PREDICTION OF ROCK TENSILE FRACTURE TOUGHNESS: HYBRID ANN-WOA MODEL APPROACH

Abstract

Various techniques are used in rock engineering to evaluate tensile fracture toughness, which is a critical parameter in assessing and designing stable rock structures. These methods typically involve laboratory investigations and statistical analysis. Nevertheless, artificial neural networks can also establish correlations among different data sets. Artificial intelligence approaches are becoming increasingly essential in all engineering fields, including the ones that study rock fracture mechanics. In this work, an artificial neural network with a hidden layer and eight neurons as well as a hybrid artificial neural network with a whale optimization algorithm were utilized to determine the tensile fracture toughness of rocks. In order to develop accurate models, this study has carefully selected four fundamental parameters to serve as inputs. These parameters include radius, thickness, crack length, and mean tensile strength of specimens. Also, 113 rock datasets were collected for models. The results show that utilization of the optimization algorithm enhances the precision in estimating the tensile fracture toughness of rocks. The R2 improved to 0.93 when the whale optimization algorithm was used. On the other hand, the correlation factor reached 0.81 when the whale optimization algorithm was not implemented.