Date of Award


Degree Type


Degree Name

Master of Science


School of Engineering


Faculty of Computing, Health and Science

First Advisor

Associate Professor Ute Mueller

Second Advisor

Dr Steven Richardson


Geostatistical methods are currently used by mining companies to determine a resource model of the tonnage and head grade that may be obtained from a potential orebody, making it one of the first and most vital operational stages in any mining project. Currently long term mine planning is based on the estimated head grades model, which provides vital information on the quality of the ore. The risks associated with mining a particular ore may be reduced if geometallurgical information, such as material types, is incorporated into the operational flow model. Material type proportions are obtained from evaluated reverse circulation (RC) drillholes which are estimated directly into a long term geological model. However this causes smoothing of the estimates material types, unlikely combinations of material types within the blocks and large differences between the theoretical head grades and estimated head grades (OK HG). The aim of this study was to determine the best way to model the six grouped MTPs and reconcile the estimated proportions per block with the estimated head grades from the resource model using the direct block simulation (DBS) algorithm and the LSSOL optimisation program. One of the main decisions was to determine the best way to model and simulate the MTPs. Three different simulation options, all using DBS, were implemented. The first option modelled and simulated the MTP variables independently and the second option modelled and simulated the MTP variables jointly. As the spatial structure of the HGH attribute was quite different to those of the remaining five variables, the final option was to jointly simulate the five MTPs whose sample MTPs have similar spatial structures with the sixth block MTP making up the sum difference to one. A variety of different baseline methods, which comprise computation of MTPs from the simulation only and MTPs obtained from the optimisation alone, clearly demonstrates the need for a method that incorporates both the optimisation program and DBS to calculate reasonable MTPs. Seven methods which combined both the DBS and optimisation program were examined and compared, in the hope to obtain a method which calculated optimal MTPs that captures the sample MTPs and OK HGs. The optimisation program ensured that the optimal proportion of each material type within each block was calculated by minimising the difference between the head grades which have been estimated using ordinary kriging (OK HG) and HGs calculated using from the MTPs. Different bounds were applied to the maximum and minimum MTPs, obtained from the DBS, to determine a suitable method to obtain constraints which ensured that the optimal MTPs reflected the sample MTPs. For the given data set, the quadratic program which used the joint DBS simulation resulted in the most suitable representation of MTPs which was consistent with the OK HGs.