Research On Prediction Method And Space-time Characteristics Evolution Of Breakout In Continuous Casting

As the key link of iron and steel metallurgy,continuous casting not only directly affects the quality of steel,but also relates to the production efficiency and economic benefits of steel mills.Due to the complexity of continuous casting process,various breakout accidents often occurred,which not only affecting the smooth production of continuous casting,damaging the equipment of continuous casting machine,but also bringing huge economic losses to the steel plant.In all kinds of breakout accidents,sticker breakout is the main one.Therefore,if the space-time characteristics of sticker breakout can be captured and the identification method and prediction system of continuous casting breakout can be developed,the occurrence of breakout accidents can be effectively prevented,and the frequency of false alarm can be reduced to ensure the smooth progress of continuous casting production.So this paper researched from the following four aspects,as follows:First of all,based on the pouring data of a domestic steel plant,analyzed the temperature change rate of the continuous casting,used interpolation algorithm and computer graphics technology,established the corresponding relation of the temperature and color space,and reflected the changing rate of the temperature of each coin to a flat surface,realized the “thermal imaging” of mould temperature change rate,so as to obtain the two-dimensional spatial distribution of the temperature of the mould.Then,the two-dimensional space-time evolution characteristics of the temperature of the mould are captured.Secondly,the characteristics of hot and cold region of sticker breakout were extracted,and the sticker breakout was divided into four types from the perspective of space-time evolution:typical sticker breakout,one-sided of left propagation of sticker breakout,one-sided of right propagation of sticker breakout and long-span propagation sticker breakout.The area,temperature change rate and the position of the sticker region characteristics of the various types of breakout were analyzed,and the further understanding of the relationship between space-time evolution characteristics when the sticker breakout occurred,then judged the true and false sticker breakout.Again,based on continuous casting mould breakout visual monitoring system,expanded the true and false sticker database,and there were true sticker image 60 cases,false sticker image 60 cases.Then obtained the region of interest of the thermal imaging,and combined with the propagation behaviors of sticker breakout in the mould,used clustering method to choose 4templates,template matching method was used to recognize the true and false sticker breakout.The results showed that the reporting rate of true sticker breakout is 100%.Finally,based on Python language,the continuous casting breakout prediction system was designed,which included some modules such as region of interest selection,different types of template selection and template matching result information.The system could realize the off-line measurement of temperature thermal imaging in continuous casting process and accurately identify the breakout in continuous casting.