A method for predicting uniaxial compressive strength of rock mass based on digital drilling test technology and support vector machine

Uniaxial compressive strength (UCS) is an important index for classification of the surrounding rock and determination of the supporting parameters in underground engineering. The commonly used standard UCS tests are expensive, time-consuming and difficult to quantitatively evaluate the UCS of fragmented rocks because these rocks cannot be effectively cored. To solve the above problems, a method for predicting UCS of rock mass based on digital drilling test technology is introduced in this paper. The key to implementing this method is to establish a quantitative and universal relationship between the drilling parameters and the UCS. Therefore, the digital drilling tests and standard UCS tests of rock specimens with different strength values are carried out based on the laboratory rock mass digital drilling test system developed by the authors. A relational model between the drilling parameters and the UCS is established by support vector machine (SVM). The drilling parameters as model input parameters include the rotation speed N, drilling rate V, torque M and thrust F measured by the sensors, and the unit cutting energy eta(c) deduced from energy analysis by the authors as well. The research shows that the UCS predicted by the relational model in validation set is close to the UCS measured by the uniaxial compression test, the coefficient of determination R-2 is 0.977, and the mean absolute error (MAE) is 3.037 MPa. These results indicate that the relational model between the drilling parameters and the UCS based on SVM method is successful in the rock UCS prediction, and the digital drilling test technology could realize the effective prediction of UCS of rock mass.