Estimating The Temperature Distribution Of Molten Steel In Ladle Based On Inverse Heat Transfer Problem

The ladle is a indispensable equipment in the produce of iron and steel, thetemperature of the molten steel is an important impact in the ladle steel refining andcontinuous casting process. Therefore, how to measure the temperature distribution ofmolten steel, accurately, in the ladle is very important. In this paper, the feasibility andeffectiveness of the indirect measurement of molten steel temperature distribution inladle with IHTP is approved.In this paper, finite element nonlinear problem solving software ABAQUScombined conjugate gradient method (CGM) and distributed fuzzy inference (DFI)inverse the temperature distribution of molten steel in the ladle, The main work of thepaper are as follow:1) Creating a model of positive problem. With ABAQUS the paper established80tladle two-dimensional heat transfer system and calculated ladle thermal systemtemperature distribution.2) Combined with the model of the forward problem, the temperature distributionof the ladle of molten steel in the ladle system first boundary condition, and as aparameter to be inversed with conjugate gradient method (CGM) and distributed fuzzyinference (DFI) method inversion model.3) Establishing the connection between ABAQUS and MATLAB program toconstruct the integrated inverse model.4) Simulation tested the method in the use of measuring the temperaturedistribution of molten steel in the ladle is effective. The effect of initial guesses, thetemperature measuring points, and temperature measurement errors to the inversionresults are discussed.The research has shown that the use of conjugate gradient method (CGM) anddistributed fuzzy inference (DFI) the inversion the ladle boundary conditions, are ableto get satisfactory results: DFI method of comparison of the inversion results better thanCGM. Studies have shown that: measurement error, the number of measurement pointshave a certain influence on the inversion results, the smaller the error, the correspondinginversion results better; greater the number of measuring points, the more accurate thecorresponding inversion results; initial guess of parameters to be inversion resultsalthough the impact relative to the first two terms of influencing factors, but much smaller, the inversion results are not sensitive to the initial guess.