The development of a robust belt sampling strategy and protocol in
underground gold mines is a fundamental requirement for effective grade quality control, ore accounting and mine-mill reconciliations. As part of the ore management system, a belt sampling system has been developed by the author, which is used to predict monthly gold grade and its error for underground/belt/mill reconciliation and grade control.
This paper presents the step-by-step development of the belt sampling
strategy and sample preparation protocol at Target, an underground gold mine in South Africa. The belt sampling system at Target consists of an underground belt transporting run-of-mine material into an underground transfer point, a weightometer, a two stage crushing system, a tube splitter and a catchment device. The design allows run-of-mine ore on the belt to be sampled automatically by a go-belt sampler.
Gy's sampling theory of particulate materials forms the basic
framework of the belt sampling strategy. The theory is used to quantify the optimum sample mass of broken material to be taken from the belt and to predict its reliability. It is also employed to determine the Fundamental Sampling Error (FSE), time interval between samples and particle size requirements of the final underground belt sample. Further, Gy's sampling theory provides the means of analysing the crushing processes, splitting, segregation error, analytical variance, and the best sample preparation protocol.
Although the belt sampling strategy developed in this study was for
Witwaterstrand gold ore, the process is entirely general and may be very successfully used in any underground or surface mine.