Estimation of hardgrove grindability index (HGI) based on the coal chemical properties using artifical neural networks

S. khoshjavan¹, M. mazlumi², B. rezai¹ and M. Rezai²

¹Department of Mining and Metallurgy Engineering, Amirkabir University of Technology, Hafez Avenue, Tehran (Iran). ²Department of Mining Engineering, Faculty of Engineering, University of Tehran, Tehran (Iran).

ABSTRACT:

In this research, the effect of different parameters of coal composition (coal chemical properties) were studied, to estimate the coal HGI values index. To estimate the HGI values artificial neural networks (ANNs) and linear multivariable regression methods were used for 400 data. In this work, ten input parameters, such as moisture, volatile matter (dry), fixed carbon (dry), ash (dry), total sulfur (organic & pyretic) (dry), Btu/lb (dry), carbon (dry), hydrogen (dry), nitrogen (dry) as well as oxygen (dry), were used. For selecting the best method to predict HGI values, the responses of aforementioned methods were compared. The results of ANNs, show that the training and test data’s square correlation coefficients (R2 ) achieved at 0.962 and 0.82 respectively. The equation of linear multivariable regression for HGI values were produced. Square correlation coefficients, (R2 ), from regression achieved at 0.76. Sensitivity analysis showed that volatile matter (dry), Btu/lb (dry), carbon (dry), hydrogen (dry), fixd carbon (dry), nitrogen (dry) and oxygen (dry) are the most effective parameters on the HGI.

KEYWORDS:

Coal; HGI; Artificial neural network; Linear multivariable regression; Coal chemical properties

Download this article as: 

Copy the following to cite this article: Khoshjavan S, Mazlumi M, Rezai B, Rezai M. Estimation of hardgrove grindability index (HGI) based on the coal chemical properties using artifical neural networks. Orient J Chem 2010;26(4).

Copy the following to cite this URL: Khoshjavan S, Mazlumi M, Rezai B, Rezai M. Estimation of hardgrove grindability index (HGI) based on the coal chemical properties using artifical neural networks. Available from: http://www.orientjchem.org/?p=11608