The Modeling And Optimization Control Study Of Product Quality Of Coke Production Process

The production process of coke, is a complex industrial process, with the characteristics of serious nonlinear, time-varying, multiple parameters and uncertainty, its main product is coke. Coke is the main fuel and raw materials in the process of iron production in the blast furnace, plays an important role of a chemical reducing agent and permeable support in the skeleton. The quality of coke directly affects the reliability and stability of subsequent industrial production. It is necessary to establish the quality model of coke to realize the product quality optimal control in the coke production process. And there exists numerous interference factors and uncertainty factors in the coke production process, to establish the quality model and achieve the product quality control and optimization is of great difficulties. According to the actual situation of coke production process, this paper develops depth research in the modeling and optimizing method of coke quality. The results have certain practical significance, the main work in this paper is concluded as follows:Based on the coke production process and analysis mechanism of coke formation, this paper puts forward the problem of quality modeling and control of coke production process, coke quality indexes and the factors affecting the coke quality are analyzed and discussed, thus determines the input and output of the quality model. The data of coke production process are screened, selects and preprocesses the sample data.Based on the analysis and modeling method of fuzzy neural network, presents a modeling method of coke quality based on fuzzy neural network. The simulation results of actual data show that the fuzzy neural network modeling method is effective, of quicker convergence, with higher prediction accuracy and hit rate.Based on the analysis of wavelet neural network, establishes a coke quality model based on adaptive genetic algorithm to optimize wavelet neural network, the simulation results show that the method has high prediction accuracy and good prediction effect.Finally, based on the established quality model, this paper analyzes and determines target function and constraints of coke quality optimization model, establishes the optimization model of coke quality, and uses the generalized gradient algorithm to find the optimal solution of the optimization model, namely, to find the optimal blend coal performance index. The results lay a foundation for implementing the quality control of coke production process.