VANCOUVER 2015

A TABLETOP COST ESTIMATE REVIEW OF SEVERAL LARGE HPGR PROJECTS

B. Costello and J. Brown
Fluor Canada Ltd
700-1075 West Georgia Street
Vancouver, BC, V6E 4M7
(*Corresponding author: [email protected])

1|P age

VANCOUVER 2015

A TABLETOP COST ESTIMATE REVIEW OF SEVERAL LARGE HPGR PROJECTS

ABSTRACT
The objective of the paper is to evaluate at a conceptual level the implementation of high pressure
grinding roll technology in the crushing and milling circuit of a 140 ktpd copper concentrator. The paper
focuses on a conceptual design of the HPGR circuit, and a relative capital and operating costs estimation to
allow evaluation of the concept against Semi-Autogenous, Ball Milling and Pebble Crushing (SABC) style
circuits.
A robust reference set of projects and studies were chosen as the basis for the conceptual level
benchmarking analysis and economic evaluation. Documents and reference source data provided by the
owner of this project were reviewed and relevant information extracted from this data was incorporated
into the analysis. The HPGR design concept selected for the present study follows the principles of the
circuit adopted for a similar, existing operating plant using HPGR technology.
Based on the outcomes of the conceptual level review, there appears to be an economic case
supporting consideration of HPGR technology. The existing site chosen for the project is also physically
able to accommodate the installation of a HPGR circuit, with little or no adjustment required. Furthermore,
the stable operating characteristics of an HPGR-based plant are better able to deal with variability in the
ore feed characteristics and Run-Of-Mine (ROM) size distribution. If the economic advantage attached to
HPGR is attractive enough to warrant further investigation, then the next logical stage would be to
undertake the necessary HPGR-specific testwork required to establish a good design basis for the project,
and to develop the comparative analysis to the next level of detail and accuracy.
KEYWORDS
HPGR, SABC, Trade-Off, B. Costello, J. Brown, CIM, SAG 2015
INTRODUCTION
The objective of the paper is to evaluate, at a conceptual level, the possibilities for using high
pressure grinding rolls (HPGR) in the crushing and milling circuit of a 140 ktpd copper concentrator. The
evaluation is intended to simplify the analysis associated with the determination of whether the
implementation of HPGR technology in a circuit would generate value. Particular emphasis is placed on
the advantages and disadvantages both general and specific to a 140 ktpd test case, Project A. The study is
based on benchmarking and reference project information, thus allowing the analysis to be conducted at
early stages in the project development, even when little or no specific HPGR test data is available.
Typically, in many projects HPGR technology is not always considered during the early stages of
project development. As such, specific ore competency data allowing for the viability of HPGR technology
to be considered is not always established.
As part of this current evaluation, traditional SABC comminution data were taken from a robust
set of reference projects, operating sites and studies with a view to identifying other project signatures
comparable to the equivalent Project A ore data, thus enabling comparisons to be made between flowsheets
and for technology requirements to be quantified.

2|P age

VANCOUVER 2015

METHODOLOGY
Fluor developed the set of reference projects from its own database and published information.
Documents and reference source data associated with the test case project were reviewed and relevant
information extracted from this data was incorporated into the analysis. The HPGR design selected for the
present analysis follows proven design principles successfully applied to an existing, operating HPGR
facility.
Reference Projects
Fluor has contributed to several recent large copper porphyry projects. Each of the projects did
undertake a study to identify the most beneficial comminution circuit design, primarily comparing high
pressure grinding roll installations to conventional SAG / Ball Mill / Pebble crushing (SABC) installations.
The names of the projects have been changed in order to respect client confidentiality.

Project B:
Project C:
Project D:
Project E:
Project F:
Project G:
Project H:

120 ktpd medium-hard rock Cu/Mo operation, operating (HPGR)

240 ktpd medium-hard rock Cu/Mo operation, under construction (HPGR)
110 ktpd hard rock Cu/Mo, operating (HPGR)
120 ktpd soft rock Cu operation, construction currently on hold (SABC)
150 ktpd medium-hard rock Cu/Mo project (HPGR)
240 ktpd medium-hard rock Cu/Mo project (HPGR)
17 ktpd medium-hard rock Cu/Mo project (HPGR)

Of specific interest is the fact that the majority of the projects referenced above are associated
with high tonnage processing of copper porphyry ores several of which proceeded with HPGR circuits,
whilst others selected conventional SABC designs, depending on the relative technical and economic
merits in each specific case.
Project A Input Data
Documents and reference source data describing Project A were reviewed and any relevant
information was incorporated into the analysis. This data included:

Process Design Criteria

Comminution testwork results
Throughput projections
PFS level capital cost estimates
PFS operating cost estimates

As shown in Table 1 Project As key parameters, projects of interest when considering HPGR vs.
SABC (such as UCS, WiBM, A*b), are similar to Project B, therefore, Project B represents a realistic real
world confirmation for the potential application of HPGR technology to Project A.

3|P age

VANCOUVER 2015

Table 1 Ore Characteristics Observed in Other HPGR vs. SABC Evaluations

Average
Value

Units

Project A

Project B

Project D

WiC

kWh/t

8.3

13

10.2

WiRM

kWh/t

14.7

UCS

MPa

135

140

140

15.07

17.5

0.225-0.766

0.24

0.33

6.06
47.12

47.3

0.478

0.522

WiBM

kWh/t

16.3

Ai
DWi
A*b

kWh/m

ta
3

Project F

Project G

17.09

14.0

260

262

ts/m h
260
2604
2501
m-dot
Notes:
1 Estimated from specific press force
2 Pivot table analysis of grinding intervals - 557 samples
3 Estimated value from comparison with other projects
4 Calculated from site data
CONCEPTUAL HPGR CIRCUIT DESIGN

This section describes the technical considerations supporting the development of the conceptual
level HPGR circuit design for Project A. The conceptual design provides the basis for the capital and
operating cost estimates, and the comparative economic evaluation of HPGR versus SABC technology.
Project A Process Design Criteria
A review of the existing Project A Process Design Criteria (Table 2) yields the following
important process parameters, considered throughout this study document:
Table 2 Project A Project Design Criteria
Description
Daily throughput
Ore lithologies
Solids specific gravity
Bulk density (volumetric)
Impact crushability index (CWi)
Bond ball mill work index
Bond rod mill work index
Abrasion index (Ai)
Primary crusher open side setting
Coarse ore stockpile live capacity
SAG mill feed rate (balance)
SAG Mill feed rate (design)

Unit
ktpd

t/m
t/m3
kWh/t
kWh/t
kWh/t
mm
tonnes
t/h
t/h

Value (mean)
140
Breccia, Granodiorite, Porphyry
2.6
1.6
4 34
14 22.6 (16.3)
12.3 20.5
0.225 0.766
173
80,000
6,341
7,608 (120% of balance)

4|P age

VANCOUVER 2015

Additional HPGR-Related Criteria

ROM Size Distribution
The ROM size distribution estimates for Project A were reviewed to identify qualitative and semiquantitative insights on how Project A plant feed compares with other ore types. Although there was not a
wide variation in prediction of ROM size distribution through the three dominant Project A lithologies
(Breccia, Granodiorite and Porphyry) for the purposes of establishing a design basis to allow for mass
balancing and equipment selection both the Breccia and Granodiorite distributions were selected for
comparison.
A brief review was conducted to provide context to this information (relative to other existing
operations having similar / same mechanical rock parameters). Data from Project B was plotted against
Project As ROM blast model outputs. These are shown graphically in Figure 1.
100
Project A Zone 1

90

Project A Zone 2

80

Project B

% Passing

70

Kuz ram + Kanchibotla correction

60
50
40
30
20
10
0
1

10

100

1000

Size (mm)

Figure 1 Project A ROM predictions vs Project B

The ROM size distribution projections for Project A were reviewed and compared with a range of
reference data. As Figure 1 demonstrates, the minus 100 mm size fractions compare favourably with the
reference data set, specifically Project B. Above the 150mm size fraction Project As data differs from the
reference set. Regardless, the current Project A ROM size distribution projections are used for this
evaluation without any correction.
Naturally, the ROM size distribution will dictate not just the primary crushing requirements, but
also the primary milling requirements. In practice and as is evident in several of the references, the
operation of an HPGR circuit will be little influenced by variability in the primary crusher product sizing,
unlike the SABC circuit where the impact on performance can be much more significant.
In addition to providing a comparison between the expected Project A ROM size distribution and
Project B, a check of the blast prediction calculation was carried out using the same blast hole parameters.
The method used is the combined Kuz Ram with the Kanchibotla fines correction (also shown in Figure 1).
Most open pit practices would expect distributions to be closed (p100) at 1000-1200 mm with a
medium-hard and homogenous ore type. Project As estimates indicate that most particles would be
relatively fine (92-97% -300 mm). This degree of fineness might be explained by two possibilities:

5|P age

VANCOUVER 2015

(a) An unusually tight blast pattern combined with a high powder factor or
(b) A degree of in-situ fracturing or weathering.
Additional data is required to properly understand the underlying cause behind the deviation in the
coarser end of the size distribution. Regardless, the projected primary crusher product size P80 of 108mm
based on Project As forecast ROM distribution for Zone 1 has been used in this study for the purposes of
developing the HPGR circuit mass balance and for selecting the equipment for the HPGR circuit.
Ore Variability
Another important consideration is that of ore variability. Project As information showed
definite levels of inter- and intra-lithological variation in the ores comminution properties.
Significantly, and in marked contrast to the observed behaviour of SAG milling circuits, previous
studies and technical papers relating to actual HPGR plant operation have noted that circuits employing
secondary crushing and tertiary HPGRs are able to accommodate ore variability without significant
adverse effects on downstream throughput or finished ore quality.
It is well understood that variability in the comminution parameters can exist between the various
zones and lithologies in an orebody. However, to examine this level of variability and to fully quantify the
effects is considered beyond the scope of this paper. For the purposes of the evaluation, a single Bond ball
mill work index has been used (16.3kWh/t: the mean of the hundreds of Bond ball mill work index values).
HPGR Flowsheet Concepts
Typically, for hard rock applications, a controlled HPGR feed top size is required to avoid damage
to the roll wear surface and this implies the use of closed circuit secondary crushing upstream. In a similar
way, a controlled ball mill feed top size is normally required to avoid milling inefficiencies, thus
mandating closed-circuit operation of the HPGR itself.
Ore characterization data for Project A suggests that the HPGR product flake competency might
be equivalent to that currently observed at Project B, which as a comparison does not present any excessive
downstream ore handling requirements. Flake competency test procedures have been developed by the
HPGR manufacturers to quantify this effect. KHDs method entails treatment of the HPGR product in a
tumbling device for a fixed period. The proportion of surviving flake is then assessed and the flake product
given a score on a scale of 0-100; the lower the score, the higher the de-agglomeration energy required. A
score above 70 indicates no special measures are required, with adequate de-agglomeration occurring in
normal treatment on conveyors and in bins. However, in practice, flake competency is only a real problem
on ores with a high smectite or clay content (often connected with diamondiferous ores such as kimberlite).
None of which has been identified in the Project A ore types.
In the absence of ore-specific HPGR and flake competency testwork (which Fluor would typically
recommend), Project As comminution data suggested wet screening would be a requirement (as at Project
B and Project D).
HPGR Circuit Basis For Project A
The HPGR design concept selected for the present study follows the principles of the circuit
adopted for Project B. Whilst many possible permutations of HPGR based circuits can be conceived, some
involving quite subtle differences from the others, the basic Project B circuit represents the simplest and
lowest cost circuit available and is one that has already stood the test of time.

6|P age

VANCOUVER 2015

The HPGR circuit schematic flowsheet for this study is given in Figure 2, and comprises the
following process stages.

Primary crushing
Screening
Closed circuit secondary cone crushing
Closed circuit HPGR (tertiary crushing)
Wet screening
Ball milling

The outcomes of Project Bs trade-off studies validated the use of HPGR over SABC, as have
more recent studies for Project C. The combination of a medium-hard rock application with a long mine
life contributed to placing a focus on operating costs and on securing reliable, consistent plant production,
irrespective of ore variability, and it was these factors that fundamentally drove the decision making
process for these projects.
Base Case
- CV1 Circuit
HPGR
(combined
screening and secondary crushing)

Secondary Crushing
Circuit Feed Bin

Tertiary Crushing Circuit

Feed Bin

Coarse Ore
Stockpile
Secondary Screens

u/s

HPGRs

o/s
Grinding Circuit Feed
Bin

Secondary Crushers

o/s

Wet Tertiary Screens

u/s
Ball Mill Grinding Circuit

Figure 2 HPGR Circuit Flowsheet - based on Project B

Design Development
To develop a suitable basis for design, equipment consistent with previous projects has been
selected. The main items are as follows:

Primary crushers1:
Secondary cone crushers:
Secondary screens:
HPGR:
Wet screens:

60-89 Gyratory crushers

1300 HP
3.66 x 8.5m (12x27) double deck banana screens
24/17 or 26/18 units
3.66 x 8.5m (12x27) double deck banana screens

Notes:
1. For cost estimating purposes, all equipment installed upstream from the coarse ore stockpile (COS) is
considered to be equivalent in installed cost for the SABC and HPGR flowsheet options. On this basis,

7|P age

VANCOUVER 2015

cost elements for primary crushing are not included in the analysis however it is necessary to know
details of the primary crushing and primary crusher product size distributions in order to correctly size and
design the HPGR circuit.
Primary Crushing
It is conceivable that Project As primary crusher will produce a p80 of 120-130mm, primarily as a
result of a somewhat coarser ROM size distribution than is predicted from JKSimblast simulations.
Crusher modelling conducted by Fluor (and compared with a prominent Vendors database) strongly
indicated that the target p80 of 108mm through a 60-89 unit (relying on the JKSimblast prediction) is
unlikely to be achievable at an OSS of 173mm.
Notwithstanding this inconsistency, the SAG Mill feed f80 of 108mm stated above for Zone 1 has
still been carried throughout the study for the sake of consistency and used in developing the HPGR circuit
mass balance, and for selecting and sizing the equipment for the secondary and tertiary (HPGR) circuits.
Secondary Crushing
The secondary crushers will be fed in a typical arrangement each crushing line consisting of a bin,
feeder, screen, and cone crusher. The crushers would be set at 30-40 mm closed side setting (CSS) and
would produce a final product of approximately 70%-40 mm. Crushed product would be returned to the
screen bin through the return conveyors discharging onto the main coarse ore storage (COS) reclaim belt.
Screening
Screening will be performed using conventional double deck banana screen technology. For
Project B and Project D, the arrangement utilizes individual lines consisting of a crusher that is fed using a
double deck banana screen and feeder arrangement. As a result, the screens have to be matched in capacity
to the cone crusher.
For Project C, the throughput requirements presented materials handling challenges that required
the separation of the crushing and screening facilities. For Project A, the Project B design is suitable as a
basis for analysis.
Deck media selected for this study consist of a protection top deck with 75 x 100 mm apertures
and a bottom deck aperture size of 45 x 68 mm.
Tertiary Crushing (by HPGR)
In the absence of HPGR test data for Project A ore types, HPGR performance characteristics were
extracted from published full scale operational data, supplemented by theoretical relationships developed
through HPGR pilot databases.
The in-house Fluor database allows for the identification of the
relationships between the various key HPGR operating parameters, namely specific press force, specific
energy, and specific throughput. This section describes some of those assumptions, considering the key
HPGR operating parameters captured in Fluors database, as follows:

Specific press force in N/mm2

Specific energy in kWh/t
Specific throughput in ts/(m3 h)
D50 in microns

The specific press force is defined as the total piston force divided by apparent (or projected) area
of the rolls; i.e. the length times the diameter. The required specific press force is generally ore specific and

8|P age

VANCOUVER 2015

is the most important parameter in determining overall wear rates of the studs. Important aspects to
consider include:

The HPGR equipment at Project B was originally designed for specific press forces of 3.5 N/mm2.
An HPGR installation at a large copper mine in Indonesia was designed for specific press forces of
between 4.0 and 4.5 N/mm2.
The Polysius Atwal abrasion test is generally performed at 4.0 N/mm2.
For a given feed size and ore type, the higher the specific press force the finer the product size will be.
This is illustrated in Figure 3.

Whilst higher specific press forces generate finer products, higher press forces also lead to
increased stud wear, and therefore lower expected roll lives. The actual specific press force of an operation
should be decided by cost-benefit analysis on incremental maintenance costs versus the incremental
benefits of a finer HPGR product.
For the purpose of this study we have selected 3.5 N/mm2 specific press force, based again on
Project B. Breakage characteristics were modeled from the operation at Project B (Por 3 in Figure 3) and
the resulting d50 can be seen in Figure 3.

7000
CuAu Por 1 (Indo.)

6000

CuMo Por 1 (Chile)

CuMo Por 2 (Chile)

d50 (microns)

5000

CuMo Por 3 (Peru)

Mo Por 1 (N.Am.)

4000
3000
2000
1000
0
0.0

1.0

2.0

3.0

Press Force

4.0

5.0

6.0

7.0

(kN/mm2)

Figure 3 Relationship between press force and product d50 (m)

The specific energy is dependent on the specific press force (see Figure 4 after Klymowsky et
al.) and it can be seen there is a strong relationship between the press force and the specific energy applied.
For example at Project B with a press force of 3.5 N/mm2, the recorded corresponding specific energy is
1.80 kWh/t (in open circuit) which also confirms the relationship.

9|P age

VANCOUVER 2015

6
y = 0.4x + 0.4
R2 = 0.9146

Specific Energy Input (kWh/t)

0
0

10

12

Press Force (kN/mm2)

Figure 4 Relationship between press force and specific energy input

For the purposes of assessing the HPGR equipment requirements for Project A, a comparison with
Project B can be used as a starting point, understanding that the treatment rate requirements are the main
differential between Project B and Project A. As such two options would exist, if all other items are
considered equal, (a) firstly to maintain a similar scale of equipment as Project B and produce a coarser
ball mill feed or (b) secondly to consider larger HPGR equipment to maintain similar levels of performance.
For the purposes of this study, the latter approach has been adopted.
The HPGR equipment requirements for Project A are as defined in Table 3.
Table 3 Expected Project A HPGR equipment requirements
Parameter
Specific force
Specific energy
Specific throughput
p80
p50
HPGR roll diameter
HPGR roll width
HPGR peripheral roll speed
Unit throughput
Circuit throughput
Circulating load
Total HPGR duty throughput
No. of duty units
Power
Total installed power

Units
(N/mm2)
(kWh/t)
(ts/m3h)
(m)
(m)
(m)
(m)
(m/s)
(t/h)
(t/h)
(%)
(t/h)
(kW)
(kW)

Project B

Project A

3.5
1.8
260
10,337
3,000
2.4
1.65
2.61
2,687
5,056
186%
9,396
3.5
4,837
5,000

3.5
1.8
260
10,337
3,000
2.6
1.75
2.61
3,088
6,554
186%
12,180
3.9
5,558
6,600

10 | P a g e

VANCOUVER 2015

At the time of writing, several suppliers (Metso and Polysius) have developed large HPGR
geometries with increased connected power. For the purposes of this study, units established and proven in
existing operations have been used (Project B, Project D).
Ball Milling
Since the initial industrial application of HPGR technology it has been suggested and largely
accepted by the industry that additional weakening of the ore is induced through the breakage mechanism
associated with HPGR technology. This weakening is variable and is not only ore dependant, but also
specific to the machine configuration, feed size distribution and arrangement. Tests and operational
observations made at Project B demonstrate that this reduction in ore hardness can reduce the required
applied grinding power by between 5 and 10%.
Data from other projects and studies may also be mentioned, as follows:

North American molybdenum project (unpublished) 11.0%

Los Bronces (published) 12.7%
North American Cu-Mo porphyry (unpublished) 5.7%
South American Cu-Mo porphyry (unpublished) 7.5%

Recent studies show using ore taken from a mine in British Columbia (Wang, 2013), that the
reduction in the Bond ball mill work index can be as high as 10-20%. Some samples from the Project B
program showed more than 30% reduction in work index. For the purposes of this study a conservative
position has been adopted and a figure of 7.5% has been applied.
As indicated earlier, a mean Bond ball mill work index of 16.3 kWh/t has been selected and
applied throughout this study. As a check, applying this mean value to the calculation of the ball mill
specific energy for the Project As SABC circuit design gives a calculated value for the ball milling
specific energy that correctly aligns with the figure stated in the projects Design Criteria (Table 2).
Table 4 Project A Ball mill selection
Parameter

Units

SABC

HPGR

Mill throughput
Operating time
Mill circuit capacity
Ball mill f80
Ball mill p80
Bond ball mill work index
Specific energy (uncorrected)
Power required (uncorrected)

(tpd)
(%)
(t/h)
(m)
(m)
(kWh/t)
(kWh/t)
(kW)

140,000
92
6,341
4,300
200
16.3
9.040
57,320

140,000
92
6,341
3,000
200
16.3
8.550
54,211

1
1
57,320
9.040
52,734
4
26
42
78
33-38
16,400

0.925
0.925
50,145
7.909
46,134
4
24
38
78
33-38
13,000

HPGR Factor
Total Factor
Corrected ball mill power
Specific energy required
Total BM power (running)
No. of ball mills
Inside shell diameter
Effective grinding length
Percent critical speed
Charge
Selected drive size per ball mill

(kW)
(kWh/t)
(kW)
(ft)
(ft)
(%)
(%)
(kW)

11 | P a g e

VANCOUVER 2015

Coarse Ore Stockpile

6341 t/h

Secondary Cone
Crushers

11498 t/h

Double Deck Banana

Screens

5157 t/h

Fine Ore Bins

12180 t/h

HPGR

Wet Screens
5839 t/h

Ball Mill Discharge

6341 t/h

Ball Mill Cyclones

Figure 5 Project A HPGR Materials Handling Mass Balance

12 | P a g e

VANCOUVER 2015

Major Equipment Selections

The following major equipment quantities and loads were defined from the mass balancing
exercise.
Table 5 Project A Major equipment selections (HPGR circuit)
Qty
2
4
4
4
8
4

Equipment

Parameter

Units

Value

Average Load /
unit
(%)
84% (100%)
61%
77%
78%
75%
53%
79%
89%

mm
Primary crushers1
OSS
178 (150)
mm
Secondary screens
TD aperture
75
mm
(3.66 x 8.25m DD)
BD aperture
45
mm
Secondary crushers
CSS
38
mm
HPGR (26/18)
P80
10.3
mm
Wet screens
TD aperture
14
mm
(3.66 x 8.25 DD)
BD aperture
4.5
kW
Ball mills (24x38)
Mill motor
13,000
Note:
1. To achieve the required 108 mm (SAG mill equivalent F80), Fluor would suggest the use of a
tighter OSS (150 mm) to achieve a 120-130 mm F80 and with consideration given to larger
primary crushers (60-110) which would of course provide additional capacity.

Layout Considerations
In many respects, the layout and infrastructure requirements for an HPGR machine are similar to
those for conventional crushers, usually including a feed bin, belt feeder, product conveyor, dust control
facilities and tramp metal management. However, the nature of the HPGR mechanism is such that certain
additional measures are necessary to ensure satisfactory performance and successful operation.
The general layout of the HPGR circuit is distributed in different building modules with plant
conveyors linking one building with another. For this study it is assumed that the primary crusher station
configuration will be the same for either the HPGR or SABC comminution circuit.
The coarse ore stockpile has the same live capacity for either circuit, however for the HPGR
circuit a simple conical stockpile with a single reclaim tunnel is proposed, similar to Project B, in place of
the longitudinal pile with tripper or shuttle head and two (2) reclaim tunnels required for the SABC circuit.
For the purposes of cost estimation, however, these distinctions are considered to be equal and are not
applied in the analysis.
From the coarse ore stockpile, the ore is reclaimed and conveyed to the coarse ore surge bins,
located in the secondary screening and crushing building. Oversize from the secondary screens is fed
directly to the secondary crushers. Product from the secondary crushers is recirculated to the coarse ore
surge bins for re-screening. The secondary screen undersize product is conveyed to the HPGR feed bins in
the tertiary crusher building.
It is important to avoid segregation in the feed to an HPGR to avoid roll skew, which impairs
comminution performance and in the extreme will trip the machine's main drives. With this in mind, the
feed stream should be presented symmetrically to the feed hopper, and this is achieved by arranging the
belt feeder head pulley with its axis parallel to the roll axes. The main motors and drive shafts of an HPGR
typically occupy about as much floor area as the machine itself and this has to be considered when
developing the overall HPGR layout.

13 | P a g e

VANCOUVER 2015

For this particular study, four (4) HPGR units are considered with a layout that is similar to the
Project B and Project D arrangements. Product from the HPGR section is conveyed to the fine ore bins
adjacent to the ball mills.
Tertiary crushed product is reclaimed from the fine ore bins to four (4) pairs of double deck wet
banana screens (two (2) screens per mill line). The oversize from the screens is returned by conveyors to
the HPGR feed bins. Undersize from the screens is collected in the ball mill cyclone feed sump for feed to
the downstream circuit which is the same for either SABC or HPGR.
Three (alternative HPGR configurations were overlaid at the correct scale onto the existing
Project A layouts for an SABC circuit
Of these three configurations, Option 2, shown at greatly reduced size in Figure 6, is considered to
be the best for initial comparisons with the SABC layout While not optimised this arrangement provides a
sound basis for this conceptual analysis.
As may be seen from Figure 6 the HPGR circuit layout fits easily onto the present SABC circuit
site plan the only layout accommodation required being a minor relocation of the coarse ore stockpile,
along the existing alignment of the overland plant feed conveyor.
The common point of connection between the HPGR circuit and the existing downstream sections
of the SABC based site plan is at the mill building (the existing ball mills being shown in grey outline in
Figure 6, positioned at their present location) and the location of the mill building and all downstream
facilities will remain the same in either case.
ECONOMIC EVALUATION
Basis of Evaluation
In addition to developing full scale EPCM projects through to production, Fluor has conducted or
has been exposed to many trade-off studies specifically targeted at defining the cost benefits (or otherwise)
of circuits that utilize HPGR technology. These include the projects that constitute the reference set for the
conceptual level benchmarking analysis, applied in this section of the report for economically evaluating
the potential application of HPGR technology to Project A.
An observation common to all studies of HPGR versus SABC comminution is that an HPGRbased circuit always incurs a capital cost penalty relative to SABC circuits, despite the fact that the HPGR
machines themselves are typically less costly than the equivalent SAG mills. This is because the HPGR
and secondary crushing stage are normally operated in closed circuit, and it is the cost of these closed
circuit facilities crushers, screens, conveyors, bins, feeders, dust control, tramp metal management that
increases the cost of the HPGR-based circuit. They also contribute an additional power demand that offsets
some of the comminution energy savings from HPGR.
Each study Fluor has been involved with has examined the specific aspects relevant to each
particular project, but of course it is the operating cost elements (e.g. power and consumables) that provide
the main points of leverage in the assessment. In some studies a high degree of estimating accuracy has
been applied (AACE Class II) in developing these costs while other studies have relied on a less rigorous
estimating approach (AACE Class IV or V).
Fundamentally, it is the relative costs of the comminution circuit that form the basis of the capital
and operating cost comparison. In each of the projects examined; the boundary limits of the plant design
tends to encompass the coarse ore stockpile discharge through to the ball milling cyclone overflow (the
flotation circuit feed).

14 | P a g e

VANCOUVER 2015

Figure 6 Conceptual HPGR layout shown superimposed on the existing site plan
To leverage the available body of information for this paper Fluor has extracted the capital cost
and operating cost information from the studies developed for each of the projects listed in the
benchmarking reference set.
The capital cost estimates in each case have been baselined to include only the direct costs. Also
the mass earthworks costs associated with site specific topography have been adjusted to normalize the
comparisons.
Differential Capital Costs: HPGR versus SAG
Differential Capex Costs - Derived from the Reference Set of Studies
For each trade-off study the ratio of the capital cost of the HPGR circuit to the cost of the SABC
circuit has been determined. These ratios are summarised in Table 6. For each of calculated ratios, the
same set of boundary limits applies, i.e. the underside of the coarse ore stockpile to the ball mill cyclone
overflow.

15 | P a g e

VANCOUVER 2015

Table 6 Direct Capital Cost comparisons (HPGR vs. SABC circuits) from recent studies
Project

HPGR/SABC
Ratio

Project B
Project C
Project D
Project E
Project F
Project G
Project H
AVERAGE

(%)
119%
114%
115%
124%
125%
130%
115%
121%

Only direct capital costs have been considered in making the initial comparisons shown in Table 6,
in order to prevent any project-specific indirect cost anomalies from distorting the analysis.
The results of all the referenced studies fall within a fairly tight band, with the average HPGR
circuit direct capital cost being 121% of the cost of the equivalent SABC circuit. This benchmarked
average factor will be considered in the analysis that follows, to establish the likely economics of applying
HPGR technology to Project A.
It is worth noting that (with the single exception of the Project H) all of the comparative studies
listed in Table 6 were carried out for plant throughputs in the 100-150 ktpd range.
Differential Capital Cost Estimate for Project A Case
For consistency of estimating a standard capital cost template (covering the standard set of
comminution circuit battery limits considered in the present study was developed. Project As SABC prefeasibility capital cost estimate was adopted as the basis for the economic analysis of the present study.
This data is summarized in Table 7.
Table 7 SABC Circuit Direct Capital Cost Estimate for Project A
Area
Coarse Ore Reclaim & Surge Bin Feed
SAG Mill circuit
Pebble stockpile, reclaim ad conveying
TOTAL

Direct Capital Cost

($M)
67.1
476.7
42.5
586.3

It is expected that the equivalent HPGR based comminution circuit would incur an additional 21%
of direct capital cost. In addition, the associated indirect costs need to be applied to both cases in order to
establish the total difference in capital cost.
The results of this are shown in Table 8. An estimate of the Indirect Capital Costs is made by
applying a 30% factor to the Direct Capital Cost Estimate. This is a typical figure for projects of this nature.

16 | P a g e

VANCOUVER 2015

Table 8 Capital Cost Comparison for Project A

Costs
Direct Capital Cost (Factor 1.21 for HPGR)
Indirect Capital Cost (@30% of direct)
Total
Difference

SABC Circuit
($M)
586
176
762

HPGR Circuit
($M)
709
211
920
158

Differential Operating Costs: HPGR versus SAG

Differential Operating Cost Data Benchmarked from Reference Projects and Studies
The operating costs associated with the reference set of projects and studies have been developed
from many sources:

Actual operating sites

Combinations of actual test data and practical extrapolations
A more theoretical perspective

In deriving the differential operating cost estimates for this study operating cost data from the
public domain has been referenced openly, in addition to tapping into sensitive client information, the
details of which have been protected in accordance with the relevant confidentiality requirements.
The main operating cost items of relevance to a study of this nature would typically include:

Power
Grinding media
Crusher and Mill liners
Screen panels
Other maintenance items (including repair supply costs)
Labour

Of these, it has been found in previous studies that screen panels and minor maintenance supplies,
either the absolute costs themselves or the cost differentials between the comparative cases are relatively
small (typically 0.1-0.5% of the total circuit operating cost) and these are generally disregarded in the
comparisons. However, to demonstrate consistency, these items have also been accounted for in the present
study.
An equivalent HPGR based operation will also require specific additional spares and maintenance
associated with secondary crushers and HPGRs, such as crusher liners and replacement rolls. This is a
significant item for which relevant information has been extracted from previous assessments and is
accounted for under the repair supply cost item in the operating cost tables that follow.
In addition, it is recognized that subtle differences in labour functions and manning levels may
contribute to the discussion; however, previous studies demonstrate these to be minor. For the purposes of
this preliminary study, the operating and maintenance labour component of the operating cost estimate is
defined as a distinct line item, but is held constant for both cases.
Of the seven reference cases labour is included in the operating cost estimate of three of them. In
each of these cases the cost of labour is a relatively minor contributor to the overall operating expenses and
has little to no impact on the relationship between the HPGR and SABC operating costs. It is recognised
that the data set selected for the desktop study is heavily weighted towards South American projects and

17 | P a g e

VANCOUVER 2015

that this assumption would need to be validated for other locations where the cost of labour may be a more
significant contributor to the overall operating costs.
When considering the relevance of HPGR technology, it is the energy efficiency of the HPGR
comminution versus SABC circuits that typically forms the focus of any evaluation or trade-off. However,
other equipment associated with HPGR circuits also contributes an additional power demand that offsets
some of the comminution energy savings from HPGR, and this too is an important consideration.
It is important to note that while the workings are not displayed in any detail in this report (for
reasons of confidentiality), all of the energy cost information derived from the reference set of studies was
normalised to the same energy input cost basis for the purposes of making the operating cost comparison.
A cost of electrical energy has been assumed to be $0.11/kWh.
Second to energy in cost are the aspects of consumable materials including HPGR roll surfaces
(and studs), mill liners and grinding media.
Operating costs for the HPGR and SABC options are compared and a ratio calculated in a similar
fashion to the approach adopted for the capital costs. Table 9 summarises the results.
Table 9 Operating cost comparisons (HPGR vs. SABC circuits)
Project
Project B
Project E
Project C
Project C
Project F
Project G
Project H
AVERAGE

Operating Cost Ratios

SABC
HPGR
HPGR
SABC
136%
74%
115%
87%
127%
79%
109%
92%
117%
85%
132%
76%
132%
76%
124%
81%

Project A Pre-Feasibility Study Operating Costs for SABC Case

In order to forecast the likely differential operating cost for HPGR and SABC at Project A, and to
incorporate this into any economic analysis, it is necessary to have the operating costs for the SABC
comminution circuit. These costs are summarised in Table 10.
The costs of the basic set of big ticket items that drive the differential in operating costs (such as
power, grinding media, crusher and mill liners) were extracted and consolidated.
In principle, at a conceptual level of analysis, it is now possible to apply the general operating
cost ratios to Project As SABC comminution circuit costs, in order to establish a differential operating
cost to use in an economic analysis.

18 | P a g e

VANCOUVER 2015

Table 10 Project A Comminution Circuit Operating Cost (SABC case)

based on analysis of the big ticket items
Cost Item

Equipment

Liners

SAG Mill liners

0.207

Screen panels

0.006

Pebble crusher liners

0.010

Ball Mill liners

0.122

SAG Mill balls

0.372

Ball Mill balls

0.627

Materials Handling

0.069

SAG Mill

0.715

Screens/minor equipment

0.003

Pebble crushing

0.014

Ball Mills

0.994

Media
Power

$/t

Repair supply costs

0.522

O&M labour costs

0.410

TOTAL COST

4.070

However, a set of Project specific unit rates is available from the pre-feasibility study for the
SABC case (for many items, such as power, grinding media, crusher and mill liners). These may also be
applied to the calculation of the HPGR comminution circuit operating costs, in order to put the Project A
comparison onto a more apples and apples basis than would result from just applying the global
benchmarks.
Project A Operating Costs Projections for HPGR
Certain specific operating cost elements are unique to circuits that include HPGR technology (e.g.
the costs of roll refurbishment and replacement) and the costs for these, in the present study, have been
developed by applying established criteria and known costs data from similar circuits in the reference set of
HPGR projects and studies. However, as far as possible the major generic cost components are based upon
the estimates from the existing pre-feasibility study operating cost estimate, summarised in Table 11.

19 | P a g e

VANCOUVER 2015

Table 11 Project A Comminution Circuit Operating Cost (HPGR case)

based on analysis of the big ticket items
Cost Item

Equipment

Liners

Secondary crusher liners

Screen panels
HPGR rolls
Ball Mill liners
Ball Mill balls
Secondary crushers
Screens/minor equipment
Materials Handling
HPGR
Ball Mills

Media
Power

Repair supply costs

O&M labour costs
TOTAL COST

$/t
0.020
0.019
0.263
0.099
0.510
0.028
0.008
0.097
0.386
0.870
0.783
0.410
3.493

Therefore, comparing the total costs given in Table 10 and Table 11, the derived HPGR/SABC
operating cost ratio for the Project A is 86% ($3.493 per tonne / $4.070 per tonne), after having applied the
specific Project A unit rate information available from the existing PFS level work done for the SABC case.
This compares with the average value of 81% derived earlier for HPGR projects in general, and falls well
within the observed range of the data.
Differential Discounted Net Present Cost (NPC) for HPGR versus SABC
Figure 7 summarizes the results of the analysis and benchmarking carried out so far in developing
the basis for the conceptual economic analysis of HPGR technology for the Project A project.

Figure 7 Capital and Operating Costs (HPGR vs. SABC) for Project A

20 | P a g e

VANCOUVER 2015

For the purposes of conceptual economic evaluation the average capital cost ratio of 121% can be
applied. The operating cost ratio for Project A is represented by the solid triangular data point in Figure 7,
is 86%. This is somewhat higher than the average ratio (81%) for the reference set of data, but fits within
the range (74% - 92%) of previous studies and is likely to be on the conservative side.
The results of the economic analysis are presented in Table 12, after applying typical discounted
cash flow parameters.

Discount rate:
8%
Mine life: 22 years
Table 12 Net Present Cost difference between equivalent HPGR and SABC circuits for Project A
Item

Cost

SABC CIRCUIT
Direct Capital Cost

$586 million

Indirect Capital Cost

$176 million

Total Capital Cost

$762 million

Operating Cost

$4.070/t

Operating Cost

$208 million/year

HPGR CIRCUIT
Direct Capital Cost

$709 million

Indirect Capital Cost

$211 million

Total Capital Cost

$920 million

Operating Cost

$3.493/t

Operating Cost

$178 million/year

NPC in favour of HPGR

$127 million

The results of the conceptual level analysis confirm that the HPGR circuit economics are
substantially more favourable than for the SABC base case. The HPGR alternative for Project A can be
expected to deliver a net present value advantage of $127 million, relative to the SABC case.
The result is sensitive to the operating cost difference between HPGR and SABC. Applying the
average operating cost ratio derived from the industry reference set of projects (81%) immediately
improves the NPV advantage for HPGR to $224 million, relative to the SABC case. This points to the
strong need to develop a Project A-specific HPGR study, that is based on actual test data and HPGRspecific operating costs information, in order to fully confirm the scale of the opportunity.

21 | P a g e

VANCOUVER 2015

SUMMARY
An observation common to all studies of HPGR versus SAG mill comminution is that an HPGRbased circuit always incurs a capital cost penalty relative to SAG, despite the fact that the HPGR machines
themselves are typically less costly than the equivalent SAG mills. The benefits of the HPGR technology
lie in its superior efficiency of energy utilization, which in the majority of cases more than outweighs the
capital cost difference.
The results of all the referenced projects and studies fall within a fairly tight band - with the
average HPGR circuit direct capital cost being 121% of the cost of the equivalent SABC circuit. The
operating cost for the SABC circuits in general is 24% greater than the operating costs of the equivalent
HPGR circuits.
Based on this information, the derived HPGR/SABC operating cost ratio for the Project A, and
applied in the economic analysis, is forecast to be 86% ($3.493/t for HPGR, versus $4.070/t for SABC).
The results of the conceptual level analysis confirm that the HPGR circuit economics are
substantially more favourable than for the SABC base case, and that, according to the information
developed during the course of the present study, the HPGR alternative for Project A can be expected to
deliver a net present value advantage of $127 million, relative to the SABC case.
As may be expected, the result is sensitive to the operating cost difference between HPGR and
SABC. Applying the average operating cost ratio derived from the industry reference set of projects (81%)
immediately improves the NPV advantage for HPGR to $224 million, relative to the SABC case. This
points to the strong need to develop a Project A-specific HPGR study, that is based on actual test data and
HPGR specific operating cost information, in order to fully confirm the scale of the opportunity.
REFERENCES
Amelunxen, P., Mular, M., Vanderbeek, J., Hill, L., & Herrera, E., (2011). The Effects of Ore Variability
on HPGR Tradeoff Economics. Proceedings SAG2011 International Autogenous Grinding,
Semiautogenous Grinding and High Pressure Grinding Roll Technology. Vancouver, Canada, 2528 Sept. 2011 .
Bolles, T., (2011). Pioneering the Use of HPGRs at Cerro Verde. Proceedings Mill Optimisation
Conference. March 30-31, 2011, Perth, Australia.
Bos, L., van de Vijfeijken, M., & Koponen, J., (2011). Insurability of Large Gearless Mill Drives,
Proceedings SAG2011 International Autogenous Grinding, Semiautogenous Grinding and High
Pressure Grinding Roll Technology. Vancouver, Canada, 25-28 Sept. 2011.
Burchardt, E., Patzelt, N., Knecht, J., & Klymowsky, R., (2011). HPGRs in Minerals: What do Existing
Operations Tell Us for the Future? Proceedings SAG2011 International Autogenous Grinding,
Semiautogenous Grinding and High Pressure Grinding Roll Technology. Vancouver, Canada, 2528 Sept. 2011.
Fuerstenau, D., & Abouzeid, A.-Z.M., (2007). Role of feed moisture in high-pressure roll mill
comminution, Int. J. Miner. Process, 82 (2007) 203-210.
Gerrard, M., Costello, B., & Morley, C., (2004). Operational Experiences and Performance Assessment of
HPGR Technology at Argyle Diamond Mine, Proceedings of the Rio Tinto Comminution Workshop
2004. Perth, Australia.

22 | P a g e

VANCOUVER 2015

Hart, S., Parker, B., Rees, T., Maneshand, A., & Mcgaffin, I., (2011). Commissioning and Ramp Up of the
HPGR Circuit at Newmont Boddington Gold. Proceedings SAG2011 International Autogenous
Grinding, Semiautogenous Grinding and High Pressure Grinding Roll Technology. Vancouver,
Canada, 25-28 Sept. 2011.
Heinicke, F., (2010) The advantage of half scale to full scale HPGR modelling. Comminution '10, Cape
Town, South Africa, April13-16, 2010.
Herbst, J., et al., (2011). Detailed Modeling of an HPGR-HRC for Prediction of Plant Scale Unit
Performance. Proceedings SAG2011 International Autogenous Grinding, Semiautogenous Grinding
and High Pressure Grinding Roll Technology. Vancouver, Canada, 25-28 Sept. 2011.
JKTech Laboratory Services, HPGR High Pressure Grinding Rolls brochure.
Kalala, J.T., Gong, H., & Hinde, A.L., (2011). Using Piston Die press to Predict the Breakage Behaviour of
HPGR, Proceedings SAG2011 International Autogenous Grinding, Semiautogenous Grinding and
High Pressure Grinding Roll Technology. Vancouver, Canada, 25-28 Sept. 2011.
Klymowsky, R., Patzelt, N., Knecht, J. and Burchardt, E. Selection And Sizing Of High Pressure
Grinding Rolls, Mineral Processing Plant Design, Practice and Control, Proceedings SME
Conference, Vancouver, 2002, Volume 1, 636-668.
Koski, S., Vanderbeek, J., & Enriquez, J., (2011) Cerro Verde ConcentratorFour Years Operating
HPGRs, SAG2011 Proceedings International Autogenous Grinding, Semiautogenous Grinding and
High Pressure Grinding Roll Technology 2011. Vancouver, Canada, 25-28 Sept. 2011.
Lane, G., (2010). What is a competent ore and how does this effect [sic] comminution circuit design?,
Proceedings 2010, 7th International Mineral Processing Conference, 8-10 December 2010,
Santiago, Chile.
Lozano, C., & Bulled, D., (2010). A method for the selection of SAG or HPGR grinding circuits for a
project involving the hardness variability of ore body, Proceedings 7th International Mineral
Processing Conference. 8-10 December 2010, Santiago, Chile.
Morley, C., (2011), HPGR Trade-off Studies and How to Avoid Them. Proceedings SAG2011
International Autogenous Grinding, Semiautogenous Grinding and High Pressure Grinding Roll
Technology. Vancouver, Canada, 25-28 Sept. 2011.
Morrell, S., & Valery, W., (2001). Influence of Feed Size on AG/SAG Mill Performance. Proceedings
SAG2011 International Autogenous Grinding, Semiautogenous Grinding and High Pressure
Grinding Roll Technology. Vancouver, Canada, 25-28 Sept. 2011.
Nadolski, A. et al (2011). Investigation into Laboratory Scale Tests for the Sizing of High Pressure
Grinding Rolls, Proceedings SAG2011 International Autogenous Grinding, Semiautogenous
Grinding and High Pressure Grinding Roll Technology. Vancouver, Canada, 25-28 Sept. 2011.
Polysius: HPGR Brochure_1603_gb_POLYCOM.
Polysius (2013) 5-TK-Workshop-Denver_High Pressure Grinding Rolls HPGR POLYCOM.
Polysius (2013) 6-TK-Workshop-Denver_HPGR Service, Service Centers and POLRED.

23 | P a g e

VANCOUVER 2015

Pontt, J., Rojas, F., Valderrama, W., Olivares, M., Albayay S, F., Ramos, U., & Robles, H., (2011).
Mechanical Forces of Electromagnetic Origin Causing Stress Conditions in Gearless Mill Drives,.
Proceedings SAG2011 International Autogenous Grinding, Semiautogenous Grinding and High
Pressure Grinding Roll Technology. Vancouver, Canada, 25-28 Sept. 2011.
Valery, W., (2002) The Future of Comminution, Proceedings 34th IOC on Mining and Metallurgy. 30
Sept-3 Oct 2002, Yugoslavia.
Van der Meer, F., Dicke, R., (2008). High Pressure Grinding: How High Can you Go? Proceedings 2008,
9th International Mineral Processing Conference. 21-24 October 2008, Santiago, Chile.
Van der Meer, F., Quincke, L., Fernndez, R. (2012). Pilot HPGR test programs and parameters for design,
Proceedings 5th International Mineral Processing Conference, 21-23 November 2012, Santiago,
Chile.
Vanderbeek, J., Linde, T., Brack, W., &Marsden, J., (2006). HPGR Implementation at Cerro Verde,
SAG2006. Vancouver, Canada 2006.
Wang, C., (2013). Comparison of HPGR - ball mill and HPGR - stirred mill circuits to the existing
AG/SAG mill - ball mill circuits, MASc thesis, University of British Columbia, 2013,
https://circle.ubc.ca/handle/2429/43896
Weir Minerals KHD HPGR Brochure (English).
Weir Expands in Comminution Market (2013). Womp 2013 Vol 6, www.womp-int.com.

24 | P a g e