Research on the Prediction of Blasting Vibration Velocity in Open Pit Mines Based on the SSA-BP Model

A research system for the propagation of blasting vibrations and their impact on buildings was established based on a loose blasting project in the Halwusu Open-pit Coal Mine. A traditional prediction model based on the Sadovsky formula was constructed using Python and monitoring data. Based on the gray relational analysis, the main factors influencing the peak vibration velocity of particles were found to be in the following sequence: total explosive quantity, hole depth and spacing, and the blast-center distance. The main vibration frequency was controlled by the number of holes and the maximum dosage of the explosives in each section. The SSA-BP intelligent prediction model was established by combining the sparrow search algorithm (SSA) and the back-propagation (BP) neural network. The average error of this model in predicting the peak vibration velocity was only 4.22%, and the goodness-of-fit R ² value reached 0.9836. The average error in predicting the main vibration frequency was 5.84%, with a R ² value of 0.9398. This result represents a significant improvement in prediction accuracy. A four-hole differential blasting numerical model was established using LS-DYNA, and the vibration velocity analysis revealed the highest simulation accuracy in the horizontal radial direction (6.65% error), whereas errors in the tangential and vertical directions were 32.37% and 15.79%, respectively. Moreover, the prediction advantage of the SSA-BP model was verified for blast-center distances in the range of 800-1200 m (error <10%), providing a reliable theoretical basis and numerical evaluation method for optimizing blasting parameters and ensuring the safety of surrounding buildings.