3
CASE STUDY: OPEN PIT GOLD MINE
A case study was developed to demonstrate the use of sto-
chastic evaluation to compare the value and risk charac-
teristics of different strategic mine plan alternatives. Mine
plans were developed with stochastic and deterministic
optimization methods with various cut-off grade and stock-
piling strategies. For perspective, orebody uncertainty risks
are compared to the plan’s mining rate and vertical advance
rate sensitivity.
McLaughlin Mine Overview
From 1985 to 2002, the McLaughlin operation produced
about 3.4 million ounces of gold from a hot-springs type
epithermal deposit. The mine’s drillhole database serves as
the input for creating conditional simulations for the case
study. Ultimate pit and initial phase designs inside the ulti-
mate pit were already provided as inputs [Aras et al, 2019].
The initial five phase designs were split to create nine
phases—four south, four north and a final phase beneath
the northern and southern phases. Plans were created with
and without stockpiling for mill-only and mill+leach pro-
cessing scenarios.
Key constraints for the case study were defined as 10Mt/
yr max mining, 8 20' benches/yr max vertical advance for
each layback, 2.25Mt/yr max milling and 2.25Mt/yr max
leaching. Mining geometry prerequisites require that a
panel (defined as two benches) cannot start mining until
mining is completed for its overlying panel in the same
phase and until its inner phase(s) have been mined out to
at least one panel deeper. Note: vertical advance constraints
were not applied to small panels with less than 50,000t.
Key economic and cost parameters for plan optimiza-
tion and evaluation were: 12.5% discount rate $1250/oz
gold price $1.70/t +$0.01/t/bench mining cost $1.00/t
stockpile reclaim cost $12/t milling cost for 90% recov-
ery $6/t leaching cost for 70% recovery. Gold grade dis-
tribution statistics were tracked for each bench of each
phase using 22 increments based on Au cut-off grades.
Production scheduling was configured for up to 15 single
year time periods.
Gold grade conditional simulations were created
using 20' composites as input to Hexagon’s MinePlan min-
ing software’s Sequential Gaussian Simulation module.
Ordinary kriging of the simulated points was used to trans-
form each point simulation into a conditional simulation
of selective mining units—the “CS” model [Hoerger and
Dagdelen, 2024]. For this presentation, nine simulations
were used to enable easier graphing of blocks and valida-
tion of results—more realizations can be used to provide
smoother output metric distributions.
To allow testing of deterministic plans, two addi-
tional models were derived from the CS model [Hoerger
and Dagdelen, 2024]. The “IK” model, approximating an
Figure 1. Stochastic Production Scheduling data framework to solve for W, X, Y, Z and V
variables. For deterministic scheduling, remove all k subscripts. For stochastic evaluation,
treat W variables as constants
CASE STUDY: OPEN PIT GOLD MINE
A case study was developed to demonstrate the use of sto-
chastic evaluation to compare the value and risk charac-
teristics of different strategic mine plan alternatives. Mine
plans were developed with stochastic and deterministic
optimization methods with various cut-off grade and stock-
piling strategies. For perspective, orebody uncertainty risks
are compared to the plan’s mining rate and vertical advance
rate sensitivity.
McLaughlin Mine Overview
From 1985 to 2002, the McLaughlin operation produced
about 3.4 million ounces of gold from a hot-springs type
epithermal deposit. The mine’s drillhole database serves as
the input for creating conditional simulations for the case
study. Ultimate pit and initial phase designs inside the ulti-
mate pit were already provided as inputs [Aras et al, 2019].
The initial five phase designs were split to create nine
phases—four south, four north and a final phase beneath
the northern and southern phases. Plans were created with
and without stockpiling for mill-only and mill+leach pro-
cessing scenarios.
Key constraints for the case study were defined as 10Mt/
yr max mining, 8 20' benches/yr max vertical advance for
each layback, 2.25Mt/yr max milling and 2.25Mt/yr max
leaching. Mining geometry prerequisites require that a
panel (defined as two benches) cannot start mining until
mining is completed for its overlying panel in the same
phase and until its inner phase(s) have been mined out to
at least one panel deeper. Note: vertical advance constraints
were not applied to small panels with less than 50,000t.
Key economic and cost parameters for plan optimiza-
tion and evaluation were: 12.5% discount rate $1250/oz
gold price $1.70/t +$0.01/t/bench mining cost $1.00/t
stockpile reclaim cost $12/t milling cost for 90% recov-
ery $6/t leaching cost for 70% recovery. Gold grade dis-
tribution statistics were tracked for each bench of each
phase using 22 increments based on Au cut-off grades.
Production scheduling was configured for up to 15 single
year time periods.
Gold grade conditional simulations were created
using 20' composites as input to Hexagon’s MinePlan min-
ing software’s Sequential Gaussian Simulation module.
Ordinary kriging of the simulated points was used to trans-
form each point simulation into a conditional simulation
of selective mining units—the “CS” model [Hoerger and
Dagdelen, 2024]. For this presentation, nine simulations
were used to enable easier graphing of blocks and valida-
tion of results—more realizations can be used to provide
smoother output metric distributions.
To allow testing of deterministic plans, two addi-
tional models were derived from the CS model [Hoerger
and Dagdelen, 2024]. The “IK” model, approximating an
Figure 1. Stochastic Production Scheduling data framework to solve for W, X, Y, Z and V
variables. For deterministic scheduling, remove all k subscripts. For stochastic evaluation,
treat W variables as constants