Development of a model to predict vibrations induced by transient release of in-situ stress

During the process of deep rock mass excavations by drill and blast, transient release of in-situ stress (TRIS) induced vibrations have a negative effect on the safety and stability of nearby structures. However, little attention has been focused on the prediction of peak particle velocity (PPV) of TRIS induced vibrations. To forecast the PPV of TRIS induced vibrations, a new model was set up by dimensional analysis method based on an analysis of PPV influencing factors. To train and test the prediction model proposed in this paper, firstly, TRIS induced vibrations were identified and separated from the recorded vibrations during a blasting excavation of deep rock masses in the Jinping II hydropower station by the methods of amplitude spectrum analysis and finite impulse response digital filter, and then, unknown coefficients in the proposed model were calculated via multivariate regression analysis from the collected data of upper part excavation. Finally, for the lower part excavation, the separated and predicted PPV of TRIS induced vibrations were compared. In addition, the correlation coefficient and root mean square error (RMSE) were compared between the proposed model and the other commonly used model. Results seem to indicate that the proposed model with a higher correlation coefficient and a smaller RMSE is the better option for predicting the PPV of TRIS induced vibrations.