摘要
以首云露天铁矿爆破震动实测数据为基础,运用BP神经网络原理,建立了以总药量、最大段药量、爆心距和高程差为输入层的神经网络模型,预测爆破震动峰值振速。将爆破震动监测数据与BP神经网络预测结果和萨氏公式预测结果比较,BP神经网络预测爆破震动峰值平均误差为17.16%,远低于萨氏公式预测的平均误差44.12%,表明BP神经网络预测爆破震动强度精度更高,结果更可靠。
According to the blasting vibration measured data of Shouyun open-pit mine,parameters of total charge,maximum explosive,distance of explosive source and evaluation difference are regarded as the input layer to establish neural network model based on the BP neural network principle to predict the peak velocity of blasting vibration.The measured data of blasting vibration,the prediction results of BP neural network model and Sadaovsk formula are conducted contrast analysis.The results show that the average error of the peak velocity of blasting vibration of BP neural network model is 17.6%,it is far be-low the average error (44.12%)of Sadaovsk formula.The research results further indicated that the pre-diction precise of BP neural network model is higher than the others,the prediction results of BP neural network model is more reliable.
出处
《现代矿业》
CAS
2016年第1期13-16,共4页
Modern Mining
关键词
BP神经网络
爆破振速
预测
BP neural network
Blasting vibration
Prediction
