Predicting Blast-Induced Area of Tunnel Face in Tunnel Excavations Using Multiple Regression Analysis and Artificial Intelligence

In underground construction, the drilling and blasting method is widely used due to its advantages, such as low cost, simple implementation, and applicability under various geological and hydrogeological conditions. One parameter that significantly affects the effectiveness of drilling and blasting is the post-blast tunnel cross-sectional area. In this study, multiple linear regression analysis (MLRA) and multiple nonlinear regression (MNLR) models were used to predict the area of a tunnel face after blasting, utilizing 136 datasets containing parameters measured from the tunnel face area after blasting during the Deo Ca tunnel construction project. Three deep learning models, an artificial neural network (ANN) and two hybrid models combining an ANN with the particle swarm optimization (PSO) algorithm and an ANN with a genetic algorithm (GA), were then developed to predict the tunnel face area after blasting. The input variables for the calculation and prediction models included the designed tunnel face area (Sd), the specific charge (SC) of the explosion, the average borehole length (L), and the rock mass rating (RMR) of the rock mass on the tunnel face. The GA-ANN model’s results, including determination coefficient (R2) and mean square error (MSE) values of R2train = 0.9562, R2testing = 0.94, MSEtraining = 0.0156, and MSEtesting = 0.0302, indicate that it provides a better prediction of the tunnel face area after blasting than the other models.