Prediction of coal response to froth flotation based on coal analysis using regression and artificial neural network E. Jorjani a , H. Asadollahi Poorali a , A. Sam b , S. Chehreh Chelgani a, * , Sh. Mesroghli a , M.R. Shayestehfar b a Department of Mining Engineering, Science and Research Branch, Islamic Azad University, Poonak, Hesarak, Tehran, Iran b Department of Mining Engineering, Shahid Bahonar University of Kerman, Kerman, Iran article info Article history: Received 14 November 2008 Accepted 7 March 2009 Available online 7 April 2009 Keywords: Coal Neural networks Froth flotation Modeling abstract In this paper, the combustible value (i.e. 100-Ash) and combustible recovery of coal flotation concentrate were predicted by regression and artificial neural network based on proximate and group macerals anal- ysis. The regression method shows that the relationships between (a) ln (ash), volatile matter and mois- ture (b) ln (ash), ln (liptinite), fusinite and vitrinite with combustible value can achieve the correlation coefficients (R 2 ) of 0.8 and 0.79, respectively. In addition, the input sets of (c) ash, volatile matter and moisture (d) ash, liptinite and fusinite can predict the combustible recovery with the correlation coeffi- cients of 0.84 and 0.63, respectively. Feed-forward artificial neural network with 6-8-12-11-2-1 arrange- ment for moisture, ash and volatile matter input set was capable to estimate both combustible value and combustible recovery with correlation of 0.95. It was shown that the proposed neural network model could accurately reproduce all the effects of proximate and group macerals analysis on coal flotation system. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction Coal is used for coke making in steel industry of Iran. Ash has highly adverse effects on the productivity and coke consumption in blast furnace. An increase in the ash content of a coal concen- trate by 1% over a critical limit results in an increase in coke con- sumption by about 4–5% as well as decreasing in blast furnace productivity by about 3–6%; therefore there is always a pressure on coal preparation plants to supply coal with very low ash (Dey and Bhattacharyya, 2007). Beneficiated coal is usually produced in coal washery plants where mineral matter is decreased and organic matter is increased. Conventionally, coarse coal is processed through gravity separation systems and fine coal through flotation. Froth flotation is the well- established process for fine coal cleaning. Flotation is based on exploiting differences in the surface properties of the organic material in coal and in mineral matter. The organic matter is hydrophobic, particularly in bituminous coals, and is not readily wetted as the mineral surfaces. Flotation is also dependent on coal rank, floatability increasing with increasing rank through the bitu- minous range (Hower et al., 1984). Aplan (1977) reported that among the minerals most common in coal, pyrite has hydrophobic surface property, while the silicates are generally hydrophilic. The three maceral groups are vitrinite, inertinite (fusinite) and liptinite (exinite) and are present in varying proportions in differ- ent coal samples, along with minerals. The properties of coal mac- erals such as elemental analysis (Kessler, 1973), moisture content, hardness and density (Laskowski, 2001) differ widely and change during coalification. It was reported that with increasing rank, the coal macerals become more similar in chemical properties and their petrographically distinction decrease (Laskowski, 2001). Aplan and Arnold (1986) measured contact angles on several US coal samples to quantify the hydrophobicity of individual coal maceral. It was found that an order of liptinite > vitrinite > inerti- nite in hydrophobicity, have a typical ranges of contact angles of 90–130°, 60–70° and 25–40°, respectively. Hirt and Aplan (1991) examined Eastern Kentucky coal and arranged the macerals according to decreasing of floatability in order of pseudovitrinite (high Rmax) > pseudovitrinite (low Rmax) > vitrinite (high Rmax) > vitrinite (low Rmax) = micrinite = exinite = semifusinite > resinite > fusinite. Because coal ’ s organic and mineral matters have different hydrophobicity values, the combustible value (CV) and combusti- ble recovery (CR) of coal flotation concentrate can be affected where the coals from different seems are fed to the flotation circuit. Artificial neural network (ANN) is a powerful tool and have been applied successfully in numerous fields, for example, model- ing the greenhouse effect (Seginer et al.,1994), simulation N 2 O emissions from a temperate grassland ecosystem (Ryan et al., 2004), bioleaching of metals (Laberge et al., 2000), prediction of coal microbial and chemical desulphurization (Acharya et al., 2006; Jorjani et al., 2007, 2008a) as well as coal Hargrove 0892-6875/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.mineng.2009.03.003 * Corresponding author. Tel.: +98 912 3875716; fax: +98 21 44869744. E-mail address: sos4552@gmail.com (S. Chehreh Chelgani). Minerals Engineering 22 (2009) 970–976 Contents lists available at ScienceDirect Minerals Engineering journal homepage: www.elsevier.com/locate/mineng