PTA Process Online Monitoring And Optimization Based On Rigorous Model And Data Mining Techniques

Monitoring and optimization are the keys to process system engineering, as well as to the safe production and energy savings for the industry. Process monitoring and optimization has a great significance for improving the production performance, reducing economic losses and enhancing the competitiveness of enterprises. In this paper, rigorous model of PTA is used to get rid of raw process data. Then a series of data mining algorithms is combined to establish the data model, which is used as a core of the development of real-time monitoring and optimization system for PTA plant of 60 million tons production. At last on-line monitoring and optimization system of material consumption and quality indicators is realized.The main research contents and contributions are as follows:1. In order to improve the quality of modeling data, rigorous model is used to get rid of raw data. Then the processed data is compressed and clustered to establish a data model of multi-model format using hybrid PLS-BP algorithm. Compared with general single model, hybrid multi-model data model is more effective.2. Based on the above hybrid multi-model data model and the actual operation situation of the chemical plant, optimization models are established for material consumption indicators and quality indicators respectively. Through analyzing field operation case, the effectiveness of the optimization model is proved.3. Local modeling technique is used to analyze the PTA process from another aspect. The key local modeling sample, collected through KNN algorithm, is used to establish on-line local data model with PLS algorithm is to construct the local modeling sample. As analyzing large-scale process data, the local model proves a higher modeling efficiency and accuracy.4. Based on Process Monitoring and Operating Platform, real-time monitoring and operation system, integrated with data model and optimization model, is developed. The system presents calculated information to site operators using rich data visualization technology. Through field test, the stability, accuracy and effectiveness of the system is proved.In the end, the dissertation is concluded with a summary and discussions of the prospective research on open problems.