Research on explosive mass inversion method based on vibration response of underground structure

Abstract In order to improve the real‐time perception and recognition of blast intensity in underground engineering, model test research on structural vibration responses under varying explosive masses was conducted, and an inversion method for explosive mass estimation based on Sadowski's formula was proposed. The accuracy of this algorithm was verified using model test results, and an improved explosive mass inversion method was developed by combining the particle swarm optimization algorithm with the DBSCAN clustering algorithm. The results indicate that, within a certain distance range, the peak vibration acceleration is negatively correlated with the distance from the explosive source, exhibiting a strong power‐law relationship. In the near field, larger explosive masses result in greater peak vibration accelerations. Conversely, in the middle and far fields, the peak vibration acceleration of stress waves generated by blasts with different explosive masses converges to a similar value. An objective function for explosive mass inversion was developed based on Sadowski's formula, and the particle swarm optimization algorithm was employed to invert field test data, confirming the accuracy of the method. To address the limitation of the particle swarm optimization algorithm caused by its dependence on initial values, it was combined with the DBSCAN clustering algorithm, and the field test data inversion was revalidated. This study demonstrates that the combined optimization approach effectively improves the accuracy and reliability of explosive mass estimation.