Prediction of HGI of South African coalfields: a comparative application of ANN, SVR and LSTM models

ABSTRACT

Hardgrove grindability index (HGI) is a significant index used to determine a mill’s capability and overall efficiency in the grinding of coal. There are several factors that affect HGI due to the complex nature of coal. The influence of proximate analysis, calorific value, total sulfur, and pyrite content on the HGI values of 292 coal samples from South African coalfields were examined. Predictive models of HGI are developed by using soft computing techniques such as Long Short-Term Memory (LSTM), support vector regression (SVR), and Artificial Neural Network (ANN). This study shows that ANN is the most effective predictive model for all the three coalfields, with SVR models being second and LSTM models being the least effective. The correlation between the predicted value and the input data is established by using the cosine amplitude method (CAM). It was found that HGI is most influenced by the fixed carbon content, calorific value, ash content and volatile matter content while pyrite has the least influence.

KEYWORDS:

• Hardgrove grindability index
• machine learning
• coalfields
• coal properties

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.