Yachen Xie

Accurately determining the uniaxial compressive strength (UCS) of rocks is crucial for various rock engineering applications. However, traditional methods of obtaining UCS are often time-consuming, labor-intensive, and unsuitable for fractured rock sections. In recent years, using Measurement-while-drilling data to identify UCS has gained traction as an alternative approach. To develop a method that can rapidly, efficiently, and economically estimate UCS across different rock types and engineering conditions based on while-drilling tests, this study compiles a comprehensive dataset from existing literature. The dataset includes drilling parameters and their corresponding UCS values, collected under varying lithologies, strength levels, drill bit types, and drilling conditions. Five machine learning models-multilayer perceptron (MLP), support vector regression (SVR), convolutional neural networks (CNN), random trees (RT), and long short-term memory networks (LSTM)-were trained and evaluated. Among these, RT demonstrated superior predictive performance, achieving a root mean square error (RMSE) of 15.851, a mean absolute error (MAE) of 4.449, a standard deviation of residuals (SDR) of 15.292, and an R² value of 0.959 on the test set. SVR also performed well, with an RMSE of 21.905, an MAE of 17.962, an SDR of 21.144, and an R² value of 0.922. While CNN and LSTM exhibited slightly higher errors, they showed better generalization capabilities across validation and test datasets. Furthermore, the models were validated on an unseen independent dataset, where RT achieved the best results, followed by SVR, while the other methods performed relatively poorly. This study indicates that RT and SVR demonstrate superior suitability for UCS prediction.

Measurement while drilling (MWD) data reflect the drilling rig–rock mass interaction; they are crucial for accurately classifying the rock mass ahead of the tunnel face. Although machine-learning methods can learn the relationship between MWD data and rock mechanics parameters to support rock classification, most current models do not consider the impact of the continuous drilling-sequence process, thereby leading to rock-classification errors, while small and unbalanced field datasets result in poor model performance. We propose a novel deep neural network model based on Bi-directional Long Short-Term Memory (BILSTM) to extract information-related sequences in MWD data and improve the accuracy of the rock-mass classification. Two optimization modules were designed to improve the model’s generalization performance. Stratified K-fold cross-validation was used for model optimization in small and unbalanced datasets. Model validation is based on the MWD dataset of a highway tunnel in Yunnan, China. Multiple metrics show that the prediction ability of the network is significantly better than those of a multilayer perceptron (MLP) and a support-vector machine (SVM), while the model exhibits an improved generalization performance. The accuracy of the network can reach 90%, which is 13% and 15% higher than the MLP and SVM, respectively.