Hot Forming of Aisi A2 Tool Steel

In order to optimize the hot forming process of tool steels, further improvement of flow stress prediction and additional knowledge on hot deformability, especially at lower and upper limits of the hot working range, are required. The hot working ranges for these steels are very narrow as a consequence of the role of alloying elements, which form carbides. Carbides improve tool steel’s hardenability, strength, hardness, etc. but at the same time decrease hot deformability. The lower limit of the hot working range is determined by the shape, size and type of the carbides which precipitate in grain interior and on grain boundaries. On the other hand, the upper limit is determined by the occurrence of incipient melting of eutectic carbides and by segregation of Cu, Sn, Se or their phases with low melting point on grain boundaries 1-7 . At present time hot forming (rolling) of AISI A2 tool steel is becoming more and more oriented towards larger initial dimensions of roll-pieces and towards smaller outlet dimensions (below =10. mm) which consequently require lower final deformation temperatures (below 900°C). For an achieving of optimal cross-section reductions the accuracy of prediction of flow stress should be within 5% range 8 . For tool steels, such accuracy can not be obtained using various empirical and semi-empirical models, because of complex role of carbides 9-13 . Neural networks are nowadays used at solving of such complex systems 14-18 . In the past BP NN have been already used for the prediction of flow curves but the authors 17-19 did not present the flow curves for intermediate values of strain rates and temperatures simultaneously. On the other hand this has been done in 20-21 but the accuracy was questionable or the NN were too clumsy. Therefore the aims of the present contribution are firstly to apply advanced adaptive methods which are suitable for handling complex systems, and secondly to study the possibility of expanding of hot working area.