Study On The Basic Oxygen Steelmaking Endpoint Control Based On The Flame Multi-spectral Analysis

With high production efficiency and low energy consumption,converter steelmaking has been one of the most important steel production methods,in which the accurate control of the blowing endpoint has a substantive influences on reducing energy consumption and production cost,improving the steel product quality and the environmental protection capability as well as the expansion of the product range.However,due to the variation of the raw material ingredients and the complex physical and chemical reactions involved in the converter bath of high temperature,the endpoint composition of the molten steel is influenced by numerous factors,consequently,it is difficult to achieve accurate online control of the converter steelmaking endpoint up to now,which is also a difficult problem to be solved in the metallurgical industry all over the world.In this thesis,the problems mentioned above were investigated and the main works are as follows:With the constructed multi-spectral radiation information acquisition system,the long-distance and non-contact acquisition of the flame multi-spectral radiation information can be achieved stably under the adverse environment of the steelmaking shop.Besides,the acquisition procedure is rarely influenced by variation of the blowing process or the converter size.The relationship between the flame at the converter mouth and the converter bath status was analyzed by combining the distribution of the flame spectrum during different blowing periods and the experience of the converter operators.Using spectral features of the flame at the converter mouth which are related to the carbon content of the molten bath and converting parameters as input variables,two endpoint carbon content forecast methods based on the hybrid models of the support vector machine(SVM)classification and regression were proposed for two different endpoint carbon content ranges,respectively.Based on the mechanism analysis,one binary classification model of the endpoint phosphorus content which employs the flame spectral features of the middle period of the blowing process as input variables was proposed.Molten steel was selected as the radiation source to obtain the responsivity of the proposed system.Then the flame temperature was measured based on the dual wavelength radiation thermometry(DWRT)and the multi-wavelength radiation thermometry(MRT)respectively.In consideration of the relationship between the flame temperature and the bath temperature,endpoint temperature prediction model was constructed based on the support vector regression model which can achieve online,non-contact and accurate prediction of the bath temperature.All the data used were collected during the actual blowing procedures.Furthermore,the proposed system and the endpoint carbon content and temperature prediction models were applied to the actual blowing procedures so as to verify the practical application performance.All of the experimental proceedings were selected randomly from the regular blowing procedures.Experimental results indicate that the constructed endpoint prediction model can completely meet the demand of the blowing endpoint at the steelmaking field.Besides,it is the first time that the multi-spectral analysis of the flame at the converter mouth is applied to realize the precisely endpoint prediction of the converter blowing at home and abroad,and the endpoint prediction accuracy was improved by about 20%compared to that of the artificial experience method whose accuray is around 73%.The research methods presented here can be used for different kinds of converter steelmaking processes,and it is scarcely affected by the complicated bath reactions taking place during the converting process.The proposed method has innovative significance and numerous application values for the development of the converter steelmaking endpoint control technology.