Flow Of Bulk Solids In Chute Design

John Rozentals

Acknowledgements : The Bionic Research Institute, Chute Design Conference 1991

A clear understanding of the flow of bulk solids within a chute is essential for rational design. This paper follows product flow within the chute from the discharge point on the feeder belt to the loading point on the receiving belt.

John Rozenthals graduated in Mechanical Engineering at the University of West Australia. Since graduating he has consulted on many materials handling projects including Hammersly Iron. Mt. Newman Mining Co. and Sishen Iron Ore. He is a partner in BRI Tec Consulting. an associate with Hamilton Associates and Director of the Bionic Research Institute.


THE DESIGN of any type of conveying system must meet two basic requirements:

  1. the system objectives - in terms of capacity, conveying distance, product distribution, and so on.
  2. the characteristics of the bulk solids product in terms of flow properties, degradation limitations, abrasion resistance of chute materials, and so on.

To meet the process objectives the conveyor system will require a number of transfer chutes.


Chutes are used at conveyor transfer points for a number of reasons:

  1. to control the direction of flow of the product
  2. to control the shape of the flow stream
  3. to control spillage
  4. to control dust and environmental pollution
  5. to reduce product degradation
  6. to retard or control flow
  7. to provide surge control

The simple conveyor belt to conveyor belt transfer of product is the single largest contributor to atmospheric contamination, and to the loss of valuable product within the plant. For this reason we must ask the question - why use chutes?

Current trends in the mining, and in the conveyor, industry have severely magnified the problem of dust generation in bulk solids handling systems. Developments in mining methods and equipment have resulted in product having smaller top lump size and a significant increase in the quantity of fines transported.

In addition the trend is to use narrower belts running at faster speeds. Higher speeds result in a significant increase in dust generation.

It makes good sense therefore to design conveyor systems using fewer chutes.


The prime thrust in the First International Chute Design Conference is to focus on the logical design of chutes. The aim is to help designers approach chute design with clearer understanding of the main issues.

After the conveyor design engineer has conceived the conveyor system he will go on to size critical items such as belts, idlers, motors, speed reducers, and the like. Often the engineer will leave chute detailing to someone else.

If the chute detailer has plenty of field experience the resulting transfer point will work well. It not, the result can be a poor system. The chute may not look radically different from a successful design, but performance differences may be sufficient to create significant problems and loss of profit.


Let us consider what happens as product flows through the chute. Firstly flow through a chute is a transient phenomenon happening in about 1 second. If the passage of product through the chute takes 2 seconds then it is along chute of some 20 m.

What happens to the product coming off the feed belt?

  1. some of the fines float into the surrounding airstream.
  2. some of the product sticks to the sides of the chute.
  3. some of the product gets carried back on the return belt.
  4. some of the product escapes from the chute and settles on the floor.
  5. at times product will block the chute and cause overflow.
  6. some of the fines end up in the head pulley and bearings.
  7. the balance of the product continues the journey to the next transfer point (Fig.1 ).

The essential aim of good chute design is to successfully transfer as much of the incoming product as possible from one belt to the next.

To assume that 100% of the incoming product will continue onto the next belt ignores the reality of transfer point design. Each of the above points need to be given serious consideration at the design stage.


How much of the incoming product will actually leave the belt?

To answer that question the chute designer must have a clear knowledge of the product characteristic properties. It is not sufficient to be given the name of the product and a grading analysis down to 6 mm. Much of the product behaviour is determined by its fines and powder components. And these are directly affected by the moisture content.

Consider a 0.1 mm thick film of iron ore on a return belt, say 1200 mm wide and running at 5 m/s. The quantity of fines discharged from the return belt is a 24 hour period would be approximately 100 tonnes.

Belt cleaning is an essential element of successful transfer point design.


The simplest discharge from a conveyor belt is to let the product pass over an end pulley and fall onto a pile. 

Air blowing through the product stream will have a winnowing effect- generating a dust cloud which, in many cases, may travel for hundreds of meters. Dust suppression measures may be required.

By adding a suitable chute the discharge may be directed as desired- to a stockpile, a bin, or to another conveyor. A fork at the discharge chute, with a gate, will permit the product to flow simultaneously in two directions, or alternatively in either direction.

Whenever the belt is discharged over an end pulley, the speed of the belt and the diameter of the end pulley are factors which determine the path of the discharged product. This path is called the primary trajectory. The shape of the discharged product trajectory is important when designing chutes.

There are several methods available to the designer for making trajectory predictions. These have been studied by Professor P.C. Arnold and G.L. Hill at the University of Wollongong, Australia. A copy of their paper is included in the Conference Manual. The salient points are as follows:

The product is carried toward the head pulley by the conveyor belt. If the belt is running at a high enough speed inertia will carry the product over the pulley in the same direction, and at the same speed, as the belt. The diameter of the pulley has no influence on the ejection velocity of the product. Formulae which take into account pulley diameters and material profile on the head pulley tend to overestimate the trajectory. 

If the belt is running more slowly the product tends to roll down the face of the pulley. Calculation methods which ignore this effect tend to underestimate the trajectory.

The study at the University of Wollongong concludes that for granular product the Dunlop/Booth method for predicting the product trajectory provides the more reliable calculation method.

Although many products exhibit some degree of adhesion, most formulae for predicting trajectories do not take this into account. The techniques described by Korzen provide the most .thorough and detailed analytical method of all the choices available for predicting trajectories for bulk solid flow.

The analysis carried out by Korzen strives to take into account the effects of friction (both static and dynamic) as well as inertia and adhesion of the product on the belt. All of these factors play an important role in determining the separation point and discharge velocity of the product from the belt.

Material which exhibits high adhesive stress characteristics, of the order of 0.3 kPa and above, will start to carry back further around the head pulley - significantly affecting the final impact area after falls of more than 2 m.

How much information do we give our chute designers to predict the primary trajectory?

If the product under investigation displays a large amount of adhesion, tests should be carried out to determine the magnitude of the adhesive stress and the value of the friction coefficient between the product and the belt.


Consider belt to belt transfer shown in Fig. 2.

The feed belt comes in from the side and should deposit product centrally onto the receiving conveyor. If the product trajectory is not as originally designed off-center belt feeding will take place.

A close examination of the original designs of many transfer points gives the impression that the designer was counting on some form of magic to induce the product to go in the right direction. Either that, or it was hoped that the last rebound of product off the chute would, just by luck, be in the right direction and on the receiving belt centerline.

An alternative school of thought assumes that no flow deflectors can be placed correctly in any case, and the chute would require field correction. This leaves the problem in the lap of the field engineer.

The trouble with much of this reasoning, or lack of it, is that troubled chutes dramatically delay system commissioning at a time when nerves are ragged, and threats of penalties are high. Possibly the commissioning is already behind schedule.

Assuming that field staff, who may have even less experience in chute design that the original designer, are lucky enough to position and weld-in a flow deflector plate which deals adequately with the flow problem of the moment. Any change in product flow characteristics, due for example of moisture variations or flow rate, could again result in off-center loading and a new set of problems. (Ask the plant maintenance personnel when they're trying to train the belt to guess what shape and where the last site-installed deflector plate is located.)

It is essential to study very carefully the flow path the product might take. For this characteristic properties must be established at design stage as accurately as practical. How much product information do we give our chute designers? Or do they have to look up CEMA tables, take a guess, and leave the rest for site modifications?

When handling very cohesive product the layout of the conveyor transfer point must, where possible, be designed to use belt-to-belt transfer of product with a minimum height of drop. The transfer point should be designed so that the product stream, wherever possible, centers on the receiving belt with little or no impact on the chute walls.

Should it be necessary to have a transfer point with a receiving belt at right angles to the feeder belt, it is advisable to ensure that the product is turned into the direction of travel of the receiving belt by flow directors or kick plates.


The chute designer can determine where the first impact of product against chute wall will take place, and the nature of that impact.

Take a lump of product and drop it to the ground. The potential energy can be measured by the height above ground at point of release. By the time the lump reaches ground level all the potential energy has been converted into kinetic energy. If the lump rebounds after impact to 10% of the drop height, then 90% of the kinetic energy has been dissipated at impact.

Consider now the question of degradation. We have all seen egg-throwing contests. Throwing the egg is easy. The trick is in how you catch the egg. If degradation is a major factor in design the chute must be designed in such away that impact energy is minimised.

If degradation is not a prime concern impact plates, or rock boxes, may be used to obtain a controllable vertical ore stream. Such devises should have a fine adjustment -which can be adjusted easily even under ore flow conditions.


When product falls onto an incline, it will generally be coming in on some trajectory. The flow will have a horizontal velocity component, as well as a vertical one.

Consider first the effect of the horizontal component. Sliding across the incline will be resisted by friction. The force component normal to the incline will be (see Fig.3):

N = mg.cos Q

where m is the mass of the particle
g is acceleration due to gravity
Q is the angle of incline to the horizontal

The sliding motion will be resisted by friction on the surface:

F = m.a = u.N


m.a = u.mg.cos Q

a = g.(u.cos Q) (1)

Once the product starts sliding on the inclined surface of the chute bottom, the horizontal component of velocity will rapidly decellerate in accordance with equation (1). 

Friction rapidly prevents significant sliding across the surface. Thus any transverse trajectory effects are lost after impact. The product slides down the incline following the line of least resistance. 


If product falls onto a horizontal surface flow cannot take place. The bulk solids accumulate in a heap until, if they are non-cohesive, they will reach the angle of repose. Then flow resumes. The velocity of flow down the incline will remain constant irrespective of the length of the chute.

If the angle of the chute bottom to the horizontal is increased, the velocity of flow will be increased. What happens if the chute angle is decreased? Will flow cease? In a static case - yes. But not necessarily in a dynamic case. Thus the discharge angle can be less than the angle of repose if the flow has an initial entry velocity.

For steady , fully developed, constant velocity flow the chute angle should be between the angle of repose and the dynamic internal friction angle. Studies on millet seed and polythene particles have shown that the lower and upper chute inclination bounds exist within which constant velocity flow occurs. These angles of inclination differ by about 4 degrees. Outside this narrow range flow is either accelerating or slowing down.

Flow in most chutes is a transient phenomenon. It is neither steady, nor fully developed. But the same basic principles apply. The flow is either accelerating or slowing down.

The velocity of flow down the inclined surface will affect the level of abrasion of the chute surface. The rate of erosion is approximately related to the square of the velocity. Thus doubling the velocity down the chute increases erosion by a factor of 4. A similar relationship exists with regard to product degradation.

Most product comprises various blends of fine and coarse components. The fine components usually behave in a different manner to the coarse components.

Depending on the smoothness of the surface, moist bulk product may exhibit an adhesive component. Adhesion often occurs with a smooth surface. Cohesion and adhesion can cause serious flow blockages to occur.

We know that a granular particle will slide down an incline if the angle of inclination of the plane is greater than the angle of friction.

Let the forces on the particle be represented by Fig.4. 

To obtain the velocity down the slope we must first find the acceleration. The weight mg has been resolved into two components: mg COS a perpendicular to the plane and mg sin a parallel to the plane.

Now apply Newton's Second Law, and equate force to mass times acceleration. Along the plane:

mg sin Q - F = ma

Perpendicular to the plane:

N = mg cos Q

F = u N = u.mg cos Q

By substitution:

a = g (sin Q - u.cos Q) (2)

If the initial velocity is v, the exit velocity V can be calculated as follows:

V = v - at (3)


While a substantial amount has been written regarding the prediction of product trajectory at the head pulley, very little attention seems to have been paid to the prediction of secondary trajectories the flow of product within the transfer chutes (Fig.5). The results of such conceptual neglect can often be seen by watching the flow stream path in the lower sections of the chute. 

The first step in calculating a secondary trajectory is to estimate the exit velocity at the point of departure. If the bottom plate is at the angle of repose of the material, the exit velocity will be the same as the velocity at entry. The entry velocity is determined from the previous trajectory calculations .

If the chute plate is vertical, product contact along the chute wall will have little effect on velocity. The product will accelerate in accordance with the demands of gravity.

Between these two angles, the flow will accelerate in accordance with formula 2 above. This depends on the chute angle, and on the value of the friction coefficient.


The friction coefficient between the product and the chute lining material is not a constant, however. It increases with reduced depth of flow in the chute, and changes with variations in moisture content. The friction coefficient may also include a velocity dependant, or viscous, component. Many products also exhibit some cohesive and adhesive properties.

Fig.6 shows that for coal at varying moisture contents high friction coefficients can occur at low depths of flow. The decrease in friction is considerable as depth of flow increases.

It is therefore not reasonable to expect chute designers to guess the values of friction coefficients and adhesive stresses. Tests are relatively inexpensive and provide far better guidelines to critical chute design parameters.

In some cases drag curtains need to be provided to increase friction and limit flow surges at the chute feed point.


While the bulk density D at any section of the flowing stream would show some variation with depth, this variation is small and is usually related to local turbulence at impact points. It is convenient to accept a constant average value. Thus:

W = D.A.v = constant

where W is the mass flow rate, A the cross sectional area, and v the velocity of flow.

In flowing down the incline the depth of product will decrease as the product accelerates down the slope. If the slope is long enough a terminal velocity may be reached where steady-state flow is achieved. Surge waves and granular jumps may be formed when obstructions are placed at the downstream end of the chute.


In most practical chute design flow is neither steady nor fully developed. The length of contact may be short, so the deceleration may well be negligible. Exit velocities can therefore be easily calculated using equations 2 and 3.

Once the exit velocity has been estimated the calculation of the secondary trajectory is fairly straight forward.

In some cases the flow stream is deflected by further impact surfaces - resulting in further secondary trajectories (Fig .5) .

One approach to side loaded chutes, where the feed belt comes in at an angle to the receiving belt, is to use impact plates to obtain a controllable vertical ore stream. A lower kicker plate is then used to direct the vertical ore stream into the direction of the belt travel. Use adjustable 'V' plates in the bottom of the chute to centralise the product on the receiving belt.

In cases of fine powders or bulk solids containing a high percentage of fines attention needs to be given to design details which ensure that during flow within the chute aeration, which leads to flooding problems, is minimised. For this to be achieved zones of free-fall and zones of high acceleration must be kept to a minimum.


Since impact dissipates kinetic energy in a violent way, thereby increasing product degradation, it makes sense to redirect the kinetic energy in a usefull manner. Thus the concept of using curved chutes is good where a gentler deceleration of the product is warranted. (See Fig.7) 


From the previous discussion it becomes clear that the chute designer can predetermine the velocity of product down the chute, and therefore can control it. 

Velocity control leads logically to questions of surface friction, bottom slope angles, chute shapes and so on. One solution is to use segmented chutes. Another is to use mini rock boxes. Still another is to use cascade chutes.

It also becomes clear that even with straight bottom chutes, in many cases, the chute angle should not be a constant value but should decrease as the velocity down the chute increases.


A major contributor to the dust problem is energy dissipation. Due to the velocity of the product and the drop height, it possesses both kinetic and potential energy. If these energies are dealt with properly, product degradation and dust generation will be minimised.

Kinetic energy is imparted to the product by the carrier belt. The higher the belt speed, the higher will be the resulting impact forces in the chute. Gradual deceleration of the flowing stream is quite difficult, and requires special chutes.

The usual method with high speed belts is to simply allow the stream to collide with a deflector plate, or the back wall of the transfer chute, in such a way that all the kinetic energy is dissipated quickly and violently. The result is obvious dust and deqradation. The first step in controlling dust at the design stage is to carefully look at reducing belt speeds. A small increase in belt width can have a large effect on reduced belt speed and resultant dust generation.

The next logical step in minimising dust generation is to minimise the height of fall (the potential energy).

It is all too easy to specify, for example, a 60 degree minimum slope angle on all chutes, as a means of totally avoiding product hang-ups in the chutes. For most cases, 60 degree is much steeper than necessary - especially since the product is almost always in motion in the chutes. Testing the dynamic angle of friction of the product is the best guide available. But test at different moisture contents to allow for wet, sticky, carryback fines. 

For example if a 50 degree chute angle is adequate, increasing this to 55 degree increases the chute height by 21%, and 60 degree results in a 45% increase. And these increases relate directly to increases in potential energy and to increased dust generation. 

When dealing with dust control at the transfer point make sure the chute has plenty of volume. This will allow free re-circulation of air before the point of contact of the falling stream with the chute wall.


Turbulence in the stream flow occurs at the impact zone as the falling product strikes the receiving belt. The receiving belt is moving more or less at right angles to the direction of the falling flow stream. The particles which touch the belt are accelerated by it the most. The other particles are accelerated by shear.

The length of the lone of turbulence along the belt line depends on the relative speed of the belt and the tangential velocity of product in the direction of the belt travel. If the belt and the incoming product both travel in the direction of the belt at the same speed the zone of turbulence will be minimal. The greater the difference the longer is the zone of turbulence, and the greater the difficulty in achieving an efficiently sealed, clean, loading zone.

The depth of turbulence depends on the bulk density at the impact point, which in turn is proportional to the amount of potential energy being dissipated. The greater the effective height of fall of the product the greater the degree of turbulence.

The degree of turbulence can be greatly reduced by providing a curved kicker plate, or adjustable gate, which will change the direction of the falling flow stream. The gate should be adjusted so that the velocity component in the direction of belt travel is equal to the speed of the receiving belt. This will reduce turbulence, reduce belt wear, degradation, and even power requirements for the drive motor.

With high falls it is possible to angle the curved kicker plate so sharply that the velocity component of the falling stream exceeds the belt speed. This would cause the flow of product to push the belt and possibly cause the belt to sag between idlers, complicating the sealing and drive power requirements.


Buildup in the chute, or changing material characteristics, will alter the falling stream of product. This may tend to pile deeper on one side of the belt than the other, as shown in Fig.8. Such off-center loading can often be seen in practice.

Flow training gates should be provided which can be adjusted from outside the chute. Such a feature allows easy load centering during commissioning, and during the life of the plant when the product and its characteristics change.

Alternatively special idlers can be used to form the belt into a 'U' shape and vibrate the load back into the center.


Skirting design is an essential part of chute design. This important subject is discussed in detail by Richard Stahura in his written contribution to this Conference.


If the transfer chute becomes blocked with product due to an abnormal condition, spillage will occur when the belt is restarted. As there is usually no means to control product flow, the belt will be greatly overloaded.

When the chute plugs at shutdown - normal or emergency - it becomes a hopper. When the belt is restarted it draws product to the full depth of the skirts. When the product reaches the end of the skirt-boards, the excess product spills over the edges of the belt onto walkways, floors, and between the idlers. But, as the belt is still brim-full, it will continue to dribble material between carrying idlers.

The conveyor system should be designed so that, when the system shuts down under normal or emergency stop conditions the chutes do not plug. This can be accomplished during normal shutdown by stopping the system sequentially - beginning with the most upstream belt, and allowing sufficient time for each belt to purge itself before stopping.

Sequential shutdown cannot work during emergency conditions. If at any transfer point the receiving belt tends to coast for a shorter time than the feed belt, product will pile up in the chute. 

If chute plugging cannot be avoided, then the transfer chute must be considered as a hopper, and designed accordingly. Design must then make allowance for plugged chute restarts. A control gate may be fitted to manually adjust the amount of product fed onto the belt during a full chute restart.


Cleanup of spillage can be both difficult and, in some cases, dangerous if attempted while the conveyor is running.

To reduce the cost of cleanup of spillage the following points should be taken into consideration:

  1. Raise the tail end of the conveyor high enough to allow cleanup by a front end loader. Spillage around a tail pulley can cause serious belt wear if the belt is left to run through the accumulated product.
  2. Provide a well drained concrete slab under the transfer area.
  3. Fit 45 degree shedder angles under the troughing idlers at the loading points.
  4. Fit a profile plant inside the chute to prevent a flooded belt situation occurring when restarting a loaded belt (fine material only).

A simple cleanup door should be located near the tail pulley as shown in Fig.9. If the door is correctly located, and the tail pulley safety screen designed for easy removal, cleanup of loading zone spillage will be made easy. 

Where the tail pulley is not located near the floor, or it is impractical to return the spilled product to the tail section, fixed or portable cleanup conveyors may be used. It is important to take cleanup measures into consideration when designing the load transfer point.


Very often flow problems within chutes could be more easily solved if the flow stream could be observed. Product flow within the chute cannot always be accurately calculated. Therefore observation is necessary to enable appropriate flow adjustments to be made.

Many chutes have only one inspection door - and that usually near the head pulley. This does not permit a view of the actual flow path in the lower chute and in the skirt area. This is where problems usually develop.

Provide at least two inspection or access doors - one in the head pulley area, and one in the loading zone. If the chute has more than a 3 m drop, locate another door at about the midpoint and out of the expected product trajectory. In cases where drops of 6 m or more occur, a door every 3 m is recommended. This will greatly simplify any corrective action required inside the chute - Including liner replacement, flow correction, removing tramp material, and unblocking chutes. Such doors must be dust tight and located on the non-wearing side of the chute. Depending on their location they must also be capable of withstanding the side pressure from a plugged chute condition.


'Rock boxes' are one method of absorbing the energy and abrasive wear of falling ore. However, if the product is wet, cohesive and/or adhesive, or, as in the case of lignites or sub-bituminous coals, tends to ignite spontaneously, the rock box solution should not be considered.

When the rock box shelf builds up with material and then continues to build-up beyond the anticipated design slope, the deflected product may rebound onto the belt cleaners, affecting their performance. The rebound lumps may also hit the end of the dribble chute, causing it to build-up and plug.

The rock boxes may need to be cleaned periodically to eliminate excessive buildup.

Most chute designers will provide for wear by the use of abrasive resistant material at points where the product impacts upon, and slides on, the chute surfaces.

Sooner or later the liner material will have to be replaced. Access to the inside of many chutes and skirt areas is extremely difficult, if not impossible.

Chutes with belts 900 mm or wider should have ample room for maintenance men to enter the chute and replace liners. Even so, extra access doors should be included in chutes 3 m or more in length. These doors should be large enough to pass liner plates through, and must be bolted and gasketted. The doors should be located in pairs, directly opposite one another, and be no smaller than 450 mm square. By locating two doors opposite each other the openings make a convenient place to insert planks or scaffolding.

For smaller chutes on narrow belts careful design and detailing is required to make adequate provision for liner replacement. Fabricating the chutes in short lengths for easy disassembly is one approach. The non wear side of the chute can be flange bolted to facilitate maintenance access. 

Good access does not just happen. It takes careful thought during design. Such consideration at the design stage can be stimulated by appropriate specification by the conveyor end users.


In view of their apparent simplicity transfer chutes have all too often received insufficient attention to their design.

The aim of this paper has been to focus attention on the flow of product within chutes. Flow visualisation is only one aspect of chute design. Even this one-day Conference cannot do more than awaken a more vivid interest in the subject of chute design.

Many instances can be given where chutes are the weakest link in the bulk handling system. Lack of understanding and lack of attention to design details have caused problems of environmental pollution, spillage, accelerated belt wear, and flow blockages, to name but a few. 

This First International Chute Design Conference provides an opportunity for designers and conveyor users to pool their knowledge. It is hoped this Conference has made a positive contribution toward more successful transfer point design.


Although he has not been able to attend this Conference, the contribution of Richard P. Stahura to the success of this First International Chute Design Conference, and to the understanding of flow in transfer chutes, is greatly appreciated.


Roberts, A.W. and Scott, O.J., "Flow of Bulk Solids through Transfer Chutes of Variable Geometry and Profile", Bulk Solids Handling Vol 1, No 4, Dec 1981.

Rober ts , A .W ., "Modern Technological Developments in the Storage and Handling of Bulk Solids", The University of Newcastle, NSW, Australia, 1990.

Rober ts , A .W ., "Design and Application of Feeders for the Controlled Loading of Bulk Solids on to Conveyor Belts", Supplementary Paper 3.

Rober ts , A .W ., "Transfer Chute Performance and Design ", Supplementary Paper 4.

Rozentals, J.J., "Rational Design of Conveyor Chutes", Proc. Beltcon 2, International Conference on Materials Handling, South Africa, May 1983.

Goodwin, P.J., and Ramos, C.M., "Belt Conveyors in Bulk Terminal Applications", Proc. Beltcon 3, International Conference on Materials Handling, South Africa, Sep 1985.

Korzen, Z. , "Mechanics of Belt Conveyor Discharge Processes", Bulk Solids Handling Vol. 9 (1989) No 3, p 289-297.

Sabina, W .E ., Stahura, R .P ., and Swindeman, R .Todd "Conveyor Transfer Stations -Problems and Solutions " , Martin Engineering Co., (1988).

Arnold, P.C., and Hill, G.L., "Predicting the Discharge Trajectory from Belt Conveyors", Bulk Solids Hand ling , Vol. 10 (1990) No 4, p 379-382.

CEMA, "Belt Conveyors for Bulk Materials" , Cahners Books .

Roberts, A.W., Ooms, M., and Wiche, S.J., "Concepts of Boundary Friction, Adhesion and Wear in Bulk Solids Handling Operations", Bulk Solids Handling, Vol 10 (1990), No 2, May 1990.

Beckley, D.E. , "Belt Conveying Systems in West Australian iron ore industry", Day 3, Paper 1 .

Wright, H., "Large scale handling of bulk materials", Bulk Solids Handling, Vol. 10, No 2, 1990.

Mills, Mason and Agarawal, "Some Aspects of Bend Erosion in Pneumatic Conveying System Pipelines", First National Symposium on Hydraulic Transportation of Solids through Pipelines, Ranchi, India, (1985).