Reserve estimation using aeromagnetic survey and cluster analysis based on multiple-point geostatistical ore modeling at the iron deposit, Korea

A methodology using geostatistical approach for the reserve estimation is proposed. It uses multiple-point geostatistics (MPS) to perform the ore modeling and uses geophysical data and cluster analysis to select the realistic and optimal model. MPS can derive simulation results that reflect complex geological features using pattern model called a training image (TI). The TI of this study represents ore deposit, which is a complicated pattern due to the intrusion of later igneous rocks and metamorphism, but it appears as a lens-shaped. The TI was composed of using three dimensional geological modeling based on geology map, geology section, and borehole data. MPS was applied to the single normal equation simulation (SNESIM) algorithm using TI and generated several simulation results. Cluster analysis was applied in probabilistic models to find ore model that reflect the ore deposits of the study area. This method used distance-based multidimensional scaling and the radial basis function kernel. For each model cluster classification, various algorithms such as k-means and k-medoids are applied to select the optimal model. Finally, forward modeling is used to compare clustering simulation models with aeromagnetic data. To this end, the clustered facies model is converted to a physical model through inverse transform sampling of the magnetic susceptibility of the rock physical properties test. It is possible to select the most suitable model by analyzing the magnetic anomaly of the aeromagnetic survey and forward model. Reserve estimation was carried out by applying traditional two-point geostatistical techniques to the grade data of the borehole. AS a results, the ore modeling using MPS represents the shape of a realistic ore body, and reasonable result could be derived from integrated interpretation of the geophysical data. Also, this methodology could be applied to various ore deposits for mine evaluation.