Investigating the Sensitivity of Different Estimation Techniques over Block Dimensions

It cannot be said that accurate and reliable ore reserve estimation is the fundamental of any mining operation success. Estimation techniques which are used to estimate the grade distribution of an orebody together with the assessment of uncertainty and risk have a vital influence on the success of any optimisation algorithm. Traditional estimation techniques such as kriging or inverse distance weighting are frequently used for assigning grade values to blocks. Various attempts have been made in order to compare geostatistical estimated methods for ore reserve estimation. This paper aims to evaluate the sensitivity of different estimation methods over changing the block dimension. For this reason, several block models with different dimensions were derived in Chador Malu iron ore mine which is located in 180 km northeast of Yazd in the centre of Iran. Each block model was estimated by different estimation methods such as nearest neighbours (NN), inverse distance (ID) and ordinary kriging (OK). The results of each estimation methods were compared against a control block file created by using the assay results of all blastholes drilled in the ore over four benches of the pit. The results signify that by estimating the large block size, no significant improvement occurred due to the employment of different estimation techniques. Furthermore, the grades estimated based on ID method is more reliable, but OK methods estimated more accurate tonnage.