Redesign of A Tata Steel Transfer Chute PDF

Redesign of a Tata Steel transfer chute
with dust liberation problems

by
W.P.J. Vreeburg
MSc Thesis
Mechanical Engineering, Transport Engineering and Logistics
at the Delft University of Technology.
Report number: 2018.TEL.8255

Student number: 1526901
University supervisor: Dr. ir. D.L. Schott, TU Delft
Company supervisor: D. Oudhoff, TATA Steel IJmuiden
Company supervisor: J. de Lange, TATA Steel IJmuiden
Abstract
Dust liberation in transfer chutes has been a persistent problem in the bulk handling industry, with an in-
creasing demand for solutions due to the tightening of environmental regulations. An investigation into the
root causes of dust liberation and a chute redesign with improvements in dust liberation potential has been
performed in this research. The biggest challenges of this redesign being that the case study chute is a multi-
material, movable chute with differing flow rates and a redirection of material by 90◦ . The main, root causes
of dust liberation have been found to be material impact, air entrainment and compact containment of the
granular flow. Measurements at the current transfer chute have been performed to investigate and quan-
tify the problems of the case study chute. Direct measurement of dust in and around transfer chute is often
circumvented to measure related issues such as material degradation and air flow measurements, since this
makes localizing dust liberation sources and quantification difficult. Therefore, stopped-belt sampling has
been performed to find material degradation in a case study chute at Tata Steel, given that impacts cause
degradation as well as dust liberation. No conclusive evidence was found due to a lack of samples. Iron ore
pellets and sinter were dynamically calibrated using an inclined surface wear tester and simulations in EDEM
software were performed for 800 t/h and 1600 t/h flow rates. Three problem areas were found where impacts
and bulk density increases provided indications of potential dust liberation. A redesign proposal that uses
a hood and spoon concept with movable hood, where the cut-off of the movable chute head needed to be
heightened to fit a spoon that can improve flow conditions, was proposed. Improvements in the identified
problem areas were found in simulations of the redesign.
iii
Contents
1 Introduction 1
1.1 Introduction transfer chute dust liberation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Problem definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Research objective and research questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Report structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Literature study chute design and dust control 5
2.1 Transfer chute definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Chute design objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2.1 Blockage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.2 Spillage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.3 Wear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.4 Material degradation and dust generation . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2.5 Belt scrapings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2.6 Serviceability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3 Dust liberation theory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3.1 Dust definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3.2 Dust liberation principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.4 Dust liberation minimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.5 Chute configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.5.1 Dead box chutes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.5.2 Hood and spoon chutes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.5.3 Multi-material chutes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.6 Chute design theory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.6.1 Material trajectory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.6.2 Hood design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.6.3 Spoon design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3 Methodology selection for problem assessment at Tata Steel 21
3.1 Selection criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.2 Measurement methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.3 Numerical modeling methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.4 Analysis Tata Steel chutes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.5 Case study at Tata Steel and methodology choices. . . . . . . . . . . . . . . . . . . . . . . . 27
3.5.1 Case study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.5.2 Methodology choices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4 Quantifying material degradation 31
4.1 Material degradation sampling methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.2 Sampling results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.3 Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5 Calibration dynamic material model 35
5.1 Experimental setup and working principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.1.1 Inclined surface wear test setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
5.1.2 Performance indicators and test procedure . . . . . . . . . . . . . . . . . . . . . . . . 36
v
vi Contents
5.2 Calibration setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

5.2.1 Calibration parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.2.2 Reference case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.2.3 Calibration procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.3 Calibration process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
5.5 Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
6 DEM analysis Tata Steel chute 43
6.1 Key performance indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
6.2 Simulation setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
6.3 Methodology and results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
6.3.1 Velocity fluctuations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
6.3.2 Impact velocity and force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6.3.3 Bulk density along material trajectory . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6.3.4 Belt loading and segregation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
6.3.5 Chute wear. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
6.4 Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
7 Redesign proposals 53
7.1 Conceptual redesign proposals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
7.1.1 Redesign proposal I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
7.1.2 Redesign proposal II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
7.1.3 Redesign proposal III. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
7.1.4 Redesign choice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
7.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
7.3 Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
7.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
8 Conclusions and recommendations 61
8.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
8.2 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
A Research paper 65
A.1 Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
A.2 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
A.3 Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
A.4 Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
A.5 Case study DEM analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
A.6 Redesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
A.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
B Appendix 71
C Appendix 73
Bibliography 75
Nomenclature
List of abbreviations
AoR Angle of repose
CFD Computational Fluid Dynamics
CI Confidence Interval
DEM Discrete Element Modeling
KPI Key Performance Indicator
MFR Mass flow rate
PDR Personal Data Ram
PIV Particle Tracker Velocimetry
PSD Particle size distribution
α Angle of discharging belt conveyor
x̄ Average of samples
δn Overlap in m
ṁ Material mass flow rate in t/h
v̇ Acceleration in m/s2
² Eccentricity in degrees
µ The sample mean
µE Actual friction coefficient between chute surface and bulk solid
Φ Angle of repose
Ψ Self-cleaning spoon angle, angle from belt up to spoon exit angle
ρb Bulk density of the material in kg/m3
σ0 Adhesive stress in Pa
θ Angle of discharge in degrees
θc Angle of impact point of radius of curvature
A or A co Minimum chute area in m2
Ap Cross-sectional area of the incoming particle stream
CR Coefficient of restitution
Cd c Belt design capacity in t/h
d 50 50% of cumulative particle mass in a PSD
vii
viii Contents
dt Tangential distance moved in m

D av g Average material particle diameter in m
E∗ Equivalent Young’s Modulus
Fn Normal force in N
F nd Damping force in N
g gravitational constant 9.81 m/s2
H Material height on belt in m
h Mean height of material in m
H /B Height to width ratio of material cross-section
k Conversion factor
ms Mass flow rate in kg/s
n Number of samples
Qm Material throughput in kg/s
Qd i s Displaced air in m3 /s
Q i nd Volume of induced air in m3 /s
R Radius in m
R∗ Equivalent radius in m
Rc Radius of curvature in m
S Material stream speed in m/s
Sn Normal stiffness in N
s between samples Distance between two sample locations
t startup Start-up time
ta Student t-test value
U Non-dimensional cross-sectional area factor
v∞ Terminal velocity
Va Outgoing stream velocity after impact
vb Belt speed in m/s
vi Velocity corresponding to drop height ’h’ at point of impact with chute
−→
v nr el Normal component of the relative velocity in m/s
Vp Incoming stream velocity before impact
vf 0 Vertical component of velocity of bulk solid discharging from feeder
Vpxn Normal component to incoming stream velocity before impact at impact plate
Vpx x-component of incoming stream velocity before impact
W Dimensionless wear constant
xc x-position of radius of curvature circle
xi Sample output
yc y-position of radius of curvature circle
1
Introduction
This first chapter will introduce the goal of this research paper, as well as an elaboration on its context and
scope.
1.1. Introduction transfer chute dust liberation

In the steelmaking process raw iron ore is smelted in blast furnaces to produce liquid pig iron, which is then
cast into steel slabs. The raw iron ore is first processed to create iron ore pellets and sinter before being
fed into the blast furnace together with cokes that are used as burning fuel. At steelmaking plants these
raw and processed materials are transported across conveyor belts from the ships that import material to
pelletization and sintering factories and finally the blast furnaces. At Tata steel in IJmuiden there is a vast and
complex network of belt conveyors of over 44 km in length that provides this transportation. With general belt
conveyor designs being restricted to a straight trajectory, lots of transfer points are needed. To guide material
from one belt to another, metal constructions called transfer chutes are employed.
1.2. Problem definition

When material is dropped from a belt conveyor onto a surface below, the free fall in combination with the
resulting impact forces and air displacements causes dust particles to shoot away from the bulk material.
The resulting side effects of this dust liberation have made this a major issue in the design of transfer points
from its inception [1]. A schematic picture of the dust liberation phenomenon at a transfer chute is shown in
Figure 1.1.
Figure 1.1: Dust liberation phenomenon at a transfer chute [2]
Dust liberation has many unwanted effects, from health concerns to equipment malfunctioning. A sum-
mary of the main negative side effects are listed below.
Environmental effects
Liberated dust originating from steelmaking plants pose a significant threat to the environment, being one
the biggest contributors to fine particles in the atmosphere. Despite significantly reducing fine particle mat-
ters (P M 2.5 ) from the environment since the early 90s (see Figure 1.2, left), ever stricter environmental reg-
1
2 1. Introduction
ulations are forcing companies like Tata Steel to reduce their air pollution practices further. Environmental
inspections by the government over the past seven years have fined Tata Steel for their problematic dust lib-
eration at transfer points. Over 13 fines have been imposed for two transfer points, from belt A312 to A509 or
A510, which mainly transports sinter, and from belt A330 to A652, which mainly transports iron ore pellets.
A screenshot of a video made by an environmental inspector which illustrates the problem is shown on the
right in Figure 1.2.
Figure 1.2: Environmental inspection video screenshot of dust liberation at Tata Steel transfer from belt A312 to the A509
Occupational hazards
Respirable dust, particles smaller than 10 microns in diameter, are not filtered out by the natural defenses of
the human respiratory system and so penetrate deeply into the lungs [1]. Here they can get trapped and lead
to serious health problems such as asthma, lung cancer, respiratory diseases, cardiovascular disease, prema-
ture delivery, birth defects, low birth weight, and premature death [3]. Besides the negative effects of dust
inhalation, dust also settles on and around the premises. This causes a slippery and dirty work environment,
which can be potentially dangerous and decreases worker morale. Dust emissions also increase the risk of a
fire or explosions [4].
Production loss
Apart from environmental concerns and occupational hazards, fugitive dust can cause significant production
loss with the amount of material that is spilled. Relatively small amounts of fugitive materials can accumulate
to large quantities over time. A study by the Royal Institute of Technology in Sweden of fugitive materials in
40 bulk handling plants found that 0.2% of material was lost [1][5]. When taking into account not only the
value of lost material, but also the cost of labor devoted to cleaning up fugitive material, the cost of parts and
labor for additional maintenance arising from fugitive material and the cost of required medical checkups
for personnel due to dusty environments, this value becomes even higher. An investigative report done on
eight UK bulk material handling plants took these factors into account and arrived at an average loss of one
percent per ton of throughput [1][4]. These factors combined makes it clear that dust control in transfer
points is crucial for an efficient and environmentally friendly bulk handling process and a subject worthy of
study.
1.3. Research objective and research questions 3
1.3. Research objective and research questions

This research focuses on finding the causes for the dust formation in transfer chutes and proposing solutions
to prevent or minimize this effect, taking into account the different materials that are being funneled through
these chutes. The main research question that is answered through this research is
What are the causes and solution methods in minimizing dust creation and liberation at trans-
fer chutes and how can these be applied in the redesign of a multi-material Tata Steel transfer
chute with differing flow rates?
In order to answer this question, several smaller research questions are formulated that guide the investi-
gation. These questions are
• What is known in literature on the influencing factors on the creation and liberation of dust in transfer
chutes and what are known transfer chute design considerations that provide solutions to these problems?
• What are suitable assessment methodologies for identifying transfer chute problems at Tata Steel that
involve dust creation and liberation?
• What is the extend of material degradation at a Tata Steel transfer chute?
• How to calibrate bulk materials for dynamic simulations like transfer chutes?
• Where along the material stream in a transfer chute at Tata Steel is potential dust creation and liberation
and what is the influence of different materials and mass flow rates in DEM software on this problem?
• What does a redesign for a Tata Steel transfer chute to minimize dust creation and liberation look like
and what can be the quantifiable improvements to the original design?
1.4. Report structure

This report consists of eight chapters, each answering one of the research questions listed above. This leads
to the answer of the main research question, which will finally be answered in summary in Chapter 8 - Con-
clusions and recommendations. Chapter 2 will be a literature study on the known theory behind the design
of chutes, the known causes of dust creation and liberation and the response of designers and the industry
into the handling of these problems at transfer chutes. In the next chapter, Chapter 3, the literature on the
methodologies for identifying and possibly quantifying dust liberation is gathered. Methodologies for trans-
fer chute design that can aid in the minimization of dust liberation is also discussed. In this chapter, based
on the literature found, methodology choices for the further investigation of transfer chute problems at Tata
Steel are made. These choices involve the selection of a case study transfer chute at Tata Steel and experi-
mental sampling of material at this chute to investigate material properties that are possibly affected by it.
Also a choice is made for a numerical methodology to study the granular flow through the case study transfer
chute and use this to later test redesign concepts. Chapter 4 will expand on the methodology of experimental
sampling of material at a case study transfer chute at Tata Steel. The results and conclusions drawn from this
methodology can also be found in this chapter. In Chapter 5, the calibration of a material model for two ma-
terials that are transported by the case study chute, sinter and iron ore pellets, for the usage in a simulation
software to study the granular flow is performed. The focus of the calibration lies in the dynamic behavior of
the material. With these material models, a simulation is setup to investigate the granular flow through the
case study transfer chute in Chapter 6. The causes for dust liberation at this chute are identified based on
the knowledge gained through literature from Chapter 2. The knowledge on transfer chute design from this
chapter is used to calculate the characteristics of a redesign concept in Chapter 7. Two concepts are made,
one where the current dimensional constraints of the case study transfer chute are used and one where a
conceptual design of an ideal case is made for minimization of dust liberation. These concepts are simulated
and the potential relative improvement in granular flow is studied. The final chapter, Chapter 8, will summa-
rize the conclusions that are drawn from this research and provide recommendations for further research or
implementation of a new design.
Literature study chute design and dust
2
control
This chapter the literature that is available on transfer chute design is examined, focusing in particular on the
minimization of dust liberation. The research question that will be answered in this chapter is
• What is known in literature on the influencing factors on the creation and liberation of dust in trans-
fer chutes and what are known transfer chute design considerations that provide solutions to these
problems?
After a brief section on the definition and research scope of transfer chutes, the design objectives for
transfer chutes design are listed and discussed in Section 2.2, along with some "rules of thumb" to meet these
objectives. Section 2.3 will commence the investigation into dust liberation by expanding on the theory be-
hind the dust liberation phenomenon. Section 2.4 will explain how this theory is used to try and minimize
dust liberation at transfer chutes. Then in Section 2.5 chute configurations that deal with the design objec-
tives in differing ways are laid out, discussing their advantages and disadvantages and areas of application.
In Section 2.6 the more detailed, analytical models that govern chute design for a so called hood and spoon
design (which is particularly useful for dust liberation minimization) are explained after which the research
question for this chapter is answered in Section 2.7.
2.1. Transfer chute definition

According to BS2890 - Specification for Throughed Belt Conveyors by the British Standards Institution the
definition for a transfer chute is "A straight, curved or spiral, open topped or enclosed, smooth through, by
which materials are directed and lowered by gravity."[6][7]. This means that when bulk material that is being
transferred from one bulk handling equipment to another, a transfer chute will help guide the direction of
the material. A distinction can be made between feed chutes, which are build to accelerate materials, and
transfer chutes, which are used in the transfer of bulk material from a belt, bin or hopper [7][8]. Belt to belt
transfers are the most common, which usually involves a change in flow direction [9], as exemplified by the
illustration in Figure 2.1. Belt to belt transfer chutes will be the main focus of this research.
Figure 2.1: Illustration of a belt to belt transfer chute involving a 90◦ change in material flow direction [8]
5
6 2. Literature study chute design and dust control
2.2. Chute design objectives

Conventional transfer chute design is normally done by an experienced designer or bulk materials handling
engineer using industry-accepted "rules of thumb" [1]. Many engineering firms establish their own design
rules and with a large application variability, different industries have developed consistent approaches to
chute design that solves issues particular to their applications. Still there exist general guidelines to properly
designing conventional transfer chutes.
A conventional transfer chute generally can be separated into four basic parts, which are listed below [1]
and depicted in Figure 2.2.
• Head chute The area surrounding the head pulley of the feeding conveyor
• Drop chute The area where the material is guided in its free fall
• Loading chute The area where the material is guided onto the receiving belt conveyor below
• Settling zone An extension area of the chutework to settle airborne dust, not technically part of the
transfer chute
Figure 2.2: Schematic of conventional transfer chute consisting of four basic parts (edited from [1])
Each of these parts have to be designed to fit the design parameters given by the developer such as the
geometric parameters of the belt conveyors or the flow capacity. Besides the adherence to these basic require-
ments, there are several main objectives in transfer chute design that need to be taken into account. These
design objectives are [1][7][8]:
• No blockage of the material flow in the chute, while facilitating the required capacity
• No spillage, prevent the escape of fugitive materials
• Minimize wear, on all components (belt, chute, material) to provide the optimal value life cycle solu-
tion
• Minimize degradation of the transported material and the generation of dust
• Return belt scrapings to the main material flow
• Minimize segregation of the bulk material
• Be service friendly and protect personnel from injury
• Proper loading of material onto receiving conveyor, such that
– the difference in velocity of the conveyor and the velocity component of the material in the direc-
tion of the conveyor is as close as possible (at least within 10%)
– the velocity component of the material perpendicular to the receiving belt is as close to zero as
possible
– the flow is loaded in the center of the receiving belt to prevent the belt from misaligning and to
avoid spillage
2.2. Chute design objectives 7
2.2.1. Blockage
One of the biggest design aspects in the design of a transfer chute is the prevention of blockage. If material will
not flow reliably though the chute, then meeting any or all other objectives is irrelevant. Bulk materials should
maintain an even and consistent flow when travelling through the chute. Surge loadings can cause several
problems such as belt tracking sliding off-center, over-stressing of conveyor system components (particularly
the drive motor[10]) and of course chute blockage. The body of the chute should have an area that is at least
2.5 to 3 times the area of the material in order to prevent blockage from a chute that is too narrow [11]. The
minimum area of the chute is given by the equation in Figure 2.3.
Figure 2.3: Equations for calculating the minimum area a chute should have to prevent blockage[11]
In order to prevent blockage close attention should be paid to the particle and bulk properties of the
transported material and the nature and characteristics of the application [8]. The sliding surface inside the
chute must be sufficiently smooth to allow the material to slide and to clean off the most frictional bulk solid
that it handles. The steepness of the impact between the bulk material and the chute is also a factor. If the
impact angle is not steep enough, sticky material will stay attached to the chute surface and cause plugging
of the chute. The relationship between these variables is shown in Figure 2.4. When V2 is significantly low to
impede the flow of material, plugging will occur and when it is zero, the material will stick to the surface.
Figure 2.4: Equations for calculating Impact pressure and particle velocity along chute after impact[11]
2.2.2. Spillage
Similarly to preventing blockage, spillage is countered by focusing on a reliable and consistent flow through
the chute. The objective of avoiding spillage is linked to many of the other before-mentioned design objec-
tives. If the chute blocks, or a hole is created in the chute material due to wear, material will spill. Some of
these problems are depicted together in Figure 2.5. Many of the causes of spillage are related to the proper
loading of the receiving belt. Off-center loading, which is placing the cargo predominantly on one side of
the belt (as can be seen in Figure 2.5), is a problem at many transfer points that contributes to generation of
fugitive materials [1]. This problem is most prevalent in transfer chutes with a change in flow direction. The
displacement of the material causes belt tracking problems where the belt can suffer edge wear damage and
may result in material spilling over the belt edge, both shown in Figure 2.5. Since training idlers are limited in
their ability to counter the effects of off-center loading, it is imperative to design the chute to properly align
the material onto the receiving conveyor belt. Besides loading the material in the center of the receiving belt,
it is also desirable to keep the velocity component of the material perpendicular to the belt as close to zero as
possible.
Figure 2.5: Overview of common problem when chute design objectives are not met (edited from [10][11])
2.2.3. Wear
The abrasive nature of particulate bulk solids such as iron ore, large scale handling commonly causes high
wear on chute surfaces [12][13]. Wear causes volume loss on the surfaces that are in abrasive or high im-
pact contact with bulk solids and consequently accelerate damage to these areas. Wear introduces lots of
economic costs such as maintenance related costs (replacement parts, productivity loss due to downtime,
maintenance personnel), spillage or development and procurement of abrasive resistant materials. Wear
can be divided in abrasive wear and impact wear. Abrasive wear is caused by relatively sliding and rolling of
particles against equipment surfaces, whereas impact wear is interpreted as a process of material removal
from equipment surfaces by succeeding particle impacts [12]. Impact wear in transfer chutes may occur at
points of entry or at points of sudden changes in direction. For ductile materials the greatest wear occurs
when impingement angles are low, which is around 15◦ to 30◦ . For hard, brittle materials like iron ore the
greatest impact damage occurs at steep impingement angles of the order of 90◦ [13]. To combat excessive
wear, wear-resistant materials can be applied to wearing surfaces. A more significant contribution to wear
reduction is made by controlling the flow of material with the optimization of chute profiles. This is done
based on analytical predictions of material trajectories and chute surface wear rates. The theory behind this
optimization is elaborated on in Section 2.5.2. Even the introduction of bionic designs on chute surfaces are
being investigated and show promising results [12]. While wear on chute surfaces are most significant, wear
on receiving conveyor belts can also be problematic. In order to minimize impact wear on conveyor belts,
loading material directly onto belts from a certain height is avoided. Abrasive wear on belts is countered by
attempting to match the material stream velocity and direction to that of the belt at the loading point. The
calculations that accompany this theory is presented in Section 2.5.2.
2.2. Chute design objectives 9
2.2.4. Material degradation and dust generation

The impact force (see Figure 2.4) that is exerted when the material comes into contact with the chute not only
affects the chute itself, but also the transported material. This can lead to the degradation of the material,
which will generate dust particles in the stream that can get released as fugitive dust, among other problems
such as inefficient further use of the material. There are different types of material degradation that can
occur. The types of breakage that exist in the handling of bulk materials is illustrated and explained in Figure
2.6. A detailed analysis of the causes of material degradation and the resulting dust generation is performed
in Section 2.3.
Figure 2.6: Breaking mechanisms in bulk material [14]
2.2.5. Belt scrapings

When conveying sticky bulk material, adhesion to the belt can occur when it is being dropped from the belt
at the head pulley. At this discharge pulley belt cleaners are installed to remove residual material that has
adhered to the belt beyond the discharge point. The material that gets scraped off these cleaners, called
carryback, has a risk of build-up and eventual spillage if it is not returned to the main material flow properly,
as depicted in Figure 2.5. The chute is usually extended underneath to catch this carryback. To ensure this
already highly adhesive carryback returns to the main flow, the angle the chute makes to the ground has to be
very steep. This angle, called the valley angle, has to be at least 70◦ [10]. In many cases additional remedies
are required such as over-sizing the chute, low-friction chute liners and auxiliary devices like vibrating dribble
chutes, air cannons and scavenger conveyors.
2.2.6. Serviceability
As noted before the enclosure of the chute is vital in minimizing spillage and dust liberation, as depicted
in Figure 2.5. It is also imperative that the transfer chute is easily accessible for personnel to accommodate
efficient maintenance. Inspection doors should be placed away from material trajectories to prevent spillage.
With chute wear damage happening in predictable areas, the chute needs to be designed to easily access
these areas and make them easily replaceable. Many suppliers provide service-friendly arrangements of their
components to have these features cancelled out by the design of the structure or by the placement of utility
piping and conduits or other components [1].
2.3. Dust liberation theory

2.3.1. Dust definition
To understand dust liberation theory, first a definition of dust needs to be determined. Definitions of dust and
the size range of dust can vary depending on the field of study. According to the International Standardization
Organization (ISO 4225 - ISO, 1994) dust is defined as "Small, solid particles, conventionally taken as those
particles below 75 µm in diameter, which settle out under their own weight but which may remain suspended
for some time"[3]. A definition of dust by the ’Glossary of Atmospheric Chemistry Terms’ is given as "small,
dry, solid particles in the size range from about 1 to 100 µm in diameter, projected into the air by natural forces
and by mechanical or man-made processes"[3][15]. The ’Handbook of Chemistry and Physics’ distinguishes
three categories of dust, shown in Figure 2.7 [2][15].
Figure 2.7: Table of dust categorizations and characteristics
Most handbooks dealing with ventilation and dust control for human health studies focus on particles
ranging from 0.1 to 100 µm [3][16], while research papers that analyze dust tend to model dust particles that
are 0.1 mm in diameter [2][17][18]. This is mainly since the smaller the diameter, the deeper particles can
penetrate into the lungs and cause health problems, while in performing dust simulations, a minuscule par-
ticle diameter is difficult to work with.
2.3.2. Dust liberation principles

Dust control = air control
Dust liberation is when dust gets separated from the bulk material stream and released or expelled into the
air. Since dust particles are light enough to travel through the air, there is great benefit in understanding the
air flows in and around transfer chutes. According to [1], there exists a generalized, intuitive relationship for
dust liberation, namely
Air velocity
Dust liberation ∝ (2.1)
Particle Size × Cohesiveness
If air velocity is increased, but particle size and cohesiveness remain constant, then airborne dust will
increase. If velocity remains constant, but particle size or cohesiveness is decreased, then the amount of
airborne dust will increase. It is therefore important to not only control the air, but also prevent material
degradation (which decreases particle size), in order to minimize dust liberation.
Dust liberation regions

According to [19], three regions in a transfer chute can be indentified as sensitive for dust liberation, as illus-
trated in Figure 2.8 [4].
2.3. Dust liberation theory 11
Figure 2.8: Graphical representation of dust liberation regions with belt to belt material transfer (edited from [4])
• Region 1 – Material discharge

The point of material discharge from the belt head pulley, emissions are driven by turbulence and de-
layed detachment of particles from the belt
• Region 2 – Free fall
Region where the material is in free fall, which is when dust particles get separated through a phe-
nomenon called air entrainment [4][20]
• Region 3 – Impact
The point where the material suffers an impact after being in free fall, which is noted to be the biggest
contributor to dust formation and emission [19]. The point and angle of impact and the number of
impacts that the material endures in a transfer point are dependent on the chute design configurations.
Air displacement
When material is transported it pushes the air that is in its path. Bulk material entering a transfer chute will
displace the air that is inside and this air will be pushed out. The displaced air can be roughly calculated using
the following equation
k × ṁ
Qd i s = (2.2)
ρb
where Q d i s is the displaced air in m3 /s, k is a conversion factor of 0.277, ṁ is the material mass flow rate
in t/h and ρ b is the bulk density in kg/m3 .
The movement of air inside and out of the transfer is related to the volumetric size of the transfer chute
enclosure and the openings in the enclosure. A slightly negative pressure is desirable inside the chute, so
that air is ’sucked’ into the enclosure and keeps fines and airborne particles from liberation. Unfortunately
when material enters the enclosure it pushes into the air already inside and brings in extra air through its
movement, which inevitably creates an unwanted positive air pressure inside the enclosure. If this positive
pressure is not addressed, dust liberation will occur. The positive pressure can be altered with proper control
of material flow, adequate pressure relief or dust collection systems.
Air entrainment
Bulk materials that are on the move have a certain amount of entrapped air and carry a small amount of air
with them as they travel on the conveyor belt. When the material gets dropped from the discharging conveyor
pulley, the material stream expands. This creates new and expanding voids in between the particles that gets
’induced’ inside the stream. This induced air can be roughly calculated by
s
3 ṁh 2
Q i nd = k × A co × (2.3)
D av g
where Q i nd is the volume of induced air in m3 /s, k is a conversion factor of 0.078, A co is the area of head
chute opening in m2 , ṁ is the material mass flow rate in t/h, h is the height of material free fall and D av g is
the average material particle diameter in m.
A boundary layer of dust forms around this core stream with expanding radius as the drop height increases
[4]. Smaller particles have a lower fall speed than larger particles, which means they have a higher chance of
floating off to the sides of the core stream of material[16]. The study of air entrainment is very complex, since
it involves the analysis of the aerodynamic properties of particles, which are extremely variable. The irregular
size of particles and the presence of surrounding particles gives each particle a different air flow pattern and
drag force, not to mention particle collisions in free fall. Air entrainment by free falling bulk materials can
be influenced by material properties, such as particle size distribution and particle density. It can also be
affected by process parameters, such as drop height and bulk material mass flow rates [4].
Impact
As explained in the previous paragraph, air gets entrained inside the particle stream when it is in free fall.
When the material lands and compresses back into a pile, the entrained air is released in every direction, as
illustrated with arrows at the bottom belt in Figure 2.8. This released air at material point of impact causes a
pressure buildup[10][16].
Another phenomenon that occurs surrounding impact is material degradation. When particles are in free
fall the gravitational force causes them to gain kinetic energy. This energy is dissipated when the material
impacts the bottom belt or the chute wall. The energy dissipation happens through particle degradation
(see also Section 2.2.4). Since the impact of material on a chute surface or belt both creates dust as well
as liberates it, material impact should be avoided as much as possible in transfer chutes to minimize dust
liberation [4][16][21][22].
2.4. Dust liberation minimization

Based on the knowledge gathered from the physical theories underlying dust liberation, basic design princi-
ples can be distilled to control dust. All the measures that are taken in this area revolve around:
• Minimize dust creation at transfer
• Minimize spreading of dust at transfer
The first of which means to prevent material degradation, while the second revolves around the control of air
flow in and around the transfer chute.
The design objectives that align with these ideas are summarized below.
Minimize drop height and impact force

In order to minimize both the creation and liberation of dust, minimizing the drop height and impact force
is vital. The first remedy that could come to mind is to place the discharging and receiving conveyors as close
together as possible. Once a conveyor is installed this is an incredibly complicated process to implement,
but in the early design stages this can be relatively easy [1]. Risk of blockage, spillage and difficulties with air
control using a smaller enclosure should be taken into account. Also the incorporation of a hood and spoon
design can become impossible if the height difference between conveyors is too low. The added benefit of a
hood and spoon chute is far greater than a reduction in drop height without proper flow control, so in this
case it is not recommended.
Drop height reduction means limit the free fall of the particle stream. Therefore a hood and spoon chute
design can be a solid solution since this will keep the material in contact with the chute walls. When designed
properly, it can also keep the material flow speed more constant which will further improve the limiting of
impact forces.
When a hood and spoon design is not in place, impact forces can be reduced by lowering the belt speeds.
When material is being launched at an impact plate and suffers an impact force, this force can be lowered by
reducing the speed of the particle stream.
Minimize material stream spreading As explained earlier air displacement inside transfer chute is when
material acts on the air it is traveling through, carrying some of the air along with it. The larger the material
stream, the more air gets swooped along and cause dust liberation problems. It is therefore recommended
to keep the materials in a consolidated stream as they leave the head pulley and move through the chute.
Deflectors can be used to do this, but they can cause flow problems and increase the number of impact
2.4. Dust liberation minimization 13
forces. Hood and spoon designs can be a more likely solution to this problem. They can also be shaped to
converge the flow as the velocity increases in order to keep the flow compact, as shown in Figure 2.9
Figure 2.9: Illustration of a converging hood design [8]
Another method of keeping the material stream compact is increasing the cohesion of the material. This
can be done by installing water hoses along the belts to spray the material on the material. This can however
provide problems in the eventual use of the material or difficult to implement if the material is heated in
some cases and the water quickly vaporizes. The added cost of the water spray installation and water usage
make this solution not favourable.
Minimize air entering enclosure

One of the most inexpensive solutions to dust liberation that do not require sophisticated models and com-
putations is to minimize the inlet of air into the chute enclosure. This will reduce the amount of induced air
inside the chute, as can be derived from A co in equation 2.3. Conventional rubber curtains can be installed at
the entrance and exit of the chute enclosure (Figure 2.10, left), skirtboards can be installed along the receiving
conveyor (Figure 2.10, right) and inspection doors should remain closed during normal operations.
Figure 2.10: Rubber chute entrance curtains (left) and skirtboards (right) [1]
Minimize air speed inside chute

When the air speed inside the enclosure is kept low, the likelihood of dust particles getting picked up and
carried outside the chute is lowered. The easiest method to achieve a lower air speed inside the chute is low-
ering the belt speed. When this is not an option, the enclosure can be extended along the receiving conveyor
to permit a significant reduction in the speed of air currents and, therefore, allow airborne particles to settle
back into the load before the conveyor leaves the enclosure. An extension of the enclosure along the receiv-
ing belt can be made called the settling zone or ’calming tunnel’[23] for this purpose (see Figure 2.2). A larger
enclosure also means the relative volume of air displacement caused by the incoming material is lowered,
which reduces the positive pressure that can carry airborne particles out of the enclosure.
Water curtains and air filtration systems Other methods that have been applied in the industry are the
implementation of a ’water curtain’ inside the chute. This is a spray of water along the material stream that
catches dust particles inside the water stream in order to minimize dust liberation. The water is then either
returned onto the material stream or channeled to a filtration system.
Chutes can also be fitted with a complete air filtration system that controls the air flow inside the chute
enclosure by extracting air from the chute and filtering it. Considering the air flow moves in the same direc-
tion as the material stream, the most efficient location for the air extraction is at the end of the enclosure, see
Figure 2.11. After the dust is filtered from the device it is taken away from the material stream or is re-entered.
Figure 2.11: Air filtration system to extract dust particles [10]
2.5. Chute configurations

As stated before, a wide variety of material properties, system layouts and other specific requirements results
in many different transfer chute designs. However, since the design objectives that have been laid out are
always the same, some general design configurations for transfer chutes can be distinguished [7]. The most
common chute designs are dead box chutes and hood and spoon chutes or approximations of these designs.
A variation on the hood and spoon concept taken into account different materials is also examined.
2.5.1. Dead box chutes

A dead box chute has a slanted surface inside the drop chute where a pile of the transferred material ac-
cumulates, as illustrated in Figure 2.12. A small bar called a wear bar is placed on the ledge of this surface
to encourage the accumulation of material, indicated in yellow in Figure 2.12. Subsequent material moving
through the chute flows over the material bed that is formed on the slanted surface. By shifting the abrasive
forces caused by the transferred material away from the chute material and onto the accumulated bed of
material, wear on chute materials is significantly reduced [1].
Figure 2.12: Cross section illustration of a typical dead box chute design[10]
The dead box configuration is applied in handling relatively dry bulk material, since sticky material could
provide blockage problems[7][10]. Similarly, care must be taken to accurately judge more cohesive character-
istics of the material, for example under wet conditions. Fragile bulk materials are not suitable to be guided
through dead box chutes, since they might suffer material degradation. Dead boxes are also avoided when
the conveyors carry more than one material, or when the bulk material may contain large enough lumps to
potentially block the material flow. Dead boxes may be designed for belts of relatively high speed (belt speeds
in excess of about 2.0 m/s) that carry washed and sized material of lump size greater than 30 mm [11]. The
design should be self-draining, meaning that the slanted surface should be inclined at a minimum angle (10◦ ,
as depicted in Figure 2.12). A small gap is left between the wear bar and the dead box surface that allows the
passage of water underneath when cleaned. In situations where a high drop of material is needed, configura-
tions with several dead boxes stacked on each other can be implemented. This can be used to reduce impact
and control material velocity over drops of greater distance. This configuration is called a cascading chute.
2.5. Chute configurations 15
2.5.2. Hood and spoon chutes

A hood and spoon chute is a design configuration where a curved hood catches the material as it comes off
the top belt conveyor and guides the material path downwards in a more controlled manner than free fall (see
Figure 2.13). The spoon is a similarly curved surface that catches the material that slides down off the hood
and guides the material onto the bottom belt.
Figure 2.13: Software simulation screenshot of a hood and spoon design, edited from [24]
The goal of this design is to keep the material stream as close as possible to continuous flow, which has
many advantages like reductions in belt wear, material degradation, spillage and dust liberation [1]. The
curvature in the hood and spoon minimizes the material impact and reduces variations in the velocity of the
material. The hood also has a converging shape, which minimizes the expansion of the material body. As
the material falls, the gradual increase in velocity due to gravity allows for the narrowing of the material flow
without plugging of the chute (see the equation in Figure 2.3, where an increase in S causes a decrease in A).
The spoon reduces the velocity and force of material impact in the loading of the bottom belt conveyor [25].
The hood and spoon system depends upon gravity and friction to maintain the speed of the material flow
through the chute. In order to properly design this system, a lot of material characteristics and continuum
mechanics knowledge and computation is required. This can make the design of hood and spoon chutes
relatively strenuous, complex and expensive [1]. Variations in flow rates of the belt conveyors or different
materials flowing through the system add to this complexity and call for significant compromises in the chute
design [23]. The configuration also requires a sufficient drop height to implement this design. In some cases
only a spoon is used to change the direction of the stream. Spoons can also be prone to backing up or flushing
if the characteristics of the bulk materials are variable [1]. This also requires compensation to be designed into
the spoon to account for this variability. In some cases where available space or budget is limited, an impact
plate is used instead of a hood [1]. An impact plate is a flat or lightly curved plate surface placed inside the
chute to absorb the force of the moving material stream coming from the top belt and to divert it towards the
bottom belt. Some impact plates have a mechanism that allows for a change in angle the plate can make in
order to suit different materials and material flow rates.
2.5.3. Multi-material chutes

As stated before, when a transfer chute transports multiple materials, the hood and spoon design has a higher
risk of encountering flow problems. Different products will have wide variations in every attribute, like par-
ticle size, moisture content and abrasion, giving the products significant flow property variations [23]. As
an example, a product range from fine material to lumpy ore is considered and the chute angles to prevent
build up of the fine material is significantly steeper (70◦ ) than lumpy ore (50◦ ). When the highest angle of 70◦
is chosen to eliminate fine material blockage, this same chute angle used with lumpy ore will produce high
chute exit velocities, and could lead to spillage and high belt wear [23]. Rather than fighting symptons with
skirtboards or increasing maintenance, designers try to convince the end users to opt for a more expensive
but potentially far more effective solution. A redesign concept of a multi-material transfer chute is presented
by [23] at Beltcon 14 for instance, displayed in Figure 2.14. It contains an adjustable spoon that can change
the material flow angles to account for different material flow properties. It could potentially react automati-
cally to potential build up of material or chute flow anomalies by performing pressure measurements on the
hydraulic cylinder at the top of the spoon, as seen in the figure, or by positioning load cells under the chute
respectively. When a conveyor is used to feed either of two discharge points, a so-called flopper gate can be
opted for (see Figure 2.14, right). Care must be taken to adhere to general chute design guidelines to be self-
cleaning, while the gate should be placed at the apex of the double chute to avoid jamming of the gate with
rocks getting stuck.
Figure 2.14: Chute redesign with adjustable spoon using hydraulic cylinder [23]. Maximum flow (left) and minimum flow position
(middle). Flopper gate concept [10] (right)
2.6. Chute design theory

When creating a hood and spoon chute, several models are made to predict the flow of material through
the chute. With the increase in computational power, these analytical models can see a significant accuracy
increase through the use of numerical models. The different models that are used for chute design are listed
below [8]:
• Trajectory model Material flow from discharge of the top belt until chute hood impact
• Impact model Material impacting the chute hood
• Hood chute flow model Material traveling along the hood surface
• Free-fall/Trajectory model Material in free fall from hood to spoon
• Impact model Material impacting the chute spoon
• Chute flow model Material traveling along the spoon onto the receiving belt
2.6.1. Material trajectory

The path the bulk material takes as it is discharged from the delivery conveyor is called the trajectory [1]. The
trajectory model consists of:
• Discharge load profile
• Angle of discharge
• Discharge trajectory
Discharge load profile

The load profile is the cross-sectional area of the material on the discharging belt. When the belt reaches the
drive pulley it is flat, while along the rest of belt length it is supported by idler sets at a certain throughing
angle. Equations along with descriptive drawings to find the loading profile dimensions at the drive drum
is shown in Figure 2.15. The cross-sectional area of the material on the belt, A is found with a well known
equation:
Qm
A= (2.4)
ρv b
2.6. Chute design theory 17
where Q m is the material throughput in kg/s, ρ is the bulk density of the material in kg/m3 and v b is the
belt speed in m/s.
Figure 2.15: Discharge load profile equations (edited from [8][11])
The assumptions that are made for these calculations are[8]:
• Material lifts off at the tangent to the drive pulley
• No belt sag in the transition
• No cohesion
• Neglecting acceleration due to transition
Angle of discharge
The free body diagram and accompanying equations that govern the point where the material leaves the belt
is presented in Figure 2.16. Slip can occur before lift-off takes place, hence the inclusion of the acceleration v̇
and inertia force ∆m v̇, with v being the relative velocity [26]. In many cases this is insignificant and therefore
neglected.
Figure 2.16: Angle of discharge free body diagram with accompanying equation (edited from [8][11][26]
In most cases the speed of the top conveyor is fast enough to discharge the material as soon as the belt
makes contact with the drive pulley. In this case θ = −(α + ²) (shown in red in Figure 2.16), where (α + ²) is the
slope of the belt at contact point with the drum [8]. The critical case will be for the belt surface, that is, when
∆r = 0. The minimum belt speed for discharge at the first point of drum contact is
s
σo
µ ¶
vb = R g cos θ + (2.5)
ρg h
Discharge trajectory
The trajectory the material makes after it leaves the discharging belt is government by common equations of
motion. When air drag is assumed to be negligible, which holds true in most cases [26], the equation of the
path is defined by the x and y coordinates according to
1 x2
y = x tan θ + g 2 (2.6)
2 v cos2 θ
The upper and lower bounds for the trajectories may be determined for the two radii (R + h) and R for
which the angle θ is obtained from Equation 2.5.
2.6.2. Hood design

After the material trajectory from the discharging conveyor is found, the hood is the next phase of the particle
flow through the chute. The radius of curvature of the discharge trajectory is given by
· µ ¶2 ¸1.5
gx
1 + v 2 cos2 θ
b
Rc = g (2.7)
v b2 cos2 θ
When the chute hood is designed with a constant radius, R, the radius of curvature of the trajectory at the
point of contact has to be such that
R ≥ Rc (2.8)
Ideally the chute hood ’catches’ the material at a tangent, so where the chute radius matches the radius
of curvature of the material at the point of contact. This is visible in the left illustration in Figure 2.17.
2.6. Chute design theory 19
Figure 2.17: Matching radius of curvature of material and chute hood[26]
When an x-value for the hood location is picked, the position of the hood circle can be found by finding
the values from Figure 2.17
θc = cot−1 (y 0 ) (2.9)
x c = X − R c cos θc (2.10)
y c = Y + R c sin θc − (R + h) cos(α + ²) (2.11)
On the right illustration in Figure 2.17, the free body diagram of forces that are present in this phase are
depicted. The differential equation that follows from these forces is
dv gR
− + µE v = (cos θ + µE sin θ) (2.12)
dθ v
The solution to this equation is Equation 2.13, when a constant radius is assumed, as well as an average,
constant value for µE for the stream, which holds when a thin cross-sectional stream is assumed (H /B <<1).
s
2g R
sin θ(2µ2E − 1) + 3µE cos θ + K e 2µE θ
£ ¤
v= (2.13)
4µ2E +1
For v = v 0 at θ = θ0 , then
( )
2g R ¤ −2µ θ
v 02 − 2
3µE cos θ0 + (2µE − 1) sin θ0 e
£
E 0
K= (2.14)
1 + 4µ2E
Of course the velocity of the material needs to be high enough as to not lose contact with the chute. Therefore,
for the condition of positive contact to be true, equation 2.13 is only applicable when
v2
≥ sin θ (2.15)
Rg
2.6.3. Spoon design

After the material leaves the hood, it undergoes a brief moment in free fall. In this period of free fall the
velocity v i can be estimated when neglecting air resistance to be
q
vi = v 2f o + 2g h (2.16)
When air resistance is taken into account, the relationship between height of drop and velocity v i is
" vf o # µ
2
v∞ 1− v∞ vi − v0
¶
h= loge v − v∞ (2.17)
g 1 − v ∞i g
where v ∞ is the terminal velocity, v f 0 is the vertical component of velocity of bulk solid discharging from
feeder and v i is the velocity corresponding to drop height, h, at the point of chute impact.
The theory of spoon design is almost identical to the hood design theory, apart from the direction of
the gravity force. Where the hood design had a minimum speed for maintaining contact between hood and
material, the spoon chute deals with an acceleration of the material followed by a retardation. To ensure the
material slides off the spoon, a self-cleaning check needs to be performed. This determines the minimum
angle the spoon exit makes to the receiving belt, defined by
Ψ > tan−1 µE + 5◦ (2.18)
The material stream suffers energy loss along the spoon trajectory due to sliding along the chute bottom.
Energy loss due to sliding against side walls and inter-particle sliding is also present, but not as significant.
So far only constant radius hood and spoon designs have been discussed. Gains can be made using non-
constant radius chutes, however the optimization is difficult. Construction of non-constant radius plate ma-
terial can also prove challenging.
2.7. Conclusions
The research question that was answered in this chapter was
• What is known in literature on the influencing factors on the creation and liberation of dust in trans-
fer chutes and what are known transfer chute design considerations that provide solutions to these
problems?
In transfer chute design the objectives are to avoid unwanted effects that occur in bulk material transfer such
as dust liberation, blockage, spillage, excessive wear or material degradation. Many of these effects can aggra-
vate each other when they occur and it can therefore be extremely beneficial to prevent or minimize them.
This can be accomplished with a proper guidance of the granular flow through the chute (no big velocity
fluctuations, compact streams, center belt loading, etc.).
Dust liberation specifically is a result of air displacement of the material stream. In order to minimize
dust liberation, the main considerations to be implemented in the design are
• minimize material impact
• minimize free fall of material
• compact containment of flow stream
These design considerations are not merely minimizing dust liberation, but also the creation of dust
through material degradation. Methods of dust liberation minimization such as chute curtains, skirtboards,
water curtains and air filtration systems can also be employed to fight dust liberation. These methods tend
to be deployed when an existing chute has dust liberation problems. The focus lies more on the containment
of dust as opposed to the prevention of dust creation, which is ultimately a better solution due to addressing
the root cause instead of fighting symptoms.
A proven chute design configuration that addresses all these design considerations is the hood and spoon
configuration. Here curved chute surfaces guide the material in a controlled manner, where material free
fall and impacts are minimized, and the flow can be contained using a converging design. This method does
provide greater challenges for engineers, since an accurate prediction of the material flow behavior is needed.
Multiple materials transported through the same chute complicates the design even further and forces the
engineer into compromises in its design.
Methodology selection for problem
3
assessment at Tata Steel
When an existing transfer chute is suffering from problems such as excessive spillage or dust liberation, sev-
eral methodologies are usually combined to solve these problems. Measuring methodologies are used to
identify and quantify the problem, which can be used to compare with an improved design later. Modeling
techniques where the granular flow or the air streams around it are modeled can also be used to specify the
source and severity of the problem, as well as aid in finding improved designs that solve these issues. To re-
design transfer chutes at Tata Steel, a selection of assessment methodologies is made. The research question
that this chapter answers is therefore:
involve dust liberation?
In order to find the answer to this question, criteria that will define what ’suitable’ means in the context
of this research have to be determined. These criteria are made explicit in Section 3.1. Measurement tech-
niques that can be used when dust liberation problems occur at a transfer chute are discussed in Section 3.2.
Then numerical modeling as an investigative tool for granular flow problems in transfer chutes are expanded
on in Section 3.3. In Section 3.4 the situation at Tata Steel is laid out in order to look at the possibilities in
investigating the causes of their dust liberation problem to eventually find feasible solutions. This chapter
is concluded in Section 3.6 with the conclusions drawn and a summarized answer to the chapters’ research
question.
3.1. Selection criteria

Multiple measuring and modeling techniques are discussed in this chapter and choices for the most suitable
methods for this research are made. Separate criteria for measurement and modeling techniques are for-
mulated, since the measurement methodologies are used to solely assess the transfer chute problem of dust
liberation and its causes, whereas the modeling techniques are also used to perform an eventual redesign.
The criteria that are set to choose a suitable measurement methodology are
• Equipment - (Measuring) equipment has to be available or possibly constructed
• Time - Methodologies have to be implemented within time frame of research
• Manpower - Amount of manpower required should be limited
• Identification - Methodology should identify chute problem(s)
• Quantification - Methodology should have some form of quantification of the problem
The criteria for choosing a suitable modeling methodology are
• Material modeling - Methodology for modeling of material and interaction between particles
21
22 3. Methodology selection for problem assessment at Tata Steel
• Equipment-material interaction - Modeling of material and equipment interaction
• Time - Modeling and computational time required should be within time frame of research
• Identification - Methodology should be capable of aiding identification of chute problem(s)
• Quantification - Methodology should have some form of quantification of the problem
• Redesign suitability - Methodology should be applicable in testing redesign chutes for comparison of
performance
3.2. Measurement methodologies

Since the quantification of dust liberation is very challenging, most quantitative dust measurements for trans-
fer chutes are performed in controlled test setups for research purposes. Dust towers for example are test
setups used to test the effectiveness of dust suppressants or different pelletization methods on fine genera-
tion. Material in a closed off tower is dropped through cascading zig-zagging trajectory where it suffers many
impacts, while a vacuum system pulls air up through the tower and trapping dust particles in filters [27][28].
Another example is a scale model of a spoon chute made by Wheeler et al. [29] in an enclosure where in
several sections of plastic plates on the bottom the settled dust was extracted and weighed. Also Particle Im-
age Velocimetry (PIV) was used to measure the air velocity profile across the chute outlet to analyze the flow
pattern.
Accurate dust measurements on actual, operational transfer chutes can be difficult, time consuming and
prone to contamination. Therefore, when actual transfer chutes in operation suffer from dust liberation,
not only dust measurements, but also air flow and material degradation measurements can be employed to
assess the problems. An overview of these existing methodologies that can be used in assessment of dust
creation and liberation is found in Table 3.1.
Measurement methodology Description Measures
Visual inspection Visually identify dust liberation quantity Airborne dust and settled dust
and source
Area sweep Periodic cleaning and weighing dust Settled dust
Dust monitors Dust measurement equipment placed Airborne dust
in/around chute
Air velocity measurements Air velocity measuring device placed at Air velocity and direction
chute entry and exit point
Material sampling Taking samples and checking amount of Material degradation
dust before and after chute
Camera’s inside chute Record video of granular flow inside Granular flow visualization
chute to identify problem areas
Table 3.1: Overview of measurement methodologies related to dust liberation problems
Visual inspection
The accurate measurement of the amount of fugitive materials at transfer points can be difficult to quantify.
Dust liberation can usually be spotted with a visual inspection, but providing any quantitative results can be
difficult. Within the enclosure of a transfer chute the dust that gets liberated can be visible, but locating the
origin of the dust liberation along the material stream within the chute is much more challenging. Visual
inspection is however a quick, easy and cheap way to determine a transfer chute dust liberation problem.
Area sweep
A very basic technique besides a simple visual inspection for dust liberation sources is to clean a defined area
around a transfer chute and weigh the amount of material collected. This ’area sweep’ is then conducted af-
ter regular intervals in time. Whether this interval should be weekly, daily, or hourly will depend on the plant
conditions [1]. A big downside to this methodology is the added difficulty of the environmental conditions
3.2. Measurement methodologies 23
that affect measurements. Wind can carry dust from other liberation sources into the measurement area
and rain and humidity variation can cause measurement inaccuracies in the amount of airborne dust that is
measured. In order to provide a useful study based on the interval data has to take into account the many
variables that may influence these results. Common operating conditions should be recorded alongside the
fugitive material measurements that include environmental conditions, operating schedules, material mois-
ture content, among other factors that potentially contaminate the results. Therefore a study like described
only seems reasonable when meticulous records over an extended period of time are collected.
Dust monitors
Many bulk handling plants have installed towers with dust measuring equipment to monitor the amount of
dust in the surrounding air, since dust liberation has negative consequences on the environment and their
employees. The measurement equipment used for dust measurements in the air are called PDRs (Personal
Data Rams), see Figure 3.1 (left). Through emission of pulsed, near-infrared lighting, PDRs are able to perform
opacity measurements to find the amount of dust in the air in mg/m3 [30]. The origin of the dust for these
tower measurements is unknown and can be from stockpiles to handling equipment. Companies that spe-
cialize in transfer chute design use these PDRs to quantify a dust problem near a transfer chute and compare
the results before and after a design change is made. For instance, in 2014, WEBA used this technique when
replacing older coal chutes at Laramie River Station in Wyoming, USA [21]. They placed five PDRs at different
areas around a chute, including by the inlets and outlets, and measure the dust levels for eight hours under
normal operations. Then averages where used to compare the dust levels before and after chute upgrades.
Figure 3.1: Personal Data Ram, PDR, measures dust in air (left) [30], Anemometer, measures air velocity (middle) [1], camera footage
inside chute (right) [7]
Air velocity measurements

As explained in Section 2.3.2, in order to control the amount of dust that is liberated, the air streams around
the material stream are an important influencing factor. The pickup velocity of dust particles is generally
around 1 to 1.25 m/s [1], but still vary widely depending on many factors, such as particle size, material
cohesiveness, and moisture content. Listed among liberation minimization methods for transfer chutes in
Section 2.4 is the minimization of air flow entering and leaving the chute enclosure, as well as keeping the air
velocity inside the enclosure at a minimum. A method to check this air flow on operational transfer chutes
is to monitor the air flow at the entrance and exit of the chute with an anemometer, depicted in Figure 3.1
(middle).
Material sampling
As explained in Section 2.3.2, material degradation and dust liberation go hand in hand. When dust gets
formed through impact, this is also where dust has the potential to be liberated. That is why in some cases
instead of looking at dust in and around the transfer chute, material samples can be taken before and after
transfer chutes to quantify material degradation and provide a relative measure for dust liberation at a trans-
fer chute. A transfer chute upgrade performed by WEBA [22] at the Isdemir Steel Plant in Turkey used exactly
this method to find material degradation after complaints on excessive dust levels, material degradation and
noise levels. The methodology for sampling material is standardized and is outlined in Chapter 4.
Camera’s inside chute

Finding out what happens with the material stream inside a problematic transfer chute is necessary to pro-
vide a solution that properly addresses the problem. This could potentially be accomplished by placing cam-
era’s inside the working chute and examine the footage (see Figure 3.1 (right)), which could also be compared
to calculated analytical or numerical models of the flow. The darkness and dust levels inside the chute can
cause difficulties in obtaining clear footage, and any form of quantification of dust liberation is not possible
using this method.
3.3. Numerical modeling methodologies

With the increase in demand for better performing chutes in regards to performance criteria such as mini-
mum chute and belt wear or minimum dust generation and liberation, simple rules-of-thumb design method-
ologies that permits flow without blockage are insufficiently detailed. This, along with the rise in computa-
tional power, gives rise to numerical modeling methodologies that can incorporate a large number of flow
properties into simulations. Two main approaches exist for building models of granular flow, the Lagrangian
and the Eulerian approach (see Figure 3.2).
Figure 3.2: Mathematical difference between the Eulerian and the Lagrangian approach [31]
The Lagrangian approach models the material stream as discrete particles and calculates their motion
and collisions, whereas the Eulerian approach models the material stream as a artificial, continuous gas-solid
flow stream [2][32]. One widely used modeling methodology for transfer chute design is Discrete Element
Method, which uses the Lagrangian approach, but Computational Fluid Dynamics typically uses a multi-
phase simulation where the granular flow can also be modeled using an Eulerian approach. An overview of
the modeling methodologies used in transfer chute design are listed below in Table 3.2, which are expanded
on in the next section.
Modeling methodology Description Model type
Discrete Element Method Timestep based particle and geome- Granular flow by modeling
try modeling and calculating interaction motion and collisions of
forces discrete particles
CFD Eulerian-Eulerian Predicting fluid flow behavior using Granular flow as continuous
Navier-Stokes equations mass
Coupled DEM-CFD Coupling of numerical methods to pre- Granular flow as particles or
dict granular flow and air flows fluid and air flow as fluid flow
Table 3.2: Overview of modeling methodologies related to dust liberation problems
Discrete Element Method

As explained above, the basic principle of DEM is to model each individual particle as a separate entity that
undergo a range of forces as in reality. These forces typically include gravity, inter particle contact forces,
particle to wall contact forces and cohesive and adhesive forces in particles if applicable. After every time
step in a simulation these forces and their resultant displacements are calculated, making DEM simulations
very computationally intensive. Different contact models are used to define the interaction between particles
and environment, which is visualized in Figure 3.3.
3.3. Numerical modeling methodologies 25
Figure 3.3: Visualization of contact between particles in DEM, including springs, dashpots, friction slider and cohesion bond[31]
When evaluating dust emission in transfer chutes, it is often sufficient to analyze the material flow in
the system using DEM [18]. From Section 2.3.2 it was found that abrupt changes in flow speed, impacts,
as well as periods of free fall that dilate or disperses the material stream are a clear indicator of potential
dust liberation. Applying this knowledge from dust liberation theory, dust liberation causes and redesign
proposals can be made in DEM simulation software. An example of an updated chute design that compacts
the material stream and provide smooth redirection of material flow using a hood and spoon design is shown
in Figure 3.4.
Figure 3.4: Example of a redesign of a transfer chute using DEM software [18]
CFD Eulerian-Eulerain approach

When CFD is used in modeling transfer chutes with material streams including dust or air flows, a multiphase
simulation is required [18]. When the material stream is modeled as a continuous mass, as well as the sur-
rounding air, this is called the Eulerian-Eulerian approach. Typically three phases are modeled, a continuous
phase representing air and two solid dispersed phases of different sizes representing the granular particles
and the dust particles [17][32]. This method can be used to solve flow fields for densely packed solids, how-
ever a major drawback of this method is that no friction can be accounted for between the dispersed solid
phases (granular particles and dust) and the chute wall, since these flows are modeled as fluids[17][18]. The
only options for the wall boundary condition are no slip and free slip, where typically no slip is selected for
the continuous phase, the air flow, while free slip is selected for the solid dispersed phases. Additionally,
only mono-sized particles are involved in these models, while typically material streams contain a large par-
ticle size distribution [33]. An example result of a CFD model of a transfer chute flow is shown in Figure 3.5.
There are several research papers that have used the Eulerian-Eularian approach [17][18][29][32][33], but this
method is not widely used in transfer chute design.
Figure 3.5: Examples of research results using the Eulerian-Eulerian approach in CFD modeling (left[32], right[29])
Coupled CFD-DEM
Another methodology for modeling transfer chutes that can include dust particles and air flows is the Eulerian-
Lagrangian approach. Here the material stream is modeled with Lagrangian-based software such as DEM to
model discrete particles and interparticle and wall interactions, whereas the air and dust flows are modeled
as a fluid using the Eulerian approach. This involves coupling DEM software with CFD software. The infor-
mation exchange between Newton’s laws of motion in DEM and the local averaged Navier-Stokes equations
that are solved in CFD is illustrated in Figure 3.6.
Figure 3.6: Information exchange between CFD and DEM [34])
This method is potentially more accurate than the previously mentioned methodologies, since both indi-
vidual granular particles are modeled, as well as the interacting air and dust flows around the material stream.
However, in the current research literature there is a lack of validation and verification presented [35]. Addi-
tionally, long computation times can be a significant problem in its application of large numbers of particles
or long duration of simulations [2]. Simulation times greatly depend on the CFD mesh size of the calculation
volume domain, as well as the choice of either 2D or 3D meshes, which in turn affect the accuracy of the sim-
ulation [34]. Some academic research that uses both DEM and CFD will opt for first simulating granular flows
in DEM to find its properties (such as porosity[17]), which are then converted for use in Eulerian-Eulerian
multiphase CFD models. Companies have started to implement CFD into DEM simulations as well, where
ANSYS Fluent has been used in combination with RockyDEM for instance by the company Rocky [36]. Dust
flows in and around transfer chutes have been investigated with the use of the Rocky software package.
CFD-DEM coupling definitely shows promise in analyzing airflows at transfer chutes in great detail. The
implementation in the industry will surely grow in the future if accurate models will be able to be produced.
3.4. Analysis Tata Steel chutes 27
3.4. Analysis Tata Steel chutes

With the knowledge of dust liberation investigation methods, the next step is to analyze the situation at Tata
Steel and see which methodology can be applied. An investigation into all the transfer chutes at the section
’Grondstoffenlogistiek’(GSL), or raw material logistics at Tata Steel has been performed, in order to get an
overview of the types and size ranges of its transfer chutes. The results of this analysis have been summarized
in Table 3.3.
Transfer chutes
Number of transfer chutes 137
Number of belt to belt chutes 121
Impact plate, dead box,
Chute types
straight onto belt loading
Drop height range 0.8 m - 6.2 m
Average drop height 2.5 m
Material direction changes unknown
Conveyor belts
Belt speed range 1.05 m/s - 4.05 m/s
Average belt speed 2.18 m/s
Mass flow rate range 90 t/h - 23197 t/h
Most used mass flow rate 1600 t/h
Belt width range 650 mm - 1800 mm
Most used belt width 1200 mm
Throughing angles used 30◦ and 40◦
Material transported Iron ore pellets, sinter,
cokes and iron ore mixture
Support under receiving belt Guirlandes, impact idler sets
regular idler sets and impact bars
Table 3.3: Data on GSL Tata Steel transfer chutes
3.5. Case study at Tata Steel and methodology choices

3.5.1. Case study
Since it is not practical to analyze over a hundred chutes, a case study will be chosen that is the most common
design of transfer chute at GSL Tata Steel. A set of transfer chutes which have noticeable dust liberation (see
Figure 1.2) and received fines for environmental contamination as mentioned in Section 1.2 is investigated.
The chute system that is chosen for the case study is a movable chute that can load material from a top
belt, called A317, onto three belts below, named A508, A509 and A510. The belts and chute in question are
very close to the most common characteristics investigated, shown in Table 3.3. The relevant data of this case
study chute is listed in Table 3.4.
Transfer chutes
Chute type Impact plate
Bottom of chute Deflector plates
Drop height 3.04 m
Material direction change 90◦
Conveyor belts
Discharging belt name A317
Receiving belts names A508, A509, A510
Belt speed 2.19 m/s
Mass flow rate 800 or 1600 t/h
Belt width 1200 mm
Throughing angle 30◦
Material transported Iron ore pellets, sinter,
cokes and iron ore mixture
Support under receiving belts Guirlandes
Table 3.4: Case study chute system data
The material that flows over the chutes usually comes from an outside storage area called EO1, ’Ertsopslag
1’. Many different materials are being transported over this section. A series of eight bunkers with a capacity
of 1600 tonnes are located with their exits on the top belt, A317. When material gets loaded onto the belt, one
or two bunkers can be unloaded onto the belt simultaneously. This results in a mass flow rate that is usually
either around 800 t/h or 1600 t/h, as can be seen in Figure 3.7.
Figure 3.7: Mass flow rates over case study chute (source: Tata Steel control panel)
3.5.2. Methodology choices

From Section 3.2 and Section 3.3, the known measurement and modeling methodologies for this research
have been discussed. After checking the availability of these methodologies, as well as the advantages and
disadvantages of each methodology, grading tables are created, shown in Table 3.5 and Table 3.6.
3.6. Conclusions 29
Measurement methodology Equipment Manpower Time Identification Quantification

Visual inspection ++ ++ ++ − −
Area sweep + + −− −− +
PDR measurements − + + + ++
Air velocity measurements − ++ + −− +
Material sampling + ++ + + +
Camera’s inside chute − + + + −
Table 3.5: Grading of measurement methodologies according to criteria
Modeling
methodology Material Equipment Time Identification Quantification Redesign
DEM ++ ++ + + + ++
CFD + −− ++ + + −
Coupled DEM-CFD ++ ++ − ++ ++ +
Table 3.6: Grading of modeling methodologies according to criteria
Material sampling proves, although not a direct measurement of dust liberation such as PDR measure-
ments, to be the most suitable method in combination with DEM simulations to identify the source of the
dust liberation in Tata Steel’s A317 transfer chute. For measurements in and around the chute, resources for
an area sweep and material sampling are available. However, time constraints eliminate the possibility of the
former. Like many chute redesigns that suffer from flow problems such as blockage, dust liberation or spillage
for example, the investigation of the problems and eventual redesign is performed using granular flow simu-
lations with DEM software [7][17][23]. CFD can model the material as a continuum, but this lacks the crucial
material-equipment interaction calculations to accurately model the material flowing through the chute in
enough detail. Coupled DEM-CFD is not chosen due to the complexity of the simulation setups, where the
analysis and redesign of the entire transfer chute is deemed too large for DEM-CFD models to simulate within
a reasonable, computational time frame.
3.6. Conclusions
This chapter has answered the following research question
involve dust liberation?
Actual measurement of dust is challenging due to the difficulty of pinpointing the source of liberation.
Therefore other measurement tools can also be used besides dust monitoring equipment such as area sweeps,
air velocity measurements and material sampling to find dust creation and prove the source of dust liberation.
Numerical modeling techniques can be a very helpful tool in the assessment of dust liberation at transfer
chutes. DEM, CFD and the coupling of the two methods are capable to varying degrees at modeling granular
and/or dust and air flows inside transfer chutes.
For this particular research, the following conclusions regarding the suitability of these methodologies is
summarized below.
• Dust measurement techniques such as dust monitors or periodic dust collection are not possible for
this research due to a lack of resources
• Material degradation will be investigated through sampling to attempt to gain physical evidence and
quantification of material flow problems leading to degradation and liberation of material
• Modeling chute material flow in DEM software is chosen as investigative method and for eventual chute
redesign steps
• After investigating Tata Steel transfer chutes, a case study chute is chosen that represents the most
common chute design at Tata Steel
4
Quantifying material degradation
The research question to be answered in this chapter is
• What is the extent of material degradation at a Tata Steel transfer chute?
Through the use of sampling according to standardized methods, this chapter will lay out the process of
answering this research question. First the standardized sampling methodology will be explained in Section
4.1. In this section a methodology that fits within the possibilities for this research is chosen as well. Section
4.2 will show the results of the sampling performed, after which a discussion and conclusions on the findings
is expanded on in Section 4.3 and Section 4.4 respectively.
4.1. Material degradation sampling methodology

The sampling of iron ores is standardized in many countries all over the world, where in Europe the standard
ISO3082-2009 Iron ores - Sampling and sample preparation procedures[37] is used. This standard outlines
many different sampling methods, some of which are depicted in Figure 4.1.
Figure 4.1: An exerpt of moving belt (top) and stopped-belt (bottom) sampling methods outlined in ISO3082 (edited from [37])
From this figure it can be seen that large equipment is needed to take samples of material from the belt
and that a distinction is made between moving belt samples and stopped-belt samples. In order to find ma-
terial degradation in samples, the particle size distribution of each sample needs to be determined. When
31
32 4. Quantifying material degradation
samples are taken right before and right after the material passes through the transfer chute, the indication
that material degradation is present can be identified when the percentage of smaller sized particles is in-
creased in the samples taken after the transfer chute.
For this particular investigation into material degradation a stopped-belt sample methodology is chosen
over a moving belt sample. To avoid differences in the ’before’ and ’after’ sample due to other factors beside
the transfer chute, it is important to take the samples as close together along the material stream. When
moving belt samples are taken, this becomes increasingly difficult, due to sample locations being limited to
transfer points. The ’after’ sample would have to be taken at the discharge point of the bottom conveyor,
which is over a 100 m.
Figure 4.2: The sampling tools on the top belt, A317 (left), the process of sampling (top right) and a finished sample(bottom right)
The sampling tools that were used in the stopped-belt sampling are a combination of the bottom two
illustrations in Figure 4.1 and are shown in Figure 4.2. An iron tool is constructed similar to the bottom left
illustration that can isolate a cross-sectional ’slice’ of material to be sampled off the belt, and a small scoop is
available to collect the sample. The smaller particles in the sample are collected with a vacuum cleaner with
a clean vacuum bag, that is released into a bucket with the rest of the sample. The sampling tool sides are
made to match the cross-sectional shape of the conveyor belt with a throughing angle of 30◦ . The width of
the sampling tool, which equals the minimum length along the conveyor of ore to be removed, is dependent
on the nominal top size of the ore according to the ISO standard. This minimum length is at least three times
the nominal top size of the ore, subject to a minimum of 30 mm [37]. With neither pellets, sinter or cokes
being above this size, 90 mm length is sufficient. For the sampling tool a length of 100 mm is chosen. The
mass of each sample in kg is determined by [37] to be
q l2
mi = (4.1)
3.6 v b
Where q is the flow rate in t/h, l 2 is the length in m of the complete cross-section of ore removed from the
conveyor and v b is the speed of the conveyor belt in m/s. With the samples being taken when one bunker is
being emptied, the sample mass over which the PSD is determined becomes
800 t/h · 0.1 m

mi = = 10.15 kg (4.2)
3.6 · 2.19 m/s
4.2. Sampling results 33
The suitable locations for sampling from the belt before and after the chute are found to be at 13 and 12
meters from the chute respectively. First a sample from the top belt is taken, after which the belt is turned on
for a set time and the second sample at the bottom belt is taken. It is important that the second sample is:
• taken as close by the first sample as possible to minimize differences between accompanying samples
• the material that is sampled went through the chute at full speed
• taken after the first sample location has passed to eliminate contamination due to moved material
during the first sample
Using video footage taken of the material flow inside the chute after sampling the top belt, the startup
time for the belt is approximated at 6 seconds. In order to find the time when the location on the material
stream where the first sample was taken passes the second sample location, t between samples , the following
equation is formulated
s between samples − 12 · t startup · v b

t between samples = t startup + (4.3)
vb
where t startup is the startup time, s between samples is the distance between the two sample locations and v b is
the belt speed. This gives a time between sample locations when startup is included of
(13 m + 12 m) − 21 · 6 s · 2.19 m/s

t between samples = 6 s + = 14.42 s (4.4)
2.19 m/s
Therefore after every first sample the belt is turned off after 10s, in order for the belt to be stationary after 16s,
taking into account a stopping time of 6s.
The case study transfer chute transports sinter, iron ore pellets, cokes and different mixtures. The method-
ology described is tested on iron ore pellets, sinter and cokes; different mixtures were unavailable at the time
of research. The methodology proved unsuccessful for cokes only. The lump sizes of coke particles are too
large for the sampling tool to penetrate the material cross-section. Despite a sampling plan where samples
of iron ore pellets and sinter are taken until the samples show a certain measuring accuracy using student
t-tests, the heavy usage of the conveyor belt line at Tata Steel for the continuation of production made it pos-
sible for only two samples of sinter from before and after the chute to be successfully extracted. The samples
are compared and according to student t-tests with a confidence interval (CI) of 95% their accuracy is deter-
mined, as described in [38]. In the form of the student t-test used, the standard deviation, s, is unknown and
estimated based on the sample information [38]. The equations governing this form of the student t-test are
shown below.
t α/2,p
µ = x̄ ± p s (4.5)
n
n x
X i
x̄ = (4.6)
i =1 n
s
1 X n
s= (x i − x̄)2 (4.7)
n − 1 i =1
Where µ is the mean, x̄ is the average of the samples, t a is the student t-test value, n is the number of
samples, s is the estimated standard deviation and x i is a sample output.
4.2. Sampling results

Two samples of sinter from before and after the chute have been collected in the course of this research.
Unfortunately no iron ore pellets have been sampled. The PSD of each sample is determined and the results
of the comparison between them is shown in Table 4.1.
34 4. Quantifying material degradation
Particle A317 A510 Sinter A317 A510 Sinter Total

diameter ’before ’after differ- ’before ’after differ- difference
range chute’ chute’ ence chute’ chute’ ence (with 95%
[mm] m 1.1 m 1.2 m1 m 2.1 m 2.2 m2 CI)
>40 2.3% 0.0% -2.3% 2.9% 0.0% -2.9% -2.6% ±3.8

20-40 9.2% 13.9% 4.7% 9.8% 10.7% 0.9% 2.8% ±24.1
10-20 44.7% 49.0% 4.3% 42.8% 43.4% 0.6% 2.4% ±23.5
5-10 36.5% 32.4% -4.1% 37.5% 39.0% 1.6% -1.25% ±36.2
3,15-5 5.4% 3.5% -1.9% 5.0% 5.0% 0.0% -0.95% ±12.1
<3,15 1.9% 1.2% -0.7% 2.0% 1.9% -0.1% -0.4% ±3.8
Table 4.1: Difference in percentage between samples of different particle diameter ranges
From the total difference taken from the sinter samples m 1 and m 2 it can be seen that there is an increase
in particles in the 10-40 mm range, while there is a decrease in particles of smaller diameter, as well as in
particles over 40 mm. This is counter to the idea that particles are degraded through the chute, which would
lead to an increase in smaller particles.
It can be seen from the confidence intervals in Table 4.1 that is found between the difference in the sinter
samples m 1 and m 2 that the results are still very inaccurate. This is obvious with only two samples, but the
accuracy can potentially increase with more sample results being incorporated. It appears from the CI results
that they could be linked to the PSD values, but tracing the results back to the biggest anomaly in the data, it
appears that this is due to the differing values of m 1.2 in the diameter range of 5-40.
4.3. Discussion
Due to the lack of samples taken the accuracy of the sample results is not sufficient to draw pertinent con-
clusions. The methodology applied will most likely require a significant amount of samples in order to be
sufficiently accurate, since it is prone to measurement errors. In order to sample the exact slice of material
caught in between the sampling tool, no dust particles should be left on both the belt or the vacuum cleaner
bag. Placing the sampling tool onto the belt requires a significant amount of force, which could lead to po-
tentially damaging the material and contaminating the results. It is also difficult to lower the tool right onto
the belt without small particles rolling underneath the tool into the sample when material is scooped out
between the tools plates, which can happen in various degrees per sample. With a larger sampling tool and
many more samples taken the likelihood to get results within a reasonable range with CI of 95% is increased.
These result data can be used to get a good indication of the overall particle size distribution of a material,
which can be used in material simulation calibrations for instance. This is done in Section 5.2.
4.4. Conclusions
The research question to be answered in this chapter is
• What is the extent of material degradation at a Tata Steel transfer chute?
Unfortunately this question is not answered to a satisfactory degree with the methodology used in this re-
search. Conclusions that can be drawn from this chapter are
• Stopped-belt sampling according to ISO3082-2009 is possible on iron ore pellets and sinter, but not
with cokes. This is due to the large lump size of cokes, making the penetration of the material on the
belt with the sampling tool not possible.
• Using a stopped-belt sampling tool of 100 mm width to determine PSD within a reasonable range with
CI of 95% requires more than two samples. It is unclear how many samples are required.
• To minimize sample inaccuracies between samples before and after a chute the samples need to be
taken as close together as possible, while still having the material travel through the chute at full speed
and while the location on the material stream where the first sample was taken has passed the second
location before the second sample is taken.
5
Calibration dynamic material model
In order to properly redesign a chute and find current chute flow problem causes, it is essential to get a ma-
terial model that accurately represents the real life bulk material flow behavior. The research question to be
answered in this chapter is
In order to answer this research question, an appropriate experimental test setup needs to be found. The
chosen setup and the following experimental plan for the calibration is found in Section 5.1. The calibration
process and step wise calibration is described in detail in Section 5.2.
5.1. Experimental setup and working principles

5.1.1. Inclined surface wear test setup
In the process of calibration experimental setups are used to record material behavior, which can be matched
in DEM simulation software to accurately model particle and bulk behavior. On iron ore many calibrations
have been performed using a wide variety of experiments [12][31][39]. There are tests to study single particle
properties such as pin-on-disk tribometer for the measurement of friction coefficients or drop tests to find
restitution coefficients [39], but also many bulk material tests to find bulk material properties such as bulk
density tests, angle of repose tests or penetration tests for examples [31]. Usually a combination of tests is
used to find a suitably calibrated model. An important aspect of calibration is to use testing methods that
will accurately portray material behavior after calibration in the aspects that the calibrated material model is
eventually used for. For instance, when material is being tested when it is picked up by a grab, a penetration
test for calibrating the interactions between the grab and the material is very suitable. Since the material in
this research is being used in a dynamic setting where it is always moving and interacting with the belt and
chute surfaces, a dynamic test setup should be selected that can model these interactions between belt, chute
and material. For this purpose the inclined surface wear tester is chosen. The wear tester and its dimensions
shown on the geometry that is made in DEM is depicted in Figure 5.1.
Figure 5.1: Picture of inclined surface wear tester during a test with sinter (left) and a simulation of the tester in DEM (right)
35
36 5. Calibration dynamic material model
This tester rotates the top bin (see Figure 5.1) to pour the material into the bin below. This allows the
material to be studied as it collapses, slides and/or rolls from one bin to the other. Even though the chute
that is investigated transports iron ore pellets, sinter, cokes and different mixtures, for this research pellets
and sinter are chosen. Cokes consists of lumps that are too large in relation to the test set-up and different
mixtures are unavailable at the time of research. The material that is tested is taken from the top belt case
study chute and weighs 8 kg. The top bin has an interchangeable bottom surface of steel and rubber. Steel
is used to represent the chute material, while rubber is used as bottom of the top bin to represent the belt
interaction.
5.1.2. Performance indicators and test procedure

Four combinations of tests are performed using sinter and pellets and using rubber and steel as bottom sur-
face for the top bin. Every combination is tested seven times and the range of accuracy of the resulting
performance indicators is determined with a confidence interval of 95%. Each test has eight performance
indicators, which are depicted in Figure 5.2. Some picture are simulation screenshots, in order to explicitly
show the movement that is recorded as dynamic performance indicator.
As can be seen in Figure 5.2, the material is loaded as a fixed sized box shape on the left side of the top
bin. The bulk density check is the first performance indicator. The height of the box is measured and used
to visually check the bulk density when compared to the simulations. In the simulations the bulk density is
determined within this box of material and calibrated to bulk density values found in literature. Then the
plastic plate in front is swiftly taken away for the material to collapse on one side. The resulting angle of re-
pose that the material forms is the second performance indicator for the material calibration. The inclined
surface tester is turned on and the top bin starts to rotate at a speed of 1◦ /s. The material is filmed during the
rotation as the pile collapses and falls into the bottom bin. This is done to identify four dynamic performance
indicators, shown and explained in Figure 5.1 with an accompanying DEM simulation screenshot for illus-
trative purposes. These indicators are timed using the inclination angle that the top bin has at the moment
of occurrence. When finally the material is all poured in the bottom bin, two final performance indicators
in the form of angles of reposes can be determined. The most important experimental setup data is listed in
Table 5.1. The results of the experiments along with their accuracy using a 95% CI are found alongside the
simulation results in Section 5.3 and Section 5.4.
Experimental test setup data
Bulk material used • Sinter

• Iron ore pellets
Bottom surface tester used • Rubber
• Steel
Material weight 8 kg
Rotational speed 1◦ /s
Number of repetitions 7
Number of performance indicators 8
Confidence interval 95%
Table 5.1: Some basic data on the experimental setup

5.1. Experimental setup and working principles 37
Picture of performance indicator Performance indicator Description

Boxed-in material to check
• Bulk density check bulk density in simulation
and visually check volume
size
One side of box is taken

• First angle of repose (Φ1 ) away and results in an an-
gle of repose
Inclination angle when

• First material-material material starts to slide/roll
shift over each other, but not
yet sliding or rolling over
the top bin surface (notice
the blue particles on the
bottom)
Inclination angle when

• First material-plate shift material starts to slide/roll
over the top bin surface for
the first time
Inclination angle of the top

• Biggest material shift bin when the biggest shift
in material happens, al-
ways a combination of slid-
ing/rolling between parti-
cles and between particle
and bottom surface
Inclination angle of the top
• Top bin empty bin at the moment when
all the material has been
poured into the bottom bin
and the top bin is empty
When all the material is

• Second and third angle of poured into the bottom
repose (Φ2 and Φ3 ) bin, the pile that remains
forms two angle of reposes
Table 5.2: Performance indicators inclined surface wear tester

5.2. Calibration setup

Armed with the experimental data, calibration of the bulk materials inside the DEM simulation environment
can begin. Often in a material calibration, first the particle properties are tested with experiments and used
as input for the simulation model, after which the bulk behavior is checked with reality and final adjustments
are made. In this research the testing of particle properties is taken from extensive
First the simulation material parameters that will be adjusted for calibration are determined. Then using
data gathered from literature and sampling, a reference case is made as a starting point in the calibration
process. Finally the procedure and reasoning for the calibration methodology used is explained.
5.2.1. Calibration parameters

The material parameters that will be varied for the calibration of both iron ore pellets and sinter are listed in
Table 5.3. These are considered to be the values that are of influence to the bulk behavior that is examined in
the experiments and will be used for calibration.
Calibration parameters
• Particle density
• Static friction particle - particle
• Rolling friction particle - particle
• Static friction particle - steel
• Rolling friction particle - steel
• Static friction particle - rubber
• Rolling friction particle - rubber
Table 5.3: Calibration parameters
5.2.2. Reference case

The next step in the calibration process is to define a reference case. This is a working model of the bulk
material, which can then be calibrated through trial and error. The reference case consists of particle and
bulk material parameters for sinter and iron ore pellets and their interaction parameters with both rubber,
steel and plastic. The plastic properties are needed for the walls of the surface tester.
Particle shape and size distribution

The particle size distribution that is used of the simulation material models for iron ore pellets is found in
literature. An accurate representation of the PSD for pellets from Tata Steel was found in a PhD research
by Stef Lommen [31] and is depicted in Figure 5.2 (left). The approximation is normal distribution with an
average diameter of 11 mm and a normalized standard deviation of 0.1 and has an R 2 of 0.9997.
Figure 5.2: Particle size distribution of iron ore pellets [31] (left) and sinter (right)
For sinter, a similar approximation was made based on the sample results shown in Section 4.2 in Table
4.1. All the PSDs from the sample results are plotted and approximated with a normal distribution. The results
are shown in Figure 5.2 (right). The normal distribution has an average particle diameter of 10.8 mm (the
d 50 result from sampling, which is 50% of cumulative particle mass in a PSD), with a normalized standard
5.2. Calibration setup 39
deviation of 5, resulting in an R 2 of 0.9877. This method is preferred due to its significant simulation time
savings when compared to an input of the PSD manually into DEM.
Particle shape
The d 50 for sinter is taken from the average results of the sampling performed in this research, described in
Chapter 4, which is 10.8mm. The d 50 for iron ore pellets is taken from Lommen [31], which studied pellets
originating from Tata Steel. With the shape of iron ore pellets being fairly round, the pellets are chosen to be
simulated as spheres. This can prevent excessive simulation times when large bulk quantities of pellets are
to be placed in transfer chute simulations. With sinter particles being very sharp and rough and having no
round shape whatsoever, the model of a sphere is less accurate. In order to get more accurate bulk behavior,
it is decided to model sinter particles as an irregular shape, despite the added simulation computation times.
A three-spherical shape (see Figure 5.3) is chosen to have the minimum amount of shapes per particle, while
eliminating the natural inclination for rolling that a single sphere has.
Figure 5.3: Three-spherical model for sinter particle
Reference case calibration parameters

The values that are picked for the reference case along with their source is listed in Table 5.4. The remainder
of the simulation settings and material properties are listed in Appendix A. The friction coefficients for sinter
in Table 5.4 are selected slightly lower than the literature [40][41], where spherical particles were calibrated as
opposed to the irregular sinter particles in this research. No literature is found for the interaction coefficients
between sinter and rubber and are therefore estimated.
Pellets Sinter
Calibration parameters Source pellets / sinter
value value
• Static friction particle - particle [-] 0.45 0.3 [31], [42] / [40], [41]
• Rolling friction particle - particle [-] 0.15 0.2 [31], [42] / [40], [41]
• Static friction particle - steel [-] 0.5 0.5 [39], [42] / [40], [41]
• Rolling friction particle - steel [-] 0.25 0.2 [39], [42] / [40], [41]
• Static friction particle - rubber [-] 0.71 0.5 [39], [42] / *
• Rolling friction particle - rubber [-] 0.29 0.4 [39] / *
• Particle density [kg/m3 ] 3700 2650 [31], [42] / [42]
Table 5.4: Reference case calibration parameters (*explained in text)
5.2.3. Calibration procedure

A calibration with multiple variables can be complex and call for a design of experiments methodology to
find the factors of influence on the results. However, with the effects of the parameters used as variables
being known, and due to constraints in time, a design of experiments is foregone and from a reference case a
manual calibration process through trial and error is adopted.
The reference case is run through a simulation and the results are compared with the performance indi-
cators determined in the experiments. Then the calibration parameters are adjusted according to the type
and extend of the deviations found between experiment and simulation results. These parameters can be
increased when a certain behavior in simulations is found, which is described in Table 5.5. Since some ad-
justments can be made in similar situations, one has to be chosen. One type of behavior that can lead to an
increase in all variables, apart from particle density, is the biggest shift in material happening too late in the
simulation. With this knowledge, combined with proper analysis of the simulations, a calibration using trial
and error is performed.
Calibration parameters Value increased in simulation when

• Static friction particle - particle First mat-mat shift is too late and/or when AoRs are too low
• Rolling friction particle - particle First mat-mat shift is too late and/or when AoRs are too low
• Static friction particle - steel First mat-plate shift is too late and/or the top bin is empty too early
• Rolling friction particle - steel First mat-plate shift is too late and/or the top bin is empty too early
• Static friction particle - rubber First mat-plate shift is too late and/or the top bin is empty too early
• Rolling friction particle - rubber First mat-plate shift is too late and/or the top bin is empty too early
• Particle density Bulk density is too low
Table 5.5: Calibration parameters and their corresponding adjustment reasoning (mat = material, par = particle, AoR = angle of repose)
5.3. Calibration process

In Table 5.6 the trial and error calibration process steps for iron ore pellets with a steel inclined surface tester
bottom is shown. Simulation 1 starts with the reference case data and the results are compared with the ex-
perimental results, with the aim of getting the simulation results within the range of the 95% CI experimental
results. Almost all results are too high and therefore friction coefficients need to be lowered, where in this case
the static inter-particle friction and the static ore-steel friction is chosen. Simulation 2 already shows results
much closer to the experiments with regards to the ore-steel interaction. Still the AoRs and the first mat-mat
shift is too high, so the static inter-particle friction is slightly lowered for the next simulation. Simulation 3 has
most performance indicators within range, with only a notable discrepancy with the empty box inclination
angle. This phenomenon is traced back to the last stage of the emptying of the box, where the particles are
dispersed and only a single or double layer of particles is left. This most likely causes the particles to slide
easier than in the experiments. With a focus on the bulk material dynamics, this is deemed less significant
than the earlier phase of the test. Also the final AoRs vary from the experiments, which is most likely due
to an imperfect model of the the plastic container bottom of the bottom bin. Therefore the first AoR in the
simulations is taken to be more representative than the final two AoRs. The third simulation is repeated three
times to verify the calibration simulation results. The 95% CI is calculated from this and is shown in Table
5.6 under Sim 3. The thought process for the remainder of the calibrations is similar, of which the resulting
tables can be found in Appendix C.
Calibration parameters Sim 1 Sim 2 Sim 3

• Static friction particle - particle [-] 0.45 0.35 0.3
• Rolling friction particle - particle [-] 0.15 0.15 0.15
• Static friction particle - steel [-] 0.5 0.3 0.3
• Rolling friction particle - steel [-] 0.25 0.25 0.25
Performance indicators Experiments

• First angle of repose, Φ1 30.9 ±0.2 34.4 33.2 30.7 ±0.3
• Inclination angle at first mat-mat shift 8.4 ±0.3 13.8 9.6 9.0 ±0.7
• Inclination angle at first mat-plate shift 11.3 ±1.9 18.7 10.1 9.2 ±0.6
• Inclination angle at biggest mat shift 14.3 ±1.2 19.6 15.4 14.7 ±0.2
• inclination angle at empty top box 22.4 ±0.2 24.5 19.7 19.7 ±0.1
:::::::
• Second angle of repose, Φ2 22.8 ±1.6 29.5 24.6 24.6 ±1.0
• Third angle of repose, Φ3 32.2 ±0.7 41.5 36.9 34.2 ±0.3
:::::::
Table 5.6: Calibration process iron ore pellets - steel

5.4. Results 41
5.4. Results
A comparison of the experimental and calibrated model results are shown in Table 5.7. In Figure 5.4 and
visual comparison can be found between experiment and simulation of the first angle of repose.
Iron ore pellets - Steel Iron ore pellets - Rubber
Performance indicators Experiment Simulation Experiment Simulation

• First angle of repose, Φ1 30.9 ±0.2 30.7 ±0.3 32.9 ±0.3 32.8
• Inclination angle at first mat-mat shift 8.4 ±0.3 9.0 ±0.7 9.0 ±0.8 9.5
• Inclination angle at first mat-plate shift 11.3 ±1.9 9.2 ±0.6 13.5 ±2.1 13.1
• Inclination angle at biggest mat shift 14.3 ±1.2 14.7 ±0.2 14.2 ±1.0 15.3
• inclination angle at empty top box 22.4 ±0.2 19.7
:::
±0.1 23.9 ±0.6 24.2
• Second angle of repose, Φ2 22.8 ±1.6 24.6 ±1.0 19.7 ±1.8 24.1
:::
• Third angle of repose, Φ3 32.2 ±0.7 34.2
:::
±0.3 29.6 ±1.9 34.6
:::
Sinter - Steel Sinter - Rubber
Performance indicators Experiment Simulation Experiment Simulation

• First angle of repose, Φ1 35.9 ±0.4 35.4 35.2 ±0.4 35.1
• Inclination angle at first mat-mat shift 13.8 ±0.5 13.9 13.3 ±1.5 13.1
• Inclination angle at first mat-plate shift 18.2 ±0.7 20.5
:::
22.2 ±2.4 23.8
• Inclination angle at biggest mat shift 23.1 ±0.7 23.0 28.3 ±1.8 27.1
• inclination angle at empty top box 31.3 ±0.2 30.2
:::
35.1 ±0.5 35
• Second angle of repose, Φ2 32.1 ±1.5 38.0
:::
27.8 ±0.9 36.5
:::
• Third angle of repose, Φ3 34.9 ±1.3 34.2 33.6 ±0.7 41
::
Table 5.7: Comparison of experimental and calibrated model results
Figure 5.4: Comparison of first angle of repose, Φ1 , in experiment and simulation
The performance indicator results that fall outside the experimental data range are indicated with a red
underline in Table 5.7. The focus for the calibration was put on the first angle of repose and the four dy-
namic indicators, whereas the last angle of repose was secondary due to the inaccuracy of the bottom bin
parameters. The final calibrated parameters are shown in Table 5.8.
Calibration parameters Iron ore pellets Sinter

• Static friction particle - particle [-] 0.3 0.25
• Rolling friction particle - particle [-] 0.15 0.15
• Static friction particle - steel [-] 0.3 0.5
• Rolling friction particle - steel [-] 0.25 0.3
• Static friction particle - rubber [-] 0.4 0.6
• Rolling friction particle - rubber [-] 0.29 0.5
• Particle density [kg/m3 ] 3700 2850
Table 5.8: Results calibration parameters

5.5. Discussion
While good agreement with the experimental results is found in the simulations, a sensitivity analysis of the
calibration parameters and their level of influence using a design of experiments can be very useful in a more
accurate calibration. The bottom bin angle of repose is deemed not a good indicator. This is not only because
in this research the plastic bin was not calibrated. It is found that the bottom surface of the receiving bin was
very bouncy for the iron ore pellets and sinter used and caused the material to spread out very wide within
the bin, as seen in Figure 5.5. The irregular shape of the iron ore and sinter in reality caused them to bounce
in every direction, which was not possible to replicate in the simulations. Another factor making these angle
of repose measurements difficult is the speed increase in the material throughout the dropping time into
the bin, as depicted in Figure 5.6. Notice the increase in speed as well as the more forward trajectory of the
particle stream.
Figure 5.5: Difference in dispersion of iron ore pellets in bottom bin (left),
Bounciness of plastic in bottom bin not present in simulation (right)
Figure 5.6: Material trajectory difference 0,8s apart
In reality this caused the material to spread out even more, making a less distinctive material cone. In the
simulation a similar phenomenon occurred, but the cone stayed more compacted, most likely due to a higher
friction of the bottom surface. This results in higher angle of reposes recorded in the simulations compared
to the experiments. Another caveat in the validity of the angle of repose measurements is the high ratio of
the particle size to the height of the material piles which make it difficult to get an exact measurement. This
effect is visible in Figure 5.4. Finally, the camera perspective distortions could have a minor impact on the
experimental results.
5.6. Conclusions
In this chapter the following research question was posed:
A dynamic calibration of iron ore pellets and sinter is performed on an inclined surface tester with steel
and rubber surfaces. This test setup is suitable to get static indicators in the form of angles of reposes, while
simultaneously record dynamic flow behavior when time is a factor. Verification of the simulations was done
by comparing results of three repeated simulations. The trial and error methodology used leaves sensitivity
analyses and the level of influence each parameter has out of the calibration process, which is a strong rec-
ommendation for future research. Eight performance indicators are used in combination with six calibration
parameters, and while good agreement was found between experimental and simulation results, not all pa-
rameters are calibrated within the experimental range of results. The bottom bin angle of reposes are not
prioritized due to several reasons, outlined in Section 5.5.
6
DEM analysis Tata Steel chute
In this chapter the aim is to answer the following research question:
• Where along the material stream in a transfer chute at Tata Steel is potential dust creation and lib-
eration and what is the influence of different materials and mass flow rates in DEM software on this
problem?
The key performance indicators, KPIs, used to investigate the DEM chute model are explained in Section
6.1. The geometry modeling and simulation setup for the DEM simulations of the Tata Steel case study chute
is expanded on in Section 6.2. Results of the simulations performed, along with discussion and conclusions
is shown in Section 6.3, 6.4 and 6.5 respectively.
6.1. Key performance indicators

With the dynamically calibrated material models for iron ore pellets and sinter from Chapter 5, the material
flow through the case study transfer chute at Tata Steel is simulated using DEM software. The influencing
factors in the creation and propagation of dust is found in Chapter 2 to be:
• Minimize material impact

• Minimize free fall of material
• Compact containment of flow stream
To find where these factors are present in the current chute design, key performance indicators, or KPIs
for short, are defined for quantification of the dust liberation potential. These primary KPIs are listed below
along with a short description:
• Velocity fluctuations in material flow
– General overview of the ’smoothness’ of the flow, big fluctuation spikes indicate impacts
• Impact velocity and reaction force
– Calculates the impact normal reaction force for quantification

• Bulk density along material trajectory
– Measures the compactness of the material along its trajectory, where bulk density drops indicate
air entrainment potential
43
44 6. DEM analysis Tata Steel chute
From Section 2.2 many other chute design objectives should be met for a properly functioning chute.
Therefore the following, secondary KPIs are introduced for comparison with the redesign and for a better
understanding of the quality of the simulation model.
• Belt center loading

• Segregation
• Chute wear
6.2. Simulation setup

Geometry
With the aid of design drawings from Tata Steel and the use of on site measurements, a close approximation
of the case study transfer chute geometry is made. To convey the layout of the A317 movable transfer chute,
Figure 6.1 depicts the Tata Steel design drawings of the chute, along with pictures of the current state. The
design drawing is edited to represent the current situation, since many adjustments have been made on the
design throughout its operation. The head chute (blue in Figure 6.1) has a v-shaped impact plate inside and
moves with the drive pulley and the top belt, A317, over a rail (green in Figure 6.1) to locate itself above one
of the bottom belts. Stationary deflector plates above each bottom belt help to guide the material onto these
belts.
Figure 6.1: Schematic overview of the chute geometry configuration of the A317 case study chute with two of the three bottom belts.
A509 and A508 depicted underneath
The geometry created in EDEM features only the parts of the geometry that are in contact with the mate-
rial, which are all the parts indicated in red in Figure 6.1. The geometry as created in EDEM with a completed
simulation of the material flow is shown in Figure 6.2.
6.2. Simulation setup 45
Simulation
As shown in Figure 3.4, the case study chute transports material from one or two bunkers simultaneously,
which translates to a mass flow rate of 800 t/h and 1600 t/h. Table 3.4 shows that the case study chute trans-
ports iron ore pellets, sinter, cokes and iron ore mixtures, of which iron ore pellets and sinter have been
calibrated in Chapter 5. In order to assess the influence of different materials and different flow rates on the
KPIs, there will be four simulations that will be run with the following combinations:
• Iron ore pellets with a 800 t/h mass flow rate

• Sinter with a 800 t/h mass flow rate
• Iron ore pellets with a 1600 t/h mass flow rate
• Sinter with a 1600 t/h mass flow rate
To achieve a continuous flow of bulk material from the top belt onto the bottom belt with the smallest
amount of particles in the simulation to limit computation time, only a very short part of the belts is modeled.
The placement of a particle factory at the top belt is at a distance from the pulley such that the material is
settled on the belt before discharge. The high density of sinter required a larger shaped particle factory and
a higher speed at creation for the flow rate of 1600 t/h to be achieved, as depicted in Figure 6.2. Mass flow
rate sensors at the discharge and bottom belt are placed and when at both points the flow is constant at the
desired flow rate, the simulation is stopped. In Figure 6.2 this is depicted in the graph which shows the 800
t/h simulations for iron ore and sinter.
Figure 6.2: Case study chute simulation, particle factories (left), mass flow rate sensor data (right)
It can be seen from the mass flow data on the left in Figure 6.2 that both bottom bins reach the same
flow rate of 800 t/h that is measured before discharge and therefore continuous flow is achieved without
spillage, similar to what is observed at Tata Steel. The impact point on the impact plate is measured at 1.6 m
downwards from the top of the discharge pulley, which is close to what is measured inside the chute at Tata
Steel, although measuring this point in real life presented accuracy problems. The wear and material residue
found on the deflector plate, as shown in Figure 6.1 on the bottom left image, is also found in the simulations,
as shown in the results in the next section.
6.3. Methodology and results

This section presents the methodology and underlying theory for the chosen KPIs alongside the results from
the simulations.
6.3.1. Velocity fluctuations

To graphically represent the bulk material speed as it navigates through the chute, the average velocity of a
small selection of particles is tracked. The results are presented in Figure 6.3, along with a visual aid of the
side view of iron ore pellets at 800 t/h.
Figure 6.3: Graph of average velocity profile of a small material selection (left), side view of pellets, 800 t/h (right)
Figure 6.4: Vector plot comparison of cross-sectional slice of the material stream in contact with impact plate (left), angular particle
velocity comparison (right) with rotation around the horizontal axis parallel to impact
As can be seen in Figure 6.3, the particles are part of a continuous flow as it starts and ends at the same
speed of the top and bottom belt at 2.19 m/s. Indicated in the graph and in the side view image in Figure
6.3, it can clearly be seen that there are three impacts that the material stream encounters throughout the
chute, onto the impact plate, the deflector plate and the bottom belt respectively. The higher flow rate sim-
ulations reach the impact plate slightly earlier, since they have a ’taller’ material stream cross-section. Sinter
particles decelerate faster upon impact compared to iron ore, indicated by the steeper declines in the veloc-
ity profile. The high flow rate simulations, especially for sinter, show a very mild bump and no acceleration
in the velocity profile at the third impact. A better view of the velocity difference at impact is displayed in
Figure 6.4, while the difference at the third impact is shown in Section 6.3.4, where results on belt loading
aspects are presented. The vectorplots on the left in Figure 6.4 show a larger deceleration at impact, which
is less prevalent in particles furthest from the impact plate. The stream thickness after impact is significantly
6.3. Methodology and results 47
smaller for iron ore. Angular velocities (Figure 6.4, right) show little difference, besides a more defined line
where particles rotate, ’splitting’ the stream before and after impact sideways.
6.3.2. Impact velocity and force

To quantify the level of impact that occurs in the simulations, a method presented by Korzen [43] is used, also
exemplified in an impact research performed by Wypych et al. [44]. It uses the impulse-momentum equation
to calculate the magnitude and direction of reaction forces exerted by bulk material striking a flat surface.
Figure 6.5 shows a schematic drawing of the forces present at impact.
Figure 6.5: Schematic drawing of forces present at impact (left: side view, right: top view)
The impulse-momentum is given by

X ¡−
→ − →¢
F = Vp − Va m s (6.1)
where Vp and Va are the incoming and outgoing bulk velocities respectively and m s the mass flow rate. The
normal reaction force is determined from this theory to be [43][44]
R nx = ρ b A p Vp Vpx = m s Vpx
¡ ¢
(6.2)
Where ρ b is the bulk density and A p the cross-sectional area of the incoming particle stream. This equation
is adjusted for the angles α and β, shown in Figure 6.5, to yield the normal reaction force to the impact plate
R n = m s Vpx sin β = m s (Vpx n cos α) sin β (6.3)
The x-component of the stream velocity, Vpx is extracted from the simulation using bins placed right
above the impact location. The impact angle β was found to be 28◦ and the angle α from Figure 6.5 is 15.3◦ .
The reaction forces are calculated using Equation 6.3 and displayed in Table 6.1. The same calculation is
performed for the deflector plate, except the downwards velocity component, Vp y , is extracted from the sim-
ulations. The angle the deflector plate makes with the material is 35◦ , ultimately finding the normal compo-
nent of the velocity to the deflector plate, Vp y n . The same mass flow rates are found when measurements are
conducted right before impact on the deflector plate.
Impact plate Deflector plate

Vpx n [m/s] R n [N] Vp y n [m/s] R n [N]
Pellets, 800 t/h 2.19 469 4.75 738
Sinter, 800 t/h 2.10 451 4.49 698
Pellets, 1600 t/h 2.28 976 4.94 1538
Sinter, 1600 t/h 2.26 970 4.63 1440
Table 6.1: Velocities normal to impact and reaction forces from impacts
From Table 6.1 it can be seen that the higher flow rates produce higher reaction forces, as can be expected.
It is noted that the reaction forces for sinter are lower than for iron ore, with an increasing difference between
them from impact plate to deflector plate.
6.3.3. Bulk density along material trajectory

Several bulk density sensors are positioned along the material stream at locations of interest. These locations
are displayed in Figure 6.6 marked A to G. The sensors are small cubes with 70 mm sides, placed inside the
material stream, ensuring they are completely engulfed by the stream. The locations are chosen right before
and after free fall of material or before and after an impact.
A large decrease of over 60% bulk density is observed from the discharge to the impact plate, whereas the
bulk density when the material slides alongside the impact plate is measured around 15%. It is also found
that the bulk density increases shortly after the material impacts the impact plate and the deflector plate,
with a noticeable difference between sinter and iron ore pellets for both flow rates, where sinter has a bigger
bulk density increase.
Figure 6.6: Results of bulk density variantions along the material trajectory
6.3.4. Belt loading and segregation

To measure the extend of off-center loading in the simulations, the flow over the bottom belt is split through
the middle using two bins (indicated in blue in Figure 6.7) and their mass over time is traced. The bins are
placed as far away from the loading point as possible, to avoid interference of the loading phase on the mea-
surement of the loaded belt further downstream. After steady, continuous flow is reached in the simulations,
the average of the measured mass in that phase is recorded. The results, alongside side view images of each
simulation configuration are presented in Table 6.2 and Figure 6.7.
Looking at the material on the bottom belt, it can be seen that the color of the material corresponds with
the belt speed of 2.19 m/s. Figure 6.7 indicates that there is a significant off-center belt loading present in the
simulations. This is corroborated and quantified by the ratios in Table 6.2, showing a big increase in off-center
belt loading with 800 t/h simulations as opposed to the higher flow rate, while only a very minor increase in
iron ore as opposed to sinter.
The same bins are also used to measure the average particle mass in each bin over time. The average of
the continuous flow phase is taken for each result. The results are presented in Table 6.3.
6.3. Methodology and results 49
Ratio
left : right
Iron ore, 800 t/h 2.69 : 1
Sinter, 800 t/h 2.58 : 1
Iron ore, 1600 t/h 1.76 : 1
Sinter, 1600 t/h 1.42 : 1
Table 6.2: Mass

distribution ratio of left to
right bin Figure 6.7: Comparison of belt loading of each simulation
Left bin Right bin

Pellets, 800 t/h 51.8% 48.2%
Sinter, 800 t/h 51.5% 48.5%
Pellets, 1600 t/h 50.0% 50.0%
Sinter, 1600 t/h 49.9% 50.1%
Table 6.3: Percentage of average particle mass, segregation results
6.3.5. Chute wear

Using the Hertz-Mindlin geometry contact model with archard wear provides a relative measurement of the
archard wear on the chute geometry. The geometries that are checked for wear are meshed in ANSYS using
CFD-fluent mesh with high smooth sizing and a max face size of 6 mm [12]. The archard equation for the
volume of material removed, Q is defined as [45]
Q = W Fn d t (6.4)
Where F n is the normal force, d t is the tangential distance moved and W is a wear constant, which is equal to
a dimensionless constant divided by a hardness measure of the softest surface. W is the input for EDEM and
is kept constant for both materials at 1 Pa−1 , since the wear results will only be used as an indication of wear
locations and for relative quantification. The normal force F n is a function of the overlap δn and is calculated
by the Hertz-Mindlin contact model in EDEM along with a corresponding damping force F nd using
4 p 3
F n = E ∗ R ∗ δn2 (6.5)
3
r
5 p −→
F nd = −2 β S n m ∗ v nr el (6.6)
6
Where E ∗ is the equivalent Young’s Modulus given by 1/E ∗ = (1 − v i2 )/E i + (1 − v 2j )/E j , R ∗ is the equivalent ra-
dius which is equal to 1/R ∗ = 1/R i +1/R j , and m ∗ = (1/m i +1/m j )−1 is the equivalent particle mass, where E i ,
v i , R i , m i and E j , v j , R j , m j are the Young’s Modulus, Poisson ratio, radius and mass of each sphere in con-
tact with another respectively. When a particle-wall interaction is calculated, the wall radius becomes R j = ∞
−→
and therefore R ∗ = R i , which similarly makes m ∗ = m i the equivalent mass. v nr el is the normal component of
the relative velocity and β and the normal stiffness S n are given by
−l nC R
β= p (6.7)
l n 2 C R + π2
p
S n = 2E ∗ R ∗ δn (6.8)
With C R being the coefficient of restitution between the interacting entities.
Figure 6.8: Relative archard wear comparison of the chute’s impact plate and deflector plate, with a cross-section view of the material
stream
Impact plate Deflector plate Bottom belt

Pellets, 800 t/h 2.30 1.53 0.53
Sinter, 800 t/h 1.38 0.52 0.17
Pellets, 1600 t/h 3.33 1.88 0.76
Sinter, 1600 t/h 1.83 0.52 0.20
Table 6.4: Normalized wear gradients
The wear that results from these calculations in EDEM is depicted in Figure 6.8 as a comparison between
iron ore and sinter. Surprisingly the relative wear in iron ore appears more pronounced than its sinter coun-
terpart, which is contradictory to what is experienced in the industry. From the impact plate wear, it is noted
that the wear pattern is thicker in the flow direction. The cross-section of material on the deflector plate
shows that there is a layer of sinter particles in contact with the plate that is not moving (blue particles), an
effect that is much less pronounced in iron ore, if present at all.
To compare all configurations in regards to wear, the graphs for archard wear are studied. When graphed,
the archard wear progresses in a linear fashion. The gradient of this linear path is extracted and presented in
Table 6.4. The wear increases with increased flow rates, except for sinter on the deflector plate and bottom
belt, the latter showing only a very minute increase.
6.4. Discussion 51
6.4. Discussion
This section discusses the KPI results and their validity, while trying to understand what happens in the sim-
ulations.
• The results suggest that the modeled sinter tends to build up material at impact zones, where iron
ore flows more fluidly. This is evidenced by the higher deceleration of sinter at the impact plate and
deflector plate as compared to iron, as well as double the bulk density increase at these locations. The
explanation for this speed decrease and compacting can be the higher friction coefficients sinter has to
steel. Another reason can be the irregular particle shape of sinter particles that causes them to bounce
away from the bulk stream direction, creating more inter-particle bounces that obstruct the steady
material stream as well as slow the particles down by loss of energy.
• At the deflector plate the brunt of the impact shifts from particle-to-wall towards particle-to-particle,
since a stagnant layer of sinter particles resides on top of the deflector plate (shown in Figure 6.8),
protecting the plate from archard wear.
• The analytical calculation for the impact forces on the impact plate and deflector plate show that the
deflector plate suffers around 2.7 times the impact force compared to the impact plate, due to the
higher angle it makes with the material stream and the much larger impact velocity the material has at
that area. Only at the deflector plate in sinter at 1600 t/h this is less extreme, due to the relatively low
incoming velocity of the material caused by the impact at the impact plate discussed in the first point.
• The wear results at the impact plate are much lower for sinter than for iron ore, which does not reflect
reality. This seems strange, since the higher friction coefficients for sinter and the irregular particle
shape should cause more sliding along steel plates compared to the spherical pellets. Tests with one
particle show that the irregular shape of sinter lowers the wear results significantly when compared to a
spherical model of sinter. From Equation 6.5 this relationship shows a higher R ∗ leading to a higher F n ,
which increases wear. The much higher density of iron ore pellets in the model also contribute to the
higher wear, looking at the m ∗ in Equation 6.6. A final contributor can be the bulk behavior of sinter,
which slows down upon impact and could therefore have a slightly different particle trajectory with
less tangential distance moved along the impact plate (d t in Equation 6.4). This is a minor contribution
however, if at all, considering this altered trajectory is not seen in the vector plots in Figure 6.4.
• wear at deflector plate has an even larger decrease in wear in sinter compared to iron ore, especially in
the higher flow rates. This increase in difference is caused by the bulk behavior of sinter, that has a layer
of particles that almost completely stop flowing on the surface of the deflector plate. This significantly
reduces the wear, since there is less sliding per time step over the deflector plate.
• The material trajectory modeled only takes into account inter-particles forces and gravity, while ignor-
ing other influencing forces such as air resistance [34], which can lead to discrepancies between the
model and reality.
• The higher flow rate configurations cause less velocity fluctuations in belt loading, which is caused by
a higher pile of material on the belt, making the free fall of material very short.
• The wear pattern as well as the stagnation of particle velocity on the deflector plates reflects the wear
and material residu found at Tata Steel. The particle residu can be seen in Figure 6.1. The off-center
belt loading is far less pronounced in real life. This is most likely caused by
– Material buildup and wear deformation of geometry elements

– Unknown exact dimensions of this section of the chute geometry
– The material settles further along the bottom belt trajectory due to belt vibrations
– Calibration of material and/or rubber could be inaccurate due to their unrealistic particle shape
or properties
• The segregation in the model is also less pronounced when compared to reality (for reference, see Fig-
ure 2.5, top right), which can be caused by discrepancies between PSDs in the model compared to
reality, the (three-)spherical particle shape of iron ore and sinter behaving different from reality, or a
slight difference in material stream from anomalies in the geometry
6.5. Conclusions
This chapter answers the following research question:
• Where along the material stream in a transfer chute at Tata Steel is potential dust creation and lib-
eration and what is the influence of different materials and mass flow rates in DEM software on this
problem?
In this chapter simulations on the case study chute at Tata Steel have been performed with iron ore pellets
and sinter of 800 and 1600 t/h. Results of three KPIs that indicate either air entrainment or dust liberation
have been analyzed, leading to the following conclusions:
• Three impact zones that have potential dust liberation, with the material impacting the impact plate,
the deflector plate underneath the impact plate and the receiving belt.
• Free fall sections from discharge till impact plate and from impact plate till deflector plate show a low-
ering of bulk density of the material stream, suggesting potential air entrainment.
• Simulations show highest material velocity fluctuation and highest bulk density change at the deflector
plate impact, making it the most problematic area for dust liberation according to the simulation.
• The DEM model for sinter shows more compacting of material in comparison to iron ore pellets, likely
due to higher friction coefficients with the steel chute and an irregular particle shape hampering sliding
and rolling.
• A higher flow rate softens the material impact at the bottom belt due to the higher material bed on the
belt, preventing the material to speed up and spread in free fall.
With the identification of problem areas in regards to the dust liberation problem at the A317 chute at
Tata Steel using DEM simulations, the next step is to design a new chute configuration where these problems
are addressed.
7
Redesign proposals
In this chapter all the knowledge gathered from the literature study on chute design is combined with the
findings from the existing transfer chute into redesign proposals for the case study chute. Three conceptual
designs are proposed and analyzed to make a preliminary selection on a conceptual redesign.
These new designs are tested for the same performance indicators as the existing chute to compare their
performance. This will result in recommendations for the redesign of the case study chute at Tata Steel. The
research question that this chapter answers is therefore:
7.1. Conceptual redesign proposals

From the analysis of the current case study transfer chute at Tata Steel in Chapter 6, the design objectives of
the redesign are summarized as follows:
• Minimize the number of impacts and impact forces
• Minimize the amount free fall in the material stream
• Improve center belt loading at the bottom belt
• Lower chute wear
The case study chute has many complicating factors in its requirements, of which most notable:
• Multi-material chute with large lump size difference
• Two different mass flow rates of 800 and 1600 t/h
• Movable top part of chute to load different bottom belts
• Bottom belts with different directions are to be loaded
• Limited space available
Using these design objectives and considerations, three concept proposals are made. Substantiated by
analytical findings, one concept is chosen for a simulation to compare with the current case study chute.
Sinter at 800 t/h and 1600 t/h is chosen for the simulation, considering the greater differences in flowabil-
ity compared to iron ore pellets, which resulted in higher bulk density and velocity fluctuations. While the
simulation time for sinter at 1600 t/h is the highest, it is also the most critical when it comes to ensuring
flowability. Convergence in hood and spoon need to be validated along the most volumetric condition. Also
the significant velocity difference between the bottom belt speed and the material stream when guided on
this belt needs to be checked for potential spillage. The same KPIs will be used to evaluate the redesign and
to compare them to the current chute. A schematic drawing of the three concept proposals is illustrated in
Figure 7.1, which are explained in detail in the next subsections.
53
54 7. Redesign proposals
Figure 7.1: Schematic drawing of the three redesign proposal concepts
7.1.1. Redesign proposal I

Concept The first proposal that is made for the redesign is a hood and spoon concept. The hood is po-
sitioned on the movable top chute, whereas stationary spoons are fitted as a replacement for the deflector
plate. The orientation of the spoons is in the direction of belt travel, which is different for one of the three
bottom belts.
Dimensional constraints The current movable top chute has a height of 2.19m counting from the top of the
discharging belt pulley downwards and one part with a height of 0.86m, taken from the bottom belt upwards,
as seen in Figure 6.1. Also the space for a spoon along the bottom belt is limited to 1.98m, seen in the top
view in the same figure. Hood and spoon designs generally have more available space [18][23], especially the
spoon section.
Material trajectory Roberts’ theory for transfer chute design [8][26], as laid out in Section 2.6, is used to
calculate the characteristics of the hood and spoon.
Some validation calculations on the observations of the simulations are performed following Section 2.6,
of which the results are presented in Table 7.1.
Description Symbol Value Source

Material cross-sectional area A 0.062 m2 Equation 2.4
Material cross-section mean height h 101 mm Figure 2.15
Actual material height on belt H 120 mm Figure 2.15 & DEM simulations
Height to width ratio material cross-section H/B 0.24 Section 2.6.2
Angle of discharge θ 0 Figure 2.16 & DEM simulations
Table 7.1: Analytical calculation results pertaining the material trajectory
With a straight discharge, the material trajectory formula from Equation 2.6 is only dependent on the
discharge velocity of 2.19 m/s. This is the material trajectory for the bottom of the stream, as seen in Figure
7.2 on the left, which is almost identical to the material stream found in DEM simulations. The simulated
trajectory is slightly lower, which is most likely due to energy dissipation from inter-particle effects [34]. The
top of the trajectory for sinter at 800 t/h is found by trial and error fitting the curve with the simulation.
The resulting trajectories are presented in Figure 7.2 (left), along with their corresponding radii of curvature
(right).
7.1. Conceptual redesign proposals 55
Figure 7.2: Calculated and simulation material trajectory (left), with corresponding radii of curvature (right)
It is noted that the radii of curvature become relatively high after only a short horizontal distance, which
is due to the relatively low discharge speed.
Hood design The trajectories of all four simulation configurations in Chapter 6 are the same, apart from
the cross-sectional material height. Therefore the hood shape calculation is applicable in all situations, pro-
vided that the hood needs to be adjustable in horizontal distance from the drive pulley to ensure no spillage
occurs from materials flowing over the hood. This is already present in the case study chute with a manual
lever, except this process will have to be automated given to frequency of different flow rates and materials
transported through the chute. Considering cokes have the highest lump sizes, the minimum chute opening
area is calculated for cokes from the equation in Figure 2.3, resulting in A mi n = 0.39 m2 . The main goal of the
hood is
• To redirect the material stream straight downwards at the bottom for the spoon to catch
• To provide a minimal impact angle between material and chute
• To keep the stream compact for the perpendicularly orientated spoon to be able to catch all the material
The vertical space available for the hood is 2.19 m. Using this, combined with the radius of curvature and
Equations 2.9 - 2.11, the radius size and x-position of a radius of curvature that has a y-position of 2.19 m is
found, resulting in R= 2.84 m, an x-position of the radius center at x c = -1.54 m. The x-position and the angle
at point of contact are x = 0.8 m and θc = 35.5◦ respectively. The minimum chute opening is met with these
hood dimensions, but due to the low discharge speed, the chute hood still is relatively close to the drive pulley,
making an adjustable hood an absolute necessity to prevent blockage for different flow rates and materials.
Since the spoon is orientated perpendicular to the material stream, it is important that the hood has some
convergence to ensure the stream stays compact and all particles can land on the spoon. Simultaneously, the
amount of convergence is limited due to the potential threat of pluggage. The next paragraph addresses this
issue. The cross-sectional shape of the hood is chosen with a typical sharp-edged inward curve on the sides,
which ensures the material stream stays compact and is still realistic to be constructed.
Spoon design and conclusion The self-cleaning check for the spoon yields Ψ >32◦ . Given limited vertical
space for the spoon, the main problem is to ensure all the material from the hood exit will land on the spoon
and not directly onto the bottom belt, as can be seen from the schematic drawing in Figure 7.1 on the left
illustration. Considering only 0.86m is available, this is not possible in this design concept, even with an
extreme convergence of the hood shape. Therefore this concept is not chosen and options involving a higher
cut-off of the movable chute will need to be considered.
7.1.2. Redesign proposal II

Concept For the second design proposal, the possibility of a larger height for the spoon is investigated.
Spoons are again fixed on top of the bottom belts, just as the current deflector plates, but are now higher than
the rail of the movable chute (see Figure 6.1). The pink line in the side view and the hashed pink area in the
front view picture indicates the area that is altered to ensure free movement of the chute. This metal plate
is cut and outfitted with rubber flaps to keep the enclosure closed off while allowing the chute to move into
position above a spoon. This metal plate does not compromise the structural integrity of the chute, as long
as the horizontal I-beam right above this plate is left intact.
Hood design The hood height has to be shortened for this concept, where the cut-off is possible between
0.93 m and about 1.4 m from the discharge point, to ensure enough space for a spoon. The hood radius of
2.84 m from Concept I is used, but a straight vertical cut-off of the hood is not possible when the hood needs
to be shorter. Therefore after the material is caught at x = 0.8 m, the hood bends along a second, smaller
radius. A radius of R 2 = 1.35 m is chosen, which is the curvature of the material trajectory at x = 0.5 m. When
the hood is pointing the material straight downwards, this radius is cut off, which is at a height of 1.27 m
from the discharge point. It is noted that this multi-radius hood curvature will be difficult to implement in
real life due to the difficulties in manufacturing these shapes, which needs to be taken into account in the
assessment of the results compared to real life. A slight convergence of the hood shape is chosen to ensure
the material stream stays compact and will be properly guided onto the spoon. The transfer rate of 800 t/h
is not influenced by this convergence, since it is not more compact than the material on the top belt. To
investigate if the convergence is of influence on sinter at 1600 t/h will have to be investigated.
Spoon design The vertical space left for the spoons is 1.76 m, accounting for a 10 mm gap between hood
and spoon. Adhering to the self-cleaning spoon exit angle of 55◦ , trigonometry dictates a spoon radius of R =
2.15 m. Larger radii are not chosen, due to the relative short height of the spoon (despite its height increase
for this concept) which will otherwise cause the material to barely touch the spoon and have a significant
free fall height on one side. The width between the rails of 1.98 m (see Figure 6.1, top view) is another space
constraint. The length of the spoon in bottom belt direction is 0.88 m, which is well within this space, since
it is placed almost directly underneath the pulley to catch all the material from the hood. The spoon exit
velocity will be much higher than the bottom belt velocity, which is similar to the current design. Simulations
and verification and validation will have to be performed to ensure no spillage will occur from high flow
rates and anomalies like surge flows. To prevent this phenomenon, a convergence in the spoon should be
incorporated. In this design, a very minor convergence is implemented, with the spoon exit width of 0.8 m,
which is still larger than the width of sinter at 1600 t/h on the top belt.
7.1.3. Redesign proposal III

Given the main problem in the redesign is the hood and spoon being perpendicular to each other, causing
material to not fully land on the spoon slope. The third concept that is considered is therefore the incorpora-
tion of a flopper gate, as seen in Figure 2.14. The flopper gate is illustrated in blue on the right image of Figure
7.1. The difficulty in this design is the forward momentum of the material stream, which requires an impact
plate or hood to redirect the stream downwards. When a stationary impact plate is chosen, the likelihood of
material getting stuck in the gap between the flopper gate and the impact plate is very high. Another possi-
bility could be to fix the impact plate to the flopper gate, but the dimensions for this concept will not be able
to fit inside the hood and significantly increase the power required to move the gate. The complexity of this
design with many moving parts, will be costly to manufacture. This complexity makes the chute concept also
vulnerable to excessive maintenance. Therefore this concept is not chosen.
7.1.4. Redesign choice

To conclude, redesign proposal II is chosen, for the following reasons:
• The current available height between the bottom belt and the movable chute is too small to properly
fit a spoon. Concept I is therefore not chosen, since the hood would direct material straight onto the
bottom belt, instead of over the spoon.
• Concept III is not chosen due to the complexity of the design, being vulnerable to excessive mainte-
nance and potentially costly construction
7.2. Results 57
• Concept II fits a hood and spoon design where the whole material stream will be able to flow over the
spoons, due to a higher cut-off of the movable head chute. The spoons will be able to redirect and load
the material into the transport direction of the bottom belts, which is an ability that is absent in the
current design.
7.2. Results
Velocity fluctuations In Figure 7.3 the velocity profile of sinter at the two different flow rates are compared
to the case study results. It is noted that the material is caught by the hood earlier than the current impact
plate, the maximum velocities have slightly increased, but decreases in velocity are less extreme.
Figure 7.3: Comparison of velocity profile between current design and the redesign simulations
Impact velocity and force Due to the limited space available in the design of the A317 chute, a perfect
tangent between the hood and the material was not achieved. On the middle and right in Figure 7.4, the
velocities and reaction force when impulse-momentum of the impact are presented. The angle β in the figure
is found by first locating the angle the material stream radius of curvature has at the impact point using
the trajectory equations. This is then subtracted from the angle the hood radius curvature has at the same
point. The impact point is approximated from the simulations, where conservative values are used in the
calculations. Finally, the total particle velocity before impact, Vp , is used in conjunction with β and the mass
flow rate to find the reaction force, R n .
Figure 7.4: Comparison of reaction forces between current design and the redesign simulations (middle) and schematic of forces at
impact on hood (left) and spoon (right)
As can be seen from the results on the right in Figure 7.4, the hood proves an extremely significant im-
provement in the reduction of the impact force, whereas the spoon provides a smaller impact reduction.
Bulk density Using bulk density bins in EDEM, an overview can be made of the change in bulk density along
the material stream in the simulations. The results are presented in comparison with the sinter simulations
of the case study chute in Figure 7.5
Figure 7.5: Comparison of bulk density percentage changes along the material stream
Belt loading and segregation The results of the KPIs to check the level of center belt loading and segregation
are shown in Table 7.2, alongside Figure 7.6, which gives an impression of the belt loading of the redesign.
Table 7.2: Comparison of center belt loading and

segregation
Total mass Average particle mass

Case study ratio left:right Left bin Right bin
800 t/h 2.58:1 51.5% 48.5%
1600 t/h 1.42:1 49.9% 50.1%
Redesign
800 t/h 0.70:1 51.8% 48.2%
1600 t/h 0.89:1 50.3% 49.7% Figure 7.6: Isentropic view of 1600 t/h belt loading
Chute wear The normalized wear gradients are extracted in the same manner as the case study chute anal-
ysis and presented in Table 7.3
Hood / Impact plate Spoon / Deflector plate Bottom belt

Case study, sinter 800 t/h 1.38 0.52 0.17
Redesign, sinter 800 t/h 1.38 5.1 0.042
Case study, sinter 1600 t/h 1.83 0.52 0.20
Redesign, sinter 1600 t/h 1.38 6.1 0.027
Table 7.3: Normalized wear gradients

7.3. Discussion 59
7.3. Discussion
Analyzing the results of the redesign in comparison with the case study chute provides the following insights
based on the KPIs:
• The hood is effective in decelerating the material more gently than the impact plate, judging from the
less steep decline in the first ’hill’ in Figure 7.3. This is much more prominent in the low flow rate simu-
lation. The explanation to this is two-fold. The convergence of the hood causes the material to compact
in the high flow rate simulation, affecting the flow speed to some degree. Also, a miscalculation caused
the hood placement to be slightly too far away from the pulley in the high flow rate simulation, making
the impact angle higher and the impact point lower.
• The reaction forces on the hood are significantly reduced compared to the impact plate (Figure 7.4),
due to a much lower impact normal velocity. This is because the impact angle is significantly reduced,
and also the material stream velocity is lower since the stream is ’caught’ earlier.
• The convergence of the hood is extremely effective in compacting the material in free fall between
hood and spoon. The bulk density increases up to 50% for the high flow rate, instead of the decrease
experienced in the case study simulation after the impact plate is hit (Figure 7.5). The wider material
stream of the 1600 t/h simulation benefits more from this convergence than the 800 t/h simulation.
This effect also leads to a much higher bulk density increase at spoon impact for the 800 t/h stream
compared to its higher flow rate counterpart.
• The converging free fall of material causes it to be caught by the spoon lower than intended, making
the impact far from tangent and therefore only a slight decrease in reaction force is found as compared
to the case study simulations (Figure 7.4). This decrease is larger for 1600 t/h, since the converging of
material after the hood causes a relatively slightly lower velocity at spoon impact.
• The spoon is more effective in gently decelerating the material compared to the case study for the high
flow rate case, whereas with low flow rates the velocity profile looks similar to the deflector plate impact
of the case study chute. This is seen in the decline rates of the graphs between 2.3 s and 2.4 s in Figure
7.3. The spoon exit onto the receiving belt is also much smoother in deceleration (see 2.4 s - 2.6 s in
Figure 7.3). In the case study, the material even decelerated to below the belt speed, forcing the belt to
speed up the material again. The elimination of this phenomenon can potentially decrease the power
consumption of the bottom belt pulley system.
• The center belt loading is significantly improved when comparing the mass ratios at the bottom belt
from Table 7.2. However, the velocity difference between the spoon exit velocity and the bottom belt
speed causes the material to spread out wide over the belt, as can be seen in Figure 7.6. In this simu-
lation it causes no spillage, but for the actual implementation of the design, a narrower spoon exit is
advised to be implemented after testing, verification and validation.
• The segregation results are almost identical to the case study results, where no more than 2% difference
in average particle mass is found between the left and right side of the bottom belt. It is unclear what
the reason for this is, since some segregation is expected when the material is transferred, as observed
at Tata Steel. Further investigation is required to analyze segregation in the transfer chute, but this is
outside the scope of this research.
• The downside of a redirection of flow over a smooth surface is the chute wear that is increased enor-
mously in the simulation results. The hood vs impact plate show little difference, whereas the spoon vs
deflector plate has a wear gradient of up to 10 times higher.
– The hood wear is similar most likely because although the impact is decreased, the amount of
sliding of particles along the hood metal surface is increased.
– The wear increase in the spoon has multiple reasons. The decrease in impact angle causes mate-
rial to slide more; the downwards velocity before impact is larger; and finally because the archard
wear is the total wear calculated over the entire geometry surface, which is significantly larger for
the spoon compared to the deflector plate. The latter reason makes a proper comparison in total
archard wear between the case study deflector plate and the spoon in the redesign not possible.
– The already small belt wear is decreased a lot more in the hood and spoon simulations. This is
most likely explained by two reasons. The first reason is a reduction of direct belt impact and the
second reason is that the spoon directs the material in the belt travel direction instead of the belt
needing to create that momentum for the particle stream. The latter will likely cause more sliding
of particles on the belt, increasing wear.
7.4. Conclusions
The research question that was answered in this chapter is:
From the results of the redesign simulations and their resulting analysis, the following can be concluded:
• The current dimensions of the movable chute show that a hood and spoon concept for a chute with a
material direction change of 90◦ will not suffice in a working chute as the concept intends. The spoon
will be too short to catch the wide material stream and material will fall directly onto the belt, even with
a significant convergence in the hood.
• A higher cut-off of the movable chute is conceptualized and simulations with sinter at 800 t/h and
1600 t/h are performed. The simulations of this hood and spoon design show signs of decreased dust
liberation potential, being:
– A lower deceleration and reaction force in the hood compared to the impact plate (especially for
800 t/h), which suggests a lower impact and therefore less risk of dust liberation.
– The hood catches the free falling material earlier, and after it leaves the hood the material stream
is more compact, which suggest less dust liberation potential according to literature[1][4]. This
compacting is much more significant in the higher flow rate simulation which makes sense con-
sidering the same hood convergent shape and a much wider material stream at 1600 t/h. The
results of this compacting propagates to the spoon impact, which again is positive regarding dust
liberation.
• However, despite the lower reaction force of the 800 t/h simulation, this configuration shows a higher
increase in bulk density at the spoon and the stream’s downwards velocity at this point is higher than
in the case study. Therefore it is unclear if the cumulative benefit regarding dust liberation for this
configuration is positive for the redesign.
• Center belt loading is significantly improved with a spoon design as compared to deflector plates as in
the case study simulations, although care much be taken to make the spoon convergence sufficient to
avoid spillage, due to the velocity difference between stream velocity at the spoon exit and the bottom
belt speed. Segregation is almost completely non-existent in the simulations, which requires further
investigation to analyze.
• Quantification is hard to compare with radically different designs, but a strong indication is shown
towards a significant increase in archard wear in the hood and spoon design, which is conform to ex-
perience in the industry with metal plates deliberately allowing sliding of ore materials over its surface.
8
Conclusions and recommendations
In this chapter all the conclusions that are drawn from this research are collected and presented, as well as
recommendations for further research or further improvement of the redesign.
8.1. Conclusions
The main research question that is answered in this report is
What are the causes and solution methods in minimizing dust creation and liberation at trans-
fer chutes and how can these be applied in the redesign of a multi-material Tata Steel transfer
chute with differing flow rates?
In order to answer this question, several smaller research questions are answered in the previous chapters,
of which an overview is presented below:
• What is known in literature on the influencing factors on the creation and liberation of dust in transfer chutes
and what are known transfer chute design considerations that provide solutions to these problems?
– Dust liberation is a result of air displacement of the material stream and is minimized in transfer
chutes by minimizing material impact, minimizing free fall of material and a compact containment
of the flow stream.
– Air control through filtration systems or using water curtains have all been tried, but usually with only
minor improvements, if any.
• What are suitable assessment methodologies for identifying transfer chute problems at Tata Steel that involve
dust creation and liberation?
– Measuring methodologies of direct dust measurements have many disadvantages regarding accuracy
of source location, therefore other methods such as air or material degradation or numerical modeling
techniques such as DEM, CFD, and the coupling of the two can be employed as a substitution for the
identification and quantification of dust liberation.
– For this research material degradation through sampling is performed. This is chosen for a multitude
of reasons, like the availability of measurement tools and level of identification and quantification.
DEM is chosen as numerical modeling tool, due to its material-equipment interaction modeling ca-
pabilities and the level of applicability time-wise compared to CFD-DEM coupling techniques.
• What is the extend of material degradation at a Tata Steel transfer chute?
– Insufficient samples are performed in this research to answer this question to a satisfactory degree.
– Stopped-belt sampling according to ISO3082-2009 is possible for iron ore pellets and sinter, but not
for cokes. The penetration of the material with the sampling tool is not possible due to the large lump
size of cokes.
61
62 8. Conclusions and recommendations
– The inclined surface wear tester is used as a dynamic calibration test setup. The setup is capable of
providing both static indicators, like angles of repose, as well as dynamic indicators.
– Verification of the simulations was done by comparing results of three repeated simulations.
– The trial and error methodology used leaves sensitivity analyses and the level of influence each pa-
rameter has out of the calibration process, which is a strong recommendation for future research.
– Eight performance indicators are used in combination with six calibration parameters, and while
good agreement was found between experimental and simulation results, not all parameters are cali-
brated within the experimental range of results.
• Where along the material stream in a transfer chute at Tata Steel is potential dust creation and liberation and
what is the influence of different materials and mass flow rates in DEM software on this problem?
– Three impact zones are found that have potential dust liberation, at the impact plate, the deflector
plate underneath the impact plate and the bottom belt.
– Free fall sections from discharge till impact plate and from impact plate till deflector plate show a
lowering of bulk density of the material stream, suggesting potential air entrainment.
– Simulations show that the highest reaction forces and highest bulk density change is at the deflector
plate impact, making it the most problematic area for dust liberation according to the simulation.
– The DEM model for sinter shows more compacting of material in comparison to iron ore pellets, likely
due to higher friction coefficients with the steel chute and an irregular particle shape hampering slid-
ing and rolling.
– A higher flow rate softens the material impact at the bottom belt due to the higher material bed on the
belt, preventing the material to speed up and spread in free fall.
• What does a redesign for a Tata Steel transfer chute to minimize dust creation and liberation look like and
what can be the quantifiable improvements to the original design?
– A hood and spoon concept is a potentially improved design with regard to dust liberation, among
other factors.
– A cut-off of the movable top chute in the case study at Tata Steel needs to be adjusted in order to fit a
hood and spoon design that is simulated.
– Simulations show potential for dust liberation minimization due to:
¦ Less and softer impacts, as shown by velocity profiles, reaction force calculations and bulk density
differences
¦ Compacting of material by hood convergence, as shown visually and through bulk density mea-
surements
¦ Earlier catching of material from discharge improves free fall time, with the free fall after the hood
being more compact
– Off-center belt loading is significantly improved, while archard chute wear is increased.
8.2. Recommendations
Many recommendations for further research are to be made from this research, which are listed in this sec-
tion, separated in several categories for clarity.
Design
• The incorporation of a honeycomb in the hood and spoon is recommended, since the amount of wear
will most likely be a more significant problem in the implementation of the redesign than dust libera-
tion. A drawback is the contamination of materials in these multi-material chutes.
• The redesign is potentially difficult to manufacture, so approximation of the curves with tiles are a
compromise.
8.2. Recommendations 63
• Chute curvatures with higher radii could potentially improve flow conditions according to Roberts’
theories. Testing and implementing this where possible is recommended.
• The bottom belt needs to allow material to pass along the spoon. A spoon that can rotate upwards to
some degree is recommended[7].
• Make use of settling zones or ’calming tunnels’ (see Section 2.4) is an alternative to minimize dust
liberation if a complete redesign is unable to be performed.
• Most likely very little dust creation occurs at the case study transfer chute, since the drop height and
impact forces are not very high, especially in comparison to the bunker loading that happens before
the transfer chute for some materials for instance.
Measuring and modeling methodologies
• The stopped-belt method can potentially be used for the extraction of samples right before and after a
transfer chute and result in PSD data within a CI of 95%, but many samples will most likely be required.
Several recommendations for a successful outcome of this procedure are:
– Since the differences in before and after samples are grouped together to obtain one average
within 95% CI, ideally all samples are taken at on the same day. Differences in source material
(like origin or exposure to elements) can lead to differences in breakage rates, making a CI of 95%
more difficult.
– A wider tool is recommended to minimize contamination of the samples with the insertion of the
tool into the material.
– If cokes or other large lump size materials are sampled, stopped-belt is not recommended. vibra-
tory plates might be able to penetrate the material for sampling, but the risk of breakage contam-
ination is likely larger than the amount of breakage occurring along the bottom belt trajectory.
Therefore sampling directly from the top belt pulley discharge and from the end of the bottom
belt.
• Improving flow aspects in transfer chutes using DEM has been proven to be sufficient in lowering dust
liberation and solve other chute problems. Modeling dust flows with CFD-DEM coupling is only rec-
ommended if dust problems persist, but with the current lack of sufficient validated research in this
area, more fundamental, experimental research in this field is needed. Also the computational power
required for entire chute granular and dust flow analysis is a current hurdle that can be overcome in the
future when computational power increases.
Calibration
• Sensitivity analysis on dynamic calibration using design of experiments for instance
• Calibration validation is required using other test setups
• The PSD of sinter that is measured is relatively small, which can vary much more at Tata Steel and
elsewhere, depending on when samples are taken. Therefore most likely an extreme amount of samples
need to be performed to get a PSD within 95% CI, and therefore an average PSD for DEM modeling of
sinter can be recommended in most cases.
A
Research paper
A.1. Abstract found that air flows are linked to the control of dust.
Three main design objectives are found that mini-
With increasingly strict restrictions on environmen-
mize dust liberation, which are:
tal pollution, dust liberation in the bulk handling in-
dusty is a problem that needs to be minimized. This • Minimize material impacts
research attempts to find the causes of dust liberation
• Minimize free fall of material
in a case study transfer chute at Tata Steel and test a
redesign with simulations to improve the chutes’ dust • Compact containment of material flow stream
liberation problems, taking into account multiple ma- Other methods that are attempted in the industry
terials, flow rates, and a movable head chute. Ma- are to minimize the air speed inside the chute us-
terial degradation was attempted to be recorded using rubber flaps around the chute opening and skit-
ing stopped-belt sampling, since impacts cause degra- boards; extension of chute enclosure over the receiv-
dation as well as dust liberation, but no conclusive ing belt to calm the air flows (also called ’calming tun-
evidence was found due to a lack of samples. Iron nels’); water curtains that contain dust; and air fil-
ore pellets and sinter were dynamically calibrated for tration systems. These methods however seem to be
DEM software using an inclined surface wear tester. battling the symptoms rather than tackling the root
Case study simulations of these materials at 800 and causes[1][23]. Current chute configurations incorpo-
1600 t/h found three problem areas involving impacts. rate dead boxes or hood and spoon concepts to at-
These were reduced in a redesign that was tested in tempt to redirect the flow in a more controlled man-
simulations. The redesign uses a hood and spoon con- ner, where the latter seems more suitable for dust
cept with movable hood, where the cut-off of the mov- liberation minimization, due to the minimization of
able chute head needed to be heightened to fit a spoon both impacts; amount of free fall; and compact con-
that can improve flow conditions. Improvements in tainment of flow if a convergent hood and spoon is
the identified problem areas were found in simula- used. A novel concept of a movable hood and/or
tions of the redesign. spoon has been introduced that accounts for differ-
ent material flow properties, but papers on this topic
A.2. Introduction as well as industry applications have been severely
lacking[23].
Dust liberation in transfer chutes has been a persis-
tent problem in the bulk handling industry, with in-
creasing demand for solution given the tightening A.3. Sampling
of environmental regulations. Besides environmen- To assess the case study dust liberation problems,
tal contamination and employee respiratory damage; a sampling method was devised to perform on-site
the added costs dust liberation introduces by the loss measurements. Direct measurement of airborne or
of material; increased maintenance costs on dam- settled dust provide a multitude of problems, among
aged equipment; and cleanup costs can add up to which the biggest is the problem of identifying the ex-
an average loss of one percent per ton of through- act source of the airborne or settled dust that is mea-
put according to investigative reports performed in sured. Where material impacts are both the cause of
the UK[1][4]. Research into the causes and solutions dust liberation and dust creation, material degrada-
for dust liberation problems at transfer chutes have tion can also be used as a quantifying measurement
65
A. Research paper
in transfer chute problems. A sampling methodol- Calibration parameters Pellets Sinter

ogy as described in ISO3082-2009 is used, for which
• Static friction p-p [-] 0.3 0.25
a sampling tool was constructed. This standard is • Rolling friction p-p [-] 0.15 0.15
made for finding the PSD of big bulk material vol- • Static friction p-s [-] 0.3 0.5
umes, but its methodology of stopped-belt sampling • Rolling friction p-s [-] 0.25 0.3
is applied in this research to measure the PSD of ma- • Static friction p-r [-] 0.4 0.6
terial right before and after the chute to compare the • Rolling friction p-r [-] 0.29 0.5
amount of smaller particles. Samples had to be taken • Density [kg/m3 ] 3700 2850
close together along the stream and the material had
to have passed the chute at full speed to get accurate Table A.1: Results calibration parameters
measurements, for which the following formula was

devised:
s bs − 12 · t su · v b A.5. Case study DEM analysis
t bs = t su + (A.1) Methodology
vb
To analyze the case study chute to identify the main
where t su is the start-up time, v b is the belt speed, causes for its dust liberation as listed in the introduc-
and s bs and t bs are the distance and time between tion and to investigate the effects of differing mass
the two sample locations respectively. Iron ore and flow rates and materials, a geometry model is made
sinter were possible to sample using the sampling and simulations are performed. The three main key
tool, but the large lump size of cokes made it impos- performance indicators (KPIs) that are used are:
sible to penetrate with the tool. A total of two before
and after samples of sinter were extracted and found • Velocity fluctuations in material flow
to be insufficient to provide conclusive evidence of
material degradation within a 95% confidence inter- • Impact velocity and reaction force
val. • Bulk density fluctuations along material stream
Secondary KPIs to give an indication of chute perfor-

A.4. Calibration mance in other design objectives are:
Iron ore pellets and sinter were dynamically cali-
brated on a rubber and steel surface using an inclined • Belt center loading
surface wear tester, shown in Figure A.1, for the pur-
poses of dynamic simulations of transfer chute flow. • Chute wear
Eight performance indicators were used, with three • Segregation
angles of repose and four dynamic indicators. The
latter being all the inclination angles at: first material Velocity fluctuations are measured by selecting a
on material shift; first material on plate shift; biggest small box of material on the top belt and extracting
material shift; and when the top box is emptied. Nor- the average velocity. The reaction forces on the im-
mal distributions were used for the PSDs, extracted pact plate and the deflector plate are calculated from
from literature for pellets and from samples for sin- extracting the normal component of the material ve-
ter. Pellet particles are modeled as single spheres and locity to the geometry surface from the simulations
sinter as three spherical particles. From a literature and using the following equation:
based reference case, seven calibration parameters
were calibrated using a trial and error method. The R n = m s Vpn sin β (A.2)
results are shown in Table A.1, where p-p is particle to
particle and s and r stand for steel and rubber respec- Where m s is the mass flow rate in kg/s, Vpn is the nor-
tively. While good agreement with the experiments mal velocity component of the particles, and β is the
was found on most performance indicators, a sen- angle between the material stream and the geometry.
sitivity analysis using design of experiments is highy Bulk density fluctuations are measured using small,
recommended to study the level of influence of the cubic, bulk density sensors with sides of 70 mm that
calibrated parameters on bulk behavior. are tucked inside the material stream along its trajec-
tory. Center loading and segregation measurements
use the same two bins that split the bottom belt in
half and measure the total particle mass and aver-
age particle mass respectively. Total archard wear is
measured from meshed geometries (using ANSYS to
create smooth surface CFD-Fleunt meshes with 0.006
Figure A.1: Picture of inclined surface wear tester during a test m maximum face sizes) and normalized gradients are
with sinter extracted from their linear progression results.
A.5. Case study DEM analysis 67
Four simulations are performed with combinations the bottom belt, due to it being loaded much higher,
of the calibrated sinter and iron ore pellet material which eliminated the possibility of material free fall.
models and 800 t/h and 1600 t/h mass flow rates. The reaction forces for the higher flow rates were
Figure A.2 shows the geometry as modeled in DEM, more than double than the low flow rate forces, due
where it can be seen that the material hits a v-shaped to the higher flow rate and slightly increased normal
impact plate, which is attached to the movable top velocity components, which was to be expected.
part of the chute along with the top belt and pulley. The bulk densities were decreased over 60% in the
Then a deflector plate that is fixed above one of three first free fall section, while upon impact an increase
bottom belts is hit and finally the material lands on in roughly 10% was seen for pellets and 22% for sinter.
the bottom belt. This pattern repeated at the deflector plate, where
pellets and sinter had around 40% and 60% bulk den-
sity increases.
The wear results at the impact plate showed higher
wear in pellets than sinter. This wear difference was
higher at the deflector plate. This result does not re-
flex reality.
Discussion
The results suggest that sinter tends to build up ma-
terial at impact zones, where iron ore flows more flu-
Figure A.2: Case study chute simulation of iron ore at 1600 t/h idly. The reason can be the higher friction coefficients
sinter has to steel, and the irregular particle shape of
The chute geometry model is was checked for sinter particles cause them to bounce away from the
continuous flow without spillage using mass flow rate general stream direction, creating more inter-particle
sensors before discharge and on the bottom belt. The bounces that obstruct the steady material flow and
point of impact on the impact plate and the wear and slow down the stream due to energy losses for col-
material residue found on the deflector plate were all lisions. The brunt of the impact is shifted from the
indications of a model that represents reality to some particle-to-wall impact towards particle-to-particle
degree. impact, since it was observed that a layer of stagnant
sinter particles were residing on the deflector plates’
surface.
Results
The higher wear results for sinter were not expected,
The velocity profiles shown in Figure A.3 of the four
since the calibrated model for sinter has a higher fric-
simulation all showed roughly the same pattern of
tion coefficient on steel than pellets and the irregu-
three abrupt changes from material acceleration to
lar shape should encourage sliding over rolling. Tests
deceleration, suggesting an impact.
with single particles on steel plate impact revealed
that the irregular shape of sinter result in lower wear
registered, which is due to the relationship between
the particle radius in contact with the geometry and
the normal force, of which the latter is in the equation
for archard wear. A higher radius leads to a higher
normal force, which in turn increases the wear. An-
other factor that can cause the higher wear in pellets
are the much higher particle density. A final contrib-
Figure A.3: Graph of average velocity profile of a small material utor can be the bulk behavior of sinter, which slows
selection (left), side view of pellets, 800 t/h (right) down the upon impact and could have different par-
ticle trajectories that register less tangential distance
These three impacts could visually be traced back moved along a geometry. This is a minor factor if it is
to impacts on the impact plate, deflector plate and present, since vector plots showed no signs of altered
bottom belt. Iron ore showed less deceleration after trajectories.
hitting the impact plate than sinter, causing a higher The differences in bulk behavior for sinter does ex-
maximum speed, which happens before the impact plain the higher wear difference at the deflector plate,
on the deflector plate. The higher flow rates, espe- considering the material stagnation found on the sur-
cially sinter, showed almost no acceleration before face.
A. Research paper
A.6. Redesign lowered to the relatively low belt speed of 2.19 m/s,
Proposals and therefore simulations are required to check for
Three concept proposals were made for the redesign spillage risks and the status of belt wear.
of the case study transfer chute, where one new con- Concept III uses an innovative new design concept
cept was explored and two hood and spoon configu- of combining an impact plate with a flopper gate
rations were calculated. A schematic drawing of the pressed against it, directing the flow either left or
concepts are illustrated in Figure A.4. right onto a spoon below, depending on the spoons
orientation (the bottom belts have different direc-
tions, so the fixed spoons above the belts have differ-
ing orientations). This ensures the material makes an
s-curve from the discharge point left or right to the
spoon guiding it right or left respectively. A disadvan-
tage of this design is the high probability of material
getting stuck between the movable gate and the im-
pact plate when gaps are formed from eventual wear
damage, blocking the flopper gate. Fixing the flopper
gate to the impact plate raises new issues of dimen-
sional constraints and power consumption of mov-
ing a relatively heavy gate. For these reasons, together
Figure A.4: Schematic drawing of the three redesign proposal
concepts with high maintenance and costly manufacture, this
concept is not chosen.
Concept II is is chosen for the redesign, circumvent-
The first concept was to construct a hood and
ing the issue of straight to belt loading in concept I
spoon on the top and bottom sections of the chute
and risk of high maintenance and construction cost
respectively. To calculate the radii and location of the
in the complexity of concept III. From the case study
hood and spoon, Roberts’ theory was used [26]. The
analysis the flowability of sinter showed more signs
trajectory was first calculated and found in accor-
of pluggage or spillage risk for a redesign and due to
dance with the material stream modeled in DEM sim-
lengthy computational times, the redesign was simu-
ulations. The hood is designed to ensure a straight
lated using sinter with 800 t/h and 1600 t/h to com-
downwards redirection of flow at the height of the
pare with the case study results.
cut-off of the top chute, which is at 2.19 m. Using the
equations for radius of curvature, material trajectory
and the chute position together with this constraint Simulation results
a resulting chute radius of 2.84 m is found. The ma- The velocity fluctuations show an earlier catch of
terial is then caught after 0.8 m free fall in the hor- the free falling material, where especially the 800 t/h
izontal direction. When the spoon calculation was stream showed a more gently deceleration, as seen in
performed and only 0.86 m height was available, the the gentler decline in after the first ’peak’ in Figure
90◦ turn the material needs to make causes straight A.5. A similar, much more gentle deceleration was
to belt loading instead of landing on the spoon for a found in both flow rates at the spoon exit when com-
wide section of the stream, even when a large con- pared to the case study results.
vergence of the hood is incorporated. Therefore this
proposal was dropped.
The second concept is to making the cut-off between
the movable chute and the bottom section higher in
order to fit a larger spoon at the bottom. The hood re-
quires some adjustments to fit in the narrower space,
curving the hood with a second, smaller radius at the
bottom of 1.35 m, in accordance to the curvature ra-
dius of the material trajectory at x = 0.5 m. This results
in a hood directing the material downwards at 1.27 m
below the discharge point. A minor convergence is
fitted on the hood with the exit width being 0.8 m, to
ensure compact containment of the flow. The space Figure A.5: Comparison of velocity profile between current design
in height left for the spoon is sufficient to fit a radius and the redesign simulations
of 2.15m, where the spoon exit angle is 55◦ is chosen,

adhering to the self-cleaning check. The spoon exit The reaction forces are presented in Figure A.6. As
velocity of the material in belt direction can not be can be noted, they were significantly reduced in the
A.6. Redesign 69
first catch of material, whereas a smaller reduction in Discussion

the second impact was found on the spoon/deflector The convergence of the hood was proven extremely
plate. effective compacting the material and causing a gen-
tler deceleration. This was especially true for the high
flow rate stream, since it is much wider and can there-
fore compact much more with the relatively tighter
hood convergence. The effect of this compacting
propagates in the free fall section and eventually at
the spoon impact, where significantly less bulk den-
sity increase was found. This suggests less dust lib-
eration potential in this simulation compared to the
case study, since compact material containment was
achieved. The fact that the lower flow rate simula-
tion does not benefit as much from the hood com-
pacting was seen in the very high density increase
at the spoon impact, which suggests an adjustable
hood convergence could significantly improve both
Figure A.6: Comparison of reaction forces between current design flow rates.
and the redesign simulations
The reduction in reaction forces on the hood and
spoon was expected, given the very low impact an-
The bulk density result data are shown in Figure gle compared to the impact plate and deflector plate.
A.7. At the impact plate the bulk density increase was This reduction is much less at spoon impact, due to
lower, whereas in between impact the hood created the converging flow impacting lower on the spoon
a significant bulk density increase when the impact than intended. This created a higher downwards ve-
plate caused a decrease. The biggest bulk density in- locity before impact, but the impact angle was still
crease was found at spoon impact for the low flow sufficiently reduced to yield a decrease in reaction
rate stream. force.
The spoon is very effective in gently decelerating the
material to the belt speed, as seen in Figure A.5, be-
tween 2.3 s and 2.6 s, eliminating the third impact
on the belt as seen in the case study at low flow rate
in the figure. In the case study, the material even
decelerated to below the belt speed, forcing the belt
Figure A.7: Comparison of bulk density percentage changes along to speed up the material again. The elimination of
the material stream this phenomenon can potentially decrease the power
consumption of the bottom belt pulley system.
Segregation showed almost no difference be- The center belt loading is significantly improved.
tween case study and redesign, while the belt center However, the velocity difference between the spoon
loading was significantly increased. The total archard exit velocity and the bottom belt speed causes the
wear gradient results are listed in Table A.2. material to spread out wide over the belt. In this sim-
ulation it causes no spillage, but for the actual im-
Hood / Spoon / Bottom plementation of the design, a narrower spoon exit is
Impact plate Deflector plate belt advised to be implemented after testing, verification
and validation.
CS, 8 1.38 0.52 0.17 The hood wear is similar most likely because, al-
R, 8 1.38 5.1 0.042 though the impact is decreased, the amount of slid-
CS, 16 1.83 0.52 0.20 ing of particles along the hood metal surface is in-
R, 16 1.38 6.1 0.027 creased. The spoon wear can not be compared to
the case study, given the fact that the archard wear is
Table A.2: Normalized wear gradients comparing case study (CS)
with redesign (R) for 800 t/h (8) and 1600 t/h (16)
added from the entire geometry, where the spoon is
much larger than the deflector plate. The significant
reduction in belt wear is likely caused by two reasons.
The hood showed little difference in archard wear,
The first being a reduction of direct belt impact and
while the spoon has over ten times the total archard
the second that the spoon directs the material in the
wear, which is addressed in the discussion. The bot-
belt travel direction instead of the belt needing to cre-
tom belt showed a significant decrease in wear for the
ate that momentum for the particle stream. The latter
redesign.
70 A. Research paper
will likely cause more sliding of particles on the belt, • In the case study analysis, three impact zones
increasing wear. are found that have potential dust liberation,
which are at the impact plate, the deflector
plate underneath the impact plate and the bot-
A.7. Conclusions tom belt. The deflector plate is the highest dust
From this research it can be concluded that a hood liberation risk, given its highest reaction forces
and spoon concept redesign shows potential in and bulk density increase.
dust liberation minimization compared to a multi-
material case study chute at Tata Steel with differ- • The DEM model for sinter shows more com-
ing flow rates. This was concluded from DEM sim- pacting of material in comparison to iron ore
ulations using sinter and iron ore pellets at two flow pellets, likely due to higher friction coefficients
rates, 800 t/h and 1600 t/h. with the steel chute and an irregular particle
More detailed conclusions from the research per- shape hampering sliding and rolling.
formed is listed below.
• A hood and spoon concept is a potentially im-
• Although insufficient samples were taken for proved design with regard to dust liberation,
conclusive material degradation evidence, the among other factors.
stopped-belt sampling according to ISO3082-
2009 was proven possible for iron ore pellets • A cut-off of the movable top chute in the case
and sinter, but not for cokes. The penetration study at Tata Steel needs to be adjusted in or-
of the material with the sampling tool was not der to fit a hood and spoon design that is simu-
possible due to the large lump size of cokes. lated.
• The inclined surface wear tester was used as • Simulations show potential for dust liberation
a dynamic calibration test setup. The setup minimization due to:
is capable of providing both static indicators, – Less and softer impacts, as shown by ve-
like angles of repose, as well as dynamic in- locity profiles, reaction force calculations
dicators. Eight performance indicators were and bulk density differences
used in combination with six calibration pa-
rameters in a trial and error calibration pro- – Compacting of material by hood conver-
cess, and while good agreement was found be- gence, as shown visually and through bulk
tween experimental and simulation results, not density measurements
all parameters are calibrated within the exper- – Earlier catching of material from dis-
imental range of results. Sensitivity analyses is charge improves free fall time, with the
highly recommended to find the level of influ- free fall after the hood being more com-
ence each parameter has on the results. pact
B
Appendix
EDEM simulation settings

Calibration simulator settings
Time integration Euler

Fixed time step 80%
Target save interval 0.01 s
Cell size 5 R min
Simulation time 35 s
Simulation computation time 7 - 12 h
Inter-particle contact model Hertz-Mindlin (no slip) built in
Particle-geometry contact model Hertz-Mindlin with Archard wear built in
Geometries’ wear constant 1 Pa−1
Tata Steel and redesign chute simulator settings
Time integration Euler

Fixed time step 80%
Target save interval 0.01 s
Cell size 5 R min
Simulation time 3.5-5 s
Simulation computation time 40-80 h
Inter-particle contact model Hertz-Mindlin (no slip) built in
Particle-geometry contact model Hertz-Mindlin with Archard wear built in
Geometries’ wear constant 1 Pa−1
Material properties
71
72 B. Appendix
Iron ore Sinter Steel Rubber Plastic

Poisson’s ratio [12][44] 0.3 0.25 0.3 0.4 0.4
Shear modulus [12][31][39][44][46] 10 GPa 13.5 GPa 78 GPa 7.1 GPa 0.96 MPa
Density [12][44] - - 7932 kg/m3 950 kg/m3 1136 kg/m3
Coefficient of restitution with iron ore 0.45 - 0.4 0.3 0.5
Coefficient of restitution with sinter - 0.3 0.4 0.3 0.5
C
Appendix
EDEM simulation calibration steps
Iron ore pellets - Steel
Calibration parameters Sim 1 Sim 2 Sim 3

• Static friction particle - particle [-] 0.45 0.35 0.3
• Rolling friction particle - particle [-] 0.15 0.15 0.15
• Static friction particle - steel [-] 0.5 0.3 0.3
• Rolling friction particle - steel [-] 0.25 0.25 0.25

• First angle of repose, Φ1 30.9 ±0.2 34.4 33.2 30.7 ±0.3
• Inclination angle at first mat-mat shift 8.4 ±0.3 13.8 9.6 9.0 ±0.7
• Inclination angle at first mat-plate shift 11.3 ±1.9 18.7 10.1 9.2 ±0.6
• Inclination angle at biggest mat shift 14.3 ±1.2 19.6 15.4 14.7 ±0.2
• inclination angle at empty top box 22.4 ±0.2 24.5 19.7 19.7 ±0.1
:::::::
• Second angle of repose, Φ2 22.8 ±1.6 29.5 24.6 24.6 ±1.0
• Third angle of repose, Φ3 32.2 ±0.7 41.5 36.9 34.2 ±0.3
:::::::
Iron ore pellets - Rubber
Calibration parameters Sim 1 Sim 2 Sim 3 Sim 4

• Static friction particle - particle [-] 0.3 0.3 0.3 0.3
• Rolling friction particle - particle [-] 0.15 0.15 0.15 0.15
• Static friction particle - rubber [-] 0.71 0.5 0.45 0.4
• Rolling friction particle - rubber [-] 0.29 0.29 0.29 0.29

• First angle of repose, Φ1 32.9 ±0.3 35.4 34.4 33 32.8
• Inclination angle at first mat-mat shift 9.0 ±0.8 14.8 13.8 11.4 9.5
• Inclination angle at first mat-plate shift 13.5 ±2.1 14.8 13.8 13.5 13.1
• Inclination angle at biggest mat shift 14.2 ±1.0 18.5 16.2 16.3 15.3
• inclination angle at empty top box 23.9 ±0.6 25.7 25.6 25.5 24.2
• Second angle of repose, Φ2 19.7 ±1.8 23.8 25.8 26.7 24.1
:::
• Third angle of repose, Φ3 29.6 ±1.9 32.2 33.7 33.7 34.6
:::
73
74 C. Appendix
Sinter - Steel
Calibration parameters Sim 1 Sim 2

• Static friction particle - particle [-] 0.3 0.25
• Rolling friction particle - particle [-] 0.2 0.15
• Static friction particle - steel [-] 0.5 0.5
• Rolling friction particle - steel [-] 0.2 0.3

• First angle of repose, Φ1 35.9 ±0.4 39.9 35.4
• Inclination angle at first mat-mat shift 13.8 ±0.5 15.4 13.9
• Inclination angle at first mat-plate shift 18.2 ±0.7 18.0 20.5
:::
• Inclination angle at biggest mat shift 23.1 ±0.7 24.6 23.0
• inclination angle at empty top box 31.3 ±0.2 30.2 30.2
:::
• Second angle of repose, Φ2 32.1 ±1.5 37.2 38.0
:::
• Third angle of repose, Φ3 34.9 ±1.3 40.6 34.2
Sinter - Rubber
Calibration parameters Sim 1 Sim 2 Sim 3 Sim 4 Sim 5

• Static friction particle - particle [-] 0.25 0.25 0.25 0.25 0.25
• Rolling friction particle - particle [-] 0.15 0.15 0.15 0.15 0.15
• Static friction particle - rubber [-] 0.5 0.5 0.5 0.65 0.6
• Rolling friction particle - rubber [-] 0.4 0.45 0.5 0.5 0.5

• First angle of repose, Φ1 35.2 ±0.4 35.7 35.6 35.6 35.4 35.1
• Inclination angle at first mat-mat shift 13.3 ±1.5 12.8 12.7 12.6 13.4 13.1
• Inclination angle at first mat-plate shift 22.2 ±2.4 21 20.3 21 24 23.8
• Inclination angle at biggest mat shift 28.3 ±1.8 23.6 24.2 24.2 31.2 27.1
• inclination angle at empty top box 35.1 ±0.5 30.3 30.1 30.2 37.5 35
• Second angle of repose, Φ2 27.8 ±0.9 35.1 35.1 35.6 36.2 36.5
:::
• Third angle of repose, Φ3 33.6 ±0.7 40.2 35.8 36.1 41.5 41
::
Bibliography
[1] R. Swinderman, L. Goldbeck, A. Marti, D. Marshall, and M. Strebel, Foundations of Safe Bulk-Materials
Handling (Martin Engineering, 2009).
[2] S. M. Derakhshani, Modelling Dust Liberation in Bulk Material Handling Systems, Ph.D. thesis, TU Delft
(2016).
[3] Hazard Pevention and Control in the Work Environment: Airborne Dust, World Health Organization,
available at http://www.who.int/occupational_health/publications/airdust/en/, (Jan 2018).
[4] Z. Liu, Air entrainment in free falling bulk materials, Ph.D. thesis, University of Wollongong (2003).
[5] H. Wilkinson, A. Reed, and H. Wright, The Cost to the U.K. Industry of Dust, Mess and Spillage in Bulk
Materials Handling Plant, Bulk Solids Handling 9, 93 (1989).
[6] British Standards Institution 2018, BS2890 - Specification for Throughed Belt Conveyors, (1989).
[7] M. Van Aarde, The optimisation of transfer chutes in the bulk materials industry, Master’s thesis, North-
West University, South Africa (2009).
[8] G. Lodewijks, Lecture notes for course ’Characterization and Handling of Bulk Solid Materials’, (2014).
[9] F. Kessler, Recent developments in the field of bulk conveying, FME Transactions 34 (2006).
[10] J. Page, Examples of Good and Bad Chute Design, Ckit, available at http://www.ckit.co.za/secure/
conveyor/papers/bionic-research-1/f-bri1-paper05.htm, (Dec 2017).
[11] How to Design and Implement Chutes in Bulk Solids Handling Systems, Bulk Materials Handling,
available at https://bulk-solids-handling.com/?q=topics/conveying-transportation/
mechanical-belt-conveyor-systems/how-design-and-implement-chutes, (Dec 2017).
[12] G. Chen and D. Schott, Surface Wear Reduction of Bulk Solids Handling Equipment Using Bionic Design,
Ph.D. thesis, TU Delft (2017).
[13] A. Roberts, Chute Performance and Design for Rapid Flow Conditions, Chemical Engineering and Tech-
nology 26, 163 (2003).
[14] D. Oudhoff, Preventing or controlling the negative effects of material losses during pellet handling, Mas-
ter’s thesis, Hogeschool Utrecht (2017).
[15] J. Calvert, Glossary of Atmospheric Chemistry Terms, Pure and Applied Chemistry 62 (1990),
https://doi.org/10.1351/pac199062112167.
[16] C. Rowland, Dust Control at Transfer Points, Ckit, available at http://www.ckit.co.za/secure/

conveyor/papers/bionic-research-2/d-bri2-paper05.htm, (Jan 2018).
[17] A. Roberts, T. Donoheu, C. Wheeler, and W. McBride, Computer Simulations as a Tool for Investigat-
ing Dust Generation in Bulk Solids Handling Operations, Particle and Particle Systems Characterization
(2010), https://doi.org/10.1002/ppsc.200900096.
[18] A. Roberts, T. Donohue, and C. Wheeler, Effective Transfer Chute Design including Dust Control for Han-
dling Grains and other Products, (2018).
[19] A. Ullman and A. Dayan, Exhaust volume model for dust emission control of belt conveyor transfer points,
Powder Technology 96, 139 (1998).
75
76 Bibliography
[20] L. Xiaochuan, L. Qiang, Z. Dong, J. Binbin, L. Huiqing, and H. Yafei, Model for induced airflow velocity of
falling materials in semi-closed transfer station based on similitude theory, Advanced Powder Technology
26, 236 (2015).
[21] WEBA Case Study - Degradation, Tech. Rep. (WEBA Chute Systems and Solutions) avail-
able at https://d1jiktx90t87hr.cloudfront.net/145/wp-content/uploads/2016/03/
Laramie-Dust-Case-Study.pdf, (Jan 2018).
[22] WEBA Case Study - Dust, Tech. Rep. (WEBA Chute Systems and Solutions) available
at https://d1jiktx90t87hr.cloudfront.net/145/wp-content/uploads/2016/03/
Isdemir-Degradation-Case-Study.pdf, (Jan 2018).
[23] P. Staples and L. Liccardo, Developments in Chute Design for High Capacity Multi Product Conveying,
(International Materials Handling Conference - Beltcon 14, 2017) available at http://www.beltcon.
org.za/docs/b1419.pdf, May 2018.
[24] Design - From Concept to Commissioning, Rula Bulk Materials Handling, available at http://rula.co.
za/Design.html, (Dec 2017).
[25] F. Kessler and M. Prenner, DEM - Simulation of conveyor transfer chutes, FME Transactions 37, 185 (2009).
[26] Chute Design Considerations For Feeding And Transfer, Tech. Rep. (Centre for Bulk Solids and Partic-
ulate Technologies, University of Newcastle, NSW, Australia) available at http://www.beltcon.org.
za/docs/b1103.pdf, (Dec 2017).
[27] C. Copeland and S. Kawatra, Design of a dust tower for suppression of airborne particulates for iron mak-
ing, Minerals Engineering 24, 1459 (2011).
[28] J. Halt, Design of a dust tower for suppression of airborne particulates for iron making, Master’s thesis,
Michigan Technological University (2014).
[29] C. Wheeler, X. Chen, T. Donohue, R. McLean, and A. Roberts, Evaluation of dust emissions from conveyor
transfer chutes using experimental and CFD simulation, International Journal of Mineral Processing 110,
101 (2012).
[30] Personal DataRAM pDR-1500 Aerosol Monitor, Thermo Fisher Scientific, available at https://assets.
thermofisher.com/TFS-Assets/LSG/Specification-Sheets/D19971~.pdf, (Jan 2018).
[31] S. Lommen, Virtual prototyping of grabs - co-simulations of discrete element and rigid body models, Ph.D.
thesis, TU Delft (2016).
[32] P. McIlvenna and R. Mossad, Two dimensional transfer chute analysis using a continuum method, (Third
International Conference on CFD in the Minerals and Process Industries, 2003).
[33] C. Wheeler and X. Chen, Computational Fluid Dynamics (CFD) modelling of transfer chutes: Assessment
of viscosity, drag and turbulence models, International Journal of Multiphase Flow 69, 42 (2015).
[34] S. Derakhshani, D. Schott, and G. Lodewijks, Modelling dust liberation at the belt conveyor transfer point
with CFD and DEM coupling method, (11t h International Congress on Bulk Materials Storage, Handling
and Transportation At: The University of Newcastle, Australia, 2013).
[35] D. Hastie, L. Stone, P. Wypych, and S. Zigan, CFD-DEM Modelling of Powder Flows and Dust Genera-
tion Mechanisms – A Review, (12t h International Conference on Bulk Materials Storage, Handling and
TransportationAt: Darwin, Northern Territory, Australia, 2016).
[36] Rocky and ansys fluent coupling (dem-cfd), Rocky, available at https://rocky.esss.co/images/pdf/
ROCKY_and_ANSYS.pdf, (Dec 2017).
[37] ISO3082:2009(E), Iron ores - Sampling and sample preparation procedures, (2009).
[38] F. Dekking, C. Kraaikamp, H. Lopuhaa, and L. Meester, A Modern Introduction to Probability and Statis-
tics (Springer-Verlag London, 2005).
Bibliography 77
[39] G. Barrios, R. Carvalho, A. Kwade, and L. Tavares, Contact parameter estimation for dem simulation of
iron ore pellet handling, Powder Technology 248, 84 (2013).
[40] S. Ishihara, R. Soda, Q. Zhang, and J. Kano, Dem simulation of collapse phenomena of packed bed of raw
materials for iron ore sinter during charging, ISIJ International 53, 1555 (2013).
[41] H. Wei, Y. Zhao, J. Zhang, H. Saxén, and Y. Yu, Liggghts and edem application on charging system of
ironmaking blast furnace, Advanced Powder Technology 28, 2482 (2017).
[42] D. Schott, W. Vreeburg, C. Molhoek, and G. Lodewijks, Granular flow to a blast iron ore furnace: Influence
of particle size distribution on segregation of a mixture, (Traffic and Granular Flow ’15, 2016).
[43] Z. Korzen, Dynamics of bulks solids flow on impact plates of belt conveyor, Bulk Solids Handling , 689–697
(1988).
[44] A. Grima and P. Wypych, Investigation into calibration of discrete element model parameters for scale-up
and validation of particle–structure interactions under impact conditions, Powder Technology 212, 198
(2011).
[45] DEM Solutions Ltd., Edem 2018 documentation, Available at www.edemsimulation.com, (Jan 2018).
[46] H. Mio, S. Komatsuki, M. Akashi, A. Shimosaka, Y. Shirakawa, J. Hidaka, M. Kadowaki, S. Matsuzaki, and
K. Kunitomo, Effect of chute angle on charging behavior of sintered ore particles at bell-less type charging
system of blast furnace by discrete element method, ISIJ International 49, 479 (2009).

Redesign of A Tata Steel Transfer Chute PDF

Uploaded by

Copyright:

Available Formats

Redesign of A Tata Steel Transfer Chute PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Redesign of A Tata Steel Transfer Chute PDF

Uploaded by

Copyright:

Available Formats

What is the research objective and what research questions are being addressed?

What is the research objective and what research questions are being addressed?

What are some of the main causes of dust liberation according to the literature?

What are some of the main causes of dust liberation according to the literature?

Redesign of a Tata Steel transfer chute

with dust liberation problems

at the Delft University of Technology.

Report number: 2018.TEL.8255

5.2 Calibration setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

AoR Angle of repose

CFD Computational Fluid Dynamics

DEM Discrete Element Modeling

KPI Key Performance Indicator

MFR Mass flow rate

PDR Personal Data Ram

PIV Particle Tracker Velocimetry

PSD Particle size distribution

α Angle of discharging belt conveyor

ṁ Material mass flow rate in t/h

µ The sample mean

µE Actual friction coefficient between chute surface and bulk solid

Ψ Self-cleaning spoon angle, angle from belt up to spoon exit angle

ρb Bulk density of the material in kg/m3

θ Angle of discharge in degrees

θc Angle of impact point of radius of curvature

A or A co Minimum chute area in m2

Ap Cross-sectional area of the incoming particle stream

Cd c Belt design capacity in t/h

d 50 50% of cumulative particle mass in a PSD

dt Tangential distance moved in m

1.1. Introduction transfer chute dust liberation

1.2. Problem definition

Figure 1.1: Dust liberation phenomenon at a transfer chute [2]

1.3. Research objective and research questions

• What is the extend of material degradation at a Tata Steel transfer chute?

1.4. Report structure

2.1. Transfer chute definition

2.2. Chute design objectives

• No spillage, prevent the escape of fugitive materials

• Minimize degradation of the transported material and the generation of dust

• Return belt scrapings to the main material flow

• Minimize segregation of the bulk material

• Be service friendly and protect personnel from injury

• Proper loading of material onto receiving conveyor, such that

2.2.4. Material degradation and dust generation

Figure 2.6: Breaking mechanisms in bulk material [14]

2.2.5. Belt scrapings

2.3. Dust liberation theory

Figure 2.7: Table of dust categorizations and characteristics

2.3.2. Dust liberation principles

Dust liberation regions

• Region 1 – Material discharge

2.4. Dust liberation minimization

• Minimize dust creation at transfer

• Minimize spreading of dust at transfer

Minimize drop height and impact force

Figure 2.9: Illustration of a converging hood design [8]

Minimize air entering enclosure

Minimize air speed inside chute