Risk assessment with geostatistical estimation in mine planning

This thesis aims to assess the risk associated with the different techniques/methods used for geostatistical estimation and modelling of the mineral ore bodies using 2D Lerchs Grossman Algorithm (LG). 2D LG algorithm was used to find out the ultimate pit limit (UPL) and optimum sequence of mining for different estimation techniques and estimation and economical parameters. The research was performed using Microsoft excel Visual Basic Application (VBA) macros interface for approximately all procedures. Drill-hole data of a regularly drilled gold-silver deposit have been used to illustrate the methods and concepts for this research work. The spatial characteristics of the data were analysed using descriptive statistics and variogram analysis. The data then fitted to a model variogram to get a mathematical function and different variogram parameters i.e. nugget, sill and range which were later on used for kriging estimation. Then we estimated the block grades using Inverse Distance Weighting (IOW), Ordinary kriging (OK), simple kriging (SK) and indicator kriging (IK). These estimated grades and other economic parameters (metal price, recovery, mining cost and processing costs) are used to assign the Economic values to the blocks. Finally the 2D LG algorithm was used to find out the optimized UPL with the maximum Net Present Value (NPV) for each estimation technique i.e. IOW and different Kriging types. I came up with the result that with decrease in number of samples taken to compute variogram, the nugget decreases and sill increases but the range remains the constant The NPV for searching radius of 20 meter was higher than searching radius of 30 meters. Similarly the NPV for less number of samples was higher than more number of samples. The NPV for 5x5 blocks is approximately double as compared to 20x20 m (Exactly 2.075 times). From the comparison of the UPL for IOW and OK it was found that the UPL is approximately the same for both techniques but due to minute difference between the estimated block values there is a difference between the total NVP of the pit. From nested pit technique it was found out that the most valuable part of pit was same for all of the estimation method