Regression and soft computing models to estimate young's modulus of CO2 saturated coals

Young's modulus of coal is a very important deformational property which dictates how the material will behave in presence of sub- and super-critical carbon dioxide during sequestration. But this is also a difficult property to measure due to the extensive instrumental requirements, and wide compositional and structural heterogeneity of the coal. Therefore, an indirect method to measure the saturated Young's modulus of coals has been developed in the present research using artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS). Low and high rank specimens from three different basins of India and Australia have been used for the analysis. Saturation pressure and the compressive strength (UCS) of the coal specimens have been used as the input parameters to build the models. The performance of the models were evaluated against the multivariate regression model using four different types of statistical parameters such as root mean square error (RMSE), coefficient of determination (R2), mean absolute error percentage (MAPE), and variables accounted for (VAF). Results show that ANFIS model is the best performing one with the least RMSE and MAPE, and highest VAF and R2. This research demonstrates that it is possible to develop soft computing models that can successfully estimate some of the critical rock mechanics parameters essential for the technical evaluation of the sequestration projects on coal. This generalized model is also excellent in incorporating the possible effect of heterogeneity in specimens and performs well for samples that show similar data distribution. If applied, this can potentially reduce the requirement of extensive, complex and expensive instruments that are required for similar investigations.