铜矿峪矿床5号矿体建模估值参数探讨
|
摘要
利用三维矿业软件矿体建模时估值参数的确定是一项复杂且重要的工作,其不仅包含对地质统计学的理解还包含有日常工作的经验。在铜矿峪矿床5号矿体建模过程,多次变化估值的参数,应用交叉验证的统计结果(平均值误差、绝对误差、克里金方差、相关系数和线性回归系数)观察估值参数的恰当与否。在1~6次估值采用变异函数模型的产状和变程,主要变化样品数,交叉验证统计结果显示第3次最优,也就是最少样品数是10个最大样品数是20个可能是最好的样品数。7~15次是采用最少样品数是10个最大样品数是20个,主要变化单工程最大样品数,结果显示单工程最大样品数为6时估值参数最优。第16~23次是变化搜索椭球体半径,结果显示椭球体半径在112m、168m和28m以后参与估值的样品数和空间位置基本没有变化。铜矿峪估值参数中搜索椭球体的产状可以利用变异函数模型参数为最佳,单工程最大样品数为6个,最少样品数为10个,最大样品数为20个可能对于估值更为有利,搜索托球体半径推荐为112m、168m和28m。
It is a complicated and important work to determine the parameter of estimation in the modeling ore body using the 3D mining software.It includes not only the understanding of the geological statistics but also the experience of the daily work.In the No.5 ore body modeling process,applying the statistical results of cross validation(average error,absolute error,the Kriging variance,correlation coefficient and linear regression coefficients),We change the parameters of estimation,and observe changes the results of estimation.In 1~6 estimations,making use of variogram model and variable range,we changes in the number of samples.The cross validation statistics results show optimal results is the third time,in other words,the least number of sample is 10 and the largest number of samples is 20 may be the best number of samples.In 7 ~ 15 estimations,the minimum number of samples is the 10,and the largest number of samples is 20,and we change the largest number of samples in the single drill hole.The the results show that the results is optimal when the maximum number of samples is 6 in the single drill hole.In16~23 estimations,we change search ellipsoid radius,and the results show that the results is optimal when the radius of the ellipsoid is the 112 m,168m and 28 m.In the parameter of estimation in Tongkuangyu,the shape of search ellipsoid can be used the Kriging variance model parameters,and the largest number of samples in the single drill hole is 6,and the minimum number of samples is 10,the maximum number of samples is 20,he search radius of the ball body is recommended for 112 m,168 m and 28 m.
It is a complicated and important work to determine the parameter of estimation in the modeling ore body using the 3D mining software.It includes not only the understanding of the geological statistics but also the experience of the daily work.In the No.5 ore body modeling process,applying the statistical results of cross validation(average error,absolute error,the Kriging variance,correlation coefficient and linear regression coefficients),We change the parameters of estimation,and observe changes the results of estimation.In 1~6 estimations,making use of variogram model and variable range,we changes in the number of samples.The cross validation statistics results show optimal results is the third time,in other words,the least number of sample is 10 and the largest number of samples is 20 may be the best number of samples.In 7 ~ 15 estimations,the minimum number of samples is the 10,and the largest number of samples is 20,and we change the largest number of samples in the single drill hole.The the results show that the results is optimal when the maximum number of samples is 6 in the single drill hole.In16~23 estimations,we change search ellipsoid radius,and the results show that the results is optimal when the radius of the ellipsoid is the 112 m,168m and 28 m.In the parameter of estimation in Tongkuangyu,the shape of search ellipsoid can be used the Kriging variance model parameters,and the largest number of samples in the single drill hole is 6,and the minimum number of samples is 10,the maximum number of samples is 20,he search radius of the ball body is recommended for 112 m,168 m and 28 m.
引文
[1]郭怀成,周丰,刀谮.地统计方法学研究进展[J].地理研究,2008,27(5):1191—1202.
[2]孙玉建.以地质统计学为基础的矿业软件在中国的历史和现状[J].中国矿业,2007,16(11):80—82.
[3]倪平泽,刘修国,李超岭.3D矿床建模技术在数字矿产勘查中的应用[J].地球科学:中国地质大学学报,2010,35(3):445—458.
[4]孙玉建,孟伟,万会.矿产资源储量估算中工程控制程度划分的探索[J].地质与勘探,2006,42(6):81—84.
[5]Jimenez-espinosa R,Chica-olmo M.Application of geostatistics to identify gold-rich areas in the FinisterreFervenza region,NW Spain[J].Applied Geochemistry,1999(14):133-145.
[2]孙玉建.以地质统计学为基础的矿业软件在中国的历史和现状[J].中国矿业,2007,16(11):80—82.
[3]倪平泽,刘修国,李超岭.3D矿床建模技术在数字矿产勘查中的应用[J].地球科学:中国地质大学学报,2010,35(3):445—458.
[4]孙玉建,孟伟,万会.矿产资源储量估算中工程控制程度划分的探索[J].地质与勘探,2006,42(6):81—84.
[5]Jimenez-espinosa R,Chica-olmo M.Application of geostatistics to identify gold-rich areas in the FinisterreFervenza region,NW Spain[J].Applied Geochemistry,1999(14):133-145.