measurements were conducted to determine the bath/ledge heat transfer characteristics at various positions, and under different operating conditions such as anode-ledge distance, current density and bath depth. Variation of the heat transfer were also examined as a function of the anode bottom inclination, the anode slot width and the position on the side ledge relative to the anode slot. The results illustrate that the ledge profile in an operating cell will take on a different shape in compliance with the heat transfer variation. A similitude analysis was carried out to interpret the measured results in a meaningful manner for use in a reduction cell. As a result, the heat transfer coefficient in industrial cells under various operating parameters, and at different positions on the side ledge, can be estimated using the empirical correlations presented. Gas bubble behaviour and bubble impingement on the side ledge were observed in the water model. Observations made on the 2-D and 3-D water models indicate that anode gas evolution in the 3-D model cell will reflect more closely the flow pattern in actual cells and hence provide more reliable quantitative results. A simple 2-D thermal model for the prediction of ledge thickness and profile as a moving boundary was developed and solved simply and efficiently with a commercial spreadsheet software using the finite difference method. The ledge profile was predicted using the ledge heat transfer coefficients measured extensively from the full-scale 3-D physical model. The results show that the ledge shape is highly sensitive to the positional variation of the heat transfer coefficient. It is also shown that the ledge heat transfer coefficients obtained from industrial measurements assuming 1-D heat flow are much lower than the actual values in a Hall-Heroult cell. A transient thermal model derived by considering the Stefan problem$/sp[*]$ for the sidewall/ledge region was developed. A fixed-grid and deforming-grid spacing were respectively superimposed on the sidewall and the ledge region in order to track the moving front of the phase change zone. Various aspects of the process dynamics with respect to the variation of ledge thickness and sidewall shell temperature were presented. The model considered dynamic heat loss through the sidewall which results in a closer approximation to the real situation. ftn$/sp[*]$N.B.: In the strict sense the problem of the ledge is not a classical Stefan problem. The classical Stefan problem involves conduction on both sides of the interface. The ledge problem involves conduction on the side and convection on the other. / Subscription resource available via Digital Dissertations only.

Heat transfer through the sidewall accounts for a significant proportion of the energy loss from an aluminium reduction cell. At the same time, the ledge formed on the sidewall has important consequences with regards to the cell service life and the dynamic heat balance during various disturbances associated with the smelting operation such as alumina feeding, anode changing, metal tapping and anode effect. Several studies have been carried out to investigate the ledge heat transfer. However, the reported heat transfer coefficients in the literature not only vary over a wide range of values but provide insufficient information on the ledge heat transfer. A heat transfer probe and measurement techniques were developed for studying the ledge heat transfer in a full-scale 3-D air-water model. Quantitative measurements were conducted to determine the bath/ledge heat transfer characteristics at various positions, and under different operating conditions such as anode-ledge distance, current density and bath depth. Variation of the heat transfer were also examined as a function of the anode bottom inclination, the anode slot width and the position on the side ledge relative to the anode slot. The results illustrate that the ledge profile in an operating cell will take on a different shape in compliance with the heat transfer variation. A similitude analysis was carried out to interpret the measured results in a meaningful manner for use in a reduction cell. As a result, the heat transfer coefficient in industrial cells under various operating parameters, and at different positions on the side ledge, can be estimated using the empirical correlations presented. Gas bubble behaviour and bubble impingement on the side ledge were observed in the water model. Observations made on the 2-D and 3-D water models indicate that anode gas evolution in the 3-D model cell will reflect more closely the flow pattern in actual cells and hence provide more reliable quantitative results. A simple 2-D thermal model for the prediction of ledge thickness and profile as a moving boundary was developed and solved simply and efficiently with a commercial spreadsheet software using the finite difference method. The ledge profile was predicted using the ledge heat transfer coefficients measured extensively from the full-scale 3-D physical model. The results show that the ledge shape is highly sensitive to the positional variation of the heat transfer coefficient. It is also shown that the ledge heat transfer coefficients obtained from industrial measurements assuming 1-D heat flow are much lower than the actual values in a Hall-Heroult cell. A transient thermal model derived by considering the Stefan problem$/sp[*]$ for the sidewall/ledge region was developed. A fixed-grid and deforming-grid spacing were respectively superimposed on the sidewall and the ledge region in order to track the moving front of the phase change zone. Various aspects of the process dynamics with respect to the variation of ledge thickness and sidewall shell temperature were presented. The model considered dynamic heat loss through the sidewall which results in a closer approximation to the real situation. ftn$/sp[*]$N.B.: In the strict sense the problem of the ledge is not a classical Stefan problem. The classical Stefan problem involves conduction on both sides of the interface. The ledge problem involves conduction on the side and convection on the other. / Subscription resource available via Digital Dissertations only.

Heat transfer through the sidewall accounts for a significant proportion of the energy loss from an aluminium reduction cell. At the same time, the ledge formed on the sidewall has important consequences with regards to the cell service life and the dynamic heat balance during various disturbances associated with the smelting operation such as alumina feeding, anode changing, metal tapping and anode effect. Several studies have been carried out to investigate the ledge heat transfer. However, the reported heat transfer coefficients in the literature not only vary over a wide range of values but provide insufficient information on the ledge heat transfer. A heat transfer probe and measurement techniques were developed for studying the ledge heat transfer in a full-scale 3-D air-water model. Quantitative measurements were conducted to determine the bath/ledge heat transfer characteristics at various positions, and under different operating conditions such as anode-ledge distance, current density and bath depth. Variation of the heat transfer were also examined as a function of the anode bottom inclination, the anode slot width and the position on the side ledge relative to the anode slot. The results illustrate that the ledge profile in an operating cell will take on a different shape in compliance with the heat transfer variation. A similitude analysis was carried out to interpret the measured results in a meaningful manner for use in a reduction cell. As a result, the heat transfer coefficient in industrial cells under various operating parameters, and at different positions on the side ledge, can be estimated using the empirical correlations presented. Gas bubble behaviour and bubble impingement on the side ledge were observed in the water model. Observations made on the 2-D and 3-D water models indicate that anode gas evolution in the 3-D model cell will reflect more closely the flow pattern in actual cells and hence provide more reliable quantitative results. A simple 2-D thermal model for the prediction of ledge thickness and profile as a moving boundary was developed and solved simply and efficiently with a commercial spreadsheet software using the finite difference method. The ledge profile was predicted using the ledge heat transfer coefficients measured extensively from the full-scale 3-D physical model. The results show that the ledge shape is highly sensitive to the positional variation of the heat transfer coefficient. It is also shown that the ledge heat transfer coefficients obtained from industrial measurements assuming 1-D heat flow are much lower than the actual values in a Hall-Heroult cell. A transient thermal model derived by considering the Stefan problem$/sp[*]$ for the sidewall/ledge region was developed. A fixed-grid and deforming-grid spacing were respectively superimposed on the sidewall and the ledge region in order to track the moving front of the phase change zone. Various aspects of the process dynamics with respect to the variation of ledge thickness and sidewall shell temperature were presented. The model considered dynamic heat loss through the sidewall which results in a closer approximation to the real situation. ftn$/sp[*]$N.B.: In the strict sense the problem of the ledge is not a classical Stefan problem. The classical Stefan problem involves conduction on both sides of the interface. The ledge problem involves conduction on the side and convection on the other. / Subscription resource available via Digital Dissertations only.

Heat transfer through the sidewall accounts for a significant proportion of the energy loss from an aluminium reduction cell. At the same time, the ledge formed on the sidewall has important consequences with regards to the cell service life and the dynamic heat balance during various disturbances associated with the smelting operation such as alumina feeding, anode changing, metal tapping and anode effect. Several studies have been carried out to investigate the ledge heat transfer. However, the reported heat transfer coefficients in the literature not only vary over a wide range of values but provide insufficient information on the ledge heat transfer. A heat transfer probe and measurement techniques were developed for studying the ledge heat transfer in a full-scale 3-D air-water model. Quantitative measurements were conducted to determine the bath/ledge heat transfer characteristics at various positions, and under different operating conditions such as anode-ledge distance, current density and bath depth. Variation of the heat transfer were also examined as a function of the anode bottom inclination, the anode slot width and the position on the side ledge relative to the anode slot. The results illustrate that the ledge profile in an operating cell will take on a different shape in compliance with the heat transfer variation. A similitude analysis was carried out to interpret the measured results in a meaningful manner for use in a reduction cell. As a result, the heat transfer coefficient in industrial cells under various operating parameters, and at different positions on the side ledge, can be estimated using the empirical correlations presented. Gas bubble behaviour and bubble impingement on the side ledge were observed in the water model. Observations made on the 2-D and 3-D water models indicate that anode gas evolution in the 3-D model cell will reflect more closely the flow pattern in actual cells and hence provide more reliable quantitative results. A simple 2-D thermal model for the prediction of ledge thickness and profile as a moving boundary was developed and solved simply and efficiently with a commercial spreadsheet software using the finite difference method. The ledge profile was predicted using the ledge heat transfer coefficients measured extensively from the full-scale 3-D physical model. The results show that the ledge shape is highly sensitive to the positional variation of the heat transfer coefficient. It is also shown that the ledge heat transfer coefficients obtained from industrial measurements assuming 1-D heat flow are much lower than the actual values in a Hall-Heroult cell. A transient thermal model derived by considering the Stefan problem$/sp[*]$ for the sidewall/ledge region was developed. A fixed-grid and deforming-grid spacing were respectively superimposed on the sidewall and the ledge region in order to track the moving front of the phase change zone. Various aspects of the process dynamics with respect to the variation of ledge thickness and sidewall shell temperature were presented. The model considered dynamic heat loss through the sidewall which results in a closer approximation to the real situation. ftn$/sp[*]$N.B.: In the strict sense the problem of the ledge is not a classical Stefan problem. The classical Stefan problem involves conduction on both sides of the interface. The ledge problem involves conduction on the side and convection on the other. / Subscription resource available via Digital Dissertations only.

Heat transfer through the sidewall accounts for a significant proportion of the energy loss from an aluminium reduction cell. At the same time, the ledge formed on the sidewall has important consequences with regards to the cell service life and the dynamic heat balance during various disturbances associated with the smelting operation such as alumina feeding, anode changing, metal tapping and anode effect. Several studies have been carried out to investigate the ledge heat transfer. However, the reported heat transfer coefficients in the literature not only vary over a wide range of values but provide insufficient information on the ledge heat transfer. A heat transfer probe and measurement techniques were developed for studying the ledge heat transfer in a full-scale 3-D air-water model. Quantitative measurements were conducted to determine the bath/ledge heat transfer characteristics at various positions, and under different operating conditions such as anode-ledge distance, current density and bath depth. Variation of the heat transfer were also examined as a function of the anode bottom inclination, the anode slot width and the position on the side ledge relative to the anode slot. The results illustrate that the ledge profile in an operating cell will take on a different shape in compliance with the heat transfer variation. A similitude analysis was carried out to interpret the measured results in a meaningful manner for use in a reduction cell. As a result, the heat transfer coefficient in industrial cells under various operating parameters, and at different positions on the side ledge, can be estimated using the empirical correlations presented. Gas bubble behaviour and bubble impingement on the side ledge were observed in the water model. Observations made on the 2-D and 3-D water models indicate that anode gas evolution in the 3-D model cell will reflect more closely the flow pattern in actual cells and hence provide more reliable quantitative results. A simple 2-D thermal model for the prediction of ledge thickness and profile as a moving boundary was developed and solved simply and efficiently with a commercial spreadsheet software using the finite difference method. The ledge profile was predicted using the ledge heat transfer coefficients measured extensively from the full-scale 3-D physical model. The results show that the ledge shape is highly sensitive to the positional variation of the heat transfer coefficient. It is also shown that the ledge heat transfer coefficients obtained from industrial measurements assuming 1-D heat flow are much lower than the actual values in a Hall-Heroult cell. A transient thermal model derived by considering the Stefan problem$/sp[*]$ for the sidewall/ledge region was developed. A fixed-grid and deforming-grid spacing were respectively superimposed on the sidewall and the ledge region in order to track the moving front of the phase change zone. Various aspects of the process dynamics with respect to the variation of ledge thickness and sidewall shell temperature were presented. The model considered dynamic heat loss through the sidewall which results in a closer approximation to the real situation. ftn$/sp[*]$N.B.: In the strict sense the problem of the ledge is not a classical Stefan problem. The classical Stefan problem involves conduction on both sides of the interface. The ledge problem involves conduction on the side and convection on the other. / Subscription resource available via Digital Dissertations only.

Master of Science / Department of Mechanical and Nuclear Engineering / Douglas S. McGregor / Purification of cadmium, zinc and tellurium was attempted to improve the quality of cadmium-zinc-telluride (CdZnTe) crystal growth. Specifically, vacuum distillation, zone refining and H[subscript]2 gas flow assisted zone refining were all investigated as methods to purify the constituent elements of CdZnTe. A unique multi-chamber ampoule was used to enable a purification sequence starting with double vacuum distillation followed by zone refining all without sample handling after the initial step. Modifications due to unique material properties of Cd and Zn were developed. Glow discharge mass spectroscopy (GDMS) analysis was used to measure impurity concentrations of 74 elements. Cd purification using vacuum distillation proved to be an effective method to reduce the impurity level of 5N starting material to a purity between the range of 6N5 and 7N5, as measured using GDMS and laser ablation mass spectroscopy. Combined Zn double vacuum distillation and zone refining in an enclosed Ar atmosphere using 5N starting material yielded material with a purity between the range of 5N8 to 6N8. Tellurium purification using combined double vacuum distillation followed by zone refining under continuous H[subscript]2 flow of 4N specified raw material resulted in high purity tellurium between the range of 6N3 and 7N4.