You are on page 1of 8

The Design of Conveyor Pulleys



B.E. Lloyd BSc (Mech.) Eng.

Jegk urt Engineering (Pty) ltd

Up until a few years ago , when conveyors in South Africa were rather sm all com pared with their counterpart in other countries overseas, the need for
engineered system s was lim ited.
W ith the advent of the oil crisis this situation changed over night for two reasons , i.e. diesel fuel vehicles becam e alm ost too ex pensive to use and
secondly the dem and for coal as a fuel becam e increasingly im portant. In order to m ine the coal econom ically , it becam e necessary to increase the
tonnage m ined , as well as to im prove the m ethod of transporting the coal that was m ined.
The obvious answer to the problem was to use conveyor belts in conjunction with unit trains and sophisticated loading and unloading stations.
Coal m ining was not the only type of m ine effected by the rise in cost of oil based fuels , but also the base m inerals , such as iron are , copper and
uranium .
It was the base m ineral m ines , which at the tim e of the oil crisis , that were using the open cast m ethod with electro haulpack s for transporting the
ore and overburden m ore so than the Collieries.
Unfortunately or fortunately for South Africa , the change from sm all or m edium size conveyors to large installations happened virtually overnight and
the technology to undergo the change was not readily available , hence people in the conveyor field used their lim ited ex isting k nowledge to try and
m ak e the larger and m ore sophisticated conveyors.
This k nowledge was however insufficient and hence resulted in m any failures. The m ost noticeable failures being that of term inal Pulleys and belting.
The idler m anufacturers seem ed however , to have escaped the lim e light , either because they were m ore prepared for the change and sufficient
technology was available from their overseas parent com pany , or the tim e and the cost to replace idlers was not sufficient to warrant an
W ith the term inal Pulleys , a different situation arose. Just about every m ining house and other conveyor orientated com panies saw the need to
im prove the standard and the design of Pulleys. However , unlik e Germ any and other countries , each individual concern tried to develop standard
and design on their own , without consulting each other.
The end result of these standards can only be classed as disastrous and unnecessary waste of m oney and tim e , as none of the standards are
com petable with each other , ex cept for one thing , and that is the rigidity of the ex pected standard of work m anship and the am ount of paper
required to prove the soundness of the equipm ent.
By the am ount of paper required, I m ean, this so called quality assurance procedures , whereby it has now becom e necessary to prove that even the
sm allest Pulley, which is just an oversize idler , passes full pressure vessel tests.
It is not that I am opposed to quality assurance , but just that I feel that the relationship between quality control and purpose of use should be
equated so as to build a unit fit for service at the right price.
This bring m e to the definition of an engineer , which goes as follows:- " An Engineer is som eone who can build for a penny, what any old fool can
build for a pound ".
W hen our com pany went into the m anufacture and design of Pulleys lik e everyone else, we went our own way to establish a standard , but with a
difference. We first tried to find out how people overseas were tack ling the problem . During the course of this investigation we found out that
Germ any , Australia and som e parts of the United States had all had their problem s when they changed over from sm all to big conveyor belts.
Germ any in particular , where m ost of the large m aterial handling plants are designed seem ed to have spent m any years in developing their
technology in conveying with the em phasis on conveyor Pulleys. At Hannover University two doctorates were awarded for research and developm ent in
conveyor Pulleys.
By dealing with the right people , we were able to get hold of copies of both these doctorates , as well as m any other papers written about conveyor
Pulleys. By tabulating this inform ation , we arrived at the bases of our design.
The m ost com plicated part of a Pulley and the area m ost prone to failure is the end disc.
The origin of disc design :
Form ula for stress and strain of disc subject to loads acting at the centre , are to be found in m any technical hand book s .
At present we are basing our calculation on form ula for disc stress , which are obtainable from book data out of " Roark s " handbook and a thesis
written by " W. Schm oltz" The disc can either be regarded as subject to bending m om ent , while the outer periphery is im agined to be fix ed rigidly
(see condition A) or on the other hand , the disc can be regarded as being suspended in such a way , where it cannot transm it any bending m om ent


The Design of Conveyor Pulleys


(as in condition B)

Condition A

Condition B
In both cases the inner diam eter is tak en to be rigidly connected to the shaft. Both conditions of fix ing have to be considered as ex trem e cases for a
Pulley disc.
The actual loading will be som ewhere in between condition "A" and "B" . Consequently it should be necessary to calculate the stress in the inner
diam eter of the disc for condition "A" as well as condition "B" . In Pulley design we will be sim ulating condition "A" to a greater degree than condition
"B" and the stresses as per condition "A" result in higher stress than for condition "B" .
For the abovem entioned two reasons only condition "A" will be considered in this paper.
Considering now the bending m om ent in the shaft , disregarding the restraining action of the disc , the shaft would develop the following bending
m om ent diagram :

In corporating the resistance of the disc and shell to restrain the deflection of the shaft , the actual bending m om ent look s therefore lik e shown
below :

P (B - L) L
Bm disc =

8I ( L + 1/t3)

(Bm disc)is the portion of shaft bending m om ent which sets up the stress in the disc at the connection with the hub , which so often has resulted in
hub to disc failure if a hub is used.
" Sim ple solution : Don't use hubs " .
The question then arises , if hubs are not used , how do you fix the disc to shaft . ( This aspect is covered in the nex t section headed : LO CKING
From the above it will be obvious that by reducing the deflection of the shaft , one im m ediately reduces the am ount of disc deflection as well as
counteracting disc restraint.

1. Bm and deflection of the disc can be readily reduced by increasing the shaft diam eter (ex pensive m ethod)
2. Bm disc can be readily reduced by decreasing the thick ness of the disc .
A steel Pulley with a thin disc allows the shaft to deflect as it naturally would by virtue of the two loads applied to it through the disc. A Pulley m ade
with thick discs that do not bend appreciably as a result of the bending m om ent applied to them causes another type of shaft deflection , nam ely ,
from the disc outwards and for all practical purposes , the shaft could just as well be cut off between the disc.


The Design of Conveyor Pulleys


Between these two ex trem es lies a range of thick ness that causes high disc stresses , unnecessary high bending m om ent at the inner diam eter of
the disc and eventually causes Pulley break age.
Using basic data , a form ula has been developed to calculate the disc stress:

P (B-L) L
4DtI(L + 1)
2I t3

k 8/cm

P = Total resultant belt pull on the pulley

= k8

B = Bearing center distance

= cm

L = Center to center distance between disk = cm

D = Disc outer diam eter

= cm

t = Disc thick ness

= cm

I = Mom ent of inertia of shaft

= cm 4

= Disc constant
= Disc constants
The factors and are selected after calculating the ratio :
Inside diam eter of disc
O utside diam eter of disc
For derivation of form ula see Appendix "A" .
The actual disc stress versus thick ness for a typical sm all diam eter Pulley is illustrated below :
Chosen Pulley 450 Dia x 1100 Face x 90 Shaft
Pulley loading 27.25 k N resultant
Load Bearing centers = 1520
Disc centers = 1020m m
Disc outer diam eter = 430 (10m m rim )
Disc inner diam eter = 185
In this ex am ple :
= 6.6
= 1.65
After calculating the disc stresses for various disc thick ness it will be seen that the stresses increase from 3m m to 8m m and beyond this the stresses
reduced as the disc thick ness increased. Please refer to the graph below :

Thus for an allowable work ing fatigue stress there are two disc thick nesses that can be utilised. O ne a thin disc and the other a thick disc.
Between these two lim its a high stress is inevitable. This represents one of the few cases in design , where increasing m aterial thick ness results in
higher , rather then lower stress in the part.
The disadvantage of em ploying thick disc to reduce stress is that the bending m om ent on the connection from the shaft to the disc is greatly
increased , and if this connection is not securely m ade the bore of the disc wears out and the Pulley m oves or what is k nown as "walk s" on the shaft
The second item effecting end disc design is the positioning of the disc relative to the bearing centers. It is quite obvious that there are again two
positions to lim it the stresses i.e. close together in the m iddle of the Pulley or alternatively as far out as possible on the face , will result in the least
stress , while in between will result in high stress regions.
We have chosen the outer position for an obvious reason , and that is that the conveyor belt does not always strain perfectly , and if the disc was
centrally situated and the belt runs off to one side , the disc would be subjected to additional bending stresses , due to the eccentric load. W hen
using stress form ula , it has been found that the large diam eter Pulleys with the low set loading are seldom problem som e as they can be designed
with the thin disc , while relatively sm all Pulleys with big shafts becom e difficult , because the disc becom es im practically thin or the direct stresses are
too great to use a thin disc . Thus forcing the m anufacturer to use the thick disc design.
Typical ex am ples of this type of Pulley is those associated with steel cord belts , where the required diam eter of the Pulley is far sm aller than the
equivalent required for a fabric belt.


The Design of Conveyor Pulleys


W hen using the thick disc design , there are two im portant areas to consider nam ely , the fix ing of the disc to the shaft (Discussed under the section
of Lock ing Elem ents), and the disc to rim fix ing .
O n ex trem ely high loaded Pulleys , it is often an advantage not to weld in the inter-face between rim and disc, but to m ove the weld inboard of the
disc and m ak e a but weld on the rim . This type of disc is k nown as a T bottom type disc See sk etch below :

The advantage of this is that it has substantial radius both inside and outside the high stress zone and to lim it stress concentration it is always better
to use radii rather than sharp corners. The other advantage is that the weld as positioned is firstly out of the high stress zone and secondly is in a
position where it is far easier to m ak e a perfect weld and the thick ness of m aterial to be welded is thinner. The question arises "W hy are not all
Pulleys m ade lik e this?" The answer is sim ply cost. Rem em ber our definition of an Engineer.
W hen look ing at the sk etch of the T bottom , you will notice that the disc is tapered and not parallel. There are two reasons for this :

1. Although we are using the thick disc design there is still m erit in giving the Pulley certain flex ibility in order to lower the stress at the shaft to
disc connection.

2. By profiling the end disc , it ensures that the stresses throughout the disc are constant and hence lim it stress concentration areas.
As m entioned before not all Pulleys are T bottom , for the sim ple fact that they cost far m ore than other designs and hence we com e to the second
type of end disc , nam ely a Turbine End. In our term inology the T bottom is our heavy duty range and the Turbine and sem i-Turbine our m edium
duty range.
The Turbine Pulley is sim ilar in characteristic to our T bottom design but has two distinct differences:
a) W e use a different type of Lock ing Elem ent
b) The welding is perform ed at the interface of disc to rim
Although we em phasized the advantage of rem oving the weld from this position when we discussed the T bottom type Pulley , we m ust clearly state
that this type of Pulley we are now discussing , is one that does not have to with- stand the sam e tensions as the T bottom , and hence the stress
level is far lower. W hen perform ing this weld between rim and disc it is im portant to ensure that this weld is of a sound nature. W ith full penetration
and no lack of side-wall fusion. To insure full penetration and to lim it stress concentration an the inner face it is good practice to perform an inner
fillet weld if possible.
See sk etch :

The last type of disc we are going to consider is a sem i-turbine end or flat bottom design.
This Pulley for the m ajority of conveyors, is the ideal type of Pulley. It is relatively in-ex pensive due to the easy m ethod of m achining the disc. It has
m ost of the advantages of the turbine end but with the ex ception of the two zones where stress concentration occurs, but tak ing suitable stress
concentration factors into consideration these stresses are well below acceptable levels. Again using suitable welding procedures to perform the disc
to rim weld to ensure a sound weld which is fit for the duty.
See sk etch :


Many people are of the opinion that the reason for using a lock ing elem ent for fix ing the disc to the shaft is so that the shaft can be rem oved if
necessary. This in itself is a good reason for installing a lock ing elem ent, but by no m eans the only reason for fitting a lock ing elem ent.


The Design of Conveyor Pulleys


Below are three sk etches of suitable lock ing elem ents for Pulley construction and we will discuss the advantages of these elem ents collectively and
thereafter individually.



the three elem ents share the following advantage over all other k nown hub to shaft connections:

a. They by virtue of the torque applied to their bolts, have a controlled pressure on both, hub and shaft. By sim ple m athem atics, this pressure
can be calculated and hence elim inating the hit and m iss factor which is inherent in a "shrink fit" thus m ak ing it possible to calculate, by using
thick wall cylinder form ula, the size of the hub required to with-stand the pressure fairly accurately and thus elim inating unnecessary hub
m aterial.

b. Their torque, which they can transm it, is also a factor controlled by the torque applied to their bolts, thus again m ak ing it unnecessary to fit
additional safe guards, such as k eys and k eyways, to ensure that they do not slip in the direction of the applied torque. By the elim ination of
these k eys and k eyways the stress concentration associated with the above is naturally elim inated.

c. All three lock ing elem ents can withstand great ax ial thrust thus virtually elim inating the possibility of walk ing on the shaft, and in the unlik ely
event of this happening, it is relatively sim ple to rectify, as it is usually only necessary to place the drum back to its original position and retorque the bolts in the lock ing elem ent. If a sim ilar situation occurs with a shrink fit, the bore has inevitably been worn away and the entire
Pulley becom es redundant.
Lock ing elem ent type A has been on the South African m ark et for approx im ately ten years and is probably the m ost com m on lock ing elem ent used
in Conveyor Pulleys. In the unloaded condition it has undoubtedly the best stress pattern of the three lock ing elem ents. However, when load is
applied to the drum and the shaft deflects, the stress pattern changes.
See com parison sk etch between lock ing elem ent "A" and "B".
In lock ing elem ent "A" because of the angle of the tapered segm ents, it is neither self centering nor self lock ing, hence at all tim es the securing
bolted are in tension, thus it becom es necessary to use a centralising ring as shown on the sk etch, as well as having to lim it the deflection of the
shaft, to ensure that the bolts do not ex ceed their elastic lim it.
The acceptable level of deflection for this type of lock ing elem ent is in order of 1/2500 bearing centers or if preferred, in radial units 0.05 degrees or
0.0008 radians, although this deflection restriction is not totally unfavorable, it often forces the designer to use a bigger shaft than he norm ally
would, to satisfy acceptable stress levels of torque to be transm itted.
The m ost unfavorable characteristic of this type of lock ing elem ent is that it is virtually im possible to control the m ating of the surfaces of the
segm ents. This can result in high point loads initially and also m ak e it necessary to re-torque the bolts after initial running in period, after the
segm ents have settled.
Lock ing elem ent type B is the type we are currently using in our m edium duty Pulleys, i.e. the turbine or flat bottom design. The reason for using this


The Design of Conveyor Pulleys


type of lock ing elem ent and not type A, is that it has the following advantages:

a. Because of the angle of the tapers on the segm ents, these elem ents are both, self-centering and self-lock ing, therefore do not require a
centralising step behind the disc. Because of the elim ination of this step, it is always possible to withdraw the shaft from the drum , because
the inevitable build-up of rust between centralising step and shaft does not occur.

b. By having self-lock ing tapers, the bolts are not in tension to the sam e ex tent as type A, hence greater shaft deflections are possible. The
perm issible deflection being in the order of 1/1800 to 1/2000 bearing centers or ex pressed radially + 0.06 degrees or 0.001 radians.

c. Unlik e lock ing elem ent type A it is not so critical that the lock ing elem ent bolts are tightened in sequence as due to the tapers it is im possible
to tighten this elem ent unsym etrically.

d. The m ajor disadvantage of this type of lock ing elem ent, is that it cannot transm it the sam e torque as type A. However, on conventional
Pulleys this seldom causes a problem as the torque to be transm itted by the shaft is usually well within the capabilities of the chosen
elem ent.
Lock ing elem ent type C was specially designed for conveyor Pulleys. It has all the advantages of type A and B , with far lower surface pressures then
either , while at the sam e tim e being able to transm it between 2 and 3 tim es the Torque of either A or B.
This elem ent is used in all our T bottom Pulleys , with the shaft sizes up to 400m m in diam eter. The Rhein Brown Coal A.G. in Germ any have been
using them with shaft sizes up to 600m m successfully.
Many people think that by specifying a Pulley fitted with lock ing elem ents , they will autom atically get a better Pulley than a shrink fit type Pulley.
This is often not the case , as the type of hub that is used is as im portant as the type of lock ing elem ent chosen. O ften , if the hub design does not
suit the lock ing elem ent , the hub crack s under pressure from the lock ing elem ent. or alternatively the hub diam eter is increased , so as to
accom m odate the lock ing elem ent , and this hub is too large , and when using a thin disc often lim its the flex ing of the disc , hence crack ing the
weld around the hub.
A design has been introduced into South Africa which is quite unique. I have not seen anything quite lik e it anywhere in the world , and it is because
of its uniqueness I feel som e k ind of m ention should be m ade to try and analyze it.
Firstly its origin during the change from light duty conveyors to heavy duty conveyors. The m ost noticeable area of failure was the two fillet welds at
the hub sk etch A.
It then becam e apparent that if the hub was ex tended so that a double butt weld could be used , this would obviously be a far better type of hub
sk etch B than its predecessor A. Design B was used for a few years and was later replaced by design C , which instead of m achined out of solid
m aterial could easily be forged by using a single sided dye.

This hub had one great disadvantage and that is that in order to utilise the potential of its sim plicity , it has to suit various shaft diam eters , drum
diam eters and disc thick ness. This not being a sufficiently com plicated ex ercise , this hub was further m odified to be able to accept various shaft to
hub fix ing , such as , shrink fits , taper lock s and lock ing elem ent type A . The result of the ex ercise resulted in a large chunk of useless m aterial
which only added weight and cost to the Pulley, no wonder it is unique to South Africa.

During the course of the paper m any references have been m ade about deflection and its lim its. A sim ple and com m on form ula is used to calculate
this deflection :

(1 -



W here : L = Bearing centers

a = Lever arm
E = Modulus of elasticity
After calculating the deflection, the nex t step is to calculate the stress level , either for straight bending in the case of non drive Pulley, or the
com bination of bending and torsion for drive Pulley.


The Design of Conveyor Pulleys


The last step is to check the lock ing elem ents Torque characteristic against the applied Torque , in the case of a drive Pulley. Rem em bering that
during starting and break ing , this Torque increases substantially and appropriate precautions m ust be tak en into consideration to prevent the
lock ing elem ent slipping.
Due to the com plex nature of the calculations involved in rim design , we will not be discussing it in detail in this paper.
For design purposes , the drum casing is regarded as a cylindrical shell with a partially loaded surface. Particular attention m ust be paid to the
boundary conditions and the load distribution , if this approach is used for conveyor Pulleys. The rim is then regarded as a sim ply supported tube
with no m om ents or ax ial forces. The belt forces at the edge are transm itted in the direction of the plane of the drum only on the contact surface
between rim and conveyor belt. A pressure and shear stress distribution , sim ilar to operating conditions is achieved by super-im posed sine functions.
The total load of the elem ent is obtained by sum m ation of the individual sinusoidal loads. This calculation is far too tim e consum ing without the use
of a com puter to be dealt with in this paper as they are usually unk nown under eytewein boundary conditions. W hen the fatigue strength of a rim is
calculated , it is necessary to rem em ber that three load changes occur in each casing segm ent during a single drum rotation. Rim design lim iting
factor is usually based on fatigue strength.
Center stiffeners in large conveyor Pulleys are again unique to South Africa , and the only article I have seen relative to stiffeners was associated with
pressure vessels titled " Collapse of Stiffened Cylinders Under Ex ternal Pressure " by S.B. Kendrick .
The only other association I have had with them is having seen sam ples of failure which have occurred in the vicinity of these center stiffeners on
large Pulleys.


{1+(dy/dx )}3/2


But R =

= dy/dx when dy/dx is sm all

.'.dy/dx = M/EI (But between points S and T on the shaft

the bending m om ent MS is constant)
.'. dy/dx =


dx =


x +c


But when x = L/2 --> dy/dx = 0

.'. M S/EI x L/2 + c = 0 ------(1)

Then C = -

(Substitute back into (1))

2 EI
Then dy/dx =



x +


.'. dy/dx = -

(and x = 0)

2 EI
= s

2 EI
But M D + M S = M and also S = D


P (B-L)
ie M D + M S =


M D(1 -

P (B-L)


.'.M S =

t L


P (B - L)
MD =
But SD =

2M D
2 P (B - L)



The Design of Conveyor Pulleys


2 P (B - L)
.'. =
4Dt 2I/L (L/2L + 1/t)
P (b - L) L
.'. SD =
4DtI(L/2L + 1/t)