ISSN 2594-5327
56º Congresso anual — Vol. 56 , num. 1 (2001)
Título
Autoria
DOI
Downloads
Resumo
Mechanical properties such as yield strength and tensile strength of hot rolled steel are predicted employing a hybrid model. This model comprises a comprehensive physical microstructure model as well as a neural network for the final link between chemical composition and process parameters on the one hand and the mechanical properties of the hot rolled steel on the other. Results obtained at the medium strip-rolling mill of Hoesch Hohenlimburg as well as from several wide strip rolling mills show reasonable agreement between measured and calculated values for a wide range of steels. Integrated in process automation of hot strip mills, application of the model can be used to reduce the number of tensile tests, save time through release of the coils immediately after coiling and even improve steel quality by adapting process parameters like coiling temperature in order to obtain targeted mechanical properties. This latter benefit can be achieved both by in-advance off-line simulations and in-line application of the model combined with an optimizer procedure.
Mechanical properties such as yield strength and tensile strength of hot rolled steel are predicted employing a hybrid model. This model comprises a comprehensive physical microstructure model as well as a neural network for the final link between chemical composition and process parameters on the one hand and the mechanical properties of the hot rolled steel on the other. Results obtained at the medium strip-rolling mill of Hoesch Hohenlimburg as well as from several wide strip rolling mills show reasonable agreement between measured and calculated values for a wide range of steels. Integrated in process automation of hot strip mills, application of the model can be used to reduce the number of tensile tests, save time through release of the coils immediately after coiling and even improve steel quality by adapting process parameters like coiling temperature in order to obtain targeted mechanical properties. This latter benefit can be achieved both by in-advance off-line simulations and in-line application of the model combined with an optimizer procedure.
Palavras-chave
microstructure, mechanical properties, neural networks
microstructure, mechanical properties, neural networks
Como citar
Löffler, Dr. HansUlrich; Döll, Dr. Rüdiger.
PRACTICAL APPLICATION OF MICROSTRUCTURE MODELLING IN HOT STRIP MILLS,
p. 195-203.
In: 56º Congresso anual,
Belo Horizonte, Brasil,
2001.
ISSN: 2594-5327, DOI 10.5151/2594-5327-C01094
