Development Of Iron Desulfurization Expert Model

Iron desulfurization expert model is constructed to predict the amount of granulated magnesium in the process of iron desulfurization with BP neural network concerned. Iron temperature, initial and target sulfur content are three important input variables in this model.Programmable logic controller and industrial automation software are also applied to realize all kinds of control functions.The process of iron desulfurization is introduced first,which is composed of the desulfurization methods and desulfurizers.Then the advanced iron desulfurization technology with granulated magnesium is selected by the cooperative Steel Company in their steelmaking process.Because of too many influence factors in the process, traditional mathematical modeling methods can’t be used to construct the model effectively. In this thesis, expert model and BP neural network are introduced to solve the problem. After several important influence factors be analyzed, iron temperature, initial and final sulfur content are used as the inputs of the four-layer BP neural network based expert model,and unit amount of magnesium as the outputThe expert model can predict the output with three inputs after it is trained with large amount of data samples.And this accurate prediction will reduce the loss of iron and the comsumption of magnesium greatly,which plays critical role in the economical cost saving and environment protection.Besides,Siemens S7-400PLC and industrial control computer are used to construct the iron desulfurization station. Many kinds of PLC programming languages are collectively used in the coding of customer program to fulfill all the necessary functions. Siemens software WinCC are also used in the construction of human machine interface,which can realize the functions of data acquisition, supervision, control,warning and fault finding and so on. After more than one year’s practical application the model has achieved good economic benefits in the steelmaking.