Characterizing the intrinsic and the one-dimensional heterogeneities of a niobium ore based on Pierre Gy’s Theory of Sampling

Abstract

Abstract The Fundamental Sampling Error (FSE) is the main error defined by Pierre Gy’s Theory of Sampling (Gy, 1967; 1979; 1992) and is related to the constitution or intrinsic heterogeneity (IH) of the ore. Even if a sampling procedure is considered ideal, this error can never be eliminated. To calculate the FSE for a certain sample taken from a certain fragmented lot, crushed to a certain size, the intrinsic heterogeneity of the lot (IHL) must be estimated, which can be done theoretically applying the Gy’s factors, or experimentally performing heterogeneity tests. FSE calculation allows the optimization of sampling protocols, the calculation of minimum sample masses, as well as the estimation of the precision of a sampling procedure or equipment. FSE represents the zero-dimensional heterogeneity of a lot and it is of upmost importance to calculate it. However, there is another type of heterogeneity related to one-dimensional lots, i.e., the material flow on conveyor belts or in pipelines. This one-dimensional heterogeneity can be characterized with variography, by estimating the Heterogeneity Fluctuation Error (HFE). Obtaining reliable information on ore grades at the plant feed is a great challenge for mining operations. When the precision of the plant feed grade is low, incorrect decisions can be made and may decrease the process yield. In order to estimate both FSE and HFE for a Brazilian niobium ore, a sampling campaign was carried out at the plant feed. Results indicated that the 5-minute sampling interval was appropriate, resulting in a low relative standard deviation of HFE, i.e., 2.26% for Nb2O5, considering a 95% confidence interval. This article shows how to estimate the zeroand one-dimensional heterogeneities of ores and how important it is to define the precision associated with the grade estimates for process control, metallurgical accounting and reconciliation purposes.

Keywords

• pyrochlore
• niobium
• theory of sampling
• fundamental sampling error
• heterogeneity fluctuation error
• sensitivity analysis.