Predicting geomechanical, abrasivity, and drillability properties in some igneous rocks using fabric features and petrographic indexes

In this study, we have established and demonstrated the relationships between petrographic and fabric features of igneous rocks and their engineering properties experimentally. To meet this purpose, we have examined several igneous rock specimens and have investigated their engineering properties, including their drillability (drilling rate index (DRI)), abrasivity (Cerchar abrasivity index (CAI)), mechanical features (uniaxial compressive strength (UCS), point load strength index ( I S50 ), Brazilian test strength (BTS)), and their physical properties (dry density, porosity ( N ), and wave velocity (VP)). Then, we have investigated their petrographic features, including shape descriptors, size descriptors, rock fabric features, and mineralogical indexes. We tested 16 types of igneous rocks from 8 various locations in the Gelas tunnel route in Naghadeh City, west Azerbaijan, Iran. The Pearson correlation coefficient indicated a low drillability potential of fine-grained rocks compared to that of coarse-grained rocks. UCS displayed the best Pearson correlation with heterogeneity ( H ) and texture coefficient (TC) ( R = − 0.88 and R = 0.86, respectively). Although the results obtained from multilinear regression (MLR) and multilinear log-linear regression (MLLR) models proved the efficiency of such models in predicting CAI, TC, H , index of interlocking ( g ), and Feldspathic index (IF). Their determination coefficient ( R 2 ) was 0.84 and R 2 = 0.87, respectively. Nevertheless, in comparison, the artificial neural network (ANN) analysis is apparently more efficient than both MLR and MLLR ( R 2 = 0.90). The results revealed rock fabric features have a higher capability in identifying the engineering properties of igneous rocks than their mineralogical composition.