Wear 434-435 (2019) 202986

Contents lists available at ScienceDirect

Wear
journal homepage: www.elsevier.com/locate/wear

Development of design criteria for reducing wear in iron ore transfer chutes T
Dusan Ilic
School of Engineering, Newcastle Institute for Energy and Resources, The University of Newcastle, Australia

A R T I C LE I N FO A B S T R A C T

Keywords: In high throughput Australian iron ore handling operations, premature failure of conveyor belts and transfer
Transfer chute wear chutes due to wear is of great significance to the success of operations. Here, a down time in the order of a few
Conveyor belt wear hours may result in millions of tonnes of annual export losses. Optimisation of existing or brownfield high
Iron ore capacity and utilisation mine, plant and port facilities with minimial capital spend has not always reduced
Design criteria
maintenance costs. Gauging the success of such endevours allows for the establishment of design standards that
are becoming accepted by the iron ore industry. The aim of these standards is to improve functionality and
maximise the service life of installed equipment without compromising flow performance.
This paper outlines key criteria for design standards, with a focus on minimising conveyor transfer and belt
wear, that can be integrated into an existing frawework of engineering management in iron ore operations.
Research and industry experience are blended with a theoretical (continuum mechancis based) and numerical
(Discrete Element Method based) modelling assessment. A qualitative modelling case study is presented in-
vestigating the sensitivities associated with transfer chute design geometry and the resulting chute and belt wear
life. An outline of the elements for consideration towards design criteria specification in iron ore transfers is
presented. Such specifications must result in adequacy of the designed transfer to meet both technical and
functional requirements across the range of operational conditions and variation of the characteristics of the ores
handled.

1. Introduction at the boundary. Additional considerations that affect the performance

and adequacy of transfer chute designs include operational require-
Successful performance of transfer chutes in belt conveying systems ments to incorporate water addition via spray bars, blocked chute de-
requires adequate design to allow re-direction of the primary flow tectors, maintenance access hatches and platforms, sampling equip-
stream via an impact plate deflector or curved hood, and a loading ment, specific loading requirements (i.e. loading onto multiple
chute or curved spoon to accelerate flow in the direction of the out- conveyors), structural elements and other pre-existing geometrical
going conveyor belt travel. This includes accounting for fine particles limitations.
remaining on the belt following discharge, which are scraped off the In high capacity iron ore operations, high wear rates of chutes and
belt by the belt cleaners. belts can significantly contribute to plant operating costs, reaching in
In Australia, the inclination of transfer chute walls when handling excess of AUD 1 million annually per conveyor. Unanticipated pre-
iron ore is typically at minimum angles of 65–70° to the horizontal. An mature failure of equipment due to wear may result in non-optimal ad-
important reason for this categorical value widely accepted by industry hoc modifications or “patching” of the chutes just to remain operational
is to prevent build-up and minimise the propensity for chute blockage until planned maintenance shutdowns which typically occur every
and downtime during handling highly cohesive and adhesive ores. An 12–16 weeks. Losses in production from undertaking unplanned
example of this is below water table ores, containing high moisture and maintenance work can have a significant impact on the profitability of
clay. This inclination is closely related to the draw-down angle or the the operation globally, in the order of millions of lost exported tonnes of
effective angle of internal friction of iron ore, which at 70° is at the product per annum. Further to this, an underestimate of the severity of
upper end of that handled in practice. In some instances, such as when the chute damage prior to commencing maintenance work can result in
handling coarse ore where build-up is unlikely, chute inclination may unsatisfactory outcomes which in turn could present the requirement
be reduced to as low as 35–45° to the horizontal. Lower chute in- for more unplanned and ad-hoc maintenance. Such a vicious cycle of
clinations typically exhibited by soft loading spoons in the coal industry both of these outcomes can cost the operations significantly through
rely on failure between the flowing material stream and the chute wall significant losses in production. This maintenance also poses a

E-mail address: dusan.ilic@newcastle.edu.au.

https://doi.org/10.1016/j.wear.2019.202986
Received 18 December 2018; Received in revised form 19 July 2019; Accepted 21 July 2019
Available online 22 July 2019
0043-1648/ Crown Copyright © 2019 Published by Elsevier B.V. All rights reserved.
D. Ilic Wear 434-435 (2019) 202986

heightened level of risk to workers as modifications typically occur in 1.2. Applied aspects of wear in the context of this study
confined spaces.
During handling and transportation of ores in belt conveying sys-
1.1. Theoretical aspects of wear in the context of this study tems, wear arises due to impact or erosion associated with a normal
load and abrasion or shear associated with differential sliding velocity
The theory to describe the mechanisms of wear of materials is under a normal load. Generally, there is uncertainty regarding if impact
complex, being studied extensively by many authors for more than a and abrasive wear act at the same time and in equal proportions or act
century. Reye [1,2] hypothesised that wear due to a load or force oc- as individual predominant mechanisms that lead to premature failure.
curring at the interface between the two contacting surfaces is pro- A number of authors have investigated factors influencing industrial
portional to the pressure applied. He also considered that the volume of transfer chute and conveyor belt wear-life in terms of impact and
removed material is proportional to the dissipated energy within the abrasive wear.
wearing material, resulting from relative motion between the two Roberts and Wiche [9], and Wiche et al. [10] developed linear and
contacting surfaces. circular abrasive wear testers for bulk solid materials. Both methods
Archard [3,4] interrogated contact and rubbing of flat surfaces. He involve a moving granular stream relative to a suspended load.
concluded that the number and size of contact areas increase with the Knowing the density of the lining samples, wear life prediction occurs
load acting, and the rate of mechanical wear is proportional to the load through monitoring the weight reduction rate over time. Chen et al.
but independent of the model used to represent the surfaces or the [11] applied the circular abrasive wear tester approach to iron ore and
apparent area of contact. The result was a definition of a relationship found a relationship between wear rate and hardness.
between the contact area during wear and the removed volume through On the other hand, Andrejiova et al. [12] and Molnar et al. [13]
a specific distance of sliding, with removed volume being proportional investigated impact and abrasive wear damage of industrial conveyor
to the applied load and hardness of the softer material. belts respectively. Andrejiova et al. [12] concluded that the condition of
Finne [5] deduced that surface wear due to solid particle erosion the belt, the type of falling material and the drop height significantly
depends on the number, direction and velocity of particles and the influence the risk of conveyor belt damage. Molnar et al. [13] corre-
behaviour of the surface when impacted. Importantly, wear due to lated properties of conveyor belts in abrasive wear and identified fa-
erosion is difficult to predict due to difficulty in determining the con- tigue related 3-body wear mechanisms better corresponding to the
ditions under which particles affect the surface. A method to predict the damage observed in real operating conditions.
erosion of ductile materials based on the angle of impact and the ve- In addition to physical testing, the current approach in practice is to
locity dependence was proposed. estimate life of conveyor belts by extrapolating the thickness worn per
Hutchings [6] provided two approaches to reducing erosive wear by unit of throughput, as documented in the study by Webb et al. [14]. The
flowing particulates, namely altering the flow conditions and selecting study identified a number of related factors that include belt thickness,
most suited materials. He concluded that velocity has the strongest grade, strength, width, operational throughput, product type, run time,
influence on wear rate and that reducing mean flow velocity in the belt length, speed, conveyor type and transfer height. From a modelling
system and avoiding local regions of high velocity flow are important perspective, Donohue et al. [15] showed that site data obtained through
considerations. belt thickness/gauge measurements can validate some of these wear
In his principles of abrasive wear, Khruschov [7] proposed that factors during handling of iron ore such as impingement angle and belt
under an established wear regime, linear wear is directly proportional wear.
to the stress and the friction path, with wear intensity rising approxi- On the other hand, difficulty associated with quantifying the in-
mately in direct proportion to the abrasive grain size, up to a maximum fluence of the installed design on overall operational life of the equip-
value. Other aspects for consideration included grain properties as well ment increases due to the inability to accurately calculate the relevant
as any crushing of the grains and impregnation of the worn material by proportions of different wear types. There is no doubt that important
the grains. Khruschov [7] also cited impact energy as an important considerations include angle of impingement; area of contact and im-
factor when considering impact abrasive wear as well as the interac- pact; particle shape and size; moisture, surface asperities; friction;
tions between the abrasive and the wearing surface. hardness; temperature; thermal conductivity and diffusivity. Other re-
While defining wear as the loss or transfer of material during sliding search that is also of high relevance includes microscopic level mole-
contact of surfaces, with the rate of wear dependent on bearing pres- cular changes of the intrinsic behaviour of chute and conveyor belt
sure, sliding velocity, material properties and surface geometry, Ashby materials as well as the particles and their relationship to the exhibited
and Lim [8] stated that a single equation defining such a relationship is wear mechanisms. For the purpose of this paper, these factors, much
insufficient. Different wear mechanisms depend differently on con- like the simulated particle properties themselves, are microscopic. In
trolling factors, with the dominant mechanism at any given set of contrast, conditions during ore flow in transfers, namely velocity and
conditions, leading to the fastest wear rate. Some of these mechanisms pressure, govern the suitability, and provide a means for differentiation
include melting, chemical change due to frictional heating, low-tem- and optimisation of transfer designs at a macroscopic level.
perature plasticity and brittle fracture. Ashby and Lim [8] proposed The study in this paper builds on and advances the development of
physical modelling as a means of identifying individual mechanisms engineering design criteria previously presented in Ilic and Hicks [16]
and their connectivity. through an assessment of modelling wear conditions in iron ore trans-
Rather than focus on a specific individual wear mechanism, the fers. The concepts presented aim to provide the necessary information
study in this paper develops necessary information to increase under- to drive a step change in the performance of transfers through gen-
standing of the interrelationship between different design targets in eralisation of macroscopic parametric relationships. Other microscopic
order to make decisions to better control conditions likely to result in based research, such as those associated with better describing the
increased wear of transfer chute and conveyor belt equipment. The prevailing wear mechanisms and test methods to select most favourable
aspects investigated are the conditions of wear namely velocity and the chute lining and conveyor belt materials will follow and complement
resulting pressure. These properties, the designer has the greatest these findings in future.
ability to influence through qualitative modelling and optimisation
regularly practiced in the iron ore industry. The approach of controlling 1.3. Traditional design guidelines and practical application
velocity as a means of controlling wear is consistent with previous
studies of Reye [1], Archard [4], Finnie [5], Hutchings [6], Khruschov Numerous examples exist in industry of transfer chutes optimised to
[7] and Ashby and Lim [8]. handle a broad range of material types while complying with

2
D. Ilic Wear 434-435 (2019) 202986

operational and geometric requirements. Inevitably, this occurs at the particular focus on steep inclined chutes. Minimum standards for ef-
expense of performance of the chute or belt when handling the process fective criteria in the design of transfer chutes in iron ore operations,
specific material type. For this reason, designing transfer chutes ac- with a specific focus on minimising the likelihood of operational delays
cording to criteria based on performance is required. Along with the and downtime due to wear are presented.
guidelines for optimal flow in transfer chutes based on original work of
Roberts [17], summarised below are practicalities that typically ac- 2. Methodology
company industrial application:
A number of researchers have applied modelling techniques de-
• Transfer head height and geometry shall satisfy operational re- scribed in this paper to transfer chute design and optimisation including
quirements. In the iron ore industry, there is a general requirement Katterfeld et al. [18], Mascarenhas et al. [19], Kessler and Prenner [20].
to satisfy minimum chute inclination of 65–70°; Similarly, Ilic [21,22] applied continuum mechanics and Discrete Ele-
• The most favourable wall lining material shall have low friction and ment Modelling (DEM) to analyse transfer chute interactions in high
high wear resistance, confirmed through flow property testing of the throughput systems. In the continuum mechanics approach, physical
ore handled. Wear liners can be extremely expensive and flow bulk solid properties, based on the work of Roberts [23], are used. Here,
property testing may not always be possible. the macroscopic parameters include compressibility or bulk density,
• The angle of impact between the material stream and transfer chute angle of repose, wall friction angle, particle size distribution and in-
geometry shall be minimised to maintain accelerated flow condi- ternal angle of friction. Typical application of this method is accelerated
tions. Minimising the angle of impact is much more favourable in flow conditions that are characterised by shallow bed depths with
smooth wall liner transfer designs compared to those that contain limited variation of the velocity across the bed.
ledges and billets (e.g. rock-box); From the work documented in Roberts [17] and Ilic [21,22], the
• To minimise non-central loading, belt mis-tracking and spillage, the equations used in the analyses of Section 3 are presented. In a straight
flow shall be symmetric and centralised onto the outgoing conveyor. inclined chute, by assuming constant equivalent friction, μE , the velo-
As will be illustrated later in the paper, a rock-box transfer design city, V, of the material stream at a distance, s, down the incline of the
generally requires more optimisation to achieve central loading; chute is given by:
• Sufficient inclination of the transfer chute walls to ensure flow at all
V= Vo2 + 2⋅a⋅s (1)
handled operational throughputs and characteristics of the ore. Flow
retardation must be minimised and blockage prevented. Practically, Here acceleration a = g⋅(cosθ − μE ⋅sinθ) and θ is the inclination of
the chute walls must have a minimum angle of inclination and the chute wall or flowing surface taken from the vertical. When con-
therefore a minimum head height for each transfer must also be sidering flow by internal shear which is exhibited with chutes that
considered; contain ledges or rock-box type chutes, the equivalent friction, μE may
• The inline component of the ore stream velocity loading the out- be directly related to the effective angle of internal friction, δ , of the
going conveyor must match, as close as possible, the corresponding bulk solid material handled. The relationship that describes this is:
velocity of the outgoing conveyor belt. The analysis presented later
will show that this is contrary to favourable wear outcomes when μE = sinδ (2)
designing for steeply inclined transfer chutes due to the excessive Ilic et al. [24] recently showed that this approach is also reflective
velocities required to achieve this; and, of interactions during shearing related flow with stacker transfers.
• The normal or perpendicular component of the ore stream velocity When smooth liners are installed, μE is directly related to the tan-
at the point of loading onto the outgoing belt shall be minimised in gent of the wall friction angle, φw . In areas of free-fall, neglecting air
order to reduce wear due to impact. Practically, minimising velocity resistance, the material stream velocity at a drop height, hf, may be
requires specification of a maximum head height limit for the calculated by:
transfer. If the focus of the design is to match belt speed this is
difficult to achieve. Transfer orientation also further conflicts with V= Vo2 + 2⋅g⋅hf (3)
this and other design guidelines such as chute minimum inclination
angle. where g is the gravitational acceleration. The impact pressure at
loading, Pvi, calculated by:
After a quick review of the above guidelines, it is evident that cer- Pvi = ρ ⋅V2N (4)
tain design objectives must be priorities over others in order to optimise
chute designs to meet specific performance criteria. Of high relevance where, ρ, is the bulk density of the bulk solid material and, VN, is the
to the analysis and concepts discussed in this paper, in view of chutes normal component of velocity at the point of loading on the outgoing
designed with steep inclinations, is the final point, which describes a conveyor. The wear indicating parameter, Wa, expresses the rate of
clear conflict between minimising normal velocity and matching belt abrasive wear on the outgoing belt. It is calculated by the following
speed. It is also very important that prioritisation of design objectives is equation:
referred to and supported by bulk solid characteristics, verified through Wa = μ 1⋅ ρ⋅ V2N⋅ (VB − VL) (5)
laboratory tests and operational requirements specific to the applica-
tion at hand. where, μ1, is the coefficient of friction between the bulk solid material
Having contradictory and inadequate design specifications in the and the outgoing conveyor belt, VB, is the outgoing belt speed and, VL,
early stages of the project life cycle reduces control and the ability to is the inline component of the loading velocity of the discharging bulk
influence final design outcomes, and significantly increases the costs of solid material stream in the direction of the outgoing conveyor belt
doing so during the latter stages of the project. This leads to a level of travel. Taking into consideration the inclination of the outgoing con-
criticality and/or risk to the business for not introducing performance veyor belt to the horizontal, α, the above Eqn. (5), can be re-expressed
based design specifications as early as possible. as a function of a non-dimensional wear factor, NWB:
This paper provides an evaluation of design criteria through an
Wa = μ1 ⋅ ρ ⋅ V3B ⋅ NW B (6)
evaluation of the influence of altering flow conditions and corre-
sponding velocity analyses of the loads acting on the flow and wearing where NWB is:
surface of transfer chutes and the outgoing belt. An interrogation of the
NW B = [(VR ⋅ A)2⋅(1 − VR⋅ B)] (7)
methodology in which transfer chute designs are specified follows, with

3
D. Ilic Wear 434-435 (2019) 202986

Table 1
Free fall and Transfer Chute Sliding Velocities.
Case Considered Initial Velocity (m/ Velocity (m/s) at Head Height of
s)
3m 7m 10 m

Free Fall 0 7.7 11.7 14.0

4 8.7 12.4 14.6
6 9.7 13.2 15.2
Chute Inclined at 45° 0 4.2 6.3 7.6
4 5.8 7.5 8.6
6 7.3 8.7 9.7
Chute Inclined at 60° 0 5.4 8.3 9.9
4 6.7 9.2 10.7
6 8.1 10.2 11.6
Chute Inclined at 70° 0 6.2 9.5 11.3
Fig. 1. Parameters at loading (α = 0°).
4 7.4 10.3 12.0
6 8.6 11.2 12.8
Here, VR, is the ratio of the bulk solid material stream velocity at the
exit of the chute or loading point, VE, to the outgoing belt speed, i.e. VE/
VB. The parameters A and B are related to the total inclination at the 3. Continuum mechanics investigation
loading point and are given by A = Sin (ψ + α) and B = Cos (ψ + α).
The cut-off angle, ψ, is the inclination angle to the horizontal of the To highlight the association between velocities within transfer head
chute at the exit. For a straight chute, ψ, is the inclination of the chute. heights, Table 1 presents the free fall and sliding velocities for a range
The relevant velocities and chute angle is depicted in Fig. 1, for a of commencing or initial flow velocities, according to Eqn. (1) to Eqn.
horizontal outgoing belt (i.e. α = 0°). An increase in α would result in (3). The assumption here is that sliding friction is equal to the internal
an increased belt inclination to the horizontal. angle of friction and the internal angle of friction is equal to the chute
In contrast, DEM simulations depend on the contact model used and inclination for each case considered. From the data, velocities of around
microscopic (particle level) parameters that include shear (Young's) 10.0 m/s and higher are evident in transfers with head heights ex-
modulus or bulk elastic stiffness, coefficient of restitution, damping, ceeding 3.0 m. It is of note that in high throughput iron ore operations
particle/particle, particle/boundary friction, Poisson's ratio, rolling conveyor belt speeds are typically over 4.0 m/s.
friction and particle density. Input parameters are typically calibrated Taking only the free fall data from Table 1 into consideration,
to physical bulk solid properties through bench scale tests to determine Table 2 below shows calculated velocities at chute exit, VE, required for
the angle of repose, draw-down angle, static/dynamic friction and ad- matching outgoing belt speed, VB, of 4.0 and 6.0 m/s with respect to
hesion. Due to computational limitations, high throughput transfers are chute inclination at exit, or cut-off angle, ψ. The results are presented in
usually modelled using particle sizes of 20–50 mm equivalent diameter terms of VE required to load the outgoing conveyor at 100% of belt
and a level of scale-up from bench-scale tests is often required. speed (i.e. matching belt speed such that VL = VB) and VE required for
Hustrulid [25] was one of the early pioneers in applying DEM to loading velocity to be 50% of belt speed (i.e. VL = 0.5⋅VB).
assess transfer chute and conveyor belt abrasive wear based on wear The results presented in Tables 1 and 2 suggest that it is not prac-
volume and calculation of lateral forces acting on the belt. Similarly, tical to match belt speed with chutes that have high angles of inclina-
Katterfeld [26] developed an approach to assess contact pressure and tion such as 60° and 70°, particularly in high belt speed applications.
friction power as a means of using DEM to model and predict chute Unfortunately, if such a design specification to match belt speed was
wear. This included quantitative comparison to site measurements part of the design optimisation, high wear rate of the chute, conveyor
through full-scale experimental tests of loads on the chute and belt. belt or both is likely.
While these studies developed DEM wear assessment methodology and According to Ilic and Hicks [16] (and later illustrated in Fig. 3),
field verification of applied methods, Nordell's [27] practical findings abrasive wear on the outgoing belt reaches a peak when the loading
reported optimisation of chute geometry to increase belt life by 300%. velocity, VL is in the range of 50 and 90% of the outgoing belt speed, VB
Such a significant increase further illustrates the influence of transfer presented later in the document. Similarly, an increase in the inclina-
chute design on possible costs of maintenance for repairs due to wear. tion of the chute, ψ, and the inclination of the outgoing belt, α, both
The reported outcomes also emphasise that maintaining full control of result in an increase in the wear of the outgoing belt. These trends
the design process is necessary to avoid accidently sacrificing perfor- suggest that, with such steep chutes, in order to minimise belt wear, the
mance of one aspect by optimising the design of another. velocity loading the outgoing belt must be equal or very close to the
The case study example presented in Section 4 uses Rocky 2.5 outgoing belt speed, which generally will never be achievable. The
(Granular Dynamics International) DEM software package. Here, the
normal force contact model is a partially latching or hysteretic linear
Table 2
spring, with no tangential damping, based on that originally developed
Chute exit velocity required for loading at 100% and 50% of outgoing belt
by Walton [28,29]. The model specifies a different contact stiffness speed.
during loading and unloading. The rolling friction implemented in the
simulations is a variable torque spring and dashpot model, based on the Outgoing Outgoing Belt VE Required for Loading VE Required for Loading
Belt Speed, Inclination, α at 100% VB Chute Cut- at 50% VB Chute Cut-Off
Model Type C as described in Ai et al. [30] and adapted by Wensrich VB (deg) Off Angle, ψ (deg) Angle, ψ (deg)
and Katterfeld [31]. Previously, Ilic [21,22] presented a comprehensive
description of the relationship between physical and DEM simulation 45 60 70 45 60 70
parameters used using these models. The analysis procedure follows
4.0 m/s 0 5.7 8 11.7 2.8 4 5.8
that previously described in Nordell [32], where two components, 5 6.2 9.5 15.5 3.1 4.7 7.7
abrasive (shear) and impact (normal) wear, evaluate belt and chute 10 7.0 11.7 23.0 3.5 5.8 11.5
performance a similar manner as that described in Aldrich and Zhang 6.0 m/s 0 8.5 12 17.5 4.2 6 8.8
[33]. 5 9.3 14.2 23.2 4.7 7.1 11.6
10 10.5 17.5 34.6 5.2 8.8 17.3

4
D. Ilic Wear 434-435 (2019) 202986

and belt wear.

4. Discrete element modelling case study

In this section, a DEM case study example demonstrates the influ-

ence of transfer chute design on the exhibited wear. Four different
design options investigated include:

• Option 1 - initial design located at the mine site, experiencing high

chute wear;
• Option 2 - design considered as a possible replacement, utilising an
optimised rock-box deflector;
• Option 3 - design highlighting the influence of number and location
of chute ledges;
Fig. 2. Non-Dimensional Wear Factor, NWB, Compare to Chute inclination of • Option 4 - design highlighting the influence of chute geometry.
45° (base case).
While atypical of Australian iron ore operations, the transfer head
height of approximately 11.0 m (typical head heights are in the order of
results also suggest that a much slower loading velocity, in the order of
7.0 m) and the generally high corresponding velocities, accentuate the
40–50% of VB, or even lower, is necessary to minimise wear.
wear of the chute and belt.
From a slightly different perspective, to minimise belt wear for a
chute inclination of 70° the velocity of the material stream at the exit,
4.1. Transfer designs investigated
VE must either be lower than the outgoing belt speed, or at least ap-
proximately three times greater than the outgoing belt speed, Ilic and
Fig. 4 and Fig. 5 show the four different designs simulated, of a 90°
Hicks [16]. Compared to a chute inclined at 45° there is a significant
transfer (plan view). The design adjustments included:
increase in the non-dimensional wear parameter, NWB, as shown in
Fig. 2 for a range of outgoing belt inclinations. A wear increase in the
- Reducing chute inclination and incorporating a small loading chute
order of 3 times for a chute inclined at 60° and more than 10 times for a
at the bottom;
chute inclined at 70° loading an outgoing belt inclined up at 5° (to the
- Replacing the curved deflector with a rock-box deflector (shown in
horizontal) is exhibited.
blue in Fig. 4);
For a steep inclined chute, as the loading velocity, VL, approaches
- Reducing the number of ledges in the chute; and
outgoing belt speed, VB, the corresponding normal component of ve-
- Substituting the straight chute with a curved one.
locity, VN, approaches a maximum. Fig. 3 illustrates the relationship
between wear and velocity for a typical iron ore transfer chute inclined
Option 1 is the existing transfer design, incorporating a curved de-
at 70°. Wear is broken up into components associated with the chute
flector, where the vertical distance from the bottom ledge of the ex-
and belt, both for impact and abrasive wear. The figure demonstrates
isting chute to the carry side or top surface of the outgoing belt is ap-
that in order to minimise abrasive belt wear through loading at belt
prox. 1000 mm.
speed, the high velocities will result in high impact pressures on both
Option 2 and Option 3 are modified designs where the rock-box
the belt and the chute. This strongly suggests that it is not favourable to
deflector replaces the curved deflector and the chute has a reduced
optimise designs for this criterion for transfer chutes of steep inclination
inclination of 60°. Here, the vertical distance from the bottom ledge to
and increasing transfer height to achieve this will likely result in se-
the carry side of the outgoing belt is approx. 750 mm. Also integrated in
verely reduced wear life.
these two options is a 45° smooth-wall straight inclined loading chute,
In summary, a lower loading velocity, and effectively transfer head
located between the final ledge and the outgoing belt. The main dif-
height, will lead to a reduction in wear on the outgoing belt. When
ference between the options is that Option 3 has most of the ledges in
reducing wear through velocity, it is important to ensure there is ade-
the chute removed, apart from one, located approximately in the
quate material stream velocity at chute exit, VE to maintain reliable
middle, between the deflector up top and the bottom ledge and loading
flow, thereby preventing blockage due to any possible build-up. This
chute. Finally, in Option 4, a smooth wall curved spoon replaces the
area corresponds to the low chute and belt wear zone indicated in
chute section of Option 2 and Option 3, with a cut-off angle of 45° and
Fig. 3. Controlling the stream velocity to optimise both the conveyor
unchanged width. In practice, both the straight and curved loading
belt and chute performance is required to produce a balance in chute
chutes would comprise of ceramic alumina wear tiles in a rubber matrix
on a mild steel backing plate. The vertical distance between the bottom
of the 45° straight loading chute of Option 2 and Option 3, the spoon of
Option 4 and the carry side of the outgoing belt is approx. 150 mm.
Table 3 below presents a summary of the design variations investigated.

4.2. Material properties and modelling parameters

Information obtained from the belt manufacturer and mine site

operations shows the incoming conveyor rubber belt is 1.5 m wide,
ST1000 (belt modulus 72,000 kN/m) ARPM Grade I, 21.6 mm thick,
with top/bottom cover thickness’ of 13 mm/6 mm respectively and has
a mass of 62.6 kg/m. The nominal steel cord diameter and approximate
pitch are 3.6 mm and 13.9 mm with a weight of 7.4 kg/m2, having
minimum ultimate and operating tensions of 1000 kN/m and 150 kN/m
Fig. 3. Belt Abrasive and Impact Wear Relationship with Velocity, adapted respectively. Similarly, the outgoing conveyor rubber belt is 1.8 m
from Ilic and Hicks [16]. wide, ST2000 (belt modulus 14,400 kN/m) RMA Grade I, 23.4 mm

5
D. Ilic Wear 434-435 (2019) 202986

Fig. 4. Transfer Design Variations Option 1, Option 2, Option 3 and Option 4 (from left to right).

thick with similar top and bottom cover thickness and mass. Nominal Table 3
cord diameter and approximate pitch are 5.2 mm and 13.5 mm and Design variations investigated and description.
weigh 12.6 kg/m2. The minimum ultimate and operating tensions are Design Description
2000 kN/m and 300 kN/m respectively. Both belts have a Shore “A”
hardness of 55–65. Option 1 Curved Deflector, 70° Chute
The Option 1 design shown in Fig. 4 on the left consisted of a curved Option 2 Rock-box Deflector, 60° Chute, 45° Straight Loading Chute
Option 3 Same as Option 2 But Only One Ledge in The Middle
deflector (green) and ledged straight loading chute inclined at 70° to Option 4 Rock-box Deflector, Curved Loading Spoon, 45° Cut-off
the horizontal. The curved deflector comprised of 98% ceramic alumina
tile wear-liners, with a density 3.65 kg/m3 and minimum hardness of 62
HRC. Upon operational commencement, the loading chute did not The incoming and outgoing conveyors operate at speeds of 4.3 m/s
contain ledges but comprised of high-density 90% alumina ceramic tiles and 3.6 m/s respectively at a throughput of 5000 t/h. The product
having flexural and compressive strengths of 280 MPa and 2900 MPa, handled is screened iron ore lump material with a nominal particle size
moulded in a rubber matrix and fixed to mild steel backing plate. The distribution that ranges from 6.3 to 32 mm. Spherical particles
tiles are vulcanised in a high-tensile rubber matrix exhibiting minimum 16–32 mm in diameter were modelled using DEM simulations for a
tensile strength of 27 MPa and 60 Shore A hardness. In other areas duration of 50.0 s. The parameters assumed in the study are sum-
exposed to the flow of the ore stream, the chute consisted of chromium marised in Table 4. The computer used in the study was an Intel® Core™
carbide rich overlay (minimum of 60% M7C3) bonded to mild steel i7-7800X CPU @3.5 GHz, 32 GB RAM, with Nvidia GeForce GTX 1080
backing plate. The wear life of the curved deflector was approximately Ti.
11 weeks.
Due to extremely poor wear life of the loading chute, in the order of
approximately only 2 weeks, observed during initial commissioning 4.3. Selection of modelling parameters
stages, modification of the design was completed and ledges retrofitted.
The ledges comprise of rows of 300 × 100 × 150 mm of Ni-Hard 4 Typically, calibration of modelling parameters is necessary to pro-
billets, bolted to mild steel frames. The billets are an alloy of nickel, vide accurate and reliable simulations. In this process, a set of micro-
chromium and white cast iron, designation Ni-HiCr, Class I, type D (as scopic parameters that represent physical bulk solid behaviour are
per ASTM532-2014) with a minimum hardness of 600 HB. This specified through bench scale tests such as angle of repose, slump plane
brownfield design change resulted in chute wear life increasing to ap- and inclining plate wall friction/adhesion, Ilic et al. [24]. Other, more
proximately 9 weeks. The reason for the increase was a shift in the detailed, application and performance criteria specific calibration
mode of flow or failure. With the initial smooth liner design, the stream methods may include site measurements and observations, 3D laser
of fast moving abrasive ore wears the chute surface through direct scanning of worn surfaces and adhesive build-up, belt top cover
contact during sliding along the boundary, or boundary failure. thickness measurements and other similar or related downtime and
Conversely, the ledges act to restrain the moving stream and form a maintenance information. The calibration test used in this study was
granular ore surface of particles. The flowing stream shifts to flowing the slump plane test, previously detailed by Ilic [22]. The method in-
over itself instead leading to internal failure, thereby preserving the life volves filling an empty box and then emptying it by removing one of the
of the chute. This is typical of rock-box or ledged transfers and is the sidewalls to reveal a slump plane or draw-down angle once the particles
main distinction to flow over smooth liners. settle.
The particle frictional parameters, Pmts 1, previously summarised

Fig. 5. Loading chute design variations.

6
D. Ilic Wear 434-435 (2019) 202986

Table 4
Modelling parameters.
Parameter Value

Throughput (t/h) 5000

Incoming Belt Speed (m/s) 4.3
Outgoing Belt Speed (m/S) 3.6
Incoming Conveyor Belt Width (mm) 1500
Outgoing Conveyor Belt Width (mm) 1800
Particle Bulk Elastic/Loading Stiffness, E (N/m2) 1e7
Boundary Loading Stiffness (Chute and Belt) (N/m2) 1e11
Poisson's Ratio, ν 0.3
Coefficient of Restitution, εres 0.3
Particle Size Distribution
Particle Diameter 100% Finer (mm) 32
Particle Diameter 30% Finer (mm) 24
Particle Diameter 15% Finer (mm) 16
Fig. 6. Simulated Draw-down Angle, or Effective Angle of internal Friction, δ
Simulation Total Number of Particles 3,654,480
Simulation Time (s) 50
Timestep (s) 2.2e-5 shapes modelled in the investigation are smooth rounded polyhedrons
Particle Friction Coefficient, μp 0.7
Rolling Friction Coefficient, μr 0.4
and have a polygonality factor of six, with the particle horizontal and
Adhesive Distance sadh (m) 0 vertical aspect ratios of 1 and 1.5 respectively. All other particle
Adhesive Force Fraction, μadh 0 parameters remained constant, as described in Table 4. The time step
Chute Wall Boundary Friction Coefficient, μw 0.5 and number of simulated particles in this instance were 1.9e-5 s and
Belt Friction and Skirting Friction Coefficient, μw1,2 0.7
5,015,870 respectively. The computation time expensed was sig-
Particle Bulk Density, ρ (kg/m3) 1800
Particle Solids Density, ρs (kg/m3) 3273 nificantly longer, in the order of five times, compared to simulations
Solids Fraction 0.55 using spherical particles.
Chute Boundary Density (kg/m3) 7850
Belt Boundary Density (kg/m3) 1400
4.4. DEM case study modelling results - Pmts1

in Table 4 as μp and μr, were selected based on information from site Fig. 7 shows an isometric view typical of steady state flow through
operations and are reflective of an iron ore lump material exhibiting a each investigated transfer design. The same velocity colour scheme is
simulated draw-down angle in the range of 40° to 50°with no adhesion applicable in each figure, with blue indicating zero and red indicating
present between the particles (i.e. μadh and sadh are both zero). In 12.0 m/s and higher. Notably, the existing design, Option 1 exhibited
Australian iron ore operations, the draw-down angle may vary sig- much faster peak velocities. The results also clearly show that Option 2
nificantly between different products, from 45° to 70°, and is influenced and Option 3 exhibit slower stream velocity in the middle of the
by the particle size distribution, moisture and the presence of clays. In transfer.
addition to the parameters described in Table 4, a second set, Pmts 2 Furthermore, Fig. 8 shows a cross sectional close-up view of the
(μp = 0.9, μr = 0.6, μadh = 0.5, sadh = 0.0032 m), reflective of an in- material stream flowing through each chute and loading the outgoing
crease in the internal angle of friction, was also investigated. The belt. The particle vector colour scheme is unchanged. The results show
parameters were guided by both practical experience and work pre- reduced variation in the stream loading direction for Option 2 and
viously documented in Ilic [22] and Ilic et al. [24]. Option 4. Steepest loading angle is be observed with Option 1 (almost
Calibration involved simulating the slump plane test according to perpendicular to the outgoing belt travel direction).
the procedure described in Ilic [22] to reflect the range obtained from For the different options investigated, Fig. 9 shows a 0.2 m high
site functional process information describing the physical behaviour of slice through the cross-section of the discharging material stream. The
the ore handled in practice. Notably, boxes of 300 mm, 550 mm, slice is located at the exit of the chute, corresponding to the interface
1000 mm and 1500 mm size were filled with the modelled particles and immediately above the top of the skirt board arrangement. Note that
one side wall released following which the steady state draw-down the particle velocity colour scheme shown is the same as that presented
angle after the material comes to rest to the horizontal is measured. As previously but particles having velocity below 1.0 m/s are omitted to
described in Ilic [22] and Ilic et al. [24], this angle is indicative of the focus on the flowing stream. Results show slowest velocity, with biggest
physical parameter, the bulk solid effective angle of internal friction, δ. stream cross sectional area for Option 3. Option 4 exhibited fastest
The pressure due to the height of the mass was calculated in the same velocity, with the stream cross-section changing orientation, from
manner as hydrostatic pressure (i.e. ρ x g x height of mass). In each case, parallel to perpendicular with the outgoing belt.
the height of the mass was equal to the height of the box. The simulated In addition to the above, Fig. 10 shows images of the burden cross-
draw-down angle relationship to height of mass obtained from cali- section (1.0 m slice along the belt travel direction) on the outgoing belt,
bration tests, for each of the sets of parameters modelled, are sum- following the skirt board arrangement at steady state. Option 1 and
marised in the plot of Fig. 6. Option 2 show predominantly centralised burden, however, Option 3
The particle size distribution modelled, 16–32 mm, is representative and Option 4 show a shift of the burden towards the left or right hand
of that handled in practice, 6–32 mm. In the simulations, the 16 mm side respectively. Typically, non-central loading in practice will result
particles represent all particles that have an equivalent diameter of in the conveyor belt self-compensating by moving laterally. These re-
16 mm and below. sults illustrate the influence of the deflector design and position on the
To illustrate future research, also provided is a brief investigation centrality of loading. The rock-box deflector is optimised to integrate
into the influence of particle shape. Particle shape considerably pro- with Option 2, however, the burden profiles in Fig. 10 show that the
longs the computational time expensed to evaluate transfer perfor- existing position, and/or design of the deflector requires adjustment for
mance. In some instances, such as where the transfer geometry is in the Option 3 and Option 4. Centrality of loading is extremely important as
order of three times the maximum particle size handled, particle shape it minimises the probability of the belt misaligning which can lead to
is of significant importance to accurately model flow. Based on previous spillage and wear of the surrounding fixed structure.
practical experience, and as described in Chen et al. [34], the particle With focus on the chute and loading of the outgoing belt, Fig. 11 to

7
D. Ilic Wear 434-435 (2019) 202986

Fig. 7. Flow through each Transfer, Option 1 to Option 4 (Left to Right).

Fig. 14 show wear contours from the DEM analyses for Option 1 to material stream flowing through the chute formed part of further in-
Option 4, with red indicating high wear and blue indicating no wear. A vestigation. Considered was a thin, 0.2 m slice of the stream, at a
colour scheme of maximum magnitude of 1500 J/s applies in all in- number of sections. The data are summarised in the plots of Fig. 16,
stances. The expressions for impact and shear wear intensities are: relative to the vertical distance below the centre of the incoming con-
t1 veyor head pulley. The bottom of the curved and rock-box deflectors
Impact Intensity ∫ (FN ⋅δN ) dt correspond to 2.0 m and 4.5 m below the incoming conveyor head
t (8) pulley respectively. The billets experiencing greatest wear, shown
previously in Fig. 15, are located 4.0–5.0 m below the head pulley.
t2
Results show average stream velocities peak at a distance of ap-
Shear Intensity ∫ (μ⋅ FN ⋅δs) dt proximately 7.5–9.0 m below the incoming conveyor head pulley.
t (9)
Option 1 exhibits the highest maximum peak velocities of over 13.0 m/
where, normal force, FN = KN ⋅δN , normal stiffness, KN = E⋅L , s. The curved chute design of Option 4 exhibits the maximum average
L = particle size (m), δN = normal overlap (m), δs = the tangential or peak velocity at around 11.0 m/s. Due to a reduction in velocity and
shear overlap, μ = coefficient of friction during sliding (i.e. particle/ zones of concentrated localised impact, chute wear is likely to reduce
chute or particle/belt friction), t = start time of the loading cycle (s), with Option 2 and Option 3. This corresponds well with a reduction in
t1 = end time of the loading cycle (s), and t2 = end time of the un- red areas shown in the impact wear contour in Fig. 12 for Option 2 to
loading cycle (s). Both the tangential and shear overlap depend on the Fig. 14 for Option 4. The velocity plots in Fig. 16 show that a reduction
stiffness and the total force acting at the point of contact and for every in maximum velocity of 3.0–5.0 m/s is achievable through making
boundary, the values over all contacts are summed and calculated per appropriate modifications to the chute design.
unit time. The stream velocity at the exit of each chute, corresponding to the
The results show Option 2 and Option 4 exhibit lowest chute impact same slice shown previously in Fig. 9, and at loading on the outgoing
wear. The greatest chute surface area exposed to shear related wear belt, were assessed and are presented in Fig. 17. The results are sum-
appears with Option 3 and Option 4. Interestingly, the shear wear of the marised in Fig. 19 on the left. Furthermore, the shear and impact in-
45° loading chute reduces by removing the ledges from Option 2 shown tensity data of the outgoing conveyor belt for each option during the
in Fig. 12 to Option 3 shown in Fig. 13. Belt impact wear appears length of the simulation are shown in Fig. 18 and are summarised in
highest in Option 1 and Option 4 as can be observed from the left image Fig. 19 on the right. Note that the average results presented in Fig. 19
in Figs. 11 and 14 respectively. Similarly, the shear wear of the out- correspond to a period of 20.0s of steady state towards the end of the
going belt is lowest with Option 4. simulation.
In practice, the upper ledges of the chute, just below the curved The results show that maximum velocities, corresponding to max-
deflector in Option 1 and previously shown in red in Fig. 4, exhibited a imum impact related wear occur during initial stages of loading, over a
high wear rate. Fig. 15 shows the wear pattern observed during a period of a few seconds. Option 1 exhibits highest peak velocity of
maintenance shutdown, following 12 weeks of operation. As previously 13.4 m/s and Option 3 exhibits the lowest average velocity of around
reported in Ilic and Hicks [16], the outer edges of the falling stream in 2.5 m/s. Option 1 (4.9 m/s), followed by Option 4 (4.5 m/s) exhibit the
this zone continuously impact with only a localised exposed surface highest average loading velocities. Both are higher than the outgoing
area of the billets (0.02–0.03 m2) at high velocities of 9.0 m/s and belt speed of 3.6 m/s. Option 2, incorporating a rock-box deflector, a
above. reduced main chute inclination of 60° and a loading chute inclined at
The average and maximum peak velocities (total magnitude) of the

Fig. 8. Loading chute flow, Option 1 to Option 4 (left

to right).

8
D. Ilic Wear 434-435 (2019) 202986

Fig. 9. Cross section at chute exit, Option 1 to Option

4 (left to right).

45° results in a loading velocity of 3.2 m/s, which is closest to the

outgoing conveyor belt speed. The values reported are the total mag-
nitude.
Referring to the summary of wear data shown in Fig. 19 (right),
Option 1 exhibits the highest impact intensity and Option 2 the lowest
overall impact. Option 1 exhibits the highest while Option 4 exhibited
the lowest shear intensity. These results also show that reducing the
chute cut-off angle and velocity may not necessarily reduce the shear
wear on the outgoing belt, and is evident by comparing Option 1 and
Option 3. Similarly, the shear wear on the outgoing belt reduces with
cut-off angle for a concentrated and accelerated flow as shown by the Fig. 11. Option 1 wear contours, impact (left) and shear (right).
curved loading spoon of Option 4.
The analysis presented does not easily allow absolute evaluation of
transfer performance, however, interrogating further based on a few
assumptions, the design modifications appear to yield improvements in
both chute and belt wear life. Based on the average velocity at both the
chute exit and at loading presented in Fig. 19, Eqn. (7) was used to
calculated the non-dimensional abrasive wear factor, NwB. Table 5
below presents this data, expressed relative to Option 1 as the base case.
The data shows a reduction in NwB and therefore wear on the belt, with
the lowest abrasive wear exhibited by Option 4. The data is in good
relative agreement with the shear wear intensity comparison presented
in Figs. 18 and 19.
Fig. 12. Option 2 wear contours, impact (left) and shear (right).
Furthermore, using the DEM wear intensity data, indicative values
of transfer performance were calculated and are summarised in Table 6.
The reported chute life performance was determined as the sum of the
square of the average peak velocities at a number of locations within
the transfer. These locations and values were as previously calculated
and presented in Fig. 16. Such a relationship is consistent with the basic
assumption discussed in the classical approaches of Reye [1], Archard
[4], Finnie [5], Hutchings [6], Khruschov [7] and Ashby and Lim [8] in
which wear is proportional to the applied load or in our case, square of
the velocity, V2. The belt life reported in Table 6 is determined based on
an assumed linear relationship between the average energy and the
resulting wear. In this instance, and somewhat arbitrarily, the as-
Fig. 13. Option 3 wear contours, impact (left) and shear (right).
sumption is that shear energy contributes 80%, 50% and 20% of the
total wear exhibited, with the remainder caused by impact in each
considered case respectively. The results show Option 2 as the most
favourable design for balanced improvement in both chute and belt life.
Option 4 illustrates how sacrificing the wear life of the chute can lead to
increasing the life of the belt.

4.5. Influence of parameter variation - Pmts 2

Investigation extended to include the influence of particle properties

on flow of ore with a higher simulated effective angle of internal fric-
tion according to Pmts 2, as described previously in Fig. 6. The velo-
cities at chute exit, loading and the wear intensities on the outgoing belt
Fig. 14. Option 4 wear contours, impact (left) and shear (right).
are summarised in Fig. 20. Table 7 presents the estimated effect on
chute and belt wear performance calculated based on these data.
The results show that with an increased internal angle of friction, not evident. Most pronounced is a decrease in impact wear with Option
there is generally a slight reduction in the maximum velocity; however, 1. More so, the summary of the estimated effect on chute and belt wear
overall, significant differences between the two sets of parameters are performance presented in Table 7 shows similar results to those

Fig. 10. Cross section on outgoing conveyor

belt, Option 1 to Option 4 (left to right).

9
D. Ilic Wear 434-435 (2019) 202986

4.7. Summary

The DEM modelling study has demonstrated that the performance of

both the chute and the belt can be improved through optimising the
design of the transfer chute. According to the investigation, it is clear
that, when compared to Option 1, all other options will result in im-
proved wear performance. While Option 3 shows improved perfor-
mance from the perspective of wear, the significant amount of slow
moving or dead material at chute exit may be of concern, particulalry if
material characteristics align with those reflective of a difficult to
handle ore with a high internal angle of friction. Notably, the risk of
blockage may increase. Option 3 also highlights the need for additional
guidelines regarding the optimisation of the number, size and spacing
between the ledges present within the transfer. Option 2 exhibits the
best balance of improved belt and chute wear performance for the
range of shear to impact wear proportional contribution assumed. The
wear modelling evaluation for Option 4 also provides an example of
prioritisation of belt life over chute life.
The variation of modelling parameters to represent a higher internal
angle of friction most notably showed a slight reduction in impact wear
for Option 1. However, overall comparison of the chute and belt wear
life performance of all four design options, for initial parameters
modelled (Pmts 1) and parameters reflective of a higher internal angle
Fig. 15. Observed billet wear on site after 12 weeks of operation. of friction (Pmts 2), shows similar outcomes. Variation in modelled
particle shape showed generally similar velocities through the transfer,
however, localised loads may result in slightly different conclusions
presented previously in Table 6 for Pmts 1. Based on the chute exit and
regarding wear performance, notably significant variation was ob-
loading velocities in Fig. 21, the Non-Dimensional Wear Factor, NWB,
served for impact wear calculations.
also showed little variation with increased internal angle of friction.
5. Development of design criteria

4.6. Influence of particle shape Based on extensive previous industry experience and the content
presented within this study, it is evident that there are significant
Further to the variation in simulated internal angle of friction, also
benefits to introducing performance-based criteria and evaluating de-
briefly investigated was the effect of non-spherical particle shape on
signs during early project stages. At this time, the cost of design mod-
Option 1 transfer performance. Fig. 21 and Fig. 22 present the results
ifications is low but of very high influence on the likelihood of the
for smooth, rounded polyhedrons and particle frictional parameters
project meeting anticipated performance targets. In addition to any
consistent with Pmts 1. As can be observed from the results presented in
existing functional or site explicit requirements, it is suggested that
Fig. 21 (left), the chute exit and loading velocities, and in general, the
design criteria for iron ore transfers with respect to wear performance
burden through the transfer including on the outgoing belt, shown in
should include specification regarding:
Fig. 22 did not show significant variation. However, the calculated
maximum impact wear intensity did present significant variation, as
shown in Fig. 21 (right). Without further insight or investigation, this • Minimum acceptable technical supporting documentation to ac-
company designs and hold points relevant to the stage of the project
appears related to the contacting area between the particles and the
life cycle;
boundaries (i.e. chute and belt).
• Minimum ore flow property experimental test work requirements;
As variation may present in localised areas, it is evident that addi-
tional research is required, however, importantly; the average results • Quantified central loading tolerance of outgoing conveyor including
deflector type and geometry;
indeed show similar trends. The variation in calculated velocities only
slightly influences the estimated chute and belt wear performance. This • Minimum inclination angles of the chute, at key locations through
the transfer (main stream and secondary stream), with selection
is summarised in Table 8.

Fig. 16. Average peak (top) and maximum peak (bottom) velocities within the transfer.

10
D. Ilic Wear 434-435 (2019) 202986

Fig. 17. Velocity at chute exit (left) and at loading on outgoing belt (right).

Fig. 18. Impact (left) and Shear (right) intensity on Outgoing Belt.

Fig. 19. Velocity at chute exit and loading (left), wear intensity on outgoing belt (right) - Pmts 1.

Table 5 Table 6
Relative Non-Dimensional Wear Factor, NWB (Calc. using Fig. 19 Velocities), - Estimated effect on chute and belt wear performance - Pmts 1.
Pmts 1.
Wear Relationship 80% Shear 20% 50% Shear 50% 20% Shear 80%
Relative NWB Impact Impact Impact

Design Chute Exit Loading Design Chute Life Belt Life Belt Life Belt Life
Option 1 1.00 1.00 Option 1 Base Base Base Base
Option 2 0.63 0.34 Option 2 132% 139% 151% 160%
Option 3 0.57 0.37 Option 3 136% 114% 137% 156%
Option 4 0.37 0.15 Option 4 106% 166% 154% 144%

supported by tested characteristics of the handled ore; variation in the velocity profile
• Material stream average and maximum velocities at key locations • Calibration of modelling parameters and verification of simulation
through the transfer and loading the outgoing belt; outcomes, including physical scale modelling and/or field mea-
• Location of maximum impact and shear loads or energy losses for all surements; and
chute components and the belt and indicative assessment of their • Selection of particle shape, size and distribution used in the mod-
magnitude (could be relative to a base case); elling and design optimisation.
• The ore stream magnitude and direction of velocity, and extent of

11
D. Ilic Wear 434-435 (2019) 202986

Fig. 20. Velocity at chute exit and loading (left), wear intensity on outgoing belt (right) - Pmts 2.

Table 7 6. Conclusions
Estimated effect on chute and belt wear performance - Pmts 2.
Wear Relationship 80% Shear 20% 50% Shear 50% 20% Shear 80%
Geometric constraints such as inclination and head height influence
Impact Impact Impact the performance of transfer chutes. In practice, the characteristics of the
iron ore product handled do not necessarily define transfer chute per-
Design Chute Life Belt Life Belt Life Belt Life formance criteria. Without an optimisation of the chute design based on
Option 1 Base Base Base Base
Option 2 134% 136% 145% 153%
the products handled, within the bounds of clearly defined chute and
Option 3 137% 95% 120% 143% conveyor belt performance, the flow of ore through the chute may be
Option 4 107% 167% 153% 140% compromised and result in unplanned delays that can significantly
disturb operations.
A continuum mechanics investigation interrogated the velocities
Considering the above, a distinction is required between greenfield during free fall and sliding for straight chutes inclined at medium to
(new) and brownfield (existing) operations. In greenfield applications, steep angles for a range of transfer head heights relevant to the iron ore
design criteria should include an element that addresses relatively easy industry. This included an analysis of the velocities required to achieve
and quick replacement or change-out, such as rotable or modularised a material stream loading velocity matching the speed of the outgoing
components. Here, material stream velocities must be minimised in belt. The investigation showed that it is impractical to design for this
view of chute and belt wear life that may not meet the planned regular criterion with chutes of steep inclination. Proposed is a generalised
maintenance shutdown intervals. In brownfield applications, re-design relationship between velocities, abrasive belt wear and impact wear,
should allow contingency for significant modifications, in case patch-up with zones of favourable loading highlighted.
type improvement does not meet the intended design specification. In A case study using DEM investigated the effect of altering designs on
this instance, changes must be minimised and reflected in insignificant wear performance of the transfer chute and the outgoing conveyor belt
production losses that do not exacerbate the duration of downtime re- by considering the velocity at key locations through the transfer. Wear
quired for maintenance. Variable ore properties and/or operational performance was evaluated in terms of impact and abrasive compo-
conditions should warrant continuous monitoring and evaluation of nents with four different design options. The designs investigated the
relevant and previously identified equipment that presents possible influence of loading angle, deflector type and chute geometry including
operational bottlenecks. To obtain a balance in chute and belt wear number of ledges. One design option (Option 2) highlighted the design
performance it is important to introduce performance based criteria in approach in obtaining a balanced improvement in both chute and belt
parallel with other integral components such as the geometry and lo- wear life. A different design option (Option 4) illustrated optimisation
cation of conveyors and fixed structure. for performance of the outgoing belt by sacrificing the life of the chute.
The influence of simulated material properties including particle shape
showed localised areas of variation in flow, velocity and wear.

Fig. 21. Influence of particle shape - velocity (left) and wear (right) - Option 1.

12
D. Ilic Wear 434-435 (2019) 202986

Fig. 22. Influence of particle shape - loaded burden -

Option 1, spherical (left) and rounded polyhedron
(right).

Table 8 [15] T.J. Donohue, D. Ilic, T. Krull, Modelling of iron ore in materials handling appli-
Variation of chute and belt wear performance with particle shape - Option 1. cations with a focus on wear, Proceedings, Iron Ore 2013, The Australasian Institute
of Mining and Metallurgy, Melbourne, Australia, 2013, pp. pp355–362.
Wear Relationship 80% Shear 50% Shear 20% Shear [16] D. Ilic, T. Hicks, Towards the development of design criteria for reduced wear in
20% Impact 50% Impact 80% Impact iron ore transfers, Technical contribution to the 17th Brazilian Symposium on Iron
Ore - 17° Simpósio Brasileiro de Minério de Ferro, ABM Brasil, Rio de Janeiro, 2016,
Particle Shape Chute Life Belt Life Belt Life Belt Life p. 10.
Spherical Base Base Base Base [17] A.W. Roberts, Chute design considerations for feeding and transfer, Proceedings
BELTCON 11, International Materials Handling Conference, Johannesburg, 2001.
Rounded Polyhedron 102% 105% 110% 115%
[18] A. Katterfeld, T. Gröger, A. Minkin, Discrete element simulation of transfer stations
and their verification, Bulk Solids Powder Sci. Technol. 2 (2007).
[19] F.P. Mascarenhas, A.L.A. Mesquita, A.L.A. Mesquita, Simulation of transfer chute
However, overall comparison of the different design options in- operation using the Discrete element method, Proc. Of XXXIV Iberian Latin
vestigated showed similar wear performance. American Confess on Computational Methods in Engineering, ABMEC, Pirenopolis,
GO, Brazil, Nov 10-13, 2013.
The variation in presented chute flow characteristics demonstrates [20] F. Kessler, M. Prenner, DEM - simulation of conveyor transfer chutes, FME Trans. 37
how a systematic approach to transfer chute design can influence spe- (2009) pp185–192.
cific performance outcomes. A set of design criteria, with focus on wear [21] D. Ilic, Transfer Chute Analysis Techniques Including Continuum Modelling and the
Discrete Element Method (DEM), Proceedings, Bulk 2014, Materials Handling
performance is proposed. The study has also highlighted the need for Engineers Association (MHEA), North Lincolnshire, England, 2014.
additional research to investigate the influence of particle size, dis- [22] D. Ilic, Bulk Solid Interactions in Belt Conveying System, PhD Thesis The University
tribution and shape, as well as more specific criteria for the design of of Newcastle, 2013.
[23] A.W. Roberts, Characterisation for hopper and stockpile design, in: D. McGlinchey
chute and rock-box ledges such as size and placement. (Ed.), Characterisation of Bulk Solids, Blackwell Publishing, 2005, pp. pp85–131.
[24] D. Ilic, A. Roberts, C. Wheeler, A. Katterfeld, Modelling bulk solid flow interactions
References in transfer chutes: shearing flow, Powder Technol. 354 (2019) pp30–44.
[25] A.I. Hustrulid, Transfer station analysis, Proceedings of the Annual Meeting of the
Society of Mining Engineers (SME), 1998, pp. pp33–53.
[1] T. Reye, Zur theorie der Zapfenreibung, Der Civilingenieur vol. 4, (1860), pp. [26] A. Katterfeld, T. Gröger, Simulation Based Wear Prediction of Transfer Chutes, Bulk
pp235–255. Europe, Conference Proceedings (CD), Vogel Transtech Publications, Würzburg,
[2] P. Villaggio, Wear of an elastic block, Meccanica 36 (2001) pp243–249. 2008.
[3] W.O. Winer, M.B. Peterson, J.F. Archard, Wear Theory and Mechanisms, Wear [27] L.K. Nordell, Modern ore transfer chute belt feeder designs developed from Discrete
Control Handbook, New York, NY, 1980 ASME, pp.35–80. element modeling (DEM), International Materials Handling Conference BeltCon 12
[4] J.F. Archard, Contact and rubbing of flat surfaces, J. Appl. Phys. 24 (8) (1953) Papers, South Africa, July 2003 online http://conveyor-dynamics.com/index.php/
pp981–988. about/publications 13/12/2018.
[5] I. Finnie, Erosion of surfaces by solid particles, Wear 3 (1960) pp87–103. [28] O.R. Walton, Explicit particle dynamics model for granular materials, Numer.
[6] I.M. Hutchings, Wear by particulates, Chem. Eng. Sci. 42 (4) (1987) pp869–878. Methods Geomech. 3 (1982) pp1261–1268.
[7] M.M. Khruschov, Principles of abrasive wear, Wear 28 (1974) pp69–88. [29] O.R. Walton, Numerical simulation of inclined chute rows of monodisperse, in-
[8] M.F. Ashby, S.C. Lim, Wear-mechanism maps, Scripta Metallugica et Materialia vol. elastic, frictional spheres, Mech. Mater. 16 (1993) pp239–247.
24, (1990), pp. pp805–810. [30] J. Ai, J. Chen, M. Rotter, J.Y. Ooi, Assessment of rolling resistance models in
[9] A.W. Roberts, S.J. Wiche, Prediction of lining wear life of bins and chutes in bulk Discrete element simulations, Powder Technol. 206 (2011) pp269–282.
solids handling operations, Tribol. Int. 26 (5) (1993) pp345–351. [31] C.M. Wensrich, A. Katterfeld, Rolling friction as a technique for modelling particle
[10] S.J. Wiche, S. Keys, A.W. Roberts, Abrasion wear tester for bulk solids handling shape in DEM, Powder Technol. 217 (2012) pp409–417.
applications, Wear 258 (2005) pp251–257. [32] L.K. Nordell, Particle flow modelling: transfer chutes and other applications,
[11] W. Chen, S. Biswas, A. Roberts, J. O'Shea, K. Williams, Abrasion wear resistance of International Materials Handling Conference BeltCon 9 Papers, South Africa,
wall lining materials in bins and chutes during iron ore mining, Int. J. Miner. October 1997 online http://conveyor-dynamics.com/index.php/about/
Process. 167 (2017) pp42–48. publications 17/12/2018.
[12] M. Andrejiova, A. Grincova, D. Marasova, Measurement and simulation of impact [33] J. Aldrich, Y. Zhang, Minimizing Belt Wear and Damage from Optimized Chute
wear damage to industrial conveyor belts, Wear 368–369 (2016) pp400–407. Design, SME Annual Meeting, Salt Lake City, UT, USA, 2014 online http://
[13] W. Molnar, M. Varga, P. Braun, K. Adam, E. Badisch, Correlation of rubber based conveyor-dynamics.com/index.php/about/publications 13/12/2018.
conveyor belt properties and abrasive wear rates under 2- and 3-body conditions, [34] W. Chen, T. Donohue, A. Katterfeld, K. Williams, Comparative Discrete element
Wear 320 (2014) pp1–6. modelling of a vibratory sieving process with spherical and rounded polyhedron
[14] C. Webb, M. Hodkiewicz, N. Khan, R. Wilson, S. Muller, Conveyor belt wear life particles, Granul. Matter 19 (2017) pp81–93.
modelling, CEED Seminar Proceedings, Pp25-30, The University of Western
Australia, 2013.

13