Two Blasting Vibration Prediction Models Based on Optimized Machine Learning Algorithms
Download as PDF
Haokai Jiao, Xiliang Zhang
the Blasting Vibration of Large Open-Pit Mines Has a Great Impact on the Production Safety of the Mine and the Stability of the Slope Rock Mass, So Ensuring That the Prediction Results of the Blasting Vibration Speed Are Accurate is an Important Part to Guarantee the Safe and Efficient Production of the Mine. However, the Traditional Machine Learning Method is Difficult to Accurately Predict the Velocity of Blasting Vibration. Therefore, the Pso Algorithm and Gsm Algorithm Are Introduced to Optimize the Parameters of the Support Vector Machine (Svm), and Construct Pso-Svm Algorithm and Gsm-Svm Model to Improve the Prediction Accuracy. 63 Groups of Blast Monitoring Data of a Mine in Eastern Zambia is Used to Train and Verify the Two Algorithms. the Result Show That the Pso-Svm Blasting Vibration Prediction Model Has the Most Accurate Prediction (87.71%), While the Prediction Accuracy of the Gsm-Svm Blasting Vibration Prediction Model is 86.53%. by Contrast, the Pso-Svm Type Prediction Has Higher Accuracy and Stronger Generalization Ability, Which Provides a Way of Thinking for the Prediction of Blasting Vibration Speed.
Blast Vibration Prediction; Pso-Svm; Gsm-Svm