| OCR Text |
Show results of the tests, resulting in an improved assessment of optimal mining method and ore treatmGnt process. Mineralogy: the experimental study of the ore sampling parameters often give additional insights into metal/mineral grain sizes, and the mineralogical nature of ore grades. Better detection of systematic biases at the assay laboratory. Improved grade control through more reliable data, with measurable benefits in terms of produced grade and accuracy of the selection, and in some instances, improved cutoff grade policy (detection of warning signs when ore delineation at current cutoff grade is illusory). At the metallurgical plant level, control of the sampling procedures and equipment, leading to improved metallurgical balances and reduced mine-mill variance problems. Improved commercial ore sampling, improved ability to resolve commercial disputes. PRINCIPLES OF APPLICATION The numerical part of Sampling Theory is a powerful tool the objective of which is to quantify and predict the reliability of samples to be taken from a lot of broken ore, or determine the minimum sample mass and best preparation protocol permitting a desired pre-set precision. Similar sampling operations occur when taking the primary sample from a heap of broken ore as well as when taking a subsample at each stage of the subsequent sample preparation procedure (the current subsample becomes the lot to be sampled in the next stage). Provided all possible sources of biases have been eliminated and the lot has been perfectly homogenized (or the samples taken in a way that is not affected by any segregation problems, which would largely decrease the overall reliability), each one of these individual cascading sampling operations still generates a sampling error. This error simply reflects the trivial fact that the grade of the sample in general cannot exactly be the grade of the lot. The sampling error at each stage is characterized by a measure of how much on average a sample grade could depart from the true grade of its lot. The measure used here is the sampling relative variance , i.e. the relative variance of the sample values we would obtain if we had the possibility of taking a large number of successive samples from the same lot. Although in practice we actually take only one sample at each stage, Sampling Theory establishes predictive formulas for these relative single-stage variances. The relative variances incurred at each stage, from primary sample to aliquot selection are cumulative, i.e. the overall precision of the grade value obtained in the form of an assay at the end of the process is characterized by the sum of the individual single-stage variances. These variance intimately depend on the mineralization, and on the sizes of the lot. the sample and the fragments constituting the lot. These formulas must be tailored to the type of mineralization and deposit of interest by experimentally adjusting the parameters controlling the rate of change of the sampling variance with changes in sample mass and in the fragment size at which the sampled material has been crushed. In addition, and of equal importance, Sampling Theory also addresses the problems of bias elimination, especially when automatic sampling equipment is concerned, and offers several methods for decreasing the effects of segregation phenomena, which have been seen to increase the sample variances by one or more orders of magnitude . Source: http://pws.prserv.net/Goodsampling.com/sampling .html |
