东北大学学报:自然科学版 ›› 2014, Vol. 35 ›› Issue (11): 1641-1645.DOI: 10.12068/j.issn.1005-3026.2014.11.027

• 机械工程 • 上一篇    下一篇

基于RBF神经网络的采场回采工艺改进试验

杨伟1,杨珊1,张钦礼1,任少峰2   

  1. (1中南大学 资源与安全工程学院, 湖南 长沙410083; 2西南能矿集团股份有限公司, 贵州 贵阳550004)
  • 收稿日期:2013-09-23 修回日期:2013-09-23 出版日期:2014-11-15 发布日期:2014-07-03
  • 通讯作者: 杨伟
  • 作者简介:杨伟(1986-),男,山东济宁人,中南大学博士研究生;张钦礼(1964-),男,山东临朐人,中南大学教授,博士生导师.
  • 基金资助:
    国家自然科学青年基金资助项目(51404305)

Experiment Study of Mining Technology Improvement Based on RBF Neural Network

YANG Wei1, YANG Shan1, ZHANG Qinli1, REN Shaofeng2   

  1. 1. School of Resources and Safety Engineering, Central South University, Changsha 410083, China; 2. Southwest Energy and Mineral Resources Co., Ltd., Guiyang 550004, China.
  • Received:2013-09-23 Revised:2013-09-23 Online:2014-11-15 Published:2014-07-03
  • Contact: YANG Wei
  • About author:-
  • Supported by:
    -

摘要: 某铁矿地下采场回采中存在生产效率低、炸药单耗大和大块率高的问题.为此,提出了水平炮孔前进式开采的改进方案,进行了以排距、孔距、周边孔距为因素的L9(33)的爆破正交试验,建立了以排距、孔距、周边孔距为输入层因子,炸药单耗和大块率为输出层因子的RBF神经网络模型;从安全和经济的角度提出了爆破综合期望指数公式,结合模型预测结果进行最终优选,优选结果为:排距1m,孔间距14m,周边孔距1m.经过现场验证,现生产能力为原来的4倍,增加了可充填的采场数目,顶板暴露时间缩短,生产效率提高约75%,炸药单耗减少62%,大块率降低74%.

关键词: 地下采矿, 采场爆破, 回采工艺改进, 正交试验, RBF神经网络, 爆破综合期望指数公式

Abstract: The underground stoping at a metal mine was in low production efficiency, large explosives consumption and high boulder yield. To solve these problems, an orthogonal test of L9(33) on blasting parameters was proposed, based on which, the RBF neural network model, using the blasting burden spacing, borehole spacing and surroundingborehole spacing as input layer, and taking explosives consumption and boulder yield as the output layer, was established. The comprehensive expectation formula of blasting based on safety and economy is given, and it is determined that the best blasting burden spacing, borehole spacing and surroundingborehole spacing are 1 m, 1. 4m and 1m, respectively. After the mining process improvements, the stope production capacity is 4 times as before, more stopes can be filled, exposure time of stope is shorter, single stope production efficiency is increased by 75%, explosives consumption is reduced by 62%, large fragment rate is reduced by 74%.

Key words: underground mining, stope blasting, mining technology improvement, orthogonal test, RBF neural network, comprehensive expectation formula of blasting

中图分类号: