【摘要】 为了避免露天金属矿爆破后导致爆堆边缘矿石品位贫化损失,需要根据最低品位阈值重新计算矿岩边界,而影响矿岩边界发生改变因素众多,需要确定主要影响因素。因此,利用爆堆爆破前地形方向和爆堆地质数据,构建灰色关联-广义回归神经网络模型(GRA-GRNN)分析爆堆矿岩边界变化主要影响因素。首先对爆堆钻孔品位数据使用析取克里金法进行空间插值,并根据矿山工艺最低品位阈值提取爆破前的矿岩边界;其次,将爆破前后的数字DEM模型进行求差,求得爆破后的爆堆数字DEM模型,并构建爆破前后爆堆数字DEM模型空间分布椭圆,从而确定爆堆爆破后的水平形变方向;对影响爆堆爆破后形变的可能因素进行提取,并应用GRA-GRNN模型选取主要影响因素及对其强度进行分析,并将其结果与BP神经网络模型预测结果进行了对比。从实验结果可知,影响爆堆爆破后形变强度排前三的因素为:爆破前地形方向、爆孔排距和列距,强度分别为0.880、0.760和0.755,预测结果优于BP模型。更多还原

【Abstract】 In order to avoid the ore grade dilution loss caused by blasting in open-pit metal mine,it is necessary to recalculate the ore-rock boundary according to the lowest grade threshold.There are many factors affecting the change of ore-rock boundaries,and the main factors need to be determined.Therefore,the grey relationalgeneralized regression neural network model(GRA-GRNN)was constructed to analyze the main influencing factors of the boundary changes of explosive pile by using the topographic direction before the explosion and the geological data of explosive pile.Firstly,the extraction Kriging method is used to spatially interpolate the data of blasting pile borehole grade,and the ore-rock boundary before blasting is extracted according to the lowest grade threshold of mining technology.Secondly,the digital DEM model before and after blasting is used to calculate the difference to obtain the blasting model,and the spatial distribution ellipse of the DEM model of the pre-explosion reactor before and after blasting is constructed to determine the horizontal deformation direction after the blasting.The possible factors affecting the deformation after blasting are extracted,and the main factors are selected by GRA-GRNN model and their strengths are analyzed.The results are compared with the prediction results of the BP neural network model.The experimental results show that the topographic direction before the explosion,the row spacing of the explosion holes,and the column spacing are the top three factors affecting the deformation strength after explosion,and the strength is 0.880,0.760,and 0.755,respectively.The prediction results are better than BP model.更多还原