基于PSO改进的BP网络在爆破大块率优化中的应用
|
摘要
为解决地下矿山爆破开采采场大块率较高的问题,将PSO算法应用于BP网络中,生成PSO-BP模型对影响大块产生的主要参数进行优化。以参数孔底距、排距、一次炸药单耗、起爆位置为输入因子,大块率为输出因子建立PSO-BP模型,采用现场实测数据初步训练模型,通过控制变量法对模型参数的选取分别进行敏感性分析,得出最佳的大块率PSO-BP评价模型。增加模型各输入因子水平数,按L16(34)正交表组成优选样本,经评价模型的计算预测,搜索出最优的大块率影响参数值。研究结果表明:以东际金矿为例,采用孔底起爆方式,得出最佳大块率预测值9.98%,最优参数值是排距为1.6 m,孔底距为1.8 m,一次炸药单耗为0.350 kg/m~3。
In order to solve the problem of high blasting boulder yield in underground mine stope,the PSO-BP model was produced to optimize the influence parameters of boulder yield by using PSO algorithm in the BP neural network. The parameters,such as hole-bottom spacing,row spacing,specific charge and detonation position,were used as the input data and the boulder yield was set to be the output data in model building. The best PSO-BP evaluation model of boulder yield can be obtained by using field test data to train model preliminarily and using the controlling variable method to analyze the sensitivity of model parameters. According to calculation and prediction of evaluation model,the best influence parameter values of boulder yield were obtained by increasing more levels of the input data to form the forecasted and optimized samples in the way of L16( 34) orthogonal array. Take Dongji Gold Mine as an example,the results show the best predictive value of boulder yield as 9. 98%. In addition,by detonating at hole bottom,the best influence parameter values were obtained as row spacing 1. 6 m,hole-bottom spacing 1. 8 m and specific charge 0. 350 kg/m~3.
In order to solve the problem of high blasting boulder yield in underground mine stope,the PSO-BP model was produced to optimize the influence parameters of boulder yield by using PSO algorithm in the BP neural network. The parameters,such as hole-bottom spacing,row spacing,specific charge and detonation position,were used as the input data and the boulder yield was set to be the output data in model building. The best PSO-BP evaluation model of boulder yield can be obtained by using field test data to train model preliminarily and using the controlling variable method to analyze the sensitivity of model parameters. According to calculation and prediction of evaluation model,the best influence parameter values of boulder yield were obtained by increasing more levels of the input data to form the forecasted and optimized samples in the way of L16( 34) orthogonal array. Take Dongji Gold Mine as an example,the results show the best predictive value of boulder yield as 9. 98%. In addition,by detonating at hole bottom,the best influence parameter values were obtained as row spacing 1. 6 m,hole-bottom spacing 1. 8 m and specific charge 0. 350 kg/m~3.
引文
[1]古德生.地下金属矿采矿科学技术的发展趋势[J].黄金,2004,25(1):18-22.[1]GU De-sheng.The development tendency of mining science and technology of underground metal mine[J].Gold,2004,25(1):18-22.(in Chinese)
[2]林大泽.降低地下矿深孔爆破大块率的技术措施[J].中国安全科学学报,2007,17(1):86-90.[2]LIN Da-ze.Technical measures on lowering boulder yield during deep-hole blast in underground mine[J].China Safty Science Journal,2007,17(1):86-90.(in Chinese)
[3]王平,李斌,许梦国.扇形孔爆破落矿大块产生机理及控制研究[J].爆破,2015,32(2):1-10.[3]WANG Ping,LI Bin,XU Meng-guo.Production mechanism and control measure of boulder by fan shaped round blasting[J].Blasting,2015,32(2):1-10.(in Chinese)
[4]钱锋.粒子群算法及其工业应用[M].北京:科学出版社,2013:29-31.
[5]赵国彦,戴兵,马驰,等.基于3DFR算法的爆破块度图像处理研究及其应用[J].中南大学学报:自然科学版,2013,44(5):2002-2007.[5]ZHAO Guo-yan,DAI Bing,MA Chi,et al.Research and application on blasting fragmentation image processing based on 3DFR algorithm[J].Journal of Central South University:Science and Technology,2013,44(5):2002-2007.(in Chinese)
[6]RATNAWEERA A,HALGAMUGE S K,WATSON H C.Self-organizing hierarchical particle swarm optimizer with time-varying acceleration coefficients[J].IEEE Transactions on Evolutionary Computation,2004,8(3):240-255.
[7]陈贵敏,贾建援,韩琪.粒子群优化算法的惯性权值递减策略研究[J].西安交通大学学报,2006,40(1):53-56.[7]CHEN Gui-min,JIA Jian-yuan,HAN Qi.Study on the strategy of decreasing inertia weight inparticle swarm optimization algorithm[J].Journal of Xi An Jiao Tong University,2006,40(1):53-56.(in Chinese)
[8]李松,刘力军,翟曼.改进粒子群算法优化BP神经网络的短时交通流预测[J].系统工程理论与实践,2012,32(9):2045-2049.[8]LI Song,LIU Li-jun,ZHAI Man.Prediction for short-term traffic flow based on modified PSO optimized BP neural network[J].System Engineering Theory and Practice,2012,32(9):2045-2049.(in Chinese)
[9]王建国,阳建宏,云海滨,等.改进粒子群优化神经网络及其在产品质量建模中的应用[J].北京科技大学学报,2008,30(10):1188-1193.[9]WANG Jian-guo,YANG Jian-hong,YUN Hai-bin,et al.Improved particle swarm optimized back propagation neural network and its application to production quality modeling[J].Journal of University of Science and Technology Beijing,2008,30(10):1188-1193.(in Chinese)
[10]王新民,赵彬,王贤来,等.基于BP神经网络的凿岩爆破参数优选[J].中南大学学报:自然科学版,2009,40(5):1411-1416.[10]WANG Xin-min,ZHAO Bin,WANG Xian-lai,et al.Optimization of drilling and blasting parameters based on back-propagation neural network[J].Journal of Central South University:Science and Technology,2009,40(5):1411-1416.(in Chinese)
[11]赵强,张建华,李星,等.降低中深孔爆破大块率的技术措施[J].爆破,2011,28(4):50-52.[11]ZHAO Qiang,ZHANG Jian-hua,LI Xing,et al.Measures to lower the large lump rate in medium-deep-hole blasting[J].Blasting,2011,28(4):50-52.(in Chinese)
[12]戴云波,张德明,高宇梁,等.基于BP网络的采场爆破钻孔参数优化[J].爆破,2014(3):57-62.[12]DAI Yun-bo,ZHANG De-ming,GAO Yu-liang,et al.Optimization of blasting parameters in underground stope based on BP network[J].Blasting,2011,28(4):50-52.(in Chinese)
[13]罗忆,卢文波,陈明,等.爆破振动安全判据研究综述[J].爆破,2010,27(1):15-22.[13]LUO Yi,LU Wen-bo,CHEN Ming,et al.View of research on safety criterion of blasting vibration[J].Blasting,2010,27(1):15-22.(in Chinese)
[14]王德永,袁艳斌,钱兆明,等.基于GA-BP神经网络矿岩爆破参数优选[J].爆破,2013,30(1):30-34.[14]WANG De-yong,YUAN Yan-bin,QIAN Zhao-ming,et al.Optimization of mine blast parameters based on GABP neural network[J].Blasting,2013,30(1):30-34.(in Chinese)
[15]陈俊桦,李新平,张家生.基于爆破损伤的岩台保护层开挖爆破参数研究[J].岩石力学与工程学报,2016,35(1):98-108.[15]CHEN Jun-hua,LI Xin-ping,ZHANG Jia-sheng.Study on blasting parameters of protective layer excavation of rock bench based on blasting-induced damage[J].Chinese Journal of Rock Mechanics and Engineering,2016,35(1):98-108.(in Chinese)
[16]周科平,翟建波.改进蚁群算法在地下矿山运输路径优化的应用[J].中南大学学报:自然科学版,2014,45(1):256-261.[16]ZHOU Keping,ZHAI Jianbo.Application of improved ant colony algorithm in route optimization of underground mine's transportation[J].Journal of Central South University:Science and Technology,2014,45(1):256-261.(in Chinese)
[2]林大泽.降低地下矿深孔爆破大块率的技术措施[J].中国安全科学学报,2007,17(1):86-90.[2]LIN Da-ze.Technical measures on lowering boulder yield during deep-hole blast in underground mine[J].China Safty Science Journal,2007,17(1):86-90.(in Chinese)
[3]王平,李斌,许梦国.扇形孔爆破落矿大块产生机理及控制研究[J].爆破,2015,32(2):1-10.[3]WANG Ping,LI Bin,XU Meng-guo.Production mechanism and control measure of boulder by fan shaped round blasting[J].Blasting,2015,32(2):1-10.(in Chinese)
[4]钱锋.粒子群算法及其工业应用[M].北京:科学出版社,2013:29-31.
[5]赵国彦,戴兵,马驰,等.基于3DFR算法的爆破块度图像处理研究及其应用[J].中南大学学报:自然科学版,2013,44(5):2002-2007.[5]ZHAO Guo-yan,DAI Bing,MA Chi,et al.Research and application on blasting fragmentation image processing based on 3DFR algorithm[J].Journal of Central South University:Science and Technology,2013,44(5):2002-2007.(in Chinese)
[6]RATNAWEERA A,HALGAMUGE S K,WATSON H C.Self-organizing hierarchical particle swarm optimizer with time-varying acceleration coefficients[J].IEEE Transactions on Evolutionary Computation,2004,8(3):240-255.
[7]陈贵敏,贾建援,韩琪.粒子群优化算法的惯性权值递减策略研究[J].西安交通大学学报,2006,40(1):53-56.[7]CHEN Gui-min,JIA Jian-yuan,HAN Qi.Study on the strategy of decreasing inertia weight inparticle swarm optimization algorithm[J].Journal of Xi An Jiao Tong University,2006,40(1):53-56.(in Chinese)
[8]李松,刘力军,翟曼.改进粒子群算法优化BP神经网络的短时交通流预测[J].系统工程理论与实践,2012,32(9):2045-2049.[8]LI Song,LIU Li-jun,ZHAI Man.Prediction for short-term traffic flow based on modified PSO optimized BP neural network[J].System Engineering Theory and Practice,2012,32(9):2045-2049.(in Chinese)
[9]王建国,阳建宏,云海滨,等.改进粒子群优化神经网络及其在产品质量建模中的应用[J].北京科技大学学报,2008,30(10):1188-1193.[9]WANG Jian-guo,YANG Jian-hong,YUN Hai-bin,et al.Improved particle swarm optimized back propagation neural network and its application to production quality modeling[J].Journal of University of Science and Technology Beijing,2008,30(10):1188-1193.(in Chinese)
[10]王新民,赵彬,王贤来,等.基于BP神经网络的凿岩爆破参数优选[J].中南大学学报:自然科学版,2009,40(5):1411-1416.[10]WANG Xin-min,ZHAO Bin,WANG Xian-lai,et al.Optimization of drilling and blasting parameters based on back-propagation neural network[J].Journal of Central South University:Science and Technology,2009,40(5):1411-1416.(in Chinese)
[11]赵强,张建华,李星,等.降低中深孔爆破大块率的技术措施[J].爆破,2011,28(4):50-52.[11]ZHAO Qiang,ZHANG Jian-hua,LI Xing,et al.Measures to lower the large lump rate in medium-deep-hole blasting[J].Blasting,2011,28(4):50-52.(in Chinese)
[12]戴云波,张德明,高宇梁,等.基于BP网络的采场爆破钻孔参数优化[J].爆破,2014(3):57-62.[12]DAI Yun-bo,ZHANG De-ming,GAO Yu-liang,et al.Optimization of blasting parameters in underground stope based on BP network[J].Blasting,2011,28(4):50-52.(in Chinese)
[13]罗忆,卢文波,陈明,等.爆破振动安全判据研究综述[J].爆破,2010,27(1):15-22.[13]LUO Yi,LU Wen-bo,CHEN Ming,et al.View of research on safety criterion of blasting vibration[J].Blasting,2010,27(1):15-22.(in Chinese)
[14]王德永,袁艳斌,钱兆明,等.基于GA-BP神经网络矿岩爆破参数优选[J].爆破,2013,30(1):30-34.[14]WANG De-yong,YUAN Yan-bin,QIAN Zhao-ming,et al.Optimization of mine blast parameters based on GABP neural network[J].Blasting,2013,30(1):30-34.(in Chinese)
[15]陈俊桦,李新平,张家生.基于爆破损伤的岩台保护层开挖爆破参数研究[J].岩石力学与工程学报,2016,35(1):98-108.[15]CHEN Jun-hua,LI Xin-ping,ZHANG Jia-sheng.Study on blasting parameters of protective layer excavation of rock bench based on blasting-induced damage[J].Chinese Journal of Rock Mechanics and Engineering,2016,35(1):98-108.(in Chinese)
[16]周科平,翟建波.改进蚁群算法在地下矿山运输路径优化的应用[J].中南大学学报:自然科学版,2014,45(1):256-261.[16]ZHOU Keping,ZHAI Jianbo.Application of improved ant colony algorithm in route optimization of underground mine's transportation[J].Journal of Central South University:Science and Technology,2014,45(1):256-261.(in Chinese)