Breakout Prediction System Of Thinslab Continuous Casting Based On SVM Algorithm

Breakout is fatal accident during continuous casting, which leads to assets loss aswell as threat to security of operators. Developing the real-time and effective breakoutprediction system for the recognition of the breakout characteristics is the main means toprevent the breakout accident during casting.At present, the research on the breakout prediction system of our country is still inearly stages. System research on the formation mechanism and the process of breakout islack. Occasionally, certain false alarms issued by the existing breakout prediction systems.Thus, the key point is to develop algorithm to improve accuracy and reduce false alarmrate. In this paper, SVM was introduced to the breakout prediction with the history data ofCSP slab casting machine in a steel plant. Series of research works on this subject werecarried out in this paper as follows.The main forms of breakout and the mechanism of sticking type breakout wereanalyzed in the paper; The typical characteristics of thermocouple temperature and itsspread in space during the process of sticking type breakout were discussed.In order to provide more accurate data for the temperature pattern identification andimprove the accurate rate, the noise in temperature signal collected by thermocouple wasreduced; The PCA method was introduced into the breakout prediction system to reducenoise data, removing redundant information and reduce the model complexity.Breakout prediction model based on SVM algorithm was established on the basis ofthe analysis of the formation process of breakout. Genetic Algorithm、Particle SwarmOptimization and Grid-Search are introduced to optimize parameters of SVM algorithm,and SVM model with better performance is applied to the breakout prediction system ofcontinuous casting. The SVM model is trained and tested by using historical data of CSPcontinuous casting production line of a steel mill, and analysis the test result.Finally, the thin slab continuous casting visual breakout prediction system wasaccomplished with the help of programming software Microsoft Visual C++6.0. Offer areference for the development of online breakout prediction system in the future.