Evolutionary Neural Network Modeling of Blast Furnace Burden Distribution

Pettersson, Frank; Hinnelä, Jan; Saxén, Henrik

doi:10.1081/amp-120022017

Cited by 20 publications

(9 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To decide the architecture of a system in a high level synthesis step of VLSI design, Mandal and Chakrabarti 164 modified the classical GA in terms of population control, cross-over control, solution completion using approximation algorithms for generating better results in solving several NP complete problems. Pettersson et al 165 applied GA to train NN weights as well as connectivity by optimising the model structures and parameters simultaneously (tackling part of the parameter estimation by linear least square techniques) for building a parsimonious network with quicker convergence while building a relationship between burden layer thickness and key charge variables in a blast furnace operation. Pettersson et al 166 also developed a simplified model based on first principles for the blast furnace burden distribution which they used to optimise several operating variables, e.g.…”

Section: Other Applicationsmentioning

confidence: 99%

Genetic algorithms in polymeric material production, design, processing and other applications: a review

Mitra¹

2008

International Materials Reviews

View full text Add to dashboard Cite

Genetic algorithms (GAs) belong to the broader family of generic population based metaheuristic optimisation algorithms called evolutionary algorithms. These are one of the biology inspired computing techniques that emulate the basic Darwinian concept of natural selection. Because of their many positive attributes such as ease of use, versatility and robustness, GAs have become important in materials research very rapidly since their inception. Though going through a major stage of theoretical development, GAs have been successfully applied in many fields such as polymer production, polymer design, polymer processing, polymer laminates and other polymers, as well as non-polymer applications in recent times. The present review attempts to present the state of the art in this area. The review is broadly divided into three sections: fundamentals of GAs, the use of GAs in polymeric materials and the use of GAs in non-polymer applications. First, the fundamentals of GAs are presented to provide some glimpse of the technique; this is followed by presentations on applications of GAs in various polymeric fields, such as in polymer production, polymer scheduling, polymer design, polymer laminates and other polymer related applications and non-polymeric fields, e.g. metallic laminates and other metallic material applications. A thorough literature review and critical analysis of the review is provided and attempts have been made to portray how evolutionary optimisation techniques, GAs in particular, are used across a variety of applications.

show abstract

Section: Other Applicationsmentioning

confidence: 99%

Genetic algorithms in polymeric material production, design, processing and other applications: a review

Mitra¹

2008

International Materials Reviews

View full text Add to dashboard Cite

show abstract

“…This problem can be relaxed by a partial training of the network, as outlined in [12]. In some recent papers the possibility to use genetic algorithms (GA) for a simultaneous optimization of network weights and structure has also been explored [13][14][15][16] as well as a multiobjective procedure for evolving radial basis networks [17].…”

Section: Introductionmentioning

confidence: 99%

“…In order to achieve both of these objectives, they should be simultaneously considered. The architecture of the lower part of the network and its corresponding weight matrix were treated as variables influencing the objective functions and were evolved by a genetic process, while the weights in the upper part of the network were determined by linear least squares [16]. Since the two objectives to be minimized, i.e., the training error, F 1 , and the required number of active connections in the lower part of the network, F 2 , are clearly conflicting, the trade off between them can be represented as a Pareto front [18,19].…”

Section: Introductionmentioning

confidence: 99%

Evolving Nonlinear Time-Series Models of the Hot Metal Silicon Content in the Blast Furnace

Saxén

Pettersson

Gunturu

2007

Materials and Manufacturing Processes

Self Cite

View full text Add to dashboard Cite

Neural networks are versatile tools for nonlinear modeling, but in time-series modeling of complex industrial processes the choice of relevant inputs and time lags can be a major problem. A novel method for the simultaneous detection of relevant inputs and an appropriate structure of the lower part of the networks has been developed by evolving neural networks by a genetic algorithm, where the approximation error and the number of weights are minimized simultaneously by multiobjective optimization. The networks on the Pareto front are considered possible candidate models that are evaluated on an independent test set. In order to consider the problem of drift in the variables, which may cause parsimonious models to perform poorly on the test set, the weights in the upper layer of the networks are recursively estimated by a Kalman filter. The method is illustrated on a data set from ironmaking industry, where time-series models of the hot metal silicon content in a blast furnace are evolved. The technique is demonstrated to synthesize models with a choice of inputs in agreement with findings presented in the literature and process know-how.

show abstract

“…Saxén H also used the genetic algorithms to train neural network weights [8] and to optimization the charging process in bell-type BF [9] . However, the burden distribution control in bell-less BF is even more complicated.…”

Section: Introductionmentioning

confidence: 99%

Burden distribution control based on adaptive genetic algorithm in bell-less blast furnace

Zhu

Yin

Cheng

et al. 2014

Proceedings of the 33rd Chinese Control Conference

View full text Add to dashboard Cite

The proper control of the ratio of ore to coke distribution (ROCD) can achieve energy saving of the blast furnace (BF). In order to achieve the desired ROCD, it is meaningful to search best charging system (especially the burden matrix). This paper deals with the problem of burden distribution control based on the adaptive genetic algorithm and multi-radar data. Firstly, the definition of the ratio of ore to coke (ROC) and the desired ROCD are given. Secondly, according to the desired ROCD, the concrete steps of searching the best burden matrices are presented based on the adaptive genetic algorithm. Then, the operators of the adaptive genetic algorithm (especially the mutation operator) are properly adjusted, according to the actual production process of the bell-less BF. Finally, three computational experiments demonstrate the effectiveness of the method.

show abstract

Evolutionary Neural Network Modeling of Blast Furnace Burden Distribution

Cited by 20 publications

References 11 publications

Genetic algorithms in polymeric material production, design, processing and other applications: a review

Genetic algorithms in polymeric material production, design, processing and other applications: a review

Evolving Nonlinear Time-Series Models of the Hot Metal Silicon Content in the Blast Furnace

Burden distribution control based on adaptive genetic algorithm in bell-less blast furnace

Contact Info

Product

Resources

About