Extension theory-based modeling for purified terephthalic acid solvent system

With the growing scale and complexity of industry production, modeling for industry process has become a hot issue. Purified terephthalic acid (PTA) solvent system is such a complex and nonlinear process, which is one of the most important parts in PTA production process. In order to monitor the production index and guide the production running better, the paper proposes an extension theory-based modeling method, which is composed of extension variable-description, extension variable-analysis, extension variable-selection and neural network modeling. Through the actual application in PTA solvent system of a chemical plant, the feasibility and efficiency of the proposed modeling method is proved, which expresses matters in an overall way, describes thinking process of human beings in a formalized way, models the production process precisely, thus exploits a new way to simulate and guide the industrial process.