Simulation And Optimization Of Stirred Tank Crystallizer Based On CFD-PBM Coupled Model

Crystallization is an important industrial operation technology,which is widely used in chemical,pharmaceutical and semiconductor fields.Due to greatly dependence of crystallization on material and mechanical properties,it is necessary to control quality of crystalline products.Average particle size and crystal size distribution are important indexes to evaluate final crystal quality,and stirred tank crystallizer is widely used in industrial production.In this paper,cooling crystallization process of acetaminophen-ethanol in stirred tank crystallizer was simulated,and effects of different processing conditions on crystal size distribution and average particle size were studied.Firstly,based on computational fluid dynamics(CFD),stirred tank crystallizer was studied.Euler model was used to simulate distribution of particles with specified size in crystallizer,and multiple reference frame(MRF)method was used to simulate impeller rotation.Influence of different agitator blade types and installation height was studied.The simulation results show that distribution of particle volume fraction is more uniform in stirred tank crystallizer with six straight blade opening turbine impeller,four straight blade opening turbine impeller and six straight blade disc turbine impeller.And at the bottom of crystallizer,particle distribution with four straight blade opening turbine impeller is the most uniform.Crystallizer with agitator installation height of 20 mm,30mm and 40 mm was simulated.With increase of installation height,particle distribution at the top of crystallizer becomes more uniform,but at the bottom of crystallizer,particle volume fraction increases and particle deposition increases.Then,cooling crystallization of acetaminophen was simulated based on population balance model,and PBM was solved by orthogonal moment method.The effects of different particle sizes and different cooling rates were investigated.The simulation results show that in crystallization process of paracetamol,crystal growth and initial nucleation were dominant in the early stage of crystallization(t < 4800s),and secondary nucleation was dominant in the late stage of crystallization(t ≥ 4800s).Particle size of seed has a great influence on particle number.Particle number is increased by reduction of particle size of seed.Particle size of seed has little effect on the average particle size,and there is little difference in average particle size with different grain size.Large cooling rate k of linear cooling curve can effectively avoid sudden increase of nucleation rate,thus avoiding generation of large amounts of particles,which is conducive to obtaining products with large average particle size.Finally,based on CFD-PBM coupling model batch cooling crystallization in stirred tank crystallizer was studied.population balance model considers growth,primary nucleation and secondary nucleation process,and is solved by discrete method.The effects of different stirring speed,cooling curve,seed and breakage process on crystal size distribution were investigated.The simulation results show that with increase of stirring speed,peak value of particle size distribution increases,number of product particles increases,and average particle size decreases.The higher cooling rate of linear cooling curve,the smaller peak value of particle size distribution.Stepwise cooling curve and secondary cooling process are beneficial to reduce burst nucleation and produce larger average particle size distribution.CSD with large average particle size and large number of particles can be obtained by adding seeds appropriately.The prediction ability of CSD can be improved by adopting suitable breakage model and considering more comprehensive crystallization kinetics model.