摘要
爆破所引起的震动强度在很大程度上具有随机性,它受到众多因素的影响,而且各因素之间存在着极其复杂的非线性关系。运用人工神经网络原理,以孔深、孔数、孔距、最大齐爆药量、总药量和爆源距作为影响爆破振动的主要因素,建立BP神经网络模型1,对爆破振动速度进行预测,并与仅考虑最大齐爆药量和爆源距作为输入量的BP神经网络模型2和萨道夫斯基公式预测结果以及现场监测结果进行了比较。分析结果表明,模型1的预测精度较模型2提高6%以上,较传统方法提高27%以上。
The vibration caused by blasting is chiefly random. It is affected by a lot of factors, and there is a complicated non-linear relationship between the various factors. A neural network model one based on theory of artifi- cial neural network is established to forcast the vibration velocity of blasting seism. The main factors affecting the blasting are the hole depth, number, and distance, maximum charge weight per delay interval,total charge and explo- sive distance. The maximum charge weight per delay interval and the explosion source distance are considered as the input of the neural network model two, which are compared to the predicted results of Sadowsky fomular and the field surveillent results. The results revealed that The first model forecasting precision was proved to be satisfactory with an error ratio less than 6% ,and its precision is better than the traditional method with an increase of 27%.
出处
《爆破》
CSCD
北大核心
2011年第4期105-107,共3页
Blasting
基金
福建省交通科技发展项目(No.201014
2009~2011)
关键词
爆破振动
BP神经网络
预测
blasting vibration
back-propagation neural network
prediction
