Complex System Fuzzy Logic Controller Design, Parameter Optimization And Application On Linear Low-density Polyethylene

As one of the five common synthetic resins, the polyethylene (PE), because ofits cheaper price and excellent performance, has been widely used in industry,agriculture and daily life. Polyethylene production is a typical polymerizationprocess with is highly complex, interrelated and has strong nonlinear characteristics.Till now, some important parameters are unmeasurarble online. The PID controllercombined with experience-based operation is the most popular way in industrialapplication. We are lack of advanced controls. How to improve the existing methodsto achieve better control quality and improve product quality has importantsocio-economic value.On this basis, as compared with conventional PID control program, a model offuzzy logic control applications, improved genetic algorithm optimization andsimulation. Because fuzzy logic control is based on operational experience andexpert knowledge, in the nonlinear model, model structure and interferes with theuncertainty of the model of time-varying characteristics of the face of complexmodel has unique advantages, In this paper, and the linear low-density polyethylenein the actual production is a complex, time-varying, nonlinear delay models, Sofuzzy logic control in the actual process of production has good prospect.Fuzzy Logic Control fits this situation. The selection of Fuzzy Logic Controlrules and fuzzy variable parameters of membership functions rely mainly on thechoice of experience, therefor it is necessary to optimize the original fuzzy controller.Taking into account the fuzzy controller optimization involves large-scale,multi-parameter, complex and continuous search, suited to the characteristics ofgenetic algorithm. Moreover, the genetic algorithm running only by thefitness-driven rather than numerical optimization needs of targeted local information,and fuzzy controller also depends little on the model, Therefore, genetic algorithmapplied to the optimization of fuzzy controller design is appropriate. So geneticalgorithm optimization is applied on the optimization results, and the geneticoptimization algorithm crossover operator, mutation operator of the education andoptimization.Ethylene production process modeling is a complex matter. Based on theUnipol Fluidized Bed Reactor process of UCC. This thesis studied the forlinear-low-density polyethylene (LLDPE) production. It briefly introduces thecircumstation of LLDPE production both in China and oversea, summarizes the polymerization functions, browses the production processes and generally builds themodel of LLDPE within fluidized bed reactor. Based on the kinetic and dynamicmodel of LLDPE, a simplified model for MI and density predition is summedup. With the application of production data collected from factory, the model isjustified and properly revised. In the end, a fuzzy logic control and GA method isimplied successfully on this model in simulink circumstance.