| Reserve estimation is a complex dynamic process that involves the entire life cycle of mine planning,development,and closure.It has important significance to mining engineers for resource assessment,mine design,and planning.Previous geological data used in this study are used to establish a three-dimensional geological model of the mine and realize visualization.The inverse distance weighted?IDW?method,Kriging method,and Multi-point Geostatistics?MPS?method is used to estimate the ore grade of iron ore,serpentine and porcelain clay based on the three-dimensional modeling.The Minkowski distance and its special form are used to improve the IDW.A new method to enhance the effect of the IDW is proposed.The MPS is introduced to the field of resource reserve estimation.The problems in estimation have been improved.A better estimation effect is got by the improved method.Therefore,the research contributes to estimation theory.Multiple ore body types are invoked as research objects.The same method is employed to the estimation of different ore body types.The effect of different methods on ore body estimation is examined.The results of the study can serve as a reference for similar estimation.Therefore,this study has a few application value.The study focuses on iron ore,serpentine ore and china clay.Three-dimensional ore body model and the block model of the three mines are established.The effect of the IDW,the Kriging method,and MPS are analyzed.Instead of the Euclidean distance,other special forms of Minkowski distance are used as the distance weights calculation method in IDW.The influence of Minkowski distance on the estimation grade in distance weights calculation is studied.Fractal variograms are used for experimental variogram fitting.Estimation characteristics of the fractal Kriging method and the ordinary Kriging?OK?method are studied.The multi-point geostatistics is introduced into the field of resource reserves estimation,and improved by using training images instead of variograms.A hybrid method based on MPS and IDW?M-I?is proposed.The influence of training images and the number of discrete interval intervals on the value of M-I method is analyzed.Finally,optimal estimation results of different estimation methods are compared laterally,and the estimation characteristics of different methods are analyzed.The research results have been obtained by studying the estimation of iron ore,serpentinite and porcelain ore body grade:?1?Geological databases and ore body models of iron,serpentine and porcelain clay are constructed.The principle of conversion of a two-dimensional orebody profile into a three-dimensional profile is studied.The conversion formula between two-dimensional and three-dimensional profiles is given.The lengths and grade of the original samples are counted for iron ore,serpentinite,and porcelain clay,respectively.?2?The use of other special forms of Minkowski distance to replace the Euclidean distance is feasible for distance weights calculation in IDW.The result of the calculation is stable.The study expands the types of distance weights in IDW.The estimated change is given in grade of Fe,Ni,MgO,Li2O and Ta2O5 based on Minkowski distance weights.The influence of sample quantity and grade distribution on the estimation is analyzed.The study gives the optimal estimation results.?3?The fractal variogram is used to improve the fitting effect,make the estimation process more convenient and the estimation result more accurately compared with the spherical variogram.The fractal Kriging method's estimation result is biased estimation.It has more average deviation than the ordinary Kriging method.Therefore,there are optimal grade estimation results with fractal variograms and OK method.?4?The multi-point geostatistics?MPS?are introduced to the field of resource reserve estimation.The influence of training images and the number of discrete intervals on the estimation is investigated.It is noted that improving the resolution of training images can improve the estimation effect of multi-point geostatistics to some extent.However,a huge training image seriously affects the computational efficiency.It does not completely address the issue of matching between data events and data patterns.In theory,increasing in the number of discrete intervals can improve the accuracy of estimation.In fact,it will reduce the matching rate of data events and data patterns,and decrease the accuracy of estimation.For this reason,a hybrid method based on MPS and IDW is proposed.This method can overcome the inaccuracy of low-frequency grade information in simple MPS.Optimal estimation results of TFe,Ni,MgO,Li2O,and Ta2O5 grade using M-I are given.?5?The results of the grade estimation of IDW and OK are compared.The characteristics of estimation deviations,grade trends,and variograms are analyzed.The following understanding is obtained:The average values of the TFe,Ni,MgO,Li2O,and Ta2O5 grade estimated by IDW,OK,and M-I are close to each other.Ni and MgO grade has larger minimum deviations,which are mainly affected by the sampling method and grade distribution.Reducing the number of samples participating in the evaluation can reduce the maximum and minimum deviations within a certain range.The estimation grade and sample grade are in the same trend.The Ni and MgO grade estimated by OK has greater smoothness.The smoothness of the estimation grade is also influenced by the sampling method.The number of sample points participating has a great influence on the estimation results.If there are greater the number of samples participate in the estimation,there is more difference between the sample grade variability function and the value grade variability function.The distance weights of IDW has no obvious effect on the estimation grade variogram.Exploration engineering and sample distribution also has an impact on the variability function of estimation grade.The estimation results of M-I are similar to IDW and OK,which shows that the M-I method has high stability and accuracy.The M-I is a more advanced estimation method. |
PDF Full Text Request