METHOD FOR MANUFACTURING HOT-ROLLED STEEL BY ACCURATELTY PREDICTING TEMPERATURE IN FINISH-ROLLING OF HOT STEEL SHEET

摘要

Learning to reflect both the heat transfer coefficient of direct copying and groups exist, the heat transfer learning correction coefficient, finish rolling by calculating the intervals output side temperature more accurately, which allows adjusting the roll gap at best, hot rolled steel sheet manufacturing method that improves the precision hot finish rolling temperature prediction interval is introduced. hot finish rolling temperature range of the present invention prediction A hot rolled steel sheet manufacturing method is to improve the accuracy, the first temperature measurement step of measuring the outlet temperature of the rough rolling section; A first calculation process of calculating the inlet temperature of the finish rolling temperature range on the basis of the measured interval rough rolling the outlet temperature; On the basis of the stored heat transfer coefficient and the calculated finish rolling inlet temperature range, the finish rolling temperature interval the second calculation process for calculating each stand inlet temperature and the outlet temperature of finish rolling section; Second temperature measurement step of measuring the inlet temperature and the outlet temperature of the finish rolling period of the finish rolling section; Relevant materials with the same material while you search for a previous material previously stored heat transfer coefficient learning tasks in the same rolling conditions, direct copying heat coefficient derivation learning; Relevant materials with the same material to the group existed as distinct from the direct copying tasks at the same rolling condition-based machine learning to search the stored heat transfer coefficient of heat transfer learning coefficient derivation exists; And modifying the relevant material in heat transfer coefficient learning by combining the learning coefficient of heat transfer derived directly copying and presence-based heat transfer coefficient correction learning process; include ;