Subhas Ganguly

Genetic Algorithms in Optimization of Strength and Ductility of Low-Carbon Steels

A comparative study between the conventional goal attainment strategy and an evolutionary approach using a genetic algorithm has been conducted for the multiobjective optimization of the strength and ductility of low-carbon ferrite-pearlite steels. The optimization is based upon the composition and microstructural relations of the mechanical properties suggested earlier through regression analyses. After finding that a genetic algorithm is more suitable for such a problem, Pareto fronts have been developed which give a range of strength and ductility useful in alloy design. An effort has been made to optimize the strength ductility balance of thermomechanically-processed high-strength multiphase steels. The objective functions are developed from empirical relations using regression and neural network modeling, which have the capacity to correlate high number of compositional and process variables, and works better than the conventional regression analyses.

Identification of Factors Governing Mechanical Properties of TRIP-Aided Steel Using Genetic Algorithms and Neural Networks

Mechanical properties of transformation induced plasticity (TRIP)-aided multiphase steels are modeled by neural networks using two methods of reducing the network connectivity, viz. a pruning algorithm and a predator prey algorithm, to gain understanding on the impact of steel composition and treatment. The pruning algorithm gradually reduces the complexity of the lower layer of connections, removing less significant connections. In the predator prey algorithm, a genetic algorithm based multi-objective optimization technique evolves neural networks on a Pareto front, simultaneously minimizing training error and network size. The results show that the techniques find parsimonious models and, furthermore, extract useful knowledge from the data.

Simulating Time Temperature Transformation Diagram of Steel Using Artificial Neural Network

Design and development of steel is essentially governed by the Time-Temperature-Transformation (TTT) diagram. The diagram predicts the phase evolution during isothermal transformation schedules for a given chemistry. Selection of chemistry for obtaining a desired microstructure in steel under isothermal schedule needs determination of the TTT diagrams either by extensive experimental exercise or by rigorous thermodynamic calculations. Artificial neural network (ANN) technique has recently been employed as a versatile tool to predict the CCT diagrams of steels. The present work aims to identify a favorable composition capable of yielding an ultrafine bainitic microstructure by isothermal holding of austenite at low homologous temperature. To achieve this, TTT diagrams of varying compositions have been predicted a priori to reduce the required experimental trials. The exercise has led to the development of bainitic microstructure of nanoscale dimension in steel having 0.7C-2.0Mn-1.5Si-0.3Mo-1.5Cr (wt%). Experimental trial with the predicted composition of bainitic steel resulted into attractive combination of strength and ductility.

Artificial Neural Network (ANN)-Based Model for In Situ Prediction of Porosity of Nanostructured Porous Silicon

Nanostructured porous silicon (PS) is usually formed by anodic etching in HF-based solution, and its porosity is measured by a destructive gravimetric technique. In this article, we report the development of an artificial neural network (ANN)-based model permitting in situ prediction of porosity of PS samples. The sensitive and nonlinear dependence of porosity on the formation parameters demanded a nonclassical treatment, and ANN was found suitable for handling this problem. A series of experiments were performed on p-type Si having resistivity 2–5 Ω-cm in 24% HF solution to generate the data for development of the ANN model. The voltage fluctuations across the electrodes during the formation of PS samples were recorded and used to develop an ANN model for prediction of voltage during the transient state of PS evolution. The predicted voltages were then used to predict porosity for different values of current density (J) at any time instant. Porosity was also measured by the conventional and destructive gravimetric method for different values of J and time. The predicted porosities agreed well with gravimetrically determined values.

Informatics-Based Uncertainty Quantification in the Design of Inorganic Scintillators

A soft computing platform, integrating rough sets, fuzzy inferences, and genetic algorithms, is used to develop a series of design rules as a guideline for optimizing inorganic scintillator materials in terms of light yield. The range of values for electrochemical factor, density, Stokes shift, valence electron factor, and size factor which lead to the highest light yield values are identified, with the range corresponding to the uncertainty in the data. The results presented in this article demonstrate how our approach can address the issues of approximation, vagueness, and uncertainty inherent in a relatively small database. We discuss how the results from this work can be used to enhance previously reported models for predicting light yield.

Designing the Multiphase Microstructure of Steel for Optimal TRIP Effect: A Multiobjective Genetic Algorithm Based Approach

A genetic algorithm-based analysis for the primary microstructural requirements of optimal transformation induced plasticity (TRIP) behavior of the multiphase steel is attempted. Design of primary multiphase microstructure of steel for maximum TRIP effect is perhaps the most important job in TRIP-aided steel design. The quantitative and qualitative control over the complex composite behavior of the different constituent phases present in these steels such as polygonal ferrite, bainitic ferrite, and martensite–austenite constituent is still a challenging task. In this work, the tensile behavior of the steel has been simulated using pertinent models proposed earlier. Necessary assumptions are made using domain knowledge to generalize the tensile model for the considered realm of the problem. Based upon the conflicting nature of the tensile strength and ductility of the steel, the problem is considered here as multiobjective optimization problem. The optimization constraints are inherently taken care by the typical formulation of the problem. The Pareto front results are found quite convincing for TRIP-aided steel design.

Effect of copper and microalloying (Ti, B) addition on tensile properties of HSLA steels predicted by ANN technique

Abstract The present study aims to model the composition, process and properties of Cu plus Ti, B microalloyed low carbon steels by using the artificial neural network (ANN) technique. This tool is found to be useful for modelling the effect of copper and microalloying additions along with the process parameters on the tensile properties using the experimental results generated by the present investigators. In first part of the modelling exercise, ANN was employed for prediction of the tensile properties from the input dataset, comprising composition and hot rolling parameters. Subsequently, prestrain and aging parameters were included in the dataset to predict their effects on the tensile properties. The predictions emerged from the modelling allow critical assessment of the role of Ti, B and Cu in conformance with established metallurgical principles.

Exploring the Possibilities of Development of Directly Quenched TRIP-Aided Steel by the Artificial Neural Networks (ANN) Technique

In TRIP-aided steels, generally the composition-process combination is aimed at circumventing pearlitic transformation during cooling of austenite and to retain the desired volume fraction of austenite (∼10 vol%) in the microstructure, which is amenable to stress/strain induced transformation during deformation. The purpose is achieved by individual and interactive contribution of numbers of compositional and process variables. Therefore, it is impractical to predict the best combination of most significant variables by using conventional expertise. In this regard, the artificial neural network (ANN) technique has already been established as a potential tool for composition–process–properties correlation in various materials. In the present study, the ANN technique is utilized to predict the composition–process–properties correlation with an aim to achieve the most attractive strength–ductility combination in TRIP aided steel. In the course of the aforesaid exercise, it is indicated that an attractive strength–ductility combination may be achieved without much requirement of intercritical annealing (ICA) and isothermal holding at bainitic temperature, even at lower level of carbon (say, 0.1 wt%), with judicious alloying by Cu and Ni. The hypothesis is first tested by conducting dilatometric study and microstructural examination of the dilatometric samples and subsequently ascertained by determination of mirostructure and mechanical properties of the as hot roll samples of predicted compositions.

New training strategies for neural networks with application to quaternary Al-Mg-Sc-Cr alloy design problems

Display Omitted The training of a neural network in multiple stages.A dual stage multi-resource data training scheme using multi-objective genetic algorithm.Development of efficient neural network model focusing on missing, but most informative domains of the dataset.The scheme is used for Al-Mg-Cr-Sc alloy system. This study concerns the training of a neural network in multiple stages considering minimization of errors from multiple data/pattern resources. The paper proposed a dual stage multi-resource data training scheme using multi-objective genetic algorithm. The training scheme has been used for the design and development of efficient neural network model focusing on missing, but most informative domains of the data set by means of introducing only a few patterns from missing domain treated separately during the later stage of training. The trained model has been used to design a quaternary Al-Mg-Cr-Sc alloy system, from the information subsets of binary Al-Cr and the ternary Al-Mg-Sc alloys. The validity of the proposed algorithm has been discussed in light of the evolution of the ageing characteristics of the new aluminium alloy system.

In silico Design of High Strength Aluminium Alloy Using Multi-objective GA

Multi-objective optimization is employed using genetic algorithm, for designing novel age-hardenable aluminium alloy with improved properties. Data on the mechanical properties of age-hardenable aluminium alloys is considered together for modeling the mechanical properties using artificial neural network. The models are used as objective functions to get the optimized combination of input parameters for the objectives, viz. high strength and ductility. The significance analyses of the variables on the ANN models gave a primary insight on the role of the variables. The Pareto solutions emerged from the GA based multi-objective optimization is found suitable for effective design of aluminium alloys with tailored properties. An in depth study of the role of the variables in the non-dominated solutions clearly describes the guideline for developing an alloy with improved mechanical properties.