The Gas Mixing Process Integration Modeling And Self-organizing Fuzzy Decoupling Control Method

The gas mixing process is an important part of the utility system in iron and steel enterprises. After the gas mixing process, the by-products in steel production such as blast furnace gas and coke oven gas, always serve as fuel for some steel production processes including the billet heating, sintering, coking, power generation or for civil use, etc. Making full and efficient use of the mixed gas is of great significance in saving energy, environmental protection and sustainable development for the enterprise. However, there exist numerous factors such as the changes of gas source pressure of blast furnace gas and coke oven gas, load fluctuations and the coupling between flow rate and pressure when adjusting butterfly valves, and always fail to establish an accurate model for the processes. Therefore, the study of the effective modeling and control methods have important practical significances.According to the in-depth analysis on the process mechanism and characteristics of the gas mixing process, the integrated process modeling method based on the mechanism and the subspace identification is proposed in this thesis. First, the butterfly valve flow characteristic formula is applied to establish the mechanism model of the calorific value. Then the multivariable subspace identification method is adopted to create the identification model for the gas pressure since this part of the mechanism is unknown. By analyzing the relationship between the variables of each sub-models, the integration model based on the mechanism and subspace identification is obtained for the gas mixing process.For the gas mixing process which is strongly coupled and non-linear, a self-organizing fuzzy decoupling control method based on the dynamic coupling degree is proposed in this thesis. Calorific value and pressure feedback fuzzy controller are designed to stable the fluctuations of the calorific value and pressure. According to the expert knowledge and operational experience accumulated in the production process, the basic fuzzy decoupling controller is obtained. By analyzing the dynamic coupling relations between the gas mixing process parameters, the decoupled fuzzy rules are dynamically corrected. Besides, the expert correction strategy is designed for the special conditions of the mixing process and the nonlinear characteristics of the butterfly valves.The actual industrial operating data of the gas mixing process in a iron and steel enterprise is applied to validate the effectiveness of the modeling and control method proposed in this thesis. The simulation results show that the integration model based on the mechanism and subspace identification is of high accuracy for describing the characteristics of the gas mixing process. In the basis of the integration model, the self-organizing fuzzy decoupling control method based on the dynamic coupling degree can obtain a better control effect. Then, the effectiveness and superiority of the proposed method are validated.