爆破参数的BP神经网络优选及验证  

Optimization and validation of blasting parameters based on BP neural network

作　　者：王永奇 戴兵[2] 

机构地区：[1]贵州开磷集团矿业总公司,贵州贵阳550030 [2]中南大学资源与安全工程学院,湖南长沙410083

出　　处：《矿业工程研究》2012年第1期1-3,共3页Mineral Engineering Research

基　　金：国家科技支撑计划课题资助项目(2007BA08B01)

摘　　要：爆破矿石块度大小及其均匀程度是反映爆破效果好坏的关键指标,它不仅直接影响采矿作业后续工序如装载、运输等设备工作效率和磨损程度,还严重影响采矿成本.因此利用BP神经网络对开阳磷矿的凿岩爆破参数进行优选,以排距、孔底距及炸药单耗作为输出结果,以国内同类矿山作为训练样本进行练习,计算得到了最优爆破参数,并根据最优爆破参数进行了现场试验验证,结果显示大块率有了明显的降低,因此这种方法由于良好的发展前景.Size of blasting ore and its uniform degree is a key indicator for whether the blasting effect is good or bad.It not only directly influences subsequent working procedure of the mining operations,such as the equipment work efficiency of loading as well as transportation and wear degree.Furthermore,it seriously affects the mining cost,so the blasting parameters are optimized in Kaiyang Phosphorite Group Company with BP neural network.It takes row spacing,hole bottom distance and explosive unit consumption as the output results.It also takes the domestic similar mines as the training samples for training.The optimal blasting parameters are calculated,and the optimal blasting parameters are tested in site.The results show that the optimal parameters have a good blasting effect.This method has a wide application.

关 键 词：BP神经网络 爆破参数 影响因素 预测 

分 类 号：TD235.1[矿业工程—矿井建设]