Shubhabrata Datta

Soft computing techniques in advancement of structural metals

Abstract Current trends in the progress of technology demand availability of materials resources ahead of the advancing fronts of the application areas. During the last couple of decades, significant progress has been made in computational and experimental design of materials. Among the potential computational techniques, soft computing stands in distinction due to the inherent flexibility in capturing the complexity of the problem in global scale. Since 1990s remarkable success has been achieved in soft computing activities in different facets of materials science and engineering. Extensive efforts have been devoted in design of metals and alloys based on composition–process–microstructure–property correlation. The present review aims to address the contribution of soft computing in the field of structural metals and alloys including processing and joining. The critical issues concerning applicability of particular techniques in specific materials problem have been particularly emphasised encompassing the scope of integrating the gradual progress in different techniques in hybrid and tandem framework to address greater complexities in larger length and time scale. Attempt has also been made to emphasise on the evolution of newer knowledge and materials through soft computing activities. Finally, the potential of soft computing techniques in futuristic design approaches has been critically enumerated.

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.

Dynamic discreduction using Rough Sets

Discretization of continuous attributes is a necessary pre-requisite in deriving association rules and discovery of knowledge from databases. The derived rules are simpler and intuitively more meaningful if only a small number of attributes are used, and each attribute is discretized into a few intervals. The present research paper explores the interrelation between discretization and reduction of attributes. A method has been developed that uses Rough Set Theory and notions of Statistics to merge the two tasks into a single seamless process named dynamic discreduction. The method is tested on benchmark data sets and the results are compared with those obtained by existing state-of-the-art techniques. A real life data on TRIP steel is also analysed using the proposed method.

Composition–Processing–Property Correlation of Cold-Rolled IF Steel Sheets Using Neural Network

Artificial neural network (ANN) models have been developed to establish relationships between the mechanical properties of the cold-rolled sheets of interstitial free (IF) as a function of the chemical composition of the steel, the rolling and the batch annealing parameters. The mechanical properties studied include yield and . The technology of rolling mill automation necessarily embraces a broad spectrum of interests, ranging from fundamental process analysis to the solution of special control-theoretic problems. Determining the relationship between various compositional and processing parameters and the mechanical properties of cold-rolled sheets is important but not easy to establish. The complex and highly nonlinear interactions between the variables make the system difficult to assess from the prediction point of view. ANN model is found to be capable of developing the inherent relationship between the variables and could be successfully used to develop better understanding on the metallurgical principles of cold rolled IF steel.

Optimal Design of Titanium Alloys for Prosthetic Applications Using a Multiobjective Evolutionary Algorithm

Multiobjective optimization using a Reduced Space Searching Algorithm (RSSA) is employed to optimally design titanium alloys suitable for prosthetic applications, i.e., with high strength, low elastic modulus, adequate biocompatibility, and low costs. The objectives in question are conflicting in nature, and thus multiobjective optimization is the ideal candidate for approaching this problem. The latter was formulated in such a way that it was necessary to develop three separate objective functions for strength, elastic modulus, and economic costs. The biocompatibility issue was introduced as a constraint in the optimization process. To develop the objective functions for yield strength and elastic modulus, a two-layered fuzzy inference system is used. To take into account economical factors, a weighted sum-based model of the elemental constituent is developed, including the costs of the alloying additions. The compositions of the alloy found from the Pareto solutions show that the above objectives can be fulfilled in the case of β Ti-alloys only.

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.

Rough Set Approach to Predict the Strength and Ductility of TRIP Steel

Transformation Induced Plasticity (TRIP) gives birth to new generation steels with high strength and good ductility. Both these properties of steel depend on a number of compositional and processing parameters, but till date there exist certain gaps in the understanding of the complex role of each parameters on the microstructure and thus the properties of the steel. Rough Set Theory is employed to derive decision rules that attempt to explain this complex behavior. Applying efficient heuristics, the number of attributes are reduced to form a minimal reduct, and their values are at the same time discretized into linguistic intervals. The derived rules could clearly indicate on the relative importance of the compositional and processing variables.

Age Hardening Behavior of Wrought Al–Mg–Sc Alloy

Ageing of Al-6Mg alloy doped with varying concentration of scandium ranging from 0.2 wt% to 0.6 wt% has been carried out. Cold-rolled alloy samples are isochronally aged for 60 minutes at different temperatures. The cast and hot-rolled samples are also aged isochronally for 90 minutes at different temperatures up to 450°C. Isothermal ageing of cold-rolled samples is conducted at various temperatures for different periods of time ranging from 30 to 480 minutes. Hardness values of the differently processed alloys have been measured to understand the ageing behavior of Al-6Mg alloy with scandium addition. The hot-rolled alloys after ageing do not show any hardening response due to ageing. Ageing of cold-rolled alloys envisaged precipitation of Al3Sc which is not noted to be dislocation induced. The kinetics of precipitation of Al3Sc in Al–6Mg–Sc alloys are found to be controlled by the diffusion of scandium in aluminum.

A Comparative Study for Modeling of Hot-Rolled Steel Plate Classification Using a Statistical Approach and Neural-Net Systems

Classification of flat steel products from the point of view of reaching the target property is a common practice in industries. In most classification problems, standard statistical methods generally place constraints such as continuous, differentiable, otherwise well behaved, etc. However, Artificial Neural Networks (ANN) has an ability to learn and generalize any complex system without making any model assumptions. This work emphasizes on making performance evaluation of usual statistical techniques such as general clustering like K-means, partition around medoid (PAM), classification and regression tree (CART), linear discriminant analysis (LDA) vis-a-vis usage of multilayer perceptron (MLP) learning algorithm, radial basis function (RBF) family of methods and Kohonen networks. To recommend the utility of modeling, some real-life industrial databases are used. It can be observed from the results that learning of networks through back-propagation yielded minimum misclassification of two groups of heats including minimization of train-test error. The statistical techniques such as LDA and CART provide the same results of misclassification along with the results obtained from perceptron learning, RBF network algorithm and Kohonen learning with learning-vector quantization (LVQ) algorithm.