Research On Diagnosis Of Corrosion Defects In Ladle Lining Based On Thermal Image Analysis

Ladle is not only a kind of metallurgical high temperature vessel for storing and transporting molten steel,but also a refining furnace for refining molten steel.The erosion of molten steel and the chemical reaction produced by refining will also erode the ladle lining,which will lead to premature erosion of the ladle lining or thermal stress,which will lead to the falling off of refractory bricks and failure to find and repair in time,resulting in steel leakage and even casualties.Therefore,it is of great significance to predict the residual thickness of ladle lining.In this paper,the nonlinear fitting ability of support vector regression is used to design the prediction model of ladle lining thickness.Combined with the process parameters obtained in the actual production and the temperature characteristics of ladle outer wall,the residual thickness of ladle lining is predicted to reduce the leakage accident.At the same time,according to the prediction model of ladle lining thickness,this paper designs and builds a ladle lining defect diagnosis system,which can continuously and accurately analyze and evaluate the corrosion state of ladle lining on-line and operate on site.The main contents of this paper are as follows:(1)the heat transfer model and the thermal stress field model of the ladle are studied deeply.The main factors influencing the change of the residual thickness of the ladle lining are the temperature of the molten steel,the refining time,the stirring of argon blowing,the composition of the molten steel,etc.(2)according to the properties of the ladle,the temperature field model of the outer wall of the ladle is established by using the finite element software ANSYS to explore the relationship between the outer surface temperature of the ladle,the temperature of the molten steel and the thickness of the ladle lining.The accuracy of the temperature field model is verified by the thermal image of the same half life cycle of the ladle.(3)analyze the ladle thermal image,determine the weak area of slag line in the research area of thermal image,and extract the characteristic temperature data.The grid search algorithm is used to optimize the parameters and establish the optimal regression model.Then,based on the optimal parameters,a prediction model of ladle residual thickness based on SVR is constructed.The prediction accuracy can reach 86.3%in a small error range.(4)scheme design of ladle lining defect diagnosis system.Firstly,the function and demand of the ladle lining defect diagnosis system are analyzed.The system has five functions:image real-time display,image acquisition,image processing and analysis,video playback and prediction of the residual thickness of ladle working layer.The hardware platform and software platform are set up and put into operation.