Fuzzy Modeling Method For Blast Furnace Gas System And Its Application

Blast Furnace Gas(BFG)is an important secondary energy source in the steel production process.In order to deploy the gas efficiently and rationally,the trend of gas tank level is needed to be grasped in real time.Due to the complicated production process of blast furnace gas system,there is a strong nonlinear dynamic relationship among the variables,and it is difficult to establish precise mathematical model.In view of the large amount of historical data accumulated in the process of steelmaking,data driven model of the blast furnace gas system can be effective,which is the relationship between the gas tank and gas users.Then dispatchers can develop a reasonable gas use plan.It is of great significance on the energy-saving and emission reduction.A fuzzy modeling method for blast furnace gas system based on multi-objective optimization is proposed against the problem of industrial data collected with high noise,large fluctuations,complex relationship among many variables.Meanwhile it has met the demand of accurate and stable gas tank level forecasting.By constructing model performance index including accuracy,complexity and interpretation,the membership functions and fuzzy rules base can be optimized simultaneously.Then final fuzzy model can be established.In order to build the membership functions,a two-stage method is proposed.In the first stage,the optimal clustering parameters are determined based on the multi-objective optimization density clustering algorithm,and the reasonable clustering number and parameters are obtained.The formulas are designed according to the characteristics of data distribution and membership function in the second stage.According to the clustering results,the shape and parameters of each membership function are determined.In this paper,the Wang-Mendel(WM)rule extraction algorithm is integrated into the training process in view of the fact that there is no matching rule in the rule base.In order to verify the effectiveness of this method,the actual production data of a steel enterprise in China is experimented and compared with the existing methods.The experimental results show that the method of establishing a fuzzy model of blast furnace gas system has a good effect on model accuracy,complexity and interpretation.Finally,an industrial data modeling and analysis platform is developed,and the method is used as the background algorithm and applied to the balance control of blast furnace gas system,which has an important reference value for decision scheduling optimization.