| As the modern aluminum electrolysis cells use to adapt the operating scheme involving low voltage, low superheat and low anode effect frequency, there is an increasing demand for the aluminum electrolysis cells to be operated under well managed heat and mass balances conditions. However, the current operations and managements on the aluminum electrolysis cells depend too much on experiences as well as controlling algorithm and always ignore the inherent dynamic characteristics, thus they cannot fulfill the increasing demand for applying accurate control on the electrolysis process. Therefore, a new control concept which intends to track the dynamic process in detail based on mathematically modeling the electrolysis process, may be the developing direction of the modern aluminum electrolysis technology.Aluminum electrolysis process is characterized by the heat and mass imbalances, which determines its inherent characters stamped by dynamics. In the present dissertation, based on plenty of reviewed literatures, every process with the potential possibility to affect heat or mass balances of a specified 200 kA prebaked cell was investigated in detail, and the mathematical models were constituted respectively. Further, these individual models were integrated into a system for predicting the working condition of the cell and its variation trends.The main contents and conclusions for the present study and the primary valuable innovations may be listed as follows:1) Thermoelectric balance theories for aluminum electrolysis cells were induced, and the theoretical methods for calculating the energy income and expense items of a cell were concluded specifically. Using the combination of theoretical methods above concerned and numerical simulation, theoretical model for evaluating the energy balance of a specified 200 kA cell was set up, and the primary liquid-solid heat transfer coefficients in the cell were determined. Subsequently, the effect of anode replacement and alumina feeding on the heat balance and mass balance in the aluminum reduction cell, and the effect of heat imbalance on electrolyte concentration were investigated and modeled mathematically respectively. In addition, the electrolyte losses were classified into several types, each of them as well as corresponding mathematical model was discussed in detail.2) The alumina dissolution process and mechanism were discussed based on the review of previous investigations. Considering the alumina dissolution process is very complicated and influenced by multiple factors, alumina dissolution process was investigated in two independent steps:alumina particle dispersion and dissolution. First, characteristics of alumina particles dispersion were correlated to quantity of added alumina and its physical properties, this made it come true to determine the number and size of agglomerates. In addition, heat transfer control model and mass transfer control model were put forward to describe alumina dissolution, thus the dissolution rates of alumina agglomerates with various size could be determined. Based on these results, for the first time, an alumina dissolution model, which has to be soluted numerically and can be used universally for industrial aluminum electrolysis cells, was built up. Alumina dissolution curve for the 200 kA cell was obtained and proved to be reasonable, which shows the reliability of the present model.3) Mechanism of bubble entrainment phenominom, which was defined as "jet drop generation caused by bubble bursting on the liquid surface the phentrained", was discussed in depth. Referring the literatural model interpretting turbulent flow entrainment, expression describing the quantity of entrained liquid was established by statictic way, where the droplet size distribution and initial velocity of droplet were under determined. As it is near impossible to obtain the droplet size distribution, a parameter E0(Ddr,DB), which designates the entrainment that consists of the droplets with less diameter than Ddr when the parent bubble diameter is Ddr, was introduced to be analyzed instead of droplet size distribution. By correlating literatural data, dimensionless discription of E0(Adr,DB) was obtained using dimensionless analysis. Further, in the same way, dimensionless initial velocity of entrained droplet was correlated to the parent bubble diameter and liquid physical properties. In addition, universal dimensionless expressions for evaluating the critical bubble diameter and droplet characteristics (diameter, number, initial) were obtained as well. In this way, univeral model for predicting the entrainment rate in various gas-liquid system the quantity of liquid produced by bubble entrainment was built up, and this model can be used in various gas-liquid systems. Using the model, for the first time, bath entrainment loss for the 200 kA cell was evaluated theoretically to be 3.2 kg/h. Most of this sort of bath loss will be adsorbed by the top cover.4) Based on the detailed understanding of the top crust formation and degradation mechanisms, the top cover of the alumimum electrolysis cell was assumed to be the Na3AlF6-AlF3-Al2O3 ternary phase, thus for the first time, top crust formation, heating, bath penetration, degradation and composition variation were modelled mathematically by phase equilibium analysis. Based on the model, dynamic process of the top cover was soluted numerically using finite element method. Numerical solution shows that, the dynamic process of top cover including crust formation, bath penetration, degradation and composition variation was controlled by heat transfer factors. For a new anode, it takes almost 14 hour from being new set to full load current. As time goes by, the thickness of top cover decreases gradually, and the decreasing rate is consistent to the measured rate literature reported. Degradation rate (around 1.2 mm/h)of the cover in the cell centre is proved to be slightly greater than that in the cell side (around 0.8 mm/h). In the first 24 hours after new anode being set, around 60 kg bath would be penetrated into the top cover, and the temperature and bath content in top cover increases rapidly. After that, the temperature field in top cover comes to be more and more steady, and the bath penetration rate keeps decreasing until a stable and very low level(0.25 kg bath/h for every anode). For the cell with 28 anodes assemble, this sort of bath loss rate accounts to 7 kg/h. However, it affects the bath mass primarily, and doesn't change the bath composition in the cell.5) As the first time, above these individual models were integrated into the dynamic electrolysis process prediction system using MATLAB code. The software system is composed of four module:data input, initialization, solution and data output. It is competent to track every parameter in the aluminum electrolysis cell and monitor the whole electrolysis process. The result for bath composition prediction in the 200kA cell show that, as the AIF3 content in the raw alumina decreases from 1 wt% to 0 wt%, daily AIF3 cosumption of the cell discussed above increases from 28.8 kg to more than 56.7 kg. The comparison between predicted conclusion and recorded data demonstrates that the error for the predicted temperature is below 0.6%, and for the predicted cryolite ratio is less than 5.5%. In addition, as the increase of period to be tracked, the predictive accuracy has the potential to be improved. In conclusion, the present "Dynamic Prediction System for Aluminum Electrolysis Process" has the necessary function and enough accuracy for industial application. |
PDF Full Text Request