Comparative analysis of intelligent models for prediction of Langmuir constants for CO2 adsorption of Gondwana coals in India

Abstract

An artificial neural network (ANN) is an artificial intelligence technique in which performance can be improved by adapting to the changes in the environment. The efficient manipulation of large amounts of data and the ability to generalize results are the main advantages of neural networks. Considering the advantages of this technique, this present paper aims to perform a comparison between linear methods like Multivariate Regression Analysis (MVRA) and different ANN techniques such as back propagation with regression analysis (BPNN), layer recurrent neural network (LRNN), generalized regression neural network (GRNN) and radial basis neural network (RBNN). This comparison was performed to predict the approximate values of Langmuir volume constant (LVC) and Langmuir pressure constant (LPC) for CO2 adsorption in coal using proximate and maceral properties of India’s major coalfield as input parameters. It is found that RMSE value for RBNN is least followed by GRNN, LRNN, BPNN and MVRA for both LVC and LPC models. Based on the best network, it is found that coal seams from Narayankuri coal mine has highest adsorbing capacity of CO2 (0.0019791 mol/gm) as compared to other coal seams of this study.