Research on the method of rock mechanics parameters determination while drilling based on machine learning

The measurement of rock mechanics parameters is the basis for the classification of surrounding rock and the design and optimization of supporting parameters in underground engineering. Therefore, the rapid and accurate measurement of rock mechanics parameters is of great significance in ensuring the safe and efficient construction of underground engineering. However, there are still some problems to be solved in the determination of rock mechanics parameters, such as complicated process, long time consuming and high cost. To solve the above problems, a method of rock mechanical parameters determination while drilling based on the hunter-prey optimizer (HPO)-backpropagation (BP) neural network is proposed. Laboratory rock drilling tests and rock mechanical parameters determination tests are carried out. Based on the HPO-BP neural network method, a relationship model between drilling engineering parameters and rock mechanical parameters is established. The results show that there is little difference between the rock mechanical parameters obtained by the verification centralized test and the predicted rock mechanical parameters. The determining coefficients R2 of uniaxial compressive strength, cohesion force c, and internal friction angle φ are 0.9778, 0.9772, and 0.9817, respectively, and the mean difference rates are 1.36%, 3.78%, and 0.81%, respectively. It is proved that the relationship model between drilling engineering parameters and rock mechanics parameters based on the HPO-BP neural network has a good effect on the prediction of rock mechanics parameters, and can effectively measure rock c–φ parameters quickly and accurately.