Study Of Modelling For Gas-Phase Propylene Polymerization And Soft Sensing Of Melt Index

Modelling of reactors for gas-phase propylene polymerization is one of the focuses of research on process control. Based on the mechanism of gas-phase propylene polymerization, the characteristics of material balance and the research of Fan and Jiang, studies of modelling and simulation on Amoco's Horizontal Stirred Bed Reactor (HSBR) were carried out. A steady model of gas-phase HSBR was set up, and the model was studied with two different assumptions. Based on the steady model, Melt Index (MI) of polypropylene and residence-time distribution (RTD) were calculated with some on-line measured operating data from an industrial reactor, and the results show that the model is quite satisfied. A dynamic model of HSBR was also set up. During the process of solving the dynamic model, the results of the steady model were used for the initial value of the dynamic model. Using the industrial data, the dynamic model was simulated. The results show that the performance of this dynamic model is very good with a strong adaptability during both steady state and grade transitions of the polypropylene production, and the dynamic model has an essential application value.The on-line application of the dynamic model of Continuous Stirred Tank Reactor (CSTR) and Fluidized Red Reactor (FBR) on an industrial Hypol plant was done. Ethylene-propylene copolymerization in the FBR was studied. The application of MI soft-sensing and the interface program between PC and DCS were developed. During this process, some questions about porting the simulation application, solving the differential equations, confirming the initial value, online updating parameters and autorunning the application were solved. The results of the model can follow both the steady state and grade transitions of the Hypol plant, and the prediction error was little and inside an acceptable range. At last, a general MI soft-sensing software platform with different industrial propylene polymerization processes was developed. The software pl-atform with a friendly interface can be operated simply, expanded easily and is desired to be applied on the practical industrial processes. Some shortcomings about the software platform are also analyzed to instruct the work for the future.