Academic Journal of Engineering and Technology Science, 2025, 8(1); doi: 10.25236/AJETS.2025.080101.
Zhang Wentao, Wang Haibo
School of Civil Engineering and Architecture, Anhui University of Science and Technology, Huainan, China
Accurately predicting the Peak Particle Velocity (PPV) of blasting vibrations is crucial for controlling the hazards of blasting vibrations. To improve the accuracy of PPV prediction, a model based on Bayesian Optimization Algorithm (BOA) and Extreme Gradient Boosting (XGBoost) is proposed, under feature selection conditions. First, correlation analysis combined with variance analysis is used to filter the initial features. Then, the Bayesian optimization algorithm is applied to fine-tune the hyperparameters of XGBoost, and the optimal hyperparameters are input into the prediction model for training, testing, and evaluation. Finally, the SHAP method is used for interpretability analysis of the model. The results show that optimizing the XGBoost hyperparameters through Bayesian optimization can alleviate overfitting caused by improper hyperparameter selection, improving the model's prediction accuracy and generalization ability. Compared to five other models, BOA-XGBoost demonstrates higher prediction accuracy and stronger nonlinear fitting performance. The importance ranking of features influencing the PPV of blasting vibrations is as follows: D > Qmax > T > Pf > L > N. The blast center distance and number of holes have a negative impact on PPV, while the maximum charge per blast section, stemming length, and powder factor have a positive impact.
Blasting vibration PPV prediction; interpretability analysis; Extreme gradient boosting algorithm; bayesian optimization algorithm; SHAP method
Zhang Wentao, Wang Haibo. Prediction and Interpretability Analysis of Mine Blasting Vibration PPV Based on the BOA-XGBoost Model. Academic Journal of Engineering and Technology Science (2025) Vol. 8, Issue 1: 1-7. https://doi.org/10.25236/AJETS.2025.080101.
[1] Gou Y, Shi X, Huo X, et al. Motion parameter estimation and measured data correction derived from blast-induced vibration: New insights[J]. Measurement, 2018, 135.
[2] HE L, YIN L, ZHONG D W, et al. Research review on blast vibration intensity, waveform and spectrum:prediction and active Control[J]. Blasting, 2024, 41(03):189-204+262.
[3] HU X B, CHEN Z Y, WEI G P, et al. Blasting Vibration Prediction System Based on BP Neural Network[J]. Mining Research and Development, 2020, 40(09):154-158.
[4] Yue Z W, Wu Y X, Wei Z, et al. Prediction of blasting vibration effect in open-pit mine based on PSO-LSSVM model[J]. Engineering Blasting, 2020, 26(6):1-8.
[5] WANG X Y, CAO P F, XIAO Y Q, et al. Blasting vibration prediction based on WOA-SVM model[J]. Mining and Metallurgical Engineering, 2023, 43(4):48-51.
[6] Chen T, Guestrin C. Xgboost:A scalable tree boos-ting system[C]//Proceedings of the 22nd Acm Sigkdd International Conference on Knowledge Discovery and Data Mining, 2016:785-794.
[7] Snoek J, Larochelle H, Adams R P. Practical Bayesian Optimization of Machine Learning Algorithms [J]. Advances in neural information processing systems, 2012, 4.
[8] Zhou S, Z. -X. Z, Luo X. Huang Y. Yu Z. Yang X. Predicting dynamic compressive strength of frozen-thawed rocks by characteristic impedance and data-driven methods[J]. Journal of rock mechanics and geotechnical engineering, 2024, 16(7):2591-2606.
[9] Kang J Y, Zhang S X, Zhang Q P, et al. Transformer fault diagnosis method based on ANOVA and BO-SVM[J]. High Voltage Technology, 2023, 49(05):1882-1891.
[10] Yu Z, Shi X, Zhou J, et al. A new multikernel relevance vector machine based on the HPSOGWO algorithm for predicting and controlling blast-induced ground vibration[J]. Engineering with Computers, 2020.
[11] Chelgani S, H. Nasiri, M. Alidokht. Interpretable modeling of metallurgical responses for an industrial coal column flotation circuit by XGBoost and SHAP-A “conscious-lab” development[J]. International Journal of Mining Science and Technology, 2021, 31(06):1135-1144.