The BP Neural Network Model Of Thermal Parameters And Application In Billet Heating Prediction

Temperature field of going-out billets is not only the important gist ofquality by heating, but also the primary aim in a rolling process. But refer tobillet temperature field, the inner temperature of billet now can't be reached ormeasured by using some instruments easily. So building the mathematical modelfor the temperature's distribution and controlling the going-out temperaturerationally can lead to better rolling quality and less energy consumption.Objected with the pusher-type continuous reheating furnace in the firstplate-rolling plant of Shijiazhuang Iron and Steel Co. Ltd in this paper, and withthe last goal of realizing the real-time monitor and prediction of the temperaturefield in the course of heating the billets, the one-dimensional unsteadymathematical model for the temperature's distribution of billet in a furnace wasbuilt after thinking of the furnace's configuration and heated system and the firstcondition and boundary condition were provided. In view of the system'son-line characteristic, we used the method of all-hidden finite limited difference,before the computational software for studying the heating-up process of billetwas developed by using Visual Basic language. Then the mathematics modelwas verified by the data, which had been obtained by testing on the spot and bysimulating through the finite element software of ANSYS, and the result wasverified that the model was correct and could predict the temperature correctly.At last, the real-time monitor system, which was friend, convenient andcomplete, was designed and realized by VB language. For the currency of thesystem, all data was input to the SQL server 2000 database and it could beshared in the company and could be used by the monitor system.The software system of billet in furnace had been developed and applied atthe first plate-rolling plant successfully.