A Design of New Brands of Martenzite Steels by Artificial Neural Networks
Abstract
The paper proposes a model-based approach for the design of martenzite structure steels with improved mechanical and plastic characteristics using proper composition and thermal treatment. For that purpose, artificial neural models approximating the dependence of steels strength characteristics on the percentage content of alloying components were trained. These non-linear models are further used within an optimization gradient procedure based on backpropagation of utility function through neural network structure. In order to optimizing the steel characteristics via its chemical composition, several steel brands with high values of tensile strenght, yield strenght and relative elongation were designed. A steel composition having economical alloying and proper for practical application was determined comparing several obtained decisions. The usage of that steel will lead to lightening of the hardware for automobile industry.
Downloads
References
[2]. Development of High Strength Steels for Automobiles, Nippon steel technical report No.88, July, 2003.
[3]. ULSAB-AVC, Advanced Vehicle Concepts Program Results, CD March 2002, www.ulsab-avc.org.
[4]. Farahani A. R. Kolleck - Hot Forming and Cold Forming-Two Complementary Processes for Lightweight Auto Bodies, Proceedings from The International Conference „New Development in Sheet Metal. Forming Technology,” Stuttgart, Germany, 2004, p. 235-244.
[5]. http://www.splav.kharkov.com/choose_type.php.
[6]. Koprinkova-Hristova, P., Tontchev, N., Popova, S. - Neural networks for mechanical characteristics modeling and compositions optimization of steel alloys, Int. Conf. Automatic and Informatics’10, Oct. 3-7, 2010, Sofia, Bulgaria, pp.I-49 – I-52.
[7]. Werbos, P.J. - Backpropagation through Time: What It Does and How to Do It. Proceedings of the IEEE vol. 78(10), p. 1550-1560, 1990.
