The use of artificial neural network (ANN) for prediction of removal of Co2+ and Ni2+ ions from waste water by electric furnace slag

Žmak, Irena and Ćurković, Lidija and Filetin, Tomislav (2010) The use of artificial neural network (ANN) for prediction of removal of Co2+ and Ni2+ ions from waste water by electric furnace slag. = The use of artificial neural network (ANN) for prediction of removal of Co2+ and Ni2+ ions from waste water by electric furnace slag. In: 69th World Foundry Congress 2010, WFC 2010, 16-20.10.2010., Hangzhou; China.

[img]
Preview
Text
Dummy article.pdf - Draft Version Jezik dokumenta:English

Download (79kB) | Preview
Official URL: http://www.scopus.com/inward/record.url?eid=2-s2.0...

Abstract

The objectives of this work was the study the removal of Co2+ and Ni2+ ions from aqueous solution by sorption onto five different electric furnace slag. All experiments were performed in batch conditions. The slag was obtained through the manufacturing processes of a fire-resistant cast steel( G-X40CrNiSi25-20) and a low-alloyed Cr-Mo-Ni cast steel, according to its chemical analysis. The sorption of metal ions on the slag depends on the chemical composition of the slag, initial ion concentration and type of the present metal ions. On all the examined electric furnace slag samples, sorption capacity for Ni2+ is higher than for Co2+ . This paper presents the results of application of artificial neural networks in predicting the Co2+ and Ni2+ removal from aqueous solutions. A static multi-layer feed-forward artificial neural network with the back propagation training function and LevenbergMarquardt optimization was used to predict the metal ions removal. The error-back propagation learning algorithm was used, with the assistance of Matlab 7.6.0 (R2008a) Neural network toolbox. The early stopping method was applied, in order to prevent the network from over-fitting. Data used for neural network testing were not used for network training. When experimental data and data obtained by neural network prediction were compared, it was concluded that the applied network model provides very good prediction of the quantity of bound metal ions. The mean error and the standard deviation were found to be very good.

Item Type: Conference or Workshop Item (Lecture)
Keywords (Croatian): Bound metals; Cast steel; Chemical compositions; Early stopping; Feed-forward artificial neural networks; Ion concentrations; Levenberg-Marquardt optimization; Manufacturing process; Mean errors; Metal ions removal; Network models; Network prediction; Network training; Neural network toolboxes; Overfitting; Sorption capacities; Standard deviation; Training function; Backpropagation; Electric furnaces; Forecasting; Foundries; Heavy metals; Learning algorithms; Metal ions; Neural networks; Slags; Sorption; Steel castings; Steel metallurgy; Adsorption
Subjects: TECHNICAL SCIENCE > Mechanical Engineering
TECHNICAL SCIENCE > Basic technical sciences
Divisions: 1000 Department of Materials > 1010 Chair of Materials and Tribology
Indexed in Web of Science: No
Indexed in Current Contents: No
Citations SCOPUS: 0 (17.8.2015.)
Date Deposited: 16 Apr 2015 10:15
Last Modified: 17 Aug 2015 10:51
URI: http://repozitorij.fsb.hr/id/eprint/3603

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year

Nema podataka za dohvacanje citata