Study On The Prediction Of Molten Steel Temperature Based On Case-Based Reasoning

Ladle Furnace(LF)refining is an important link in the production of steel.The temperature of LF molten steel has an important influence on the output and quality of the steel products and the smooth progress of the steel-making process.The accurate prediction of LF molten steel temperature is a precondition of LF molten steel temperature control,which is an important basis for the realization of automatic refining control.Taking 120 t refining furnace as the research object,through the analysis of the energy budget of the refining process,the temperature mechanism model of the LF refining process was established according to the principle of energy conservation,and the main influencing factors of the temperature of the molten steel during the refining process were determined.As the LF refining process is carried out at high temperature,it can be seen from the mechanism analysis that many parameters of the mechanism model are difficult to obtain or have low accuracy.Therefore,it is very difficult to establish an accurate mechanism model of the refining process.On the basis of determining the main influencing factors of molten steel temperature in the refining process,the weight value of the influential factors is determined by the indefinite analytic hierarchy process.In the research of Uncertain Analytic Hierarchy Process(AHP),a modified method of interval number judgment matrix based on optimization algorithm is proposed for dissatisfied interval number judgment matrix.The purpose of the modification is to save as much as possible the original matrix information and the preference of decision-makers,Meet the requirement of consistency.For the test of weight vector of interval number judgment matrix,the concept of probability is introduced into COP criterion of judgment matrix test of definite number,which makes it suitable for the test of interval number weight vector.According to the principle of Case-based Reasoning(CBR),we use the process of refining data to construct the case base,determine the reuse rules in different situations,establish the compensation model,and compensate the cases with lower similarity to improve the accuracy of the solution.Evaluate the reuse results,and maintain the case base to establish a complete case-based reasoning system.