To solve the problem of converter burn-through and steel leakage,a new model of monitoring and forecasting the residual thickness of converter lining was proposed.The erosion mechanism and influencing factors of the converter lining was analyzed to clarify the mechanical damage and chemical erosion effect of the smelting environment on the lining.Meanwhile,the feasibility of using infrared image analysis technology to detect the residual thickness of converter lining was clarified,and the research scheme of the residual thickness prediction model was designed based on infrared image recognition.Then,the image algorithm was adapted to realize the mutual conversion of infrared image and temperature,which effectively reduced the difficulty of analyzing infrared image data,and the factor analysis method was used to reduce the dimensionality of the network input.GA-BP neural network was finally employed to predict the lining spalling amount in a single smelting,and then the residual thickness of the lining was calculated.The proposed model is highly accurate,with 98%of the predictions having an error of less than 1mm and 85%of the predictions having an error of less than 0.5mm.This model filled the technical blank of detecting the lining residual thickness with infrared technology,expanded the application prospect of infrared technology in steel industry,and provided new ideas as well as methods for solving safety problems of converter lining.
关键词
炉衬侵蚀机制/红外测厚/图像算法/GA-BP神经网络/安全模型
Key words
erosion mechanism of lining/infrared thickness measurement/image algorithm/GA-BP neural net-work/safety model
维普
万方数据