Regression analysis of the blasting vibration effect in cross tunnels

A regression analysis method integrating signal preprocessing, dimensional analysis, and particle swarm optimization is introduced and applied to a cross tunnel blasting engineering example. First, through complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) decomposition-wavelet packet threshold preprocessing, the interference component in the original blasting signal is removed. Then, dimensional analysis is used to obtain a model equation that can reflect the attenuation law of cross tunnel blasting vibration. Finally, this model equation is used as the fitness function for particle swarm optimization to obtain undetermined parameters. In addition, the data are processed by other fitting models, and the results are compared with the fitting effects analyzed in this article. The results show that compared with the existing fitting models, the analysis method introduced in this article has a good fitting effect. Thus, this work can provide a reference for similar tunnel blasting engineering projects.