Modeling the properties of TRIP steel using AFIS: A distributed approach S. Dey a , S. Datta a, * , P.P. Chattopadhyay b , J. Sil c a School of Materials Science and Engineering, Bengal Engineering and Science University, Shibpur, Howrah 711 103, India b Department of Metallurgy and Materials Engineering, Bengal Engineering and Science University, Shibpur, Howrah 711 103, India c Department of Computer Science and Technology, Bengal Engineering and Science University, Shibpur, Howrah 711 103, India Received 15 September 2007; received in revised form 13 December 2007; accepted 13 December 2007 Available online 30 January 2008 Abstract The paper attempts to model the strength of TRIP-assisted steels using fuzzy inference system (FIS). The system is proficient to cope up with the changing environment deterministically due to its inherent modular structure and distributed approach creating a flexible and robust system. The problem domain has been divided into several sub-problems and autonomous agents are employed to execute the tasks and decision making by designing respective FIS. The mechanistic constituents of TRIP process are expressed in modular form using linguistic if–then rules developed from the fundamental physical metallurgy concept. Coordination among the distributed agents in the form of sharing knowledge has been attempted to divide the whole mechanism of TRIP in separate modules to make the system fault tolerant. The model is validated by data collected from published literature and found to be sound enough, considering the limi- tations of a rule-based model. Ó 2007 Elsevier B.V. All rights reserved. PACS: 81.05.Bx; 46.70.De; 81.20.n; 07.05.Mh; 81.40.z; 81.40.Lm Keywords: Steel; Transformation induced plasticity; Strain induced transformation; Tensile property; Imprecise knowledge; If–then rules; Distributed agent fuzzy inference system 1. Introduction Zackey et al. [1] introduced the concept of TRIP in C– Mn–Si steels with a composite microstructure of ferrite, bainite, retained austenite and some martensite [2–5]. Gradual transformation of retained austenite to martensite during plastic deformation of TRIP steels results in high work hardening as well as uniform elongation. To achieve tailored microstructure and mechanical properties of TRIP steel it is necessary to have knowledge about the complex role of the composition and the process parameters. For any given chemistry of the steel, the process parameters, viz. amount of cold deformation, intercritical annealing time and temperature, bainitic transformation time and temperature are known to exert appreciable influence on the microstructure and property of the product. The pat- terns relating the inputs (composition and process parame- ters) and outputs (mechanical properties) in TRIP steels are qualitatively recognized by the physical metallurgists. Though some attempts have been made to model the TRIP phenomenon from its physical understanding [6,7], but still to date no suitable model exists to predict the mechanical properties of TRIP steel from the independent variables, i.e. the composition and the processing parameters. This is mainly due to the lack of precise knowledge about the role of the independent variables on the microstructure and properties of the steel, which is difficult to gather due to highly complex and non-linear relationship between the variables. On the other hand, efforts have also been 0927-0256/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.commatsci.2007.12.009 * Corresponding author. Tel.: +91 33 26688140; fax: +91 33 26682916. E-mail address: sdatta@matsc.becs.ac.in (S. Datta). www.elsevier.com/locate/commatsci Available online at www.sciencedirect.com Computational Materials Science 43 (2008) 501–511