60
thickness proportions, etc (Audet and Ross, 2007; Wawruch and Betzhold, 2005 and
Dohm, 2005).
This paper presents a combination of drilling pattern simulation techniques associated with
the mineral resources classification methodology proposed by Verly at al. (2014), which uses the
geostatistical conditional simulations and the concept of confidence intervals (CI) and production increments for risk assessment. This combination of methodologies has been successfully
applied in different operating mines and commodities such as, gold, bauxite, iron and niobium
and it assisted mining companies in optimizing their drilling program, based on the risk evaluation and mineral resources classification, even during the early stages of mineral exploration.
Its main applicability is to assess the risk associated to a range of drilling patterns/densities to support decision-making for additional drilling campaigns to convert the resource into
measured and indicated.
2 METHODOLOGY
2.1 General workflow
The workflow of this study is fully performed with Isatis
® (Bleines 2012) and consists of a
three-step process:
1. Conditional simulations are performed on the available drillhole dataset to create different
plausible scenarios;
2. A representative subset of scenarios is then identified, and various sampling patterns
defined for each scenario allowing to generate virtual drillholes;
3. Conditional simulations are performed on each virtual drilling mesh in order to compute
the dispersion of different simulated attributes within a confidence interval of 90%, considering monthly/quarterly and annual production increments, according to the classification methodology proposed by Verly at al. (2014).
The resource classification mentioned above uses the distribution of the set of simulations to assess the risk associated to the deposit of interest and then the following rules are
applied:
– Measured: ± 15% with 90% Confidence Interval (CI) on a quarterly or monthly production increment;
– Inferred: ± 15% with 90% CI on an annual production increment;
This methodology indicates the risk level associated to each attribute and it is possible
to assess, locally and globally, the uncertainty (risk) related to each drillhole spacing. As a
result, the company may define the optimum drilling pattern considering mineral resources
classification and the risk that the company is willing to take.
2.2 Case study
The studied deposit consists in a bauxite deposit from one of the major bauxite regions in
the world. The sediments beneath the bauxitic formation are composed of alternating c1ayey,
silty and sandy layers of a weathered c1astic sediment consisting of kaolinite, quartz and a
small quantity of iron and titanium oxides. The terrain forms gently undulating plateaus.
The raw data consist of 388 vertical drillholes and 5801 samples (see Figure 1). The study
was carried out on a fictitious dataset of a domain ‘D’ and the variables of interest are the
grade ‘v1’ and the thickness ‘v2’. The study was set on a 2D environment, but it could be
done on 3D environment as well.
2.3 Perform conditional simulations on the available dataset
As the study was implemented on a 2D environment, the first step is to calculate the accumulation of the grade variable v1 (acc_v1) and to define the thickness v2 considering only
thickness proportions, etc (Audet and Ross, 2007; Wawruch and Betzhold, 2005 and
Dohm, 2005).
This paper presents a combination of drilling pattern simulation techniques associated with
the mineral resources classification methodology proposed by Verly at al. (2014), which uses the
geostatistical conditional simulations and the concept of confidence intervals (CI) and production increments for risk assessment. This combination of methodologies has been successfully
applied in different operating mines and commodities such as, gold, bauxite, iron and niobium
and it assisted mining companies in optimizing their drilling program, based on the risk evaluation and mineral resources classification, even during the early stages of mineral exploration.
Its main applicability is to assess the risk associated to a range of drilling patterns/densities to support decision-making for additional drilling campaigns to convert the resource into
measured and indicated.
2 METHODOLOGY
2.1 General workflow
The workflow of this study is fully performed with Isatis
® (Bleines 2012) and consists of a
three-step process:
1. Conditional simulations are performed on the available drillhole dataset to create different
plausible scenarios;
2. A representative subset of scenarios is then identified, and various sampling patterns
defined for each scenario allowing to generate virtual drillholes;
3. Conditional simulations are performed on each virtual drilling mesh in order to compute
the dispersion of different simulated attributes within a confidence interval of 90%, considering monthly/quarterly and annual production increments, according to the classification methodology proposed by Verly at al. (2014).
The resource classification mentioned above uses the distribution of the set of simulations to assess the risk associated to the deposit of interest and then the following rules are
applied:
– Measured: ± 15% with 90% Confidence Interval (CI) on a quarterly or monthly production increment;
– Inferred: ± 15% with 90% CI on an annual production increment;
This methodology indicates the risk level associated to each attribute and it is possible
to assess, locally and globally, the uncertainty (risk) related to each drillhole spacing. As a
result, the company may define the optimum drilling pattern considering mineral resources
classification and the risk that the company is willing to take.
2.2 Case study
The studied deposit consists in a bauxite deposit from one of the major bauxite regions in
the world. The sediments beneath the bauxitic formation are composed of alternating c1ayey,
silty and sandy layers of a weathered c1astic sediment consisting of kaolinite, quartz and a
small quantity of iron and titanium oxides. The terrain forms gently undulating plateaus.
The raw data consist of 388 vertical drillholes and 5801 samples (see Figure 1). The study
was carried out on a fictitious dataset of a domain ‘D’ and the variables of interest are the
grade ‘v1’ and the thickness ‘v2’. The study was set on a 2D environment, but it could be
done on 3D environment as well.
2.3 Perform conditional simulations on the available dataset
As the study was implemented on a 2D environment, the first step is to calculate the accumulation of the grade variable v1 (acc_v1) and to define the thickness v2 considering only