Numerical Simulation On Dynamics Model Of Alumina Concentration And Crystal Particle Size During Seeded Precipitation Process

The seeded precipitation is one of the key steps in the Bayer processfor the production of alumina. It has great influence on alumina product'soutputs and quality and also has indirect effect on other procedures.Owing to the difference of the bauxite with foreign one, China has itsown precipitation process. The study of the mechanics and dynamics ofthe precipitation with Chinese character is significant.The laboratory experiment, industrial test and numerical simulationof the precipitation are synchronously applied to this research. Based onthe dynamics law and population balance, the concentration model andparticle size model of the precipitation process were developed withMATLAB/Simulink. The models were used to analyze the variation ofconcentration and particle size during the gibbsite precipitation process.Based on the simulation, the temperature regulation and process flowwere optimized, which facilitate the control and prediction of theprecipitation process. The main achievements and conclusions are asfollows:(1) The gibbsite precipitation process concentration model wasdeveloped. The relative error between the model's solution and theindustrial test data is within 10%. Based on the computation, the resultswhich include decreasing the initial temperature and the end temperaturewere beneficial to increase the precipitation rate and the temperature'sdifferent phase fluctuation has more influence than the positive phasefluctuation were obtained.(2) The particle size distribution model of ideal growth process ofgibbsite precipitation from supersaturated sodium aluminate solution wasdeveloped based on the knowledge of characteristics of growth process.The simulation results of the model were well agree with theexperimental data under the following conditions: relatively high stirringrate, temperature between 55～75℃,α_k,0, 1.5～2.1, solid content about600 g·l^-1 and caustic concentration, 150～180g·l^-1. The relative errorbetween the model's solution and the experimental data is within 15%.The qualitative information of the sub steps of the gibbsite precipitation process were obtained by comparing the difference between thedistribution curves of prediction and detection.(3) With the full characteristics of the particle size variation oftwo-stage gibbsite precipitation process, considering agglomerationprocess as the key point, the combined model of crystal's nucleation,growth and agglomeration was developed to confirm the value ranges ofthe parameters at different experiment condition and to predict theparticle size distribution. The relative error between the model'sprediction and industrial test data satisfies the requirement, i.e., within10%.(4) Continuous multi-decomposer particle size model and productclassifier model were developed, which provide reference for adjustingthe process's temperature regulation and optimization of the process flow.