The Research And Application Of The Neural Network In Real-time Prediction System Of Cross Temperature In The Blast Furnace

The research and application of the neural network in the real-time prediction system of cross temperature are mainly discussed in this paper.As the basic research and key technology of Artificial Intelligence System of Gas Flow Distribution and Stuff-disposing Operation, We adopt time-sequence BP neural network and TD algorithm to predict cross temperature. Whereas the change of cross temperature doesn't belong to pure time-sequence problem, we select some other parameters according to the analysis of factors influencing gas flow distribution and experts' opinions, which influence cross temperature. The whole system is modeled by the UML over again based on Object-Oriented idea. The program is developed in C++6.0 Builder. Because the program is designed based on Object-Oriented idea, code is concise, at the same time, the system is stable and easy to maintain. This system improves real-time performance and increases the hit-the-target rate of prediction then before. At the end of the paper, we discuss the prediction results and put forward some ideas to improve the system.