PTA Residue And Solvent Systems Modeling

PTA as a basis for chemical raw materials, its application scopes are very extensive, involving all aspects of the national economy and closely related to people's daily life, so researching PTA production process is of great significance. In this paper, the focus is on the modeling of PTA production process devices which are film evaporator and azeotropic distillation colomn. Based on the course of the study, found that the quality of data collected from industrial is poor. This is harmful for selecting samples and modeling processes. In addition, as the complex of the mechanisms of film evaporation and azeotropic distillation, it is difficult to choose input variables. Due to the above problems, it is hard to model these tow equipments. So this paper, firstly, studies data preprocessing technology and sample selecting technology, and applies the appropriate data preprocessing methods and sample selection methods to the PTA industrial data processing. Then studies the mechanisms of film evaporation and azeotropic distillation, in order to get the input variables. Thirdly, uses RBF and GRNN networks to model these two equipments, and then verifies these models by single-variable analysis. Finally, implements PTA industry data preprocessing, sample selecting and GRNN network modeling software by using MFC single document technique.By the combination of data preprocessing techniques, sample selection techniques, mechanism analysis and neural network technology, this paper established the models of film evaporation and azeotropic distillation column successfully.