Wyniki wyszukiwania - BazTech

Ograniczanie wyników

1 Archives of Mining Sciences

1 2021

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Sortuj według:

Ogranicz wyniki do:

Artificial intelligence-based modeling for the estimation of Q-Factor and elastic young’s modulus of sandstones deteriorated by a wetting-drying cyclic process

Rashid Hafiz Muhammad Awais, Ghazzali Muhammad, Waqas Umer, Malik Adnan Anwar, Abubakar Muhammad Zubair

Archives of Mining Sciences

2021

Vol. 66, no. 4

635--658

In this study, a series of destructive and non-destructive tests were performed on sandstone samples subjected to wetting-drying cycles. A total of 25 Wet-Dry cycles were provided to investigate any significant change in the engineering properties of sandstones in terms of their porosity, permeability, water absorption, density, Q-factor, elastic modulus (E), and unconfined compressive strength (UCS). The overall reduction in the values of density, E, Q-factor, and UCS was noted as 3-4%, 42-71%, 34-62%, and 26-70% respectively. Whereas, the overall appreciation in the values of porosity, permeability, and water absorption was recorded as 24-50%, 31-64%, and 25-50% respectively. The bivariate analysis showed that the physical parameters had a strong relationship with one another and their Pearson’s correlation value (R) ranged from 0.87-0.99. In prediction modeling, Q-factor and E were regressed with the contemplated physical properties. The linear regression models did not provide satisfactory results due to their multicollinearity problem. Their VIF (variance inflation factor) value was found much greater than the threshold limit of 10. To overcome this problem, the cascade-forward neural network technique was used to develop significant prediction models. In the case of a neural network modeling, the goodness of fit between estimated and predicted values of the Q-factor (R2 = 0.86) and E (R2 = 0.91) was found much better than those calculated for the Q-factor (R2 = 0.30) and E (R2 = 0.36) in the regression analysis.