Prediction Of Blasting Vibration Based On Artificial Neural Network And Random Forest

Blasting vibration hazard as the most serious blasting hazard.How to predict and control it better has always been a hot topic in academic and engineering fields.For predicting the blasting vibration better,this paper will focus on the research of blasting vibration prediction technology.By applying methods of comparing the performance of two improved BP neural network algorithm,random forest algorithm and traditional empirical formula method in blasting vibration prediction,the best prediction method will be found.The main research contents and achievements are as follows:(1)This paper summarizes and analyzes the previous blasting prediction methods.Expounding shortcomings of the traditional prediction methods in the application of current engineering blasting by analyzing the principles of the traditional prediction methods.(2)Through the vibration monitoring of the actual engineering blasting for two years,a large number of blasting vibration monitoring data are collected,which provides a sufficient basis for forecasting with the machine algorithm.(3)In order to solve the typical nonlinear problem of blasting vibration prediction,this paper uses two improved BP neural network algorithm and artificial forest algorithm to train and predict the collected data and makes a comparative analysis with the traditional prediction method.The results show that although the traditional formula method is simple and easy to apply,its accuracy is obviously insufficient compared with the other two methods.In the case of small sample size,the random forest has more advantages than the neural network in the prediction of blasting vibration and can control theprediction error of blasting vibration speed and main frequency within 10% and20% respectively.Based on the analysis and comparison of the traditional empirical formula prediction method,two improved BP neural network prediction algorithms,and random forest prediction algorithm,this paper concludes that random forest algorithm has great advantages in the prediction of blasting vibration,and makes a beneficial exploration for better and faster blasting vibration prediction in the future.