Application of SVM in the estimation of GCV of coal and a comparison study of the accuracy and robustness of SVM

Abstract

Gross calorific value (GCV, HHV) is an important property of coal, but its time-consuming mensuration cannot always satisfy the practical demands. This paper investigates the application of statistics models to measure GCV quickly and accurately using coal components with mensuration that has been achieved in real time on-line in China to meet practical demands. Linear regression (LM), nonlinear regression equation (NLM), and artificial neural networks (ANN) have been developed for the estimation of GCV by researchers. In this paper, 1400 data points are used to predict the GCV of China coal. The estimating methodology progress is determined using the support vector machine (SVM), and the estimating robustness is evaluated. The comparison study manifested that the SVM model outperformed the three existing models in terms of accuracy and robustness. Meanwhile, the sampling method is improved, and the input variables are reduced to those that can be measured in real time on-line.