Development And Application Of Mould Breakout Prediction Expert System In Continuous Casting

Online detection and control to the solidification state of slab in the mould is the core of modern continuous casting production. In recent years, with the increasing of steel variety, production scale and casting speed in continuous casting, it put forward new demands for the mould monitoring technology. As the important part of monitoring to the mould, real-time prediction to abnormal phenomenon such as breakout in high casting speed, online detection to mould thermal behavior, force behavior and other important process parameters, thermal imaging and visualization technologies, etc.which have important meanings for producing none defects slab and ensuring continuous casting process going smoothly.In the paper, we take slab continuous caster as the research object, focusing on studying the breakout prediction method and mould expert’s development and application. After referring to domestic and foreign literature and patent, combined with the result of investigation to the steel mill, we reviewed the major breakout prediction method and its progress, summarized the typical breakout prediction algorithm and practical effect of the domestic and foreign breakout prediction system. We also analyzed the characteristics of the mould temperature in abnormal situations such as corner cracking, surface cracking and stick-type breakout. According to the general rules of breakout prediction, we gave some strategies and recommendations on the thermocouple selection, arrangement and maintenance.After analyzing too much real-time measured breakout samples, we summarized all possible temperature trend when the slab bonded to mould, and on the basis of which we established recognition model and algorithm in breakout prediction. According to the signal detection conditions of the slab continuous caster on a domestic steel mill site, we developed an offline mould breakout prediction software modules by means of Visual C++to achieve specifically analysing to the historical breakout data samples. We also built an online detection system to the mould temperature and developed an online mould breakout prediction software modules. The results of online synchronous monitoring to continuous casting with our system and foreign system demonstrate that our prediction algorithms meet online application demands for the rate of accurate alarm, false alarm and failure alarm. Finally, with the rapid promotion of the mould expert system in recent years, we focused on discussing its core function and application method. Combined with the concrete conditions on the continuous casting site, mould expert system have remarkable application effects in the areas such as guiding removing large flux rim from mould, judging whether the slab bonded to mould, evaluating powder lubrication effect, which further expanding and enriching the practical functions of the system.