基于神经网络的采场凿岩爆破参数优化及应用
|
摘要
针对二道沟金矿生产过程中存在的矿石贫化率大、生产效率低等问题,基于神经网络智能算法,对采场的凿岩爆破参数进行优化。将炮孔深度、孔间距及装药系数作为模型的输入因子,采幅和矿石块度合格率作为模型的输出因子,根据现场实测数据,结合遗传算法,建立最终的神经网络模型,并对红旗矿区的采场凿岩爆破参数进行优化,确定最佳爆破参数为:一分采炮孔深度0.90 m、孔间距0.40 m、装药系数0.60;二分采炮孔深度1.1 m、孔间距0.35 m、装药系数0.70。现场应用结果表明:优化的参数合理,有效地控制了采幅与矿石块度,降低了矿石贫化率,提高了矿山的经济效益。
In light of the problems existing in the operation of Erdaogou Gold Mine such as high dilution rate of ores and low production efficiency,the paper optimized the drilling and blasting parameters of the stope based on neural network intelligent algorithm. The borehole length,borehole intervals and charging coefficient are input factors of the model while the mining range width and ore block size pass rate are output factors. Based on field test data and genetic algorithm,the neural network model was established. The drilling and blasting parameters of the stope in Hongqi District were optimized. The results showed that the optimal blasting parameters were as follows: borehole length was0. 90 m,borehole interval was 0. 40 m and charging coefficient was 0. 60 for the first sub-blasting; borehole length was1. 1 m,borehole interval was 0. 35 m and charging coefficient was 0. 70 for the second sub-blasting. Field application results show that the optimized parameters are reasonable,can effectively control the mining range width and ore block size,lowers ore dilution rate and improves economic benefits for mines.
In light of the problems existing in the operation of Erdaogou Gold Mine such as high dilution rate of ores and low production efficiency,the paper optimized the drilling and blasting parameters of the stope based on neural network intelligent algorithm. The borehole length,borehole intervals and charging coefficient are input factors of the model while the mining range width and ore block size pass rate are output factors. Based on field test data and genetic algorithm,the neural network model was established. The drilling and blasting parameters of the stope in Hongqi District were optimized. The results showed that the optimal blasting parameters were as follows: borehole length was0. 90 m,borehole interval was 0. 40 m and charging coefficient was 0. 60 for the first sub-blasting; borehole length was1. 1 m,borehole interval was 0. 35 m and charging coefficient was 0. 70 for the second sub-blasting. Field application results show that the optimized parameters are reasonable,can effectively control the mining range width and ore block size,lowers ore dilution rate and improves economic benefits for mines.
引文
[1]李超亮.露天矿山爆破效果与综合成本优化研究[J].矿冶工程,2012,32(6):26-28.
[2]师雪娇,彭超.浅析黄麦岭磷矿露天爆破大块产生的原因及解决措施[J].爆破,2012,29(1):48-50,76.
[3]王玉杰.爆破工程[M].武汉:武汉理工大学出版社,2007.
[4]于宝新.齐大山铁矿爆破优化试验研究[J].金属矿山,2007(5):35-37.
[5]张燕军.盘区采场凿岩爆破参数优化试验研究[J].黄金,2001,22(10):17-21.
[6]王莉,赵彬.矿岩爆破参数预测新方法研究[J].爆破,2010,27(12):27-30.
[7]王创业,张飞天.基于神经网络的露天矿爆破参数优化研究[J].金属矿山,2011(3):57-59.
[8]唐海,袁超.基于神经网络的预裂爆破参数智能设计[J].工程爆破,2012(3):11-15.
[9]郑长青,陈庆涛.基于神经网络的台阶爆破参数优化设计[J].爆破,2008,25(9):22-24.
[2]师雪娇,彭超.浅析黄麦岭磷矿露天爆破大块产生的原因及解决措施[J].爆破,2012,29(1):48-50,76.
[3]王玉杰.爆破工程[M].武汉:武汉理工大学出版社,2007.
[4]于宝新.齐大山铁矿爆破优化试验研究[J].金属矿山,2007(5):35-37.
[5]张燕军.盘区采场凿岩爆破参数优化试验研究[J].黄金,2001,22(10):17-21.
[6]王莉,赵彬.矿岩爆破参数预测新方法研究[J].爆破,2010,27(12):27-30.
[7]王创业,张飞天.基于神经网络的露天矿爆破参数优化研究[J].金属矿山,2011(3):57-59.
[8]唐海,袁超.基于神经网络的预裂爆破参数智能设计[J].工程爆破,2012(3):11-15.
[9]郑长青,陈庆涛.基于神经网络的台阶爆破参数优化设计[J].爆破,2008,25(9):22-24.