欢迎访问《现代矿业》杂志官方网站,今天是 分享到:

现代矿业 ›› 2016, Vol. 32 ›› Issue (09): 18-20+27.

• 采矿工程 • 上一篇    下一篇

基于RBF神经网络的爆破参数优选

陈琼1,欧洪宁1,姜群1,张钦礼2,刘伟军2   

  1. 1.锡矿山闪星锑业有限责任公司;2.中南大学资源与安全工程学院
  • 出版日期:2016-09-20 发布日期:2016-11-14

Optimization Selection of Blasting Parameters Based on RBF Neural Network

Chen Qiong1,Ou Hongning1,Jiang Qun1,Zhang Qinli2,Liu Weijun2   

  1. 1.Hsikwang Shan Twinkling Star Co.,Ltd;2.School of Resources and Safety Engineering,Central South University
  • Online:2016-09-20 Published:2016-11-14

摘要: 为确定合理的爆破参数,建立了RBF神经网络模型,统计了8个矿山的样本数据,将影响矿岩可爆性的6项因素:矿石容重、弹性模量、抗拉强度、矿石坚固性系数、摩擦角、黏结力作为RBF神经网络模型的输入因子,排距、孔底距和一次炸药单耗作为影响爆破参数的输出因子,优选样本参数,得出最优的爆破参数。以某矿中深孔爆破为例,通过RBF神经网络模型优选出该矿的爆破参数:排距1.3 m,孔间距2.2 m,炸药单耗0.32 kg/t。实践证明,选择的孔网参数合理,爆破效果良好。

关键词: 爆破参数优选, RBF神经网络, 样本数据, 影响因素

Abstract: In order to obtain the reasonable blasting parameters,the RBF neural network is established,which is used to optimize blasting parameters.The sample data of 8 actual mines are counted,the six influence factors(the volume weight,modulus of elasticity,compressive strength,tensile strength,friction angle and bond strength)that affect the rock mass blastability are considered,which are taken as the input factors of the RBF neural network model,and the factors(rows space,depth of holes and once consumption of dynamite)that influenced blasting parameters are taken as the output factors,so the sample parameters are conducted optimization selection,and the optimal blasting parameters are obtained.Taking the medium-length hole blasting of a mine as an example,the blasting parameters of the mine are conducted optimization selection by the RBF neural network model,the results show that the row space of holes are 1.30 m,the space of holes are 2.20 m,and the explosives consumption is 0.32 kg/t.The practical application proves that the blasting effect is good, and the blasting parameters are reasonable.

Key words: Blasting parameters, RBF neural network, Sample data, Influence factors