Modeling Of Ethylene Gas-Phase Polymerization Process

Process simulation is an important method to improve industrial processes.An appropriate model can be applied to optimize the operation conditions and design new grade.In this thesis,appropriate mathematic model was established for an industrial gas-phase ethylene polymerization process.It involved kinetic model for dual copolymerization of ethylene withα-olefin and reactor model for fluidized bed.This model simulated the ethylene gas-phase polymerization process very well,and was used to develop new grade.More active sites kinetic model was established for dual copolymerization of ethylene withα-olefin.First,the effects of kinetic constants on polymerization yield,polymer density and molecular weight(Mw)were analyzed;it was found that chain propagation, chain deactivation and chain transfer to hydrogen kinetic constants were the most important factors for yield,density and Mw separately.The chain propagation,chain deactivation and chain transfer to hydrogen kinetic constants were adjusted to establish the single active site kinetic model to accord with plant process based on the plant data.Then, the molecular weight distribution curve(GPC curve)of plant product was fitted by Flory function,and it was found that five active sites curve was proper for fitting the GPC curve. So,five active sites kinetic model was established for ethylene copolymerization withα-olefin,and this kinetic model could simulated the yield,density and molecular weight distribution accurately.Temperature distribution is not uniform in gas phase polymer reactor,compared with slurry polymerization and solution polymerization.Multi compartments method was adopted to describe the ununiformity of temperature distribution in the reactor.It was established separately that well-mixed model,CSTRs in series model,two phase model of emulsion phase NCSTRs and two phase model of emulsion phase well mixed to describe the temperature distribution,it was found that the temperature distribution modeled in Emulsion phase NCSTRs model acceding with plant measured data:during h/D=0～0.2,the temperature increased sharply(7℃),and the temperature increased slightly(2℃)during the most areas above.Temperature distribution was overpredicted in CSTRs in series model and emulsion phase NCSTRs model:temperature grid was large between the bottom and the top of the bed(40℃).So,it was should clearly that,for large-scaled fluidized bed reactor,two phase model simulated the operation state of ethylene gas phase fluidized bed reactor better than CSTRs in series model,and the emulsion phase was well mixed.Emulsion phase well mixed model was chosen to simulate ethylene gas phase fluidized bed reactor.According to measured bed temperature,emulsion phase well mixed model was established as followed:emulsion phase was modeled by one CSTR,and bubble phase was modeled by 26 CSTRs in series.Compared the product properties calculated by circle gas composition and emulsion phase gas composition in emulsion phase well mixed model,it was found that the error is below four percent.It illuminated that controlling product properties could be realized by controlling the circle gas composition.Developed new grades in emulsion phase well-mixed model.Operation conditions of new grades were got by adjusting the circle gas composition and comonomer fresh feed, and the results accorded well with plant data.