The Slab Breakout Prediction Expert System Of Continuous Casting

Breakout is the most catastrophic incident in the continuous casting process. There are two methods to avoid or reduce it now, one is to deeply study the mechanism of breakout so as to get rid of the causes of breakout in process; the other is to develop techniques of the breakout prediction in order to detect the signs of breakout within the mold, so that measures can be taken to prevent it.By the first method, many models describing the mechanism of breakout have been reported in our country and overseas. Not all these models have the same opinion and their calculated results are also not accurate. The mechanism of breakout was analyzed from the viewpoint of the behavior of liquid steel meniscus in this thesis. Research shows that the damage of liquid steel meniscus is the direct cause of a breakout, and many countermeasures of improving the property of liquid steel meniscus are proposed.By the second method, neural network breakout prediction model are becoming popular and replacing previous logical decision ones in the world. This is due to the drawbacks of the logical decision breakout prediction, such as lack of robustness and error tolerability. In contrast to that, neural network has good self-learning ability, robustness, and can constantly increase the system performance with the increase of running time.In order to solve the problems above, neural network model and expert system technology are used in the system. Besides, statistics and fuzzy mathematics are also used. A new expert system development languageæ¡ŸM PS has been used in the system, which brings a lot of benefit to the development of the system.The front program is developed by Visual Basic 6.0 , which maintainsthe style of windows program. CLIPSActiveOCX is used to connect the front program and the CLIPS program, which fulfills the main function of expert system. To be concrete, the principle of the system is just like this. Whether breakout occurs or not is judged by expert system using two kinds of knowledge: neural network and knowledge base.Function of the system:1.To give the possibility of breakout and provide operational suggestions.2. Training new neural networks, testing the existing neural networks in the system and modifying the structure of neural networks.3.Good dialog interface for user and computer and convenient data input.