前白枣山铁矿地面爆破振动预测

Prediction of Ground Blasting Vibration inQianbaizaoshan Iron Mine

地面振动是爆破作业过程中最严重、最复杂的环境问题之一。爆破振动会对预留的边坡和构筑物产生不利影响,从而影响建筑物结构的完整性。因此,在爆破过程中准确预测地面振动具有重要意义。研究的主要目的是对矿山爆破引起的地面振动进行高精度预测,以减少地面振动对周围环境的影响。为此,提出神经网络(ANN)模型用于前白枣山铁矿爆破振动预测。研究中收集了前白枣山铁矿29次生产爆破振动数据,并建立了8个ANN模型用于PPV预测。为了评价所建立的网络,选择决定系数(R 2)、Root Mean Squared Error(RMSE)和Mean Square Error(MSE)作为网络性能评价指标。比较发现结构为2-6-1的网络性能最佳,R 2为0.92、RMSE为0.4、MSE为0.23。为了证明ANN预测方法的优越性,使用4种常用的经验模型和多元线性回归(MLR)模型预测PPV。结果表明:与经验模型和MLR模型相比,ANN模型具有更好的预测性能。

Ground vibration is one of the most severe and complex environmental problems in blasting operations.Blasting vibration can have adverse effects on reserved slopes and structures,which would consequently affect the integrity of the building structures.Hence,an accurate prediction of ground vibrations during blasting process is of great significance.The main purpose of the research is to make highly precise predictions of the ground vibrations caused by mine blasting,and to reduce the impact of ground vibrations on the surrounding environment.In this study,an artificial neural network(ANN)model was put forward to predict blasting vibrations of Qianbaizaoshan Iron Mine.Besides,vibration data from 29 blasts were collected,and 8 ANN models were established for PPV prediction.In order to evaluate the established network,the coefficient of determination(R 2),Root Mean Squared Error(RMSE)and Mean Square Error(MSE)were chosen as the network performance evaluation indexes.It is found that the 2-6-1 network has the best performance when R 2 is 0.92,RMSE is 0.4 and MSE is 0.23.In order to prove the superiority of ANN forecasting method,four empirical models and multiple linear regression(MLR)models were used to predict PPV.The results show that the ANN model has better prediction performance than the empirical and MLR model.