Analyzing the Fluid Flow in Continuous Casting through Evolutionary Neural Nets and Multi‐Objective Genetic Algorithms

Govindan, Deepak; Chakraborty, Suman; Chakraborti, Nirupam

doi:10.1002/srin.200900128

Cited by 29 publications

(7 citation statements)

References 22 publications

(30 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…According to the authors' experience, MOGA results to be well-suited for solving this kind of problems and its goodness it certified by several previous works [4][5][6]. Govindan et al [7] reports an overfitting problem in the neural network (NN) module in modeFRONTIER proposing a very interesting approach [8,9]. Despite this, the authors consider the overfitting problem negligible for the purpose of this article since MOGA and the subsequent approach does not make use of the neural networks module.…”

Section: Structural Geometrical Optimizationsupporting

confidence: 78%

Manufacturing Process Simulation for Product Design Chain Optimization

Gramegna

Corte

Poles

2011

Materials and Manufacturing Processes

View full text Add to dashboard Cite

Engineering simulation is, without any doubt, the key strategy to win in a globally competitive environment. The integrated Computer Aided Engineering (CAE) approach considers all life-cycle steps from the design stage to the in-service use. In this work, engineering simulation technologies provide detailed virtual production processes able to predict the final component quality in terms of defects and virtual structural behavior. In turn, the structural behavior provides information about the performance of the component. Furthermore, through the combined use of process and structural simulations, optimization techniques allow to define the component design by automatic design change and to verify the change's effectiveness in terms of structural strength. The most innovative aspect is the possibility to carry out a structural simulation using, as initial condition, the local mechanical properties and the prestress status due to residual stresses at the end of the manufacturing process. In this article, the redesign of the roller support (manufactured in ductile iron GJS400) is presented: conflicting objectives such as minimization of components weight and minimization of deformation. The use of the optimization software modeFRONTIER has allowed exploring different geometric configurations and, thanks to its multiobjective genetic algorithm (MOGA), finding the trade-off curve of conflicting objectives. The final results evidence some areas with high residual peak stresses, which have a decreasing effect on the fatigue life, and underline the importance of considering an extended optimization analysis that includes both the casting simulation and load-case analyses. IntroductionIn the latest years, the industry has studied and implemented innovative Computer Aided Engineering (CAE) design method to improve the performance and the weight of components. The innovative CAE procedure consists in integrating the whole aspects that characterize the component life, from the manufacturing to the withdrawal, in order to address an always increasing quality request, a complete knowledge of the component potentials, cost reduction, and time to market savings. In particular, the study of the mechanical properties and of the residual stress mapping, due to the manufacturing process, helps to improve the performance and the reliability of the component.CAE offers a large number of efficient computational codes to simulate each phase of product development from the manufacturing process to thermomechanic fatigue life. modeFRONTIER allows to optimize each phase of product development and at the same time to integrate the different design tools, offering to the designers the necessary design chain results to predict the component in service life. This new approach is a winning "weapon" for leveraging a company's knowledge base and to improve competitiveness of a new product.As an example of multidisciplinary optimization approach, an interesting case study regarding a roller

show abstract

Section: Structural Geometrical Optimizationsupporting

confidence: 78%

Manufacturing Process Simulation for Product Design Chain Optimization

Gramegna

Corte

Poles

2011

Materials and Manufacturing Processes

View full text Add to dashboard Cite

show abstract

“…This study emphasizes that multiobjective genetic algorithm should be explored in further detail in developing high‐strength, air‐hardened grade steels. Despite their significant practical applications, till date, applications of data‐driven modeling strategies,26–28 and for that matter, multi‐objective genetic algorithms are rather limited in this area 29. The adavantage of using this strtaegy is, however, quite overwhelming: not only it can capture the existing knowledge but it also generates fresh information avoiding tedious and cumbersome experimental procedures to a very significant extent.…”

Section: Discussionmentioning

confidence: 99%

Data‐Driven Pareto Optimization for Microalloyed Steels Using Genetic Algorithms

Kumar

Chakrabarti

Chakraborti

2012

steel research int.

Self Cite

View full text Add to dashboard Cite

A data base was put together for the mechanical properties of microalloyed steels, which contained about 800 entries for ultimate tensile strength (UTS), yield strength (YS), and elongation. Using an evolutionary neural network, based upon a predator–prey genetic algorithms of bi‐objective type, this information was used to construct data‐driven models for UTS, YS, and elongation. The optimum Pareto tradeoffs between these properties were obtained using a multi‐objective genetic algorithm. The results led to some hitherto unexplored steel compositions with optimum properties. Some such steels were actually cast and the experimentally observed property values were found to be well in accord with the predicted results.

show abstract

“…In principle, any multi-objective optimization routine can be used in EvoNN; the very first paper on this algorithm , however, was based upon a modified predatorÀprey genetic algorithm (Li, 2003) and the practice subsequently continued (Bhattacharya et al, 2009;Govindan et al, 2010;Kumar et al, 2012). In this algorithm two species, the predators and the prey, are introduced in a torroidal computational grid that emulates a forest.…”

Section: Evolutionary Neural Net and Pareto Tradeoffmentioning

confidence: 99%