‘Heterogeneity’ as a measure of localized variability of economic drivers such as grade at mining scale. It is a term strongly promoted by CRC ORE as the important influence on Grade Engineering opportunity.
Variability involves statistical measures of population distributions, whereas heterogeneity involves statistical measures of spatially aggregated population neighbourhoods. For mining application, population neighbourhoods are typically represented by volumes of modelled blocks at various scales of resource definition; or imposed mining volumes such as blast blocks or benches.
While the concept of ‘heterogeneity’ is widely appreciated within the CRC ORE consortium, specific definitions and methodologies suitable for routine industry application and commercial delivery are sparse. This is further complicated by ongoing uncertainty and debate around how potential measures of heterogeneity are best derived; and how these can be related to value drivers around coarse separation levers. This is a function of several factors:
Heterogeneity is scalar and varies dramatically as a function of neighbourhood volume – this ranges from inter-block attributes in long term resource models; to intra-bench attributes at mining scale; down to <50mm inter-particle scale attributes for ore sorting applications.
Measures of heterogeneity are derived from analysis of existing drilling and assay data rather than specific laboratory testing. This data ranges from more widely spaced (~25-50m) resource definition related to long term planning; to closer spaced (5-10m) ore control drilling related to short term planning.
Changes in data support directly affects resolution of heterogeneity.
Heterogeneity is typically smoothed out in current resource definition practice and is regarded as a problem for geostatistical interpolation rather than an opportunity for exploitation.
There is often significant confusion between uncertainty and heterogeneity particularly involving geostatistical interpolation of more widely spaced resource drilling data.
Heterogeneity in the context of Grade Engineering drives opportunity related to differential blasting and bulk sensor-based separation levers. While these levers share the same intrinsic drivers, exploitation involves very different approaches related to screening for differential blasting and diversion for sensor-based applications. To an extent these applications are mutually competitive and require different heterogeneity modelling and ranking methodologies. Outcomes of bulk separation opportunity modelling are embedded into robust algorithms and software routines as add-ins to available commercial platforms.
Example of Uncovering Heterogeneity:
Heterogeneity characteristics must be based on intrinsic properties of material being evaluated. By ensuring this is accurately captured any future modelling and simulation work can be confidently undertaken to determine true amenable material and eventually, in an economic evaluation, exploitable material.