Ilmenite smelter freeze lining monitoring by thermocouple measurements: Model results

Ilmenite smelting furnaces must be operated with a freeze lining of frozen slag to protect sidewall refractories from chemical attack by liquid slag. The process has two main control objectives: Firstly the slag TiO2 content must be controlled according to specification, and secondly a safe freeze lining thickness must be maintained. This text only concerns itself with the second objective. Freeze lining thickness cannot be measured directly due to conditions inside the furnace. It is therefore inferred from signals of thermocouples installed into the furnace sidewall. These signals are used as the 'controlled variables' in the freeze lining thickness control loop. The dynamic characteristics of these signals are the focus of this text. During this study a model of the process was constructed to study interactions between the slag bath and the freeze lining. This model was used to conduct simulation-type experiments to study the influence of various parameters (temperature, slag composition, ilmenite feed rate, reductant feed rate, power input, etc.) on the slag bath, freeze lining, metal bath, furnace wall and other parts of the process. In the case of the furnace wall, the modelling results were first plotted in the form of time-dependent isotherms. The dynamic thermal behavior of the furnace wall became evident from these results. To illustrate the dynamics of sidewall thermocouple signals, the results were also plotted as time lags between events occurring in the slag bath, and the time when a thermocouple registered a temperature change. The events studied included sudden changes in slag bath state (composition and temperature), changes in operating setpoints (ilmenite feed rate, reductant feed rate, power input), and operational errors (complete loss of reductant or of all feed). All of the result sets showed significant lags between occurrences in the slag bath and temperature changes in the refractory wall. The delays are such that it appears impractical to rely on sidewall thermocouple signals for warning of impending damage to the refractory wall, or for purposes of actively controlling freeze lining thickness. Some measures are suggested to improve the usefulness of these signals during furnace operation.