| Electrolytic cells are used to produce aluminum from alumina ore. These cells use a carbon anode, a cryolitic bath, and a carbon containment vessel, which also serves as the cathode. The process is not completely understood because of the limited number of measurements and its complexity.; A finite element model was developed that simulates the electrical, thermal, and chemical phenomena occurring within the cell. The model was verified by comparing its output to data presented in the literature, and by comparing its output to data gathered from operating reduction cells at Century Aluminum in Ravenswood, WV. The model was used to study the reaction of the system to various changes in process variables, such as the cell voltage, the bath ratio, the alumina concentration, the metal pad depth and the current flow. Changes in these variables had a significant impact on the operating point of the cell. Because deviations from the target level for the process variables are often temporary, short-term changes in the operating state were quantified so that these effects could be accounted for when determining appropriate control actions.; Two simplified models were also developed. The first model was used to estimate the alumina concentration in the bath, and it included the chemical model and a scalar version of the electrical model. This estimator predicts occurrences of high and low alumina concentrations in operating reduction cells with a high level of accuracy. The second model was developed to estimate the bath temperature and the bath ratio. It includes the chemical model, a scalar version of the electrical model, and a thermal model that is made up of a few one-dimensional elements. This model does not appear to predict fluctuations in the temperature with a high degree of accuracy. However, using daily temperature measurements, the model is capable of accurately predicting the bath ratio.; In conclusion, the finite element model provides a useful tool for studying various process interactions, and the alumina concentration and ratio estimators provide the ability to reduce variability in these two important process variables. |
