During continuous casting, abnormal predict is an important part for improving the production and the quality of slab. It is also an important problem that steel factories study. In this paper, the characteristic and influencing factors of mould friction are studied. The possibility of abnormal predicting with the friction is discussed. The paper gives a way using neural network to predict friction and using logical decision to judge whether abnormal happens. The way is divided into two steps, the first step is to predict the value of friction, the second step is to judge whether abnormal happens with difference between the predicted value and the measured value. A software named friction displaying and abnormal predicting is programmed. Much of measured data is analyzed with the software. Thus much abnormal friction fluctuating can be distinguished. When breakout happens this system can give an alarm several minutes ahead. The applications with friction in other ways are still to be studied in the coming days. |