DESIGN AND APPLICATION OF FEEDERS FOR

THE CONTROLLED LOADING OF BULK SOLIDS

ONTO CONVEYOR BELTS

A.W. Roberts
Departe!t o" #e$%a!&$a' E!(&!eer&!( T%e U!&)ers&t* o" Ne+$ast'e
N.S.W., A-stra'&a.

SU##ARY
The efficient operation of conveyor belts for bulk solids handling depends to a significant extent on the
performance of the gravity feed system. The attainment of controlled feeding with a minimum of spillage and belt
wear is of major importance. In addressing this problem, this paper focuses attention on the design requirements
of the mainly used feed system comprising a gravity flow hopper, feeder and chute. The specific function of these
three components is briefly outlined and the need for the hopper and feeder to be designed as an integral unit is
stressed. Various types of feeders are reviewed and methods for determining feeder loads and power
requirements are presented. The interaction between the flow pattern developed and wall pressures generated in
mass flow hoppers and the manner in which these influence feeder loads is discussed. ! simplified methodology
is presented for the design of belt feeders and feed hoppers with extended skirtplates for feeding directly onto
conveyor belts. The design procedures are illustrated by example ". #ention is made of the basic design
requirements of feed chutes with particular reference to the need for careful consideration of the bulk solid flow
properties and the friction characteristics of chute lining materials.
.. INTRODUCTION
$f the various modes at transporting bulk solids, belt conveying is clearly one of the most effective and reliable
and well suited to handling bulk solids over a wide range of tonnage rates. The success of belt conveyors depends
on a number of factors, not the least of which is the initial feeding of the bulk solid onto the belt and to the
efficient transfer of solids from one belt to another at conveyor transfer stations. %ith the future trend towards
higher belt speeds and narrower belts in order to achieve higher economic efficiency &'(), much attention will
need to be given to the design of belt feeding systems which will guarantee high feeding rates with minimum of
spillage and belt wear.
%hile the basic objectives of an ideal feeding arrangement for loading conveyor belts are fairly obvious &*''), it
is important that they be noted. +uch objectives may be summarised as follows,
-ree and uniform flow of material without segregation at a predetermined flow rate in the same
direction as the belt travel and preferably at the same speed.
.niform deposition of material about the centre of the belt,
!voidance of material spillage and dust problems.
#inimisation of abrasive wear and impact damage.
The feeding of bulk solids onto belt conveyors is normally controlled by a gravity flow hopper/feeder combination
and, in the majority of cases, the solids are finally directed onto the belt through a gravity flow chute. The feed
hopper may be a part of a surge bin as in -igure '0a1 or a part of a stockpile reclaim system as in -igure '0b1.
!lternatively it may be a separate dump hopper for unloading trucks or rail wagons as in -igures '0c1 and '0d1
respectively.
-eed rates are controlled by the hopper and feeder as an integral unit while the feed chute in the flow directing
and feed velocity controlling device. It is important that the interactive roles of these three components as an
integrated system be understood2 hence some elaboration is warranted,
i. 3ravity -low 4opper The hopper geometry and internal wall friction characteristics in conjunction with
the flow properties of the bulk solid establishes the type of discharge flow pattern and maximum
potential rate of discharge.
-igure ' 4opper/-eeder combinations for loading conveyor belts

ii. -eeder The feeder controls and meters the flow of bulk material from the hopper to meet the specified
discharge flow rate. The importance of the hopper and feeder to be designed as an integral unit cannot
be too greatly emphasised. ! well designed hopper may be prevented from functioning correctly if the
feeder is poorly designed, and vice versa.
iii. 5hute %hile primarily a flow directing device, the chute, when properly designed, can control the
velocity of material entering the belt in a way which ensures uniform distribution of bulk material on the
belt with minimum belt wear, spillage and power losses. In view of their obvious simplicity feed chutes
have all too often received little attention to their design. There have been many instances where feed
chutes are the 6weakest link in the chain7 in that lack of attention to design detail has led to major
problems such as flow blockages, spillage7s and accelerated belt wear.
$ver recent years considerable advances have been made in the development of theories and associated design
procedures for gravity flow storage and feeding systems for bulk solids handling. ! selection of relevant
references &*"8) are included at the end of this paper. The purpose of the paper is to review the overall
requirements for designing gravity flow feeding systems for bulk solids handling with particular emphasis on the
feeding operations in association with belt conveying. The paper outlines the characteristics of gravity flow
storage/feeder systems, presents an overview of the more common types of feeders used and discusses the
determination of feeder loads and power requirements. 9rief mention is made of the role of chutes and skirtplates
in directing and containing the motion of bulk solids.
/. GRAVITY FLOW OF BULK SOLIDS
%hile the general theories and design requirements for gravity storage and feeding systems are well
documented, &*':) it is useful to review those aspects of particular relevance to the design and operation of
feeding systems for belt conveying operations.
/.. Ge!era' Des&(! P%&'osop%*
The design of gravity flow storage bin/feeder combinations for controlling the flow of bulk solids onto conveyor
belts involves the following basic steps,
;etermination of the strength and flow properties of the bulk solids for the worst likely conditions
expected to occur in practice.
;etermination of the bin geometry to give the desired capacity, to provide a flow pattern with
acceptable characteristics and to ensure that discharge is reliable and predictable.
+election of most appropriate type of feeder and determination of feeder geometry to achieve a
satisfactory flow pattern with a fully live hopper outlet.
<stimation of loading exerted on the bin walls and the feeder under operating conditions.
;esign and detailing of the bin structure and feeder components.
It is important that all bin and feeder design problems follow the above procedures. %hen investigating the
required bin geometry, it should be assumed that gravity will provide a reliable flow from storage. =ot until it has
been demonstrated that the gravity forces available are insufficient to provide reliable flow should more
sophisticated reclaim methods be investigated.
/./ B&! F'o+ Patter!s
-ollowing the definitions of >enike, there are two basic modes of flow, massflow and funnelflow. These are
illustrated in -igure ?.

-igure ? 9in -low 5haracteristics
In massflow the bulk material is in motion at substantially every point in the bin whenever material is drawn
from the outlet. The material flows along the walls with the bin and hopper 0that is, the tapered section of the
bin1 forming the flow channel. #assflow is the ideal flow pattern and occurs when the hopper walls are
sufficiently steep and smooth and there are no abrupt transitions or inflowing valleys.
-unnelflow 0or coreflow1, on the other hand, occurs when the bulk solid sloughs off the surface and discharges
through a vertical channel which forms within the material in the bin. This mode of flow occurs when the hopper
walls are rough and the slope angle @ is too large. The flow is erratic with a strong tendency to form stable pipes
which obstruct bin discharge. %hen flow does occur segregation takes place, there being no remixing during
flow. It is an undesirable flow pattern for many bulk solids. -unnelflow has the advantage of minimising bin wear.
#assflow bins are classified according to the hopper shape and associated flow pattern. The two main types are
conical hoppers, which operate with axisymmetric flow, and wedgeshaped or chiselshaped in which planeflow
occurs. In planeflow bins the hopper half angle @ will usually be, on average, approximately 'AB larger than that
for the corresponding conical hoppers. Therefore, they offer larger storage capacity for the same head room than
the conical bin but this advantage is somewhat offset by the long slotted opening which can cause feed problems.
The limits for massflow depend on the hopper half angle @, the wall friction angle C and the effective angle of
internal friction D. The relationships for conical and wedgeshaped hoppers are shown in -igure E. In the case of
conical hoppers the limits for massflow are clearly defined and quite severe while the planeflow or wedge
shaped hoppers are much less severe. 5onsider, for example, a conical hopper handling coal with D F "8B. If the
hopper is of mild steel it is subject to corrosion, the angle C is likely to be approximately EAB. $n this basis, from
-igure E, the limiting value of @, F 'EB. ! margin of EB is normally allowed making @ F 'AB which is a very steep
hopper. If the hopper is lined with stainless steel, the friction angle C is likely to be approximately ?AB. $n this
basis @ F ?GE F ?EB. This results in a more reasonable hopper shape. The corresponding angles for planeflow
are @ F ??B for C F EAB and @ F E8B for C F ?AB.

-igure E Himits for massflow in conical hoppers
-unnelflow bins are characterised either by their squat hopper proportions or their flat bottoms. -or funnelflow
bins to operate satisfactorily is necessary for the opening siIe to be at least equal to the critical pipe dimension
;f. This will ensure that the material will not form a stable pipe or rathole but rather will always collapse and flow.
4owever, for many materials the minimum pipe dimension ;f is too large, rendering funnelflow bins
impracticable. This is certainly the case, for example, with most coals and mineral ores which, at higher moisture
levels, are known to have critical pipe dimensions of several metres.
%here large quantities of the bulk solid are to be stored, the expandedflow bin, as illustrated in -igure " is often
an ideal solution. This bin combines the storage capacity of the funnelflow bin with the reliable discharge
characteristics of the massflow hopper. It is necessary for the massflow hopper to have a diameter at least
equal to the critical pipe dimension ;f at the transition with the funnelflow section of the bin. This ensures that
the flow of material from the funnelflow or upper section of the bin can be fully expanded into the massflow
hopper. The expandedflow bin concept may also be used to advantage in the case of bins or bunkers with
multiple outlets.

-igure " <xpanded-low bin
The expanded flow concept may also be employed in gravity reclaim stockpile systems such as that illustrated in
-igure 8.
!s a general comment it is noted that symmetrical shaped bins provide the best performance. !symmetric
shapes often lead to segregation problems with free flowing materials of different particle siIes and make the
prediction of wall loads very much more difficult and uncertain.

-igure 8 <xpanded flow principal in stockpile reclaim
/.0 Pote!t&a' F'o+ Rate "ro #ass1F'o+ B&!s
The flow rate from massflow bins depends on the hopper geometry, the wall lining material and the flow
properties of the bulk solid. 4owever, for a well designed massflow bin, the discharge is uniform and the flow
rate predictable. To illustrate the influence of massflow hopper geometry on the potential unrestricted flow rate,
consider the example given in -igure G. This applies to a typical coal at '8.?J moisture content 0dry basis1
where the hopper lining material is stainless steel type EA" with ?9 finish. The graphs show the ,
i. variation in hopper half angle @ at a function of opening hopper dimension 9 and
ii. the corresponding variation in flow rate for unrestricted discharge.
The full lines refer to a conical axisymmetricshaped hopper while the chain dotted lines refer to a planeflow
wedged shape hopper. In the latter case the flow rate is given in tonnes/hr x 'A
E
per metre length of slot. The
graphs indicate the following,
The hopper half angle @, for massflow increases initially with increase in opening siIe 9 and then
approaches a constant value.
The hopper half angle for a planeflow hopper is approximately '?B larger than that for a conical hopper
for the same outlet dimension 9,
The potential flow rate depending on the outlet dimension is quite significant. %hile with train loading
operations where flood loading is required there is a need for high discharge rates, for the majority of
cases the potential flow rate may be excessive rendering the need to employ feeders as flow controlling
devices.


-igure G #ass-low hopper geometry7s and flow rates for typical coal
0. FEEDERS FOR BULK SOLIDS HANDLING
0.. Ge!era' Rear2s
-eeders for controlling the flow of bulk solids onto conveyor belts require certain criteria to be met,
;eliver the range of flow rates required.
4andle the range of particle or lump siIes and flow properties expected.
;eliver a stable flow rate for a given equipment setting. Kermit the flow rate to be varied easily over the
required range without affecting the performance of the bin or hopper from which it is feeding.
-eed material onto the belt in the correct direction at the correct speed with the correct loading
characteristic and under conditions which will produce minimum impact, wear and product degradation.
$ften a feed chute is used in conjunction with the feeder to achieve these objectives.
-it into the available space.
It is important that the flow pattern be such that the whole outlet of the feed hopper is fully active. This is of
fundamental importance in the case of massflow hoppers. %hen feeding along slotted outlets in wedge shaped
hoppers the maintenance of a fully active outlet requires the capacity of the feeder to increase in the direction of
feed. To achieve this condition special attention needs to be given to the design of the outlet as vertical skirts and
control gates can often negate the effect of a tapered outlet. 3ates should only be used as flow trimming devices
and not as flow rate controllers. -low rate control must be achieved by varying the speed of the feeder.
=oting the foregoing comments, the salient aspects of the various types of feeders commonly used to feed bulk
solids onto belt conveyors are now briefly reviewed.
0./ V&brator* Fee3ers
0./.. Ge!era' Rear2s
Vibratory feeders are used extensively in controlling the discharge of bulk solids from bins and stockpiles and
directing these materials onto conveyor belts. They are especially suitable for a broad range of bulk solids, being
able to accommodate a range of particle siIes and being particularly suitable for abrasive materials. 4owever
they are generally not suited to fine powders under '8A to ?AA mesh where flooding can be a problem. !lso
7sticky7 cohesive materials may lead to buildup on the pan leading to a reduction in flow rate.
9ulk solids are conveyed along the pan of the feeder as a result of the vibrating motion imparted to the particles
as indicated in -igure :.

-igure : #ovement of particles by vibration
The pan of the feeder is driven in an approximate sinusoidal fashion at some angle theta to the trough.
The conveying velocity and throughput depend on the feeder drive frequency, amplitude or stroke, drive angle
and trough inclination, coefficient of friction between the bulk solid and the pan as well as the bulk solid
parameters such as bulk density, particle density and general flow properties.
0././ T*pes o" )&brat&!( Fee3ers
In general vibrating feeders are classified as 7brute force7 or 7tuned7 depending on the manner in which the driving
force imparts motion to the pan.
!s the name implies 7brute force7 type feeders involve the application of the driving force directly to the pan as
illustrated in -igure (. These feeders have the following characteristics.
Hower initial cost but higher operating costs.
3reater forces to be accommodated in the design.
Impact loads on the pan are transmitted to bearings on which outofbalance weights rotate.
;elivery rates are dependent an the feeder load due to bulk solids.
3enerally confined to applications requiring only one feed rate.
$n the other hand 7tuned7 vibrating feeders are more sophisticated in their operation in as much as the driving
force is transmitted to the pan via connecting springs as indicated in -igure *. In this way they act essentially as
a two mass vibrating system and employ the principle of force magnification to impart motion to the pan. The
primary driving force is provided by either an electromagnet or by a rotating outofbalance mass system.

-igure ( 79rute7 force type vibratory feeders Lef. &'()

-igure * Tuned type vibratory feeders Lef. &'()
-ollowing the work of Lademacher &'*) some general comments may be made. =ormally the trough mass is
designed to be ?.8 to E times as large as the exciter mass. -igure 'A shows typical magnification curves for the
tuned feeder for two damping ratios M' and M?. =ormally the feeders operate below the resonance frequency with
N/NA F A.* where N F driving frequency and NA F natural frequency of the system. It is often claimed that the
7tuned7 feeder maintains its feed rate when the head load varies. The validity of this claim may be examined by
reference to -igure 'A. The increased head load effectively increases the pan mass lowering the natural
frequency NA and increasing the frequency ratio N/NA. This corresponds to a shift from ! to 9 in the diagram. !t
the same time, the damping increases due to the increased load resistance causing a shift from 9 to 5. Thus the
claim that the feeder maintains its performance regardless of the head load, basically, is not true. 4owever in
many instances the negative effect of the increase in damping approximately compensates for the change in
magnification factor. This means that points ! and 5 in -igure 'A are approximately of the same magnification
factor so that the trough stroke is kept approximately constant.

-igure 'A Typical magnification curve for a tuned two mass system <' O <? Lef.&'*)
! similar analysis in the case of the over critical operation with N/NA P ' indicates that increases in N/NA and M
leads to a smaller, magnification factor corresponding to a smaller stroke. -or this reason over critical operation of
this type of feeder must be avoided.
0./.0 Hopper4Fee3er Co!"&(-rat&o!s
There are several aspects to note when designing feed hoppers for use with vibrating feeders. These are
discussed at some length by 5olijn and 5arroll &?A). +ome important aspects are noted here.
-igure '' shows a typical vibrating feeder arrangement. The effectiveness of the feeder, as with all feeders,
depends largely on the hopper which must be capable of delivering material to the feeder in an uninterrupted
way.
-or a symmetrical hopper there is a tendency for the feeder to draw material preferentially from the front of the
hopper. .niform draw can be achieved by making the hopper outlet asymmetrical with the back wall at the
correct hopper half angle @ and the front wall at an angle of @ Q 08B to (B1.
0The angle @ is obtained from -igure E1. !lternatively a symmetrical hopper may be made to feed approximately
uniformly by using a rougher lining material on the front face. $ther recommendations include
;imension < to be at least '8A cm.
9 to be large enough to prevent arching or ratholing.
+lope R to be sufficient for the required flow rate
3ate height 4 to be chosen primarily to achieve an acceptable flow pattern rather than to vary the
flowrate.
-or high capacity feeders skirtplates extending to the outlet of the trough may be required as in -igure
'?.

-igure '' Typical arrangement for vibratory feeder
In the case of wedgeshaped planeflow bins problems arise when it is necessary to feed from the long slotted
outlets. It is always theoretically better to feed across the slot as in -igure 'E0a1 but the cost of wide feeders to
achieve this goal often becomes prohibitive. ! better solution may be to use a multioutlet arrangement as in
-igure 'E0b1 and employ several narrower vibrating feeders to feed across the slot as shown.

-igure '? Tapered skirts Lef. &?A)
%here it is necessary to feed along the slot as in -igure '" 0a1 , then tapering of the outlet in the direction of the
feed is required as indicated in -igure '"0b1. $nce again it is reiterated that the adjustable gate is a flow
trimming device rather than a flow rate controlling device.

-igure 'E !lternative arrangements for feeding across the slot wedgeshaped hopper

-igure '" !rrangement for feeding along slot of wedgeshaped hopper Lef. &'()
0.0 Be't Fee3ers
9elt feeders are used to provide a controlled volumetric flow of bulk solids from storage bins and bunkers. They
generally consist of a flat belt supported by closely spaced idlers and driven by end pulleys as shown in -igure '8.
In some cases, hoppers feed directly onto troughed conveyors as in the case of dump hoppers used in
conjunction with belt conveyors.
+ome particular features of belt feeders include
+uitable for withdrawal of material along slotted hopper outlets when correctly designed.
5an sustain high impact loads from large particles.
-lat belt surfaces can be cleaned quite readily allowing the feeding of cohesive materials.
+uitable for abrasive bulk solids.
5apable of providing a low initial cost feeder which is dependable on operation and amenable to
automatic control.
%ith respect to the first point, the hopper and feeder geometry for long slots are critical if uniform draw is to be
obtained. %hile normally feeders are installed horiIontally, on some occasions a feeder may be designed to
operate at a low inclination angle 8. The outlet should be tapered as shown in the plan view of -igure 'G.
Lesearch &?'?") has shown that the taper angle R and downslope angle S together with the gate opening 4 are
very sensitive as far as obtaining efficient performance is concerned. In particular, as stated previously, the gate
opening 4 should be used to train the flow pattern and not to control the flow rate. !s has been demonstrated by
experiment &?E), incorrect setting of the gate will cause non uniform draw with funnelflow occurring either down
the back wall or down the front wall. In one series of experiments using a free flowing granular type material,
merely increasing the gate setting 4 causes the flow to move progressively towards the front. The final gate
setting needs careful adjustment if uniform draw is to be achieved. Thus in belt feeders flow rate variations must
be achieved by varying the belt speed. This requirement places some limitations on belt feeders when very low
flow rates are required, especially if the bulk solid is at all cohesive or contains large lumps.

-igure '8 !rrangement for belt feeder
Karticular care is needed with the design of the hopper/feeder arrangement when handling fine powders in order
to ensure that problems of flooding are avoided. If the bulk material tends to stick to the belt, spillage may be a
problem with belt feeders. Therefore if sufficient headroom is available, it is desirable to mount the feeder above
the belt conveyor onto which it is feeding material in order that any material falling from the return side of the
belt will automatically fall onto the conveyor belt.

-igure 'G 9elt feeder sited above conveyor to minimise spillage Lef. &*)
9elt feeders can also have applications where a short speedup belt is used to accelerate the material at the
loading point of a high speed conveyor as illustrated in -igure ':. The accelerating conveyor avoids wear that
would otherwise occur to the cover of the long conveyor.

-igure ': 9elt feeder as acceleration conveyor Lef. &?8)
0.5 Apro! Fee3ers
!pron feeders are a version of belt feeders and are useful for feeding large tonnages of bulk solids being
particularly relevant to heavy abrasive ore type bulk solids and materials requiring feeding at elevated
temperatures. They are also able to sustain extreme impact loading. The remarks concerning the need for
uniform draw and gate settings applicable to belt feeders are also applicable to apron feeders. -igure '(0a1 shows
an apron feeder with parallel outlet which is inducing funnelflow down the rear wall of the hopper. !part from the
obvious flow problems, the funnelflow pattern developed will accelerate the wear down the rear wall. The
tapered outlet of -igure '(0b1, when correctly designed will induce uniform draw, minimising segregation and
minimising hopper wall wear.

-igure '( !pron feeders
0.6 P'o-(% Fee3ers
Lotary plough feeders are generally used in long reclaim tunnels under stockpiles where they travel along the
tunnel as in -igure '* or fixed under stockpiles and large storage bins, as shown in -igure ?A.
In the case of the stockpile slot reclaim system of -igure '*, it is necessary for the diagonal dimension of the slot
to be at least equal to the critical rathole diameter ;f of the bulk solid in order to prevent ratholes from forming
under the high storage pressures &'8). In this way the gravity reclaim efficiency is maximised. The tie beams
between slots should be steeply capped. -urthermore the slot width 9f must be large enough to prevent arching.

-igure '* Typical stockpile with paddle feeder reclaim system

-igure ?A -ixed plough feeder
The basic concept of the traveling plough feeder is to allow bulk solids to flow by gravity onto a stationary shelf
and then remove the solids from the shelf either with a linear drag plough or a traveling rotary plough. it is
important that high penetration of the plough is achieved and that there is a small vertical section behind the
plough to prevent material buildup on the sloping back wall. ! high penetration rotary plough feeder is shown in
-igure ?'.

-igure ?' 4igh penetration plough feeder
5olijn and Vitunac &?G) have reviewed the application of plough feeders in some detail. They recommend the
following limiting values for plough feeders,
a. Lotary Klough
o Tip speed O ?A m/s
o Tip diameter range '.( to " m
o 5arriage speed A.A' to A.A( m/s0does not contribute significantly to capacity1.
b. Hinear ;rag Klough
o Traversing speed A.'E to A.:G m/s
o 5ut width '." to ?.E m.
0.7 Rotar* Tab'e Fee3ers
The rotary table feeder can be considered as an inverse of the plough feeder. It consists of a power driven circular
plate rotating directly below the bin opening, combined with an adjustable feed collar which determines the
volume of bulk material to be delivered. ! typical rotary feeder arrangement is shown in -igure ??. The aim is to
permit equal quantities of bulk material to flow from the complete bin outlet and spread out evenly over the table
as it revolves. The material is then ploughed off in a steady stream into a discharge chute.

-igure ?? Lotary table feeder
This feeder is suitable for handling cohesive materials which require
large hopper outlets, at flow rates between 8 and '?8 tonnes per hour. -eed rates to some extent are dependent
on the degree to which the material will spread out over the table. This is influenced by the angle of repose of the
material which varies with moisture content, siIe distribution and consolidation. These variations prevent high
feed accuracy from being obtained.
Lotary table feeders are suitable for bin outlets up to ?.8 m diameter2 the table diameter is usually 8A to GAJ
larger than the hopper outlet diameter. %ith some materials a significant dead region can build up at the centre
of the table. This can sometimes be kept from becoming excessive by incorporating a scraping bar across the
hopper outlet. It is important to ensure that the bulk material does not skid on the surface of the plate, severely
curtailing or preventing removal of the bulk material.
0.8 S$re+ Fee3ers a!3 D&s$%ar(ers
0.8.. S$re+ Fee3ers
+crew feeders are widely used for bulk solids of low or Iero cohesion such as fine and granular materials which
have to be dispensed under controlled conditions at low flow rates. 4owever, as with belt feeders, design
difficulties arise when the requirement is to feed along a slotted hopper outlet, -igure ?E, !n equal pitch,
constant diameter screw has a tendency to draw material from the back of the hopper as in -igure ?E0a1. To
counteract this, several arrangements are advocated for providing an increasing screw capacity in the direction of
feed as in -igure ?E0b1 to 0f1. The arrangements shown are,
+tepped pitch
Variable pitch
Variable pitch and diameter
Variable shaft diameter.
Kitch variation is generally limited to a range between A.8 diameters minimum to '.8 diameters maximum. This
limits the length to diameter ratio for a screw feeder to about six, making them unsuitable for long slots.

-igure ?E +crew feeding along a slot
The section of the screw leading from the hopper to the feeder outlet is fundamental in determining the quantity
of material discharged per revolution of the screw. !t the point where the screw leaves the hopper, it is essential
for control purposes to cover the screw, normally by a 7choke7 section having the same radial clearance as the
trough. This choke section should extend for at least one pitch to prevent material cascading over the flights.
!s a screw feeder relies on friction to transport material it has a very low efficiency in terms of the energy
requirements. -urthermore, the volumetric efficiency is impaired somewhat due to the rotary motion imparted to
the bulk material during the feeding operation &?:).
+ince screw feeders are generally fully enclosed, relatively good dust control is achieved. 4owever due to the
high frictional losses abrasive type bulk solids can effectively reduce the life of the feeder due to abrasive wear.
-ine powders that tend to flood are difficult to control in a screw feeder in flooding situations.
0.8.. S$re+ D&s$%ar(ers
+crew dischargers are variations of the normal screw feeder. Two of the more commonly used versions are shown
in -igure ?". -igure ?"0a1 shows a single screw which is forced to circle slowly around the bottom of a flat bottom
storage silo. The screw rotates at the same time and slices the bulk material, transferring it to a central discharge
chute. In -igure ?"0b1 the whole floor of the silo rotates about a fixed axis. The bulk material is forced against
the rotating screw as the silo bottom rotates.

-igure ?" Various screw discharge arrangements
+crew discharges have been used successfully with some wet, sticky, bulk solids which have not been handled
effectively using other means. In addition to providing the necessary flow promotion, these devices also control
the feed rate. Kroblems could arise when devices of this type suffer breakdown need careful investigation when
considering a screw discharge device for use in a particular application.
!n alternative application of screw discharges is in the <urosilo &?() which is shown schematically in -igure ?"0c1.
This design of silo, originated in the =etherlands and, as indicated, the screws sweep around the top surface
drawing material to the central discharge channel. The screws are also used to distribute the bulk material during
filling. The <urosilo was originally developed as an inexpensive storage facility, for potato starch but it is now
being used for other bulk materials, notably coal. It provides a very large capacity, environmentally clean storage
facility. Its principal disadvantage is that it operates on a firstin lastout sequence and hence is not
recommended for materials that degrade with time.
0.9 Rotar* Fee3ers
Lotary feeders 0also known as drum, vane, star and valve feeders1 are generally used for the volumetric feeding
of fine bulk solids which have reasonably good flowability.
! rotary drum feeder, -igure ?80a1, might be considered an extremely short belt feeder. The drum prevents the
bulk material from flowing out but discharges it by rotation. This feeder is only suitable for materials with good
flowability which are not prone to aeration. +imilar considerations apply to the rotary vane feeder, -igure ?80b1,
which might be considered as an extremely short apron feeder2 -igure ?80c1 shows some modifications to the
vane. The rotary valve feeder, -igure ?G0a1, is completely enclosed and aims at preventing powders or fine
grained materials from flooding. The star feeder, -igure ?G0b1, provides a means for obtaining uniform withdrawal
along a slot opening.
These feeders are not suitable for abrasive bulk materials as clearances cannot be maintained and the feeders
tend to lose control especially when handling aerated powders. 5ohesive powders will tend to clog the rotor
pockets and reduce feeder capacity.

-igure ?8 Lotary drum and vane feeders with various rotating elements Lef &'()

-igure ?G Lotary valve T star feeders
0.: Fee3er Se'e$t&o!
The selection of a feeder for a particular situation is not always simple, especially if more than one satisfactory
solution appears possible. The type and siIe of feeder for a given application is primarily dictated by the
characteristics of the bulk material to be handled and the required capacity. +ome general guidelines on feeder
selection are given in Leferences &'(,'*).
5. FEEDER LOADS AND POWER RE;UIRE#ENTS
5.. Ge!era' Rear2s
-rom a design point of view it is important to be able to determine with some accuracy the loads acting on
feeders in hopper/feeder combinations and the corresponding power requirements. Uet the stateoftheart has,
in the past, been such that the loads and power requirements could not be estimated with any degree of
precision. -or instance %right &?*) has observed that the majority of formulae published are empirical in nature
and derived to predict loads and corresponding power requirements for feeders used in conjunction with funnel
flow bins. These formulae are inadequate when applied to massflow bins since, in such cases, the loads and
power requirements are often greatly underestimated. This is largely due to the fact that in massflow bins the
full area of the hopper outlet is presented to the feeder.
The loads acting on feeders can vary considerably. There are many reasons for this, some more obvious than
others. !s indicated by Leisner and Lothe &'(), the shape of the hopper outlet will influence the load on a feeder
as illustrated in -igure ?:. In -igure ?:0a1, the full load 0not equal to the hydrostatic head1 acts on the feeder. In
-igure ?:0b1 the load is partly reduced by changing the shape of the hopper. In -igure ?:0c1, the load is
completely removed from the feeder and only acts on the hopper wall. !lthough the advantages of -igure ?:0b1
and 0c1 appear obvious, the solution may not be as simple as that depicted. It is clear that the flow pattern
developed in the feeding operation must be such that uniform, nonsegregated flow is achieved at all times.
The loads acting on feeders and corresponding power requirements are influenced by several factors. These
include the following,
4opper flow pattern, whether massflow or funnelflow
-low properties of the bulk solid
The chosen hopper shape which in the case of massflow includes axisymmetric or conical, planeflow
or transition 0combination of conical and planeflow1
The actual hopper geometry
The wall friction characteristics between the bulk solid and hopper walls and skirtplates
The type of feeder and its geometrical proportions
The initial filling conditions when the bin is filled from the empty condition and the flow condition when
discharge has occurred.
The most efficient and reliable feeding performance is achieved by using a massflow hopper/feeder combination.
-or a given bulk solid and hopper/ feeder geometry the load acting on a feeder varies considerably between the
initial load, when the bin is first filled, and the load either during flow or after flow has stopped. Leisner &'() has
indicated that the initial load can be ? to " times the flow load. 4owever, research &?',??,?") has shown that the
variation is much greater than this with the initial loads of the order of " to ( times that of the flow load.
Theoretical predictions show that circumstances can arise whereby the initial/flow load variations can be much
higher than those indicated.
! procedure for estimating feeder loads for massflow/feeder combinations has been established &'E,EA). The
general procedure is now reviewed.

-igure ?: Varying the load on the feeder by varying the hopper configuration &'()
5./ Press-re D&str&b-t&o!s &! #ass1F'o+ B&!s
It is first necessary to examine the pressures acting in massflow bins under both initial filling and flow
conditions. The pressures in massflow bins are discussed in some detail in Leferences &'E,E'E:).
-igure ?( shows the bin stress fields and corresponding pressure distributions for the initial filling and flow cases.
In each case, pn represents the pressure acting normal to the bin wall while pv represents the average vertical
pressure.

-igure ?( Kressures acting in massflow bins
5./.. I!&t&a' F&''&!( Case 1 F&(-re /9<a=
In this case, vertical support is provided and the major consolidating principal pressure is nearly vertical. !
peaked stress field exists in both the cylinder and hopper as indicated.
0i1 5ylinder
The normal wall pressure pn is given by the >anssen equation with an initial surcharge pressure term included.
pn F
VL
&'e
WXj h/L

) Q pno
e
WXj h/L

0'1
W
where L F
;
F 4ydraulic radius 0?1
?0'Qm1
Y F 9ulk +pecific %eight
; F 5ylinder diameter or width
m F ' for axisymmetric or circular bin
m F A for long rectangular planeflow bin
W F tanR F coefficient of wall friction
R F wall friction angle
Kno F initial pressure at top contact point with wall due to natural surcharge of material caused by filling.
The coefficient Xj relates the normal pressure pn to the average vertical pv. That is
Xj F pn/pv 0E1
The value of Xj varies depending on the cylinder geometry. -or parallel sided cylinders with slight convergence7s
>enike suggests
Xj F A." 0"1
-or continuously diverging or stepwise diverging cylinders without convergence7s, the theoretical value of Xj may
be used
Xj F
'sinD
081
'QsinD
%here D F effective angle of internal friction for the bulk solid
-or continuously converging cylinders Xj is approximated by
Xj F '
The initial surcharge pressure pno may be estimated by
pno F Xj V hs 0G1
%here hs F effective surcharge.
hs depends on the bin shape and manner in which the bin is loaded. !ssuming central loading, then for an axi
symmetric bin a conical surcharge is assumed2 for a planeflow bin in which the length is greater than the width,
an approximate triangular shape would occur if a travelling feeding arrangement, such as a tripper, is used. %ith
these two limits
hs F
4s
0:1
msQ?
where ms F A for triangular surcharge
ms F ' for conical surcharge
4s F actual surcharge
-or a plane flow bin, some rounding of the ends of the top surface of material is likely to occur so that an
approximate intermediate value of ms can be used.
The >anssen curves for the cylinder are shown in -igure ?(0a1.
0ii1 4opper
-ollowing >enike &E:), the normal pressure acting against the wall is related to the average vertical pressure by a
parameter X such that
X F
Kn
0(1
pv
-rom an equilibrium analysis for the hopper the following differential equation is obtained.
dpn
Q
npn
F V X 0*1
dI 0hoI1
+olution of this equation leads to
pn F VX Z0
hoI
1 Q &hc
ho
)&
hoI
)
n
[ 0'A1
n' n' ho
where n F 0mQl1 ZX 0' Q
W
1 '[ 0''1
tan@
@ F hopper half angle
hc F surcharge head acting at transition of cylinder and hopper
ho F distance from apex to transition
hc is given by
hc F
\c
0'?1
V!c
where
\c
is derived from the >ansen <quation 0'1
!c
That is
\c
F
VL
&'e
WXj 4/L

) Q Vhse
WXj 4/L

0'E1
!c WXj
where \c F surcharge force at transition
!c F area of cylinder
4 F height of material in contact with cylinder walls.
-or the initial filling condition in the hopper, >enike &E:) assumes that the average vertical pressure distribution in
the hopper follows the linear hydrostatic pressure distribution. -or this condition the value of X in 0'A1 is the
minimum value. That is
X F Xmin F
tan@
0'"1
WQtan@
+ubstitution into 0''1 yields n F A.
4ence 0'A1 becomes
Kn F V Xmin0hcQI1 0'81
and
Kv F
Kn
F V0hcQI1 0'G1
Xmin
The linear relationships for pn and pv for the hopper are shown in -igure ?(0a1.
5././ F'o+ Case 1 F&(-re /9 <b=
In this case with the vertical support removed, the load is transferred to the hopper walls and the peaked stress
field switches to an arched stress field. !s flow is initiated from the hopper the switching from a peaked to arched
stress field commences at the hopper outlet and travels upward. There is evidence to suggest that the switch
becomes locked at the transition and remains there during continuous flow. In the arched stress field, the major
principal pressure acts more in the horiIontal direction.
i. 5ylinder
;uring the flow, the peaked stress field remains in the cylinder and for a perfectly parallel cylinder
without localised convergences the >anssen stress field remains. 4owever, in practice, localised
convergences are difficult to avoid, such convergences being caused by weld projections or plate
shrinkages in the case of steel bins or irregular formwork in the case of concrete bins. -or this reason
an upper bound normal wall pressure based on >enikels strainenergy stress field &E:) is computed for
design purposes. The modified upper bound cylinder normal wall pressures is shown in -igure ?(.
ii. 4opper
The normal wall pressures for flow in the hopper are given by equation 0'A1 with X set at the maximum
value
X F Xmax 0':1
where Xmax is given in graphical form in Leference &E:) or else may be computed using equations
derived in Leference &'E). That is
Xmax F &
]w
/q0
"
1
m
) 0'(1
V9 ^
%here
]w
F
U0'QsinD cos?S1
0'*1
V9 ?0x'1sin@
and q F 0
^
1
m
'
Z?0
]w
10tan@ Q tanC1
'
[ 0?A1
E "tan@ V9 'Qm
%here S F _&C Q sin
'
0
sinC
1) 0?'1
sinD
x F
?
m
sinD
&
sin0?SQ@1
Q ') 0??1
'sinD sin@
y F
&?0'cos0SQ@11)
m
0SQ@1
'm
sin@ Q sinS sin
'Qm
0SQ@1
0?E1
0'sinD1sin
?Qm
0SQ@1
C F tan
'
W where C F wall pressure angle
D F effective angle of internal friction
m F ' for axisymmetric flow
F A for planeflow.
In equation 0??1 the planeflow numerate term 0@QS1 must be in radians.
5.0 T%eoret&$a' Est&ate o" Fee3er Loa3s a!3 Po+er
The theoretical prediction of the surcharge load acting at the outlet of a massflow hopper requires consideration
of both the initial and flow consolidation pressures acting on the bulk material. -igure ?* shows diagrammatically
the loads acting in a hopper feeder combination.
-igure ?* -eeder loads in hopper/feeder combination
-ollowing the approach adopted by #cHean and !rnold &EA) the surcharge \ acting at the hopper outlet is given
by
\ F q V H
'm
9
?Qm
0?"1
where,
q F nondimensional surcharge factor
V F specific weight of bulk solid at hopper outlet
F `g
` F bulk density at outlet
g F acceleration due to gravity
H F hopper outlet length
9 F hopper outlet width or diameter
m F A for planeflow or wedgeshaped hopper
m F ' for axisymetric or conical hopper
5.0.. A!a'*t&$a' E>press&o! "or I!&t&a' No!1D&e!s&o!a' S-r$%ar(e Fa$tor
The load acting on the feeder may be estimated by assuming the average vertical pressure Kv acts over the whole
area of the hopper outlet. -rom +ection ".?.', the average vertical pressure is given by equation 0'G1, when
multiplied by the hopper outlet area and combined with equation 0?"1 yields the following expression for the
initial nondimensional surcharge factor qi
qi F 0
^
1
m
'
&
;
Q
?\c tan@
') 0?81
? ?0mQ'1tan@ 9 !cV9
where
; F bin width
9 F hopper opening dimension
@ F hopper half angle
m F A for planeflow
m F ' for axisymmetric flow
#cHean and !rnold &EA) considered the surcharge at the outlet as being the difference between the weight of the
material in the hopper section plus the surcharge \c at the transition minus the vertical wall support. They used
the original >enike method &'E) for the initial normal wall pressures in the hopper. The expression obtained is
qi F 0
^
1
m
'
&
;
Q
?\c tan@
') 0?G1
? ?0mQ'1tan@ 9 !cV;
It will be noted that equations 0?81 and 0?G1 are identical except for the denominator of the middle term of the
section contained within the square bracket2 in 0?81 the denominator includes 9 while in 0?G1 it is ;.
<vidence suggests that 0?G1 provides a good estimate of qi. +ince 0?81 will yield higher values of qi it is suggested
that it be regarded as an absolute upper bound. -or design purposes it is suggested that equation 0?G1 be used.
5.0./ A!a'*t&$a' E>press&o! "or F'o+ No!1D&e!s&o!a' S-r$%ar(e Fa$tor
-ollowing #cHean and !rnold &EA) the flow nondimensional surcharge factor qf is given directly by equation 0?A1.
That is
qf F a0
^
1
m
'
&
y
0
'sinD cos?S
10tan@ Q tanC1
'
) 0?:1
E tan@ x' sin@ 'Qm
where S, x and y are given by equations 0'*1, 0?A1 and 0?'1 respectively.
5harts for qf are also presented in Leference &'E).
5.5 Ep&r&$a' Approa$%es "or Est&at&!( Fee3er Loa3s
In this section the empirical approaches of Leisner &'(), 9ruff &E() and and >ohanson &E*) are given.
".".' Leisner7s #ethods
i. -low Hoad 9ased on ]'
Leisner postulates that, faced with higher experimental values of feeder loads, one should
consider the possibility that the stress field shifts and rotates above the feeder causing a
different vertical stress condition. The highest value would occur when the major consolidation
pressure 0]'1 acts downward. ]' is given by
]' F
V9 ff
0?(1
40@1
where ff F flow factor. 5harts for ff and 40@1 are given in &'E). %hile Leisner suggests that
this value should be used for vibratory feeders to predict feeder loads during flow, recent
research &?',??,?") suggests that the prediction is also applicable to belt feeders.
ii. -low Hoad 9ased $n ]%
Leisner suggests that the normal wall pressure at the hopper outlet 0]%1 0which is less than ]'
but greater than the mean consolidation stress ]1 provides a good approximation for belt,
apron and table feeders during flow. 0+ee comment at end of 0i1 above.1
]%/V9 is available in chart form &'E) or can be calculated from the equation 0':1 which is
repeated below
]%
F
y0'QsinD cos?S1
0?*1
V9 ?0x'1sin@
x and y are given by equations 0??1 and 0?E1 respectively. -or both the above approaches \f is
calculated from
\f F ]' 0or ]%1 x 4opper $utlet !rea.
iii. Initial Hoads
Leisner indicates that initial loads are ? to " times higher if the bin is filled from completely
empty and only '.' to '.? times higher if the bin is not completely emptied before refilling.
5.5./ Br-""?s #et%o3
9ruff &E() suggests that \ for flow conditions be approximated by taking the weight of a block of bulk
solid of height F " x L 0where L F hydraulic radius1 above the hopper outlet, -igure EA.
L F
crosssectional area of outlet
0EA1
perimeter of outlet
L for a circular outlet F 9/" 0E'1
L for slotted outlet F
H9
including end effects 0E?1
?0HQ91
F9/? neglecting effects 0EE1
\ for flow and initial conditions can then be calculated from the following equations
-or circular outlet \ F ^/" 9b Vcs 0E"1
-or slotted outlet including end effects
\ F
?Hd9d
Vcs 0E81
HQ9
-or slotted outlet neglecting end effects

\ F ?H9d Vcs 0EG1
where,
cs F " for initial filling conditions
F ' for flow conditions.
5.5.0 @o%a!so!?s #et%o3
>ohanson &E*) suggests a similar empirical approach to 9ruff, for flow conditions, except that he uses
half 9ruff7s values and always neglects the end effects for a long slotted outlet, -igure E'. >ohanson
makes no recommendations for initial filling conditions.
5.6 Po+er to S%ear B-'2 So'&3 &! Hopper
Xnowing the load acting on the feeder, the force required to shear the bulk solid tangentially at the hopper outlet
may be estimated. -or a belt or apron feeder, the force to shear the material is approximated by
- F W' \ 0E:1
Various authors assume different values of W'. -or instance
-igure EA 9ruff7s approximation for \f
-igure E' >ohanson7s approximation for \f
Leisner and 9ruff assume W F A."
#cHean and !rnold, and >ohanson assume
W F sinD 0E(1
where D F effective angle of internal friction.
The power required to the shear bulk at the hopper outlet is
p F -v 0E*1
%here v F belt or apron speed.
It is to be noted that depending on the feeder type, additional forces may be exerted on the feeder. -or example
force due to material contained within skirtplates
additional load due to material on a belt, trough or table.
It may also be necessary to determine the resistances and powers due to other factors such as
+kirtplate resistance
belt resistance
These aspects are discussed in more detail in +ection 8.
5.7 E>ap'e
5.7.. Prob'e Des$r&pt&o!
It is required to determine the loads exerted at the outlet of the planeflow wedgeshaped bin of -igure
E?.

-igure E? %edgeshaped planeflow bin for feeder load example
The relevant details are as follows,
i. 9in
$pening ;imension 9 F '.8 m
4eight 4 F (.A m
%idth ; F 8.A m
+urcharge 4s F '.8 m
4eight of hopper 4h F ?.'G m
4alf angle @ F E*B
Karallel section of bin F mild steel
4opper section F Hined with stainless steel type EA"?9
Hength of opening H F 8 m
ii. 9ulk +olid Type F coal
<ffective angle of
internal friction

D

F 8AB
!ngle of friction between
coal and mild steel

C

F EAB
!ngle of friction between
coal and stainless
steel


Cn


F '(B
9ulk density ` F A.*8 t/mb
=$T< , The coal type and hopper geometry is based on that given in -igure G.
5.7./ So'-t&o!
i. +tatic +urcharge -actor qi
!ssuming material surcharge on top of the bin is of conical shape, ms F '. Then from 0':1
hs F
'.8
F A.8m
'Q?
5ylinder is square in crosssection. Thus from 0?1
L F
8
F '.?8
?x?
+urcharge at the transition from 0'E1, assuming Xj F A."

\c
F
A.*8x*.('x'.?8
&'e
tanE* x A." x(/'.?8
)QA.*8x*.('xA.8e
tanE*xA."x(/'.?8
!c tanE* x A."
F "A.8E kKa
-rom 0?G1 with m F A for planeflow
qi F
'
&
8
Q
?x"A.8EtanE*
') F ?.E'
'x?xtanE* '.8 A.*8x*.('x8
ii. -low +urcharge -actor
-rom 0?'1
S F _&'( Q sin
'
0
sin'(
1) F ?A.(*
sin8A
iii. -rom 0??1 and 0?E1
x F
sin8A
&
sin0?x?A.(*QE*1
Q ')
'sin8A sinE*
iv.
F (."'
y F
'.A"8sinE*Qsin?A.(*xsin0?A.(*QE*1
0'sin8A1sind0?A.(*QE*1
v.
F 8.8?
vi. -rom 0?:1
qf F
'
&
8.8?
0'Qsin8Asin"'.(10tanE*Qtan'(1 ')
"tanE* :."'
vii.
F A.E"
viii. -eeder Hoads
.sing 0?"1
Initial \i F ?.E' x A.*8 x *.(' x 8 x '.8d
F ?"?.? k=
-low \f F A.E" x A.*8 x *.(' x 8 x '.8d
F E8.*E k=
ix. <mpirical Values
These have been determined using the methods outlined in +ection ".". The results, together with
those above are summarised in Table '.

T!9H< ' +.##!LU $- KL<;I5T<; -<<;<L H$!;+
So-r$e
I!&t&a' $o!3&t&o!s
;& <2N=
F'o+ Co!3&t&o!s
;" <2N=
;&
;"
;&
;"A .B
Theoretical ?"?.?8 E8.*E G.:" ?.?'
Leisner ]' "E(.E 'A*.8: ".A ".A
Leisner ]w E*A.A *8.8 ".A( E.8G
9ruff Incl. <nd <ffects G"8.? 'G'.EA ".A 8.(*
9ruff <xcl. <nd <ffects (E(.( ?A*.: ".A :.GG
>ohanson 'A".(
e Latio of \i to \f]' for Leisner computed on basis of ]'.
5.8 Soe Coe!ts
-or the given hopper geometry and coal properties, the results in Table ' show the theoretical initial and flow
values to be the lowest respectively for all the methods. The ratio of initial load to flow load, \i/\f is of significant
magnitude as shown. !s pointed out previous'y, Leisner &'() indicates that the initial load can be ? to " times
the flow load. 4owever research to date &?',??, ?") suggests a ratio of " to ( which confirms the theoretical ratio
\i/\f of G.:( given in Table '. -igure EE shows a typical filling and flow feeder load variation obtained in
laboratory tests on a planeflow stopper and belt feeder.
-igure EE -eeder load variations for planeflow hopper and belt feeder
The research &?',??,?") has also indicated that the theoretical value of \i given by equation 0?G1 provides a good
estimate. 4owever the evidence suggests that the theoretical value of \f determined using equation 0?:1 under
estimates the flow load and that Leisner7s method, based on the major consolidating pressure ]' at the outlet,
provides a better estimate. The last column in Table ' shows the ratio of \i to \f]i where \f]i is value given by
Leisner7s method.
It is useful to examine the influence of variation in hopper geometry on the magnitude of the feeder loads. -igure
E" shows the variation in the nondimensional surcharge factors while -igure E8 shows the variation of the
theoretical feeder loads as a function of hopper opening dimension 9 for the bin of the previous example. The
principal dimension of the bin are maintained with the exception of 9, @ and 4h. 9 and @, are varied in accordance
with -igure G, while 4h is adjusted to accommodate these variations. The decrease in \i and \f with decrease in @
and 9 to maintain massflow is clearly evident. !lso \f]i for the Leisner method also decreases. 4owever the
ratios \i/\f and \i/\f]' increase with decrease in 9.
-igure E" Variation of nondimensional surcharge factor for bin of example ".G.'.
5.9 Co!tro''&!( Fee3er Loa3s
The high initial loads which may act on feeders are a matter of some concern and where possible steps should be
taken to reduce the magnitude of these loads. %hile it is possible to limit the loads by a suitable arrangement of
hopper outlet as depicted in -igure ?:, it is important to ensure that the hopper outlet is fully active if massflow
is to be maintained.
+ince the initial load only ever occurs when the bin is filled from the empty condition without discharge taking
place, it is good practice to always maintain a cushion of material in the hopper &??). -rom a practical point of
view this is most desirable in order to protect the hopper surface from impact damage during filling.
-igure E8 Variations in feeder loads with hopper geometry planeflow bin of example ".G.'
4owever there is a further advantage2 the material left in the hopper as a cushion, having previously been in
motion, will preserve the arched stress field. The new material being deposited in the bin will initially have a
peaked stress field. This will provide a surcharge load an the arch field, but the load at the outlet will be of lower
order than if the bin is totally filled from the empty condition. The stress condition and reduced loading is
illustrated in -igure EG.
-igure EG 5ushioning in hopper to reduce feeder load
6. FEEDING OF BULK SOLIDS FRO# BIN ONTO BELT CONVEYOR
In the handling of bulk solids, belt feeders with skirtplates are commonly used. In other cases dump bins are
used in combination with belt conveyors as illustrated in -igure E:. -or design purposes it is necessary to
determine the belt loads under initial filling and flow conditions and the corresponding drive powers &'').
-igure E: +chematic arrangement of dump bin and belt conveyor
6.. Des&(! EC-at&o!s
6.... B&! a!3 Hopper S-r$%ar(e a!3 Correspo!3&!( Po+er
These are determined in accordance with the methods described in +ection ".
6.../ S2&rtp'ate Res&sta!$e
!ssuming steady flow, the skirtplate resistance may be determined as follows,
i. 4opper +ection
-sph F W? Xv0?\ Q `g9Hy1 y/9 0"A1
ii. <xtended +ection 0+ection beyond hopper1
-spe F W? Xv `g0HsH1yd 0"'1
where,
\ F feeder loads as determined by equation 0?"1
` F bulk density
y F average height of material against skirtplates
Xv F ratio of lateral to vertical pressure at skirtplates g F acceleration due to gravity F *.('
0m/sd1
9 F width between skirtplates
W? F skirtplate friction coefficient
Hs F total length of skirtplates 0m1
6...0 Be't Loa3 Res&sta!$e
i. 4opper +ection

-bh F 0\ Q `g9Hy1Wb 0"?1

ii. <xtended +ection

-be F `g90Hs H1yWb 0"E1
%here Wb F Idler friction.
6...5 Ept* Be't Res&sta!$e

-b F %b Hb Wb 0""1
%here,
%b F belt weight per unit length
Hb F total length of belt
6...6 For$e to A$$e'erate #ater&a' o!to Be't
-! F \m v 0"81
%here,
\m F massflow rate
v F belt speed
It is assumed that

\m F `9yv 0"G1
.sually the force -a is negligible.

6...7 &!&t&a' a!3 F'o+ Loa3s a!3 Po+ers
The foregoing loads and resistances are determined for the initial and flow conditions using the
appropriate values of the variables involved.
The power is computed from

K F 0f Lesistances1. v/c 0":1
%here c F efficiency and v F belt speed.
The condition for nonslip between the belt and bulk solid under steady motion can be determined as
follows,
WE 0\f Q w1 P 0-f Q -sp1 0"(1
%here,
WE Fcoefficient of friction between belt and bulk solid
\f Fflow surcharge at hopper outlet
% F weight of bulk material between skirtplates in hopper section of conveyor
-f Fforce to shear material at hopper outlet
-sp F skirtplate resistance.
6./ E>ap'e
6./.. Prob'e Des$r&pt&o!
!s an extension of the example of section ".G.', consider the bin being used in conjunction with a belt
conveyor. referring to -igure EG the relevant details are
i. 9in
$pening ;imension 9 F '.8 m
4eight 4 F (.A m
%idth ; F 8.A m
+urcharge 4+ F '.8 m
4eight of 4opper 4h F ?.'G m
4alf angle @ F E*B
Hength of bin H F 8 m
5ylinder F #ild steel
4opper lining F +tainless steel
ii. 9ulk +olid
Type F coal
<ffective angle of internal
friction

D

F 8AB
!ngle of friction between
coal and steel

C

F EAB
!ngle of friction between
coal and stainless steel


F '(B
9ulk density F ` F A.*8 t/mb
iii. 5onveyor
9elt velocity v F A.8 m/s
5onveyor length Hc F ?A.A m
Hength of skirtplate Hs F ?A.A m
!verage skirtplate height y F A.( m
9elt width 9b F ?.A m
6././ So'-t&o!
i. 4opper +urcharges

-rom +ection ".G
\i F ?"?.?
\f F E8.* k=
ii. -orce to +hear #aterial at 4opper $utlet
-i F \i sinD F ?"?.? sin 88 F '(8.G k=
-f F \f sinD F E8.* sin 88 F ?:.8 k=
iii. +kirtplate Lesistance 4opper +ection
<quation 0"A1
!ssume W? F tan ?AB for polished mild steel
F A.EG"
Xv F A." for initial case
Xv F A.G for flow case &"A)
Initial -sph0i1 F A.EG" x A." 0?x?"?.?QA.*8x*.('x'.8x8xA.(1 A.(/'.8
F "?.A 0k=1
-low -sph0f1 F A.EG" x A.G 0?xE8.*QA.*8x*.('x'.8x8xA.(1 A.(/'.8
F '".* 0k=1
iv. +kirtplate Lesistance <xtended +ection
<quation 0"'1
Initial -spe0i1 F A.EG" x A." x A.*8 x *.(' 0?A 81A.(d
F 'E.AE 0k=1
-spe0f1 F A.EG" x A.G x A.*8 x *.(' 0?A 81A.(d
F '*.8" 0k=1
v. 9elt Hoad Lesistance 4opper +ection
<quation 0"?1 !ssume Wb F A.AG
Initial -bh0i1 F 0?"?.? Q A.*8 x *.(' x '.8 x 8 x A.(1 x A.AG
F ':.* 0k=1
-low -bh0f1 F 0E8.* Q A.*8 x *.(' x '.8 x 8 x A.(1 x A.AG
F 8.8 0k=1
vi. 9elt Hoad Lesistance <xtended +ection
<quation 0"E1
Initial -be F 0i1 F A.*8 x *.(' x '.8 0?A 81 A.( x A.AG
F 'A.A: 0k=1
-low -be0f1 same as for initial condition
i.e. -be0f1 F 'A.A: 0k=1
vii. <mpty 9elt Lesistance
<quation 0""1
!ssume wb F
GA 9b
g F
GAx?.Ax*.('
'AAA 'AAA
F '.'( 0k=/m1
-b F '.'( x ? x ?A x A.AG
F ?.(E 0k=1
0!ctual belt weight will need to be checked after final belt selection is made1.
viii. !cceleration Lesistance
\m F A.*8 x '.8 x A.( x A.8 F A.8: t/s
-! F A.8: x A.8 F A.?(8 0k=1
ix. 5hecking =on+lip
-rom 0"(1
WE P
-f Q -sph0f1 Q -spe 0f1
\f Q w
P
?:.8 Q '".* Q '*.8"
E8.*E Q A.*8 x *.(' x ?A x '.8 x A.(
P A.?"
-or nonslip WE P A.?". -or coal on rubber belting this condition is easily satisfied.
x. Total Kower
!ssume c F *AJ
Initial case Ki F 0'(8.G Q "?.A Q 'E.A Q ':.* Q 'A.' Q ?.(E1x A.8/A.*
F '8' k%
-low case Kf F 0?:.8 Q '".* Q '*.8 Q 8.8 Q 'A.' Q ?.(E Q A.?*1x A.8/A.*
F "".( k%
The value of Kf will be larger if the Leisner value of \f is used.
The total initial power is based on the assumption that the total initial load acts with belt
velocity v during startup from the initial filling condition. In practice, the initial power is most
likely to be less due to the starting characteristics of the drive. -urthermore, once flow
conditions have been established, and the bin is kept nominally full, then startup from a
stopped condition will most likely occur at a much lower power corresponding to the flow
condition.
6.0 Be't Fee3ers 1 #ore R&(oro-s A!a'*s&s
In order to obtain satisfactory draw of material uniformly distributed over the full hopper outlet, as previously
discussed, it is usually necessary to taper the hopper bottom in the direction of feed. ! detailed study of the
forces and power requirements for belt feeders has been presented by Lademacher &'*,"A).
6.5 A$$e'erate3 F'o+ &! S2&rtp'ate Do!e
The skirtplate analysis presented in the example of +ection 8.? is based on the assumption of uniform flow. %hen
directing bulk solids onto conveyor belts through feed chutes, it is necessary to accelerate the solids to belt
speed. The motion is resisted by the drag imposed by the skirtplate. This problem has been analysed by Loberts
and 4ayes &?,'').
!s material accelerates, the depth h of the bed decreases as indicated in -igure E(.
In view of the shallow bed conditions, it may be assumed that the lateral pressure distribution is proportional to
the hydrostatic pressure distribution. -or each wall the average pressure is Xv `gh/?, where Xv is the pressure
ratio coefficient at the walls of skirtplates. 4ence for the two walls the total average lateral pressure is `gh Xv.
4ere h is the depth of the material and within the acceleration Ione varies inversely with the velocity vs. That is,
for a constant throughput \m 0kg/s1, it follows that
h F 0
\m
1
'
0"*1
`b vs
where ` is density of the material 0kg/mb1, b is width between skirtplates 0m1, and vs is velocity at section
considered 0m/s1. 4ere 0\m/`b1 F constant.
-igure E( +kirtplates in acceleration Ione
! dynamic analysis of the motion of the material in contact with both the belt and skirtplates shows that the
acceleration of the material is nonuniform and that the velocity vs as a function of distance l is nonlinear in
form. 4owever for simplicity in this case it will be assumed that the average height of the material is inversely
proportional to the average velocity based on linearity. That is, from equation 0"*1,
hav F 0
\m
1 0
?
1
`b vvo
4ence the drag force due to side plate friction can be obtained as follows,
-! F W? `g gb Xv 0hav1d
or
-a F
"W? g gb' \dm
08A1
` bd Xv0vvo1d
where W? is friction coefficient between material and skirtplates F A.8 to A.:, gb' is the acceleration length.
!ssuming, for simplicity, a blocklike motion of the bulk solid, an analysis of the forces due to the belt driving the
material forward and the skirtplates causing a resistance to forward motion shows that the acceleration of the
bulk solid is
a F W' g W? g Xv h/b 08'1
where,
W' F belt friction
W? F skirtplate friction
It is to be noted that the component W' g in 08'1 is the acceleration in the absence of skirtplates. +ubstituting for
h from 0"*1 gives
a F g0W'W?
Xv\m
1 08?1
`bdvs
where,
vs F bulk solid velocity at distance s from point of entry to belt.
<quation 08?1 shows that for a given throughput and skirtplate configuration the acceleration decreases as vs
decreases. This implies that there is a minimum value of vs to achieve the required feeding at the rate \m onto
the belt. The critical condition will be at the point of entry where s F A and vs F vo. Kutting a F A in 08?1 yields
vomin F
W? Xv \m
08E1
W' ` bd
%riting a F vs
dvs
in 08?1 and substituting from 08E1 yields
ds
dvs
F W' g0
'

vomin
1 08"1
ds vs vsd
Integration of 08"1 over the limits vs F vo to vs F v where v is the belt speed yields the following expression for
the acceleration length
gb' F
'
Z
0vdvod1
Q vomin 0vvo1 Q vod mingn &
vvomin
)[ 0881
W'g ? vovomin
<xample,
5oal of density ` F A.* t/mb is fed onto a conveyor belt at a rate of \m F (AA t/h with an initial velocity of vo F
A.8 m/s. The conveyor has skirtplates in the acceleration Ione. It is assumed that W' F A.8, W? F A." and, for
conservative design, Xv F '.A. The width between skirtplates b F '.A m. It is required to determine the
acceleration length. The belt speed v F E m/s.
+olution,
-rom 08?1 the minimum initial velocity is
vomin F
A." x '.A x (AA
F A.'*( m/s
A.8 x A.* x ' x ' x EGAA
-rom 08"1
gb' F
'
Z0
EdA.8d
1 Q A.'*( 0EA.81 Q A.'*( gn0
EA.'*(
1[
A.8 x *.(' ? A.8A.'*(
gb' F '.A'' 0m1
7. FEED CHUTES
7.. Ge!era' Rear2s
!s outlined in the introduction, the role of feedchutes is to direct bulk solids from bins and feeders onto conveyor
belts in a manner which will minimise spillage and belt wear. The chute may also be designed in a manner which
will ensure the component of the exit velocity tangential to the belt vT is matched as closely as possible to the
belt speed.
! typical chute arrangement is shown in -igure E*.
%hile the normal component V= of the exit velocity should be as small as possible to minimise impact damage to
the belt, it is necessary to ensure continuity of feed with sufficient chute slope to maintain flow and prevent
choking.
The flow characteristics of bulk solids in chutes has been the subject of considerable research, and general design
procedures have been presented &"'"8). It is beyond the scope of this paper to discuss the design of chutes2
reference may be made to the literature cited for details. 4owever it is worth mentioning the importance of the
friction characteristics of chute lining materials and the influence of friction on the selection of chute geometry.
7./ C%-te Fr&$t&o! a!3 S'ope A!('es
There are many lining materials available and these need to be selected on the basis of their frictional and wear
resistance properties. It is also important to consider any corrosive influence of the bulk solid on the hopper wall.
9y way of illustration, -igure "A shows the wall yield loci or friction characteristics for coal at '*J moisture
content 0d.b.1 on stainless steel, polished mild steel and rusted mild steel for the instantaneous condition as well
as the polished mild steel surface after :? hour storage. The increase in friction in the latter case is quite
considerable. It has been found that
-igure E* -eed 5hute for 9elt 5onveyor
-igure "A %all yield loci for coal at '*J
moisture content 0d.b.1
certain coals, for example, will build up on mild steel surfaces even after a short contact time of a few hours. The
type of behavior found to occur in practice is illustrated in -igure "'. #oist coal from a screen has been found to
adhere to vertical mild steel surfaces as indicated, particularly where the initial velocity of the coal in contact with
the surface is low.
-igure "' 9uild up of cohesive material on chute surfaces
!s indicated by -igure "?, the wall yield loci are normally slightly convex upward in shape. !lso, depending on the
surface 6roughness6 or rather 6smoothness6, moisture bulk solids may exhibit an adhesion component adhesion
often occurs with a smooth surface. The effect of the adhesion and/or the convexity in the wall yield locus is to
cause the wall friction angle to decrease as the consolidation pressure increases. This is illustrated in -igure "E.
The variation of friction angle with consolidation pressure must be taken into account when determining chute
slope angles. +ince bed depth on a chute surface is related to consolidating pressure it is useful, for design
purposes, to examine the variation of friction angle with bed depth. -igure "E shows, for a range of moisture
contents, the considerably high friction angles that can occur at low bed depths, the decrease in friction angle
being significant as the bed depth increases.
-igure "? %all yield locus and wall friction angles
-igure "E %all friction versus bed depth
-or a chute inclined at an angle R to the horiIontal, the relationship between bed depth and consolidation
pressure at the chute surface is
h F
]'
08G1
V cos R
The slope of the chute R should be at least 8B larger than the maximum friction angle measured. $ften moist
bulk solids will adhere initially to a chute surface particularly if the initial velocity tangential to the chute surface
is low. 4owever, as the bed depth increases, the corresponding decrease in friction angle will cause flow to be
initiated. In such cases flow usually commences with a blocklike motion of the bulk solid. This is depicted in
-igure "".
-igure "" 9locklike flow of bulk solid
9y far the greatest component of the drag force in a chute occurs along the chute bottom2 the side walls
contribute to a lesser extent. %here possible the side walls should be tapered outward or flared as indicated in
-igure "8 with gussets in the corners to prevent, or at least reduce, the buildup of material in the corners.
-igure "8 Lecommended chute configuration
8. CONCLUDING RE#ARKS
This paper has focused attention on the interactive role of storage bin, feeders and chutes in providing efficient
and controlled feeding of bulk solids onto conveyor belts. Various types of feeders have been reviewed and
methods for determining feeder loads and power requirements have been presented. The theoretical expressions
given by equations 0?"1 and 0?G1 appear to provide a good estimate of the initial feeder load while the
combinations of equations 0?"1 and 0?:1 give a theoretical prediction based on the radial flow stress theory for
the flow loads. 4owever the flow loads tend to be underestimated by this procedure and accordingly the method
due to Leisner based on the major consolidation stress ]' as given by equation 0?(1 is recommended as providing
a more realistic estimate. It is clear that more research is necessary to validate the predictions proposed.
!s a general comment it is worth noting that the modern theories of storage and feeding system design have
been developed over the past EA years with many aspects still being subject to considerable research and
development. It is gratifying to acknowledge the increasing industrial acceptance, throughout the world, of the
modern materials testing and design procedures. These procedures are now well proven, and while much of the
industrial development has, and still is, centered around remedial action to correct unsatisfactory design features
of existing systems, it is heartening that in many new industrial operations the appropriate design analysis and
assessment is being performed prior to plant construction and installation. It is more important that this trend
continue.
1. Loberts, !.%., 4ayes, >.%. and +cott, A.>., 6<conomic 5onsiderations in the $ptimum ;esign of
5onveyors for 9ulk +olids 4andling6, Kroc. International Kowder and 9ulk +olids 4andling and
Krocessing 5onference, Khiladelphia, ..+.!., #ay '*:*, 'A'''G.
2. Loberts, !.%. and 4ayes, >.%., 6<conomic !nalysis in the optimum ;esign of 5onveyors6, T.=L!, The
.niversity of =ewcastle, =.+.%., !ustralia, ?nd <dition, I+9= A :?8* E"A G, '*(A.
3. Loberts, !.%., 4ayes, >.%. and +cott, A.>., 6$ptimal ;esign of 5ontinuous 5onveyors6, 9ulk +olids
4andling, ' 0?1, l*('.
4. Loberts, !.%., 6<conomic !nalysis in the $ptimisation of 9elt 5onveyor +ystems6, Kaper presented at
9eltcon ', >ohannesburg, +outh !frica, +eptember '*('.
5. 4arrison, !., 4ayes, >.%. and Loberts, !.%., 6The -easibility of 4igh +peed =arrow 9elt 5onveyors for
9ulk +olids 4andling6, Trans. institution of <ngrs. !ustralia, #echanical <ngineering, Vol. #<:, =o. E,
+ept. '*(?.
6. Loberts, !.%., 6<conomic #odels in 9elt 5onveying6, +ymposium on 9elt 5onveying of 9ulk +olids,
T.=L! 9ulk +olids 4andling Lesearch !ssociates, The .niversity of =ewcastle, !ustralia, =ovember
'*(?.
7. 4arrison, !., 6<conomic !nalysis of 5onveyor 9elt +ystems !n Interaction #odel !pproach6,
+ymposium on 9elt 5onveying of 9ulk +olids, T.=L! 9ulk +olids 4andling Lesearch !ssociates, The
.niversity of =ewcastle, !ustralia, =ovember '*(?.
8. 4arrison, !., 6Transient +tresses in long 5onveyor 9elts6, +ymposium on 9elt 5onveying of 9ulk +olids,
T.=# 9ulk +olids 4andling Lesearch !ssociates, The .niversity of =ewcastle, !ustralia, =ovember
'*(?.
9. Lawlings, L.!., 6The Importance of <fficient -eeder ;esign6, 9ulk, >uly/!ugust '*::.
10. !rnold, K.5., 6-eeding of 9ulk +olids onto 5onveyor 9elts, -eeders and -eeder Hoads6, +ymposium on
9elt 5onveying of 9ulk +olids, T.=L! 9ulk +olids 4andling Lesearch !ssociates, The .niversity of
=ewcastle, !ustralia, =ovember '*(?.
11. Loberts, !.%., 6-eeding of 9ulk +olids onto 5onveyor 9elts Transfer 5hute Kerformance and ;esign6,
+ymposium on 9elt 5onveying of 9ulk +olids, T.=L! 9ulk +olids 4andling Lesearch !ssociates, The
.niversity of =ewcastle, !ustralia, =ovember '*(?.
12. >enike, !.%., 6+torage and -low of +olids6, 9ul. '?E, .tah <ngng. <xper. +tation, .niversity of .tah,
'*G".
13. !rnold, K.5., #cHean, !.3. and Loberts, !.%., 69ulk +olids, +torage, -low and 4andling6, The .niversity
of =ewcastle Lesearch !ssociates 0T.=L!1 Htd., ?nd <dition, ?nd printing '*(?.
14. Loberts, !.%., 69ulk +olids.. +torage, 4andling and -low6, Hecture =otes
for 5ourse organised by +outh !frican Institution of #echanical <ngineers, Land !frikaans .niversity in
!ssociation with the +outh !frican Institute of materials 4andling and .niversity of the %itswatersrand,
>une/>uly '*(?.
15. Loberts, !.%., !rnold, K.5., #cHean, !.3. and +cott, $.>. 6The ;esign of 3ravity +torage +ystems for
9ulk +olids6, Trans. Institution of <ngrs. !ustralia, #echanical <ngineering, Vol. #<:, =o. E, +eptember
'*(?.
16. Loberts, !.%., $oms, #. and +cott, A.>., 6Kractical ;esign and operational !spects of 9ulk +olids
+torage and ;ischarge -acilities6, Kroc. #ill $perators7 5onference, !ustralasian Inst. of mining and
#etallurgy, #t. Isa, !ustralia, +eptember '*(?.
17. $oms, #. and Loberts, !.%., 6The .se of -eeders and -low Kromotion ;evices in 3ravity +torage
+ystems for 9ulk +olids 4andling, Kroc. #ill $perators7 5onference, !ustralasian Inst. of #ining and
#etallurgy, #t. Isa, !ustralia, +eptember '*(?.
18. Leisner, %. and <isenhart Lothe, #.V., 69ins and 9unkers for 4andling 9ulk #aterials6, Trans. Tech.
Kubl., '*:'.
19. Lademacher, -.>.5., 6-eeders and Vibratory 5onveyors6, T.=L! 9ulk +olids 4andling Lesearch
!ssociates, '*(A.
20. 5olijn, 4. and 5arroll, K.>., 6;esign 5riteria for 9in -eeders6, Trans. +oc. of #ining <ngrs., !I#<, Vol.
?"', ;ec. '*G(, pp.E(*"A".
21. #urphy, K.5., 6-eeder Hoads6, 9.<. Thesis, The .niversity of %ollongong, !ustralia, '*(A.
22. #ann, 3.4., 6-eeder Hoads and -low Katterns in %edge+haped 9ins6, 9.<. Thesis, The .niversity of
%ollongong, !ustralia, '*('.
23. 4ookham, L., 6-low Visualisation of 9ulk #aterials in a 9elt -eeder6, 9.<. Thesis, The .niversity of
=ewcastle, !ustralia, '*('.
24. $rmerod, ;., 6;esign and Kerformance 5haracteristics of a 9elt -eeder +upplied with Variable +peed
;rive6, 9.<. Thesis, The .niversity of =ewcastle, !ustralia, '*(?.
25. 5onveyor <quipment #anufacturers7 !ssociation, 69elt 5onveyors for 9ulk #aterials6, 59I Kublishing 5o.
Inc., 9oston, #ass. ..+.!., ?nd <dition, '*:*.
26. 5olijn, 4. and Vitunac, <.!., 6!pplication $f Klow -eeders6, Kaper :*%!/#4 ', presented at !nnual
%inter #eeting, !+#<, =ew Uork, ?: ;ecember, '*:*.
27. Loberts, !.%. and %illis, !.4., 6Kerformance of 3rain !ugers6, Kroc. Instn. of #ech. <ngrs., Vol. ':G,
0(1, pp.'G8'*", l*G?.
28. Van der 9rock, +.<.;., 6Lecent ;evelopments and -uture Trends in Harge +cale +torage of 9ulk +olids6,
Kroc. Intl. Kowder and 9ulk +olids 5onference, Khiladelphia, ..+.!., #ay '*:*.
29. %right, 4., 69+57s 5ontribution to the ;esign and $peration of #ass-low 9unkers6, Iron and +teel
International, pp.?EE?E(, !ugust '*:(.
30. #cHean, !.3. and !rnold, K.5., 6! +implified !pproach for the <valuation of -eeder Hoads for #ass-low
9ins6, >ournal of Kowder and 9ulk +olids Technology, Vol. E,, =o. E, p.?8?(, '*:*.
31. >enike, !.%. and >ohanson, >.L., 6$n the Theory of 9in Hoads6, >ournal of <ngng. for Industry, Trans.
!+#<, +eries 9, Vol. *', =o. ?, EE*, '*G*.
32. >enike, !.%., >ohanson, >.L. and 5arson., >.%., 69in Hoads Kart ?, 5oncepts6, >ournal of <ngng. for
industry, Trans. !+#<, +eries 9, Vol. *8, =o. ', ', '*:E.
33. >enike, !.%., >ohanson, >.L. and 5arson, >.%., 69in Hoads Kart E, #ass-low 9ins6, >ournal of <ngng.
for Industry, Trans. !+#<, +eries 9, Vol. *8, =o. ', G, '*:E.
34. !rnold, K.5. and Loberts, !.%., 6! .seful Krocedure for Kredicting +tresses at the %alls of #ass-low
9ins6, Kaper presented at +ymposium on 6+olids -low and 4andling6, !I5h<, (Ath =ational #eeting,
9oston, +eptember '*:8.
35. #cHean, !.3. and !rnold, K.5., 6Krediction of 5ylinder -low Kressures in #ass-low 9ins using #inimum
+train <nergy6, >ournal of <ngng. for Industry, Trans. !+#<, +eries 9, *( 0"1, 'E:A, '*:G.
36. !rnold, K.5. and #cHean, !.3, 6%all Hoads in +teel 9ins6, +teel 5onstruction, >nl. !ust. Inst. +teel
5onstr., Vol. '?, =o. E, !ugust, '*:(.
37. >enike, !.%., 6Hoad !ssumptions and ;istributions in +ilo ;esign6, Kaper presented at 5onference on
5onstruction of 5oncrete +ilos, $slo, =orway, >anuary '*::.
38. 9ruff, %., 6Industrisiloer6, Ingeniorforlaget !/+, '*:", 0in =orwegian1.
39. >ohanson, >.L., 6+torage and -low of +olids6, E ;ay %orking +eminar =otes, !ustralian #ineral
-oundation, !delaide, >uly '*:G.
40. Lademacher, -.>.5., 6Leclaim Kower and 3eometry of 9in interfaces in 9elt and !pron -eeders6, 9ulk
+olids 4andling., Vol. ?, =o. ?, pp.?(',?*", >une '*(?.
41. Loberts, !.%., 6The ;ynamics of 3ranular #aterials -low through 5urved 5hutes6, #echanical and
5hemical <ngineering Transactions, Institution of <ngineers, !ustralia, Vol. #5E, =o. ?, =ovember '*G:.
42. Loberts, !.%., 6!n Investigation of the 3ravity -low of =oncohesive 3ranular #aterials through
;ischarge 5hutes6, Transactions of !.+.#.<. >nl. of <ng. in Industry, Vol. *', +eries 9, =o. ?, may '*G*.
43. Loberts, !.%. and !rnold, K.5., 6;ischarge 5hute ;esign for -ree -lowing 3ranular #aterials6,
Transactions of !.+.!.<., Vol. '", =o. ?, '*:'.
44. Loberts, !.%., +cott, A.>. and Karbery, L.;., 63ravity -low of 9ulk +olids through Transfer 5hutes of
Variable Krofile and 5ross+ectional 3eometry6, Kroceedings of international +ymposium on Kowder
Technology, Xyoto, >apan, +eptember '*('.
45. Loberts, !.%. and +cott, A.>., 6-low of 9ulk +olids through Transfer 5hutes of Variable 3eometry and
Krofile6, 9ulk +olids 4andling, vol. ', =o. ", pp.:'8, ;ecember '*('.