Research On Prediction Method Of Ladle Leakage Based On Infrared Thermal Image

Ladle is an important transport container for molten steel.It is also a refining furnace for ASEA-SKF process.The molten steel has the characteristics of high temperature and strong corrosiveness.Therefore,the leakage of molten steel will pose a great threat to the public property and personal life safety of steel mills.The domestic mainstream ladle leakage prevention method is a corrosion resistance test on the high temperature resistant material of the ladle wall,and the number of times of the ladle wall refractory material can be used is obtained,and sufficient space is reserved for the number of times to achieve the purpose of preventing leakage.This will not only make the entire high temperature resistant material not fully utilized,but also cannot monitor the ladle status in real time and cannot cope with unexpected situations.In view of the shortcomings of the current anti-leakage and the low utilization rate of ladle refractory materials,this work proposes a leakage prevention method based on ladle thermal image.The specific research contents are as follows:(1)Summarize and analyze the causes of ladle erosion and understand the changes of steel bag wall erosion.The thermal differential equation of the ladle wall is modeled to determine the relationship between the internal erosion degree and the outer wall temperature field.(2)Through the analysis of the thermal image and the research of the ladle wall,the weak area is determined and the positioning is segmented.Using the conclusions,the characteristics of the weak regions are analyzed,and then analysis the image segmentation algorithm.Finally,the template matching algorithm is used to locate the weak regions.Otsu and the global threshold segmentation algorithm are combined to segment the accurate hot images of the ladle weak regions.(3)A method for online leakage prevention of ladle is proposed.Aiming at the shortcomings that the current leakage prevention can not be carried out online,through the research on the erosion model and the external temperature data of the ladle collected,the estimation model between the steel bag's own property and its own lining erosion degree is proposed.Establish whether the ladle has the risk of leakage prevention by comparing the actual erosion area to the forecast area.(4)Investigate the actual situation of the site,and combine the temperature measurement principle of the infrared camera to determine the on-site installation plan of the infrared camera.This part complete part of the system of the ladle's online monitoring.In this work,the template matching is combined with two threshold algorithms.The positioning accuracy of the weak zone can reach about 92 percent,and the online positioning can be realized.The SVR is used to predict the area of the damaged area.The model's the maximum error of the model prediction is not more than 8 percent of the corrosion area,so the prediction result of the better erosion area can be obtained.