We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Application of soft computing methods to estimate uniaxial compressive strength and elastic modulus of soft sedimentary rocks.
- Authors
Shahani, Niaz Muhammad; Zheng, Xigui; Liu, Cancan; Li, Peng; Hassan, Fawad Ul
- Abstract
Uniaxial compressive strength (UCS) and elastic modulus (E) are the fundamental parameters in rock engineering and rock structure designs. The laboratory-based direct measuring of UCS and E according to the recommended International Society for Rock Mechanics (ISRM) and American Society for Testing and Materials (ASTM) standards is complicated, time-consuming, and expensive process involving sample destruction. Likewise, preparation of the required core samples in case of weak, internally damaged, thin, and highly foliated rock is quite difficult. Therefore, the indirect assessment of UCS and E by employing the rock index tests is need of the day. In this study, firstly UCS in MPa, E in GPa, wet density (ρѡet) in g/cm3, dry density (ρd) in g/cm3, and Brazilian tensile strength (BTS) in MPa were measured. Next, the adaptive neuro-fuzzy interference system (ANFIS), artificial neural network (ANN), and multiple linear regression (MLR) models were used to predict the UCS and E of soft sedimentary rocks collected from Block-IX Thar Coalfield. The estimated UCS and E results were compared with their actual values. The performance of ANFIS model for UCS and E shows the most powerful prediction with the strongest coefficient of determination (R2), minimum root mean squared error (RMSE), and the highest variance accounts (VAF) of 0.93 and 0.85, 0.16939 and 0.24702, and 0.96 and 0.91, and 0.95 and 0.91, 0.06938 and 0.03294, and 0.99 and 0.97 at the training and testing stages of the data, respectively. Finally, through the prediction index (PImod) performance test, the accuracy of the proposed models was further improved, which also shows that the prediction accuracy of the ANFIS model is the best among all the models used.
- Publication
Arabian Journal of Geosciences, 2022, Vol 15, Issue 5, p1
- ISSN
1866-7511
- Publication type
Article
- DOI
10.1007/s12517-022-09671-6