SAFETY, APPLICATION, AND SERVICE FACTORS

AS APPLIED TO SHAFT COUPLINGS

by

Michael M. Calistrat
Owner

Michael M. Calistrat and Associates

Missouri City, Texas

and
William E. (Ed) Nelson
Consultant

Dickinson, Texas

He is a registered Professional Engineer in the State of Texas, and

received a B. S. degree in Mechanical Engineeringfrom Texas A&M
University. He was given a College of Engineering Alumni Honor
Award by Texas A&M in 1992, and recently had a test cell named in
his honor in the new Turbomachinery Laboratory research facility.

Michael M. Calistrat has outstanding ex

perience in the field of power transmission
equipment, which he has accumulated work
ing with oil drilling equipment, gearing, and
flexible couplings. He also has a solid back
ground in industrial gas turbines.
He is active in engineering societies and
manufacturing associations. He chaired
many technical committees for ASME, ASTM,
and ASLE, and was Chairman of the Inter
national Conference on Power Transmis
sion and Gearing. He was associate editor of the ASME Journal of
Machine Design.
He has been a coauthor on three engineering manuals and has
written many technical papers and articles in the Unites States,
Japan, France, Canada, Holland, Italy, and Taiwan.
Mr. Calistrat was Manager of Research and Development for the
Power Transmission Division of Koppers Company, where he was
instrumental in the development of many products. Presently, he
heads a consulting company which specializes in rotating
machinery.
Mr. Calistrat received his M. S. degree in Mechanical Engineering
from the University of Bucharest and holds I 7 patents.

ABSTRACT
Shaft couplings are rated by their manufacturers as to how much
torque they can transmit, how much misalignment they can accom
modate, and the maximum speed at which they can operate. Still,
manufacturers, buyers, or various organizations often use qualifi
ers, or factors that significantly reduce the published capabilities
of couplings, when the published ratings cannot be used at face
value.
It is the authors' intent herein to help coupling users in under
standing these factors.

INTRODUCTION
Three correction factors are used, often in combination:
safety factors, which establish the ratio between the breaking
point and the maximum catalog ratings,

application (experience) factors, which establish the ratio

between the rated torque from the catalog, and the actual torque of
a given application, and

William E. (Ed) Nelson is the author of

more than 40 technical papers, a speaker at
many turbomachinery seminars, and a con
tributor to several handbooks on machinery
maintenance and operation. He is the coau
thor of The Centrifugal Pump Source Book,
published by McGraw-Hill. He has received
six patents on refinery and machinery main
tenance techniques.
Mr. Nelson was a founding member, with
20 years service, on the Turbomachinery
Symposium Advisory Committee. He was also a founding member of
the International Pump Users Symposium Advisory Committee.
Other affiliations include the Vibration Institute, IMI, and ASME.
Mr. Nelson retired from Amoco Oil Company, Texas City, Texas,
after more than 36 years in various engineering, materials manage
ment, and maintenance positions. Responsibilities in his last position
included refinery instrument, electrical, weld shop, operation of
mobile equipment such as cranes, etc., and rotating machinery
maintenance and repair.

service factors, which establish the ratio between the maxi

mum value of a fluctuating torque and the average torque as
calculated on the basis of power and speed.

All these factors are subjective, inasmuch as they are deter

mined based on experience and economic parameters. They differ
from manufacturer to manufacturer and from user to user. Organi
zations such as the American Petroleum Institute (API) and Amer
ican Gear Manufacturers Association (AGMA) have issued
standards in which values for the experience and service factors
are recommended; in the views of many people, these standards
have created as many problems as they were trying to solve. No
attempt was made so far to standardize safety factors; these are
established by coupling manufacturers. Only a limited amount of
data on safety factors was published.

SAFETY FACTORS
A safety factor is the ratio between the breaking point and the
rating value of a given device; its magnitude is a compromise
65

PROCEEDINGS OF THE TWENTY-SECOND TURBOMACHINERY SYMPOSIUM

66

between safety and economics. For equipment in public use (such

as elevators) safety factors are mandated by government agencies;
in the case of couplings, each manufacturer establishes factors
arbitrarily. It can be said that a good safety factor is a "sleeping
pill," as it allows both the manufacturer and the user to sleep well
in case any device is operated at its catalog rating. Safety comes at
a price; the larger the safety margin, the higher the cost of a
product.
In the particular case of flexible couplings, safety factors apply
to speed, torque, and operating misalignment, as any of these
parameters can cause breaking of a coupling component. Cou
plings with flexible elements (metallic or elastomeric) have an
additional parameter to consider: axial displacement. Axial dis
placement stretches flexible elements, and when resulting stresses
become excessive they can also cause failures.

Speed Parameter
Rotating speed creates both radial and tangential stresses in a
component; the tangential stresses, also known as "hoop" stresses,
are usually larger. Manufacturers publish the maximum speed at
which each coupling type and size can operate. The margin of
safety of the speed rating is easy to determine through "reverse
engineering," as the approximate hoop stress ( cr) is easily obtained
from the formula:
For cr in pounds/inch',
V"
K

cr=-

where

V =rim velocity, and

V in inches/second, and

materal being steel,

K=1500

K =coefficient that depends on material

Diameter
inch

Maximum

Hoop

Speed

Velocity_

Stress

r om

inch/sec

psi

Maximum

Diameter
inch

5.6

33,800

9962

66,160

7.4

24,500

9500

60,170

8.4

21,300

9400

58,900

9.8

18,400

9490

60,040

11.1

16,000

9300

57,700

12.4

13,400

8750

51,040

13.9

11,600

8490

48,050

Figure 2. Hoop Stresses in an European Disk-Pack Coupling Line.

Notes: The hoop stresses are not held constant; the safety factor
becomes larger as the diameter increases.
In both cases, the materials are alloyed steels with an tensile
strength of approximately 170,000 psi. The safety factor for Case
1 (Figure 1) is determined as:
s

170,000
.
= 3 86
44,000

For the coupling of Case 2 (Figure 2) the safety factor varies

according to size, and is best shown in graphical form (Figure 3).
In both cases, even the smallest safety factor is large enough to
satisfy most users.

Safety Factor

The strength of materials used in couplings can be found in the

literature; most manufacturers publish the materials used in their
products, particularly for "special purpose" couplings. Special
purpose couplings are used in high-performance applications,
which generally implies high speeds. They are made of heat treated
alloy steels, or other high strength metals.
The authors have surveyed the published speed ratings of a large
number of coupling manufacturers from many countries. The
result of the survey showed that manufacturers use either a con
stant hoop stress across a line of couplings, or that the stress is a
function of rim diameter. Two such cases are shown in Figures 1
and 2.

Outside

Peripheral

Outside

Peripheral

Hoop

Speed

Velocity

Stress

rpm

inch/sec

psi

6.56

23,800

8217

44,320

7.84

19,900

8211

44,250

9.09

17,100

8181

43,930

10.47

14,900

8211

44,250

11.94

13,100

8232

44,480

13.06

11,900

8180

43,910

14.31

10,900

8209

44,230

15.75

9,900

8207

44,200

17.00

9,200

8200

44,470

Figure I. Hoop Stresses in an American Disk-Pack Coupling Line.

Notes: The hoop stresses of this coupling line are held constant for
all sizes.

3.6

3.3

2.7 --

1----"----1

2.4 f------,----r--.....----.--1
14
12
4
6
Diamet r (inches 0

Figure 3. Speed Safety Factor of the Disk-Pack Coupling Line

from Figure 2. Notes: Even the smallest safetyfactor, as related to
hoop stresses, is quite satisfactory.
One manufacturer of diaphragm-type couplings relates the
maximum speed to the operating axial stretch of its metal
diaphragm coupling (Figure 4). Making speed and axial stretch
interdependent is logical; both generate constant stresses in the
element, and reducing one allows the increase of the other without
exceeding acceptable stresses.
A tabulation of this coupling's maximum rated speed at zero
axial stretch, and of the resulting hoop stresses is shown in Figure
5. Operating one of these couplings at the "maximum rated speed"
would probably place it in a failure mode.
The resulting safety factor is shown in the graph in Figure 6,
based on an alloy steel with a tensile strength of 170,000 psi. If the
larger couplings would be operated at the published maximum

SAFETY, APPLICATION, AND SERVICE FACTORS AS APPLIED TO SHAFT COUPLINGS

67

Safety Factor

\
\

1.5
1.3
1.1
0.9
0 -------r----
0.00
0.04
0.08
0. 12

Stretch (inches)

Figure 4. Speed Vs Axial Stretch Graph Used for a Line of

American Diaphragm Couplings. Notes: While operation at zero
axial stretch is possible and desirable, extending the ratings down
to zero speed is superfluous.

Outside

Maximum

Diameter

Peripheral

Hoop

Speed

Velocity

Stress

inch

rpm

inch/sec

psi

6.1

40,000

12,840

109,950

7.1

38,000

14,200

134,400

9.2

28,500

13,800

126,960

11.0

24,500

14,200

134,240

13.1

21,500

14,800

146,500

14.9

19,500

15,300

155,900

16.9

18,000

16,000

170,900

18.9

16,500

16,400

179,600

22.6

14,000

16,650

184,900

26.4

12,800

17,780

210,900

Figure 5. Hoop Stresses in the Diaphragm Couplings of Figure 4,

at Maximum Rated Speed. Notes: Operation of coupling sizes that
fall below the heavy line, at the maximum rated speed, will
probably cause failures, as stresses exceed ultimate strength.

acceptable" speed, failures would most likely occur. Although

making speed and axial stretch interdependent is logical, a safe
limit for maximum speeds should always be incorporated.
It is interesting to note that an older catalog of the same
manufacturer, as well as the one of their European competitor, list
speeds that result in hoop stresses of only 60,000 psi, equivalent to
a safety factor of 2.8 .

Torque Parameter
Catalog torque often determines the useful life of a coupling
rather than the breaking point of the torque-transmitting elements
of a coupling. Therefore, safety factors cannot be easily applied to
torque. Because of this, many manufacturers make the rated
torque, the maximum speed, and misalignment inter-related. For
instance, one of the most popular types of special purpose gear
couplings has a rated torque that is a function of the operating
speed. This coupling's torque derating factor as a function of speed
is shown in Figure 7.

0.7

it'

......

""'

'

r--------------- r:;:-::------

10

20

30

Diameter (inches)

Figure 6. Safety Factors, at Maximum Rated Speeds, in the

Diaphragm Coupling Line of Figure 4. Notes: Operation of cou
pling sizes that fall to the right of the heavy line, at the maximum
rated speed, will probably result in failures, as the safety factors
are smaller than one.

Torque Factor

----\------------1

0.9 t--

0.71---------,---------l

--0.5 1-------..----.---....-----1
0
5000
10000
15000
20000

Speed

Figure 7. Torque Vs. Speed Curve of a Special-Purpose Gear

Coupling Line. Notes: The manufacturer reduces the rated torque
at speeds above 3, 000 rpm, in order to maintain a satisfactory
service life at all speeds.
However, not all coupling types are designed the same way; in
all coupling types some components will eventually break if
torque reaches a certain level. For instance, in gear-type couplings,
flange bolts will break long before the teeth will break, while in
spoke-type diaphragm couplings, the spokes will break first.
Peak torque is defined as a momentary or instantaneous load that
occurs infrequently, such as startup, surge, or other occasional
events. It subjects couplings to very high stresses, which are close
to the yield strength of the components. Usually, manufacturers
list a rated peak torque that is two to three times larger than the
rated torque. One notable exception will be discussed later on.

Alignment Parameter
In couplings with flexible metal membranes, the safety factor is
related to the Goodman diagram, which establishes a zone of
infinite life, as a function of the alternating and constant stresses
in the components. Diagrams published by two different manufac
turers are shown in Figures 8 and 9. They provide a curve for the
limit conditions (at which failures start occurring), and a rated"

PROCEEDINGS OF THE TWENTY-SECOND TURBOMACHINERY SYMPOSIUM

68

curve, which represents operations at a given safety margin. While

the rated curve from Figure 8 provides a safety factor of two at
worst conditions, for both the alternating and the constant stresses,
the ratings from Figure 9 provide a safety factor of two for the
alternating stresses, but a much smaller and unusually low safety
factor of 1.3 for constant stresses at point "A."

Alternative Stresses

"'7
?

Factor of Safetv

......,

./

1'

Operating
Envelope

0.4
0.1

'-

ooll

\
\

\
\

Figure 8. Rating Method of a Special-Purpose Disk-Pack Cou

pling Line. Notes: Thisfigure illustrates a very conservative rating
method, as even the worst conditions still have a safety factor of
two for alternating stresses, and even larger for constant ones. On
the other hand, the manufacturer mandates a minimum applica
tion factor of 1. 5.

Alternative Stresses (psi)

-------. r-

60
40
20

Infinite-life zone

Failure one

0
0

40

""-...

'

i
I

i
I

-t "'k

120

0 .2

0.3

0 .4

Misalignment (degrees)

Manufacturers of high-misalignment couplings publish torque

derating factors related to misalignment. The maximum torque as
a function of misalignment, as published by an American coupling
manufacturer, is shown in Figure 11.
One Japanese gear-coupling manufacturer publishes a compos
ite torque derating graph (Figure 12), in which the catalog rated
torque is reduced as a function of both misalignment and speed
(actually, the ratio of operating speed to maximum catalog speed).
Derating torque as a function of speed is an understandable
practice, and has been used by many manufacturers. However, the
rated torque in this case can only be used if the coupling operates
at zero misalignment. As this is an impractical condition, the
validity of this manufacturer rating practice is cast into doubt.

........_

80

I
I

t--""'
Continuous and short-t rm

1- operation
ne
\

Figure 10. Speed Correction Factor of an European Gear

Coupling Line. Notes: The manufacturer uses a different method
(as compared with the one of Figure 7) to maintain a satisfactory
service life, at any acceptable misalignment.

Constant Stresses

80

0.8 1-------

0.6

Failure Line

Speed Factor

1.0

160

Constant Stresses (psi)

Figure 9. Rating Method of a Special-Purpose Diaphragm Cou

pling Line. Notes: The manufacturer uses a less-conservative
rating method, as operation at point A result in a safety factor of
only 1.3, in respect to constant stresses.

Constant stresses are generated by the combination of torque,

centrifugal acceleration, and axial stretch. The latter represents the
largest stress of the three. Therefore, any unpredictable increase in
axial shaft separation, particularly when combined with a momen
tary peak torque, could cause element failures. The combination of
large axial stretch and large torques are often encountered during
the startup of a machine.
Some manufacturers publish speed derating factors related to
operating misalignment. A European coupling manufacturer re
duces the maximum speed as a function of misalignment, as shown
in Figure 10.

Torque Factor
1.0

"'

0.8
0.6
0.4

"'

""

""

---

---

----

0.2
0.0
0

Misalignment (degrees)

Figure 11. Torque Correction Factor of an American Line of High

Misalignment Gear Couplings. Notes: The manufacturer reduces
the rated torque at misalignments larger than 1 degree, in order to
maintain a satisfactory service life at any acceptable misalign
ment.

APPLICATION (EXPERIENCE) FACTORS

Application factors, also named experience factors, establish
the ratio between the rated torque from the catalog and the actual
torque of a given application.

SAFETY, APPLICATION, AND SERVICE FACTORS AS APPLIED TO SHAFT COUPLINGS

Torque Factor

0.3 -'----'-----'---L--...1.--.l
u.2
0.4
0.6
0.8
1.0
Operating speed I Maximum speed

Figure 12. Rating Method of a Japanese Gear-Coupling Manufac

turer. Notes: The manufacturer allows operation at rated torques
only at zero misalignment! This is quite unacceptable, as at zero
misalignment a flexible coupling becomes unnecessary!
The first flexible couplings were of the gear type, and ratings
were of no concern: selection was made based on bore capacity,
not torque capability. The pitch diameter of early couplings was
twice the maximum bore. This geometry resulted in forces on teeth
that allowed couplings to operate, for a satisfactory period of time,
at whatever torque the shafts could transmit. Although current
catalogs provide torque ratings, many types of gear couplings still
have an inherent safety built in, if selected, so that the hubs can be
installed on the shaft, then they can safely transmit the shaft's
torque.
To be competitive, a few gear coupling manufacturers increased
the bore capacity of their couplings while maintaining the same
pitch diameter for the teeth. Not only was the loading on teeth
increased, but the lubricant volume that can be contained within
the coupling was drastically decreased. This increase in bore size
made torque rating, rather than bore size, the determining param
eter in the selection of these types of couplings.
Originally, application factors were established by the coupling
buyer. A number of reasons may have led a purchaser to select a
larger coupling than the one determined from catalog ratings: a
margin was left for future increase in the machine output, the shaft
did not fit in the coupling, or the purchaser may have had a lack of
trust in the catalog ratings, a fear that was fueled by field failures
of "correctly" selected couplings.
Field failures started to occur in alarming numbers about 25
years ago, when a tremendous surge forward occurred in speeds
and power of machinery used in the process industries. Along with
these advances came misapplications and frequent failures of
couplings. One of the authors was involved with nearly thirty
coupling failures in one ammonia syngas train (35,000 hp at
10,500 rpm). Similar experiences were encountered on hydrogen
compressors. Serious questions about ratings occurred when one
coupling manufacturer stated that the life expectancy of its prod
ucts is about 18 months.
Maintenance of couplings became the bottle neck of the plants'
ability to operate machines at increasingly longer periods of times.
As coupling manufacturers provided limited help, users resorted to
measures of their own, including the development of the Turboma
chinery Symposium. Discussions at the Symposium revealed how
universal coupling problems are, and the American Petroleum
Institute (API) decided to take action.
In 1979, API published Standard 671: "Special-Purpose Cou
plings for Refinery Services." Among other guidelines for special
purpose couplings was a torque application factor of 175 percent.

69

The intent of this application factor was not to address the margin
of safety on design, rather, "to mandate an adequate experience
factor in order to allow for off-design operations which may occur
at operating points requiring higher torque than the normal oper
ating point at which the coupling selection is based, as well as
equipment variation resulting in higher torque than actual equip
ment design point."
For awhile, some manufacturers thought that the standard would
put them out of business, as they could no longer be competitive.
Users accepted exceptions to this rule for some time. When new
coupling catalogs were published, they suddenly contained higher
ratings in many instances for identical products. An example of
such an increase after the issuing of API Standard 671 is shown in
Figure 13. Even more striking is a note that appears in the English
version of a European coupling catalog: "A service Factor of 1.75
as per API 671 is already incorporated in the (torque) values
indicated in the list." However, the ratings listed in this catalog
were identical to those published in a pre-API 671 catalog of the
same manufacturer.
%Increase

150
140

- Mdium duty Seris

..._,

'\_

'\.

130
120

"""

'\.
6

I v

-.,1'\.

110
100
4

'\.

I
10

;[\,_

'\.
I
\I

L ."'\
I !\

/".
12

\/

ht Duty Series
!
I
14

16
18
Size

Figure 13. Torque Rating Increases of a Diaphragm-Coupling

Line, After AP/671 Was Issued. Notes: The manufacturer elected
to increase the catalog ratings of two coupling types, without any
apparent changes in geometry or material.
Recently, three manufacturers of disk-pack couplings (two
American and one European) published catalogs that included
their own application factors, as shown in Figure 14. The minimum
application factor that must be applied to the "continuous torque
rating" is 1.5. Under this condition, the published ratings cannot be
used for even the best operating conditions, thus making the
ratings meaningless. Furthermore, all three manufacturers list a
peak torque which is 1.33 times the rated torque (most coupling

APPLICATION FACTORS
Constant Torque ...................... 1.50
API 671 .....................................1.75
Moderate Fluctuations ............ 2.00
Figure 14. Application Factor Table, as Published by Three
Manufacturers of Similar Disk-Pack Coupling Lines. Notes: The
manufacturers of these special-purpose couplings do not allow
operation at rated torques, under any conditions. Therefore, the
published ratings become meaningless.

PROCEEDINGS OF THE TWENTY-SECOND TURBOMACHINERY SYMPOSIUM

70

catalogs list a peak torque that is a minimum of two times the rated
torque). Indeed, multiplying the peak-torque factor of 1.33 by the
"minimum" application factor yields exactly 2.0. This suggests
that the real coupling ratings are 2/3 of the ones listed.
The question that some engineers might ask is: why use appli
cation factors at all? Applying one safety factor on top of the one
the manufacturer has already applied to the ratings seems unnec
essary and uneconomic.
There are at least two reasons for the use of experience factors:
A large number of machines are periodically upgraded. One
example involves a number of simple-cycle Frame 5 gas turbines
with which one of the authors has been involved for over 30 years.
Two such machines were originally installed in the early 1960s,
and could develop 14,000 hp. Through continuous improvements
in power turbine materials, blade geometry, increased compressor
section flow, and better governor systems, the same machines now
generate in excess of 20,000 hp, an increase of over 45 percent.
Because the experience factors originally used in the selection of
the load couplings were large, the same couplings are successfully
used even today.

Torque

_____

__

Experience factors should be used as a correction factor in

those cases when manufacturers use less-than-desirable margins
when they rate couplings. An indiscriminate, across-the-board
experience factor would lead to more uneconomic selections than
wise ones. This statement is particularly valid for cases where
ratings are artificially inflated.

SERVICE FACTORS
Service factors were introduced by manufacturers of gear cou
plings with curved faced teeth. At low misalignment, these cou
plings can transmit less torque than the ones with straight teeth,
because of larger contact pressures. Manufacturers of couplings
with curved teeth often publish the same ratings as the ones of
couplings with straight teeth, but apply a "service factor," a
practice that was not used with couplings having straight teeth.
Service factors are a significant selection factor for applications
where torque fluctuates cyclically. Couplings must be selected for
the maximum torque that occurs during one cycle. As this value is
seldom known, coupling manufacturers used previous experience
to establish a ratio between the maximum torque and the average
torque (as determined through calculations, using power and
speed). The actual torque (curve B) and the calculated torque
(curve A) is shown in Figure 15 for a cyclic torque application.
The areas under the two curves from Figure 15 are identical, as
they represent the power that flows through the coupling. Because
torque, not power, defines the required rating of a coupling, the
selection of a coupling's size must be made for the maximum
torque that occurs during one cycle, and not for the power
transmitted.
The maximum torque during each cycle should not be confused
with the peak torque, which occurs only occasionally.
Original equipment manufacturers know the torque curves of
their machines, and can make judicious coupling selections. With
out a torque curve, the maximum torque must be estimated as a
percentage of the average torque. This percentage factor is the
service factor.
As each coupling manufacturer has its own list of service
factors, attempts were made to make them uniform. A standard
titled "Load Classification and Service Factors" was published in
1968 by AGMA (No. 514.01). With the advent of elastomer
couplings, it was found that various materials react differently to
torque fluctuations, and that service factors must, therefore, be
made a function of coupling material, along with application.
AGMA's standard has since been withdrawn; only manufacturers'
data should be used.

1---- one cycle

Time

Figure 15. Torque Fluctuation through a Coupling. Notes: The

areas under curve A (actual torque), and curve B (calculated
torque) are identical, and represent transmitted power. Couplings
must be selected to accommodate the maximum torque within one
cycle, rather than the power transmitted.

To help with coupling selection, manufacturers publish long

lists of service factors.

Example
A reciprocating compressor rated at 38 hp is driven by an
electric motor rated at 40 hp at 1800 rpm. Assuming that the
selected coupling type has a service factor of two, the selected
coupling size should have a rating of:

Note that in selecting the coupling size, the power consumed, and
not the one of the driving machine, should be used in the calcula
tion.
Service factors vary widely between different types of cou
plings using elastomer flexible elements. When elastomers are
subjected to continuous flexing, they absorb part of the energy
transmitted through the coupling. The energy absorbed (damping)
is transformed into heat, raising the coupling's temperature, which
in turn softens and weakens the elastomer. The amount of heat
absorbed is a function of the magnitude of torque fluctuation, the
operating speed, and the type of elastomer. Generally, rubber has
a smaller damping coefficient than urethanes, and it therefore
absorbs less energy, under the same operating conditions. Without
the cooling provided by the windage (caused by rotation), the
elastomer elements would become very hot, and their strength
diminishes. The authors have seen cases where elements actually
melted, because the coupling guard did not allow for any air
circulation.
Because of the larger damping factor, manufacturers of cou
plings utilizing urethanes recommend a larger service factor than
the ones used with rubber couplings. A comparison of service
factors of similar couplings, using rubber and urethanes is shown
in Figure 16.
The torque generated by the driving machine can also be cyclic,
as is the case with reciprocating engines. Therefore, the maximum
torque transmitted through a coupling during one cycle is a
function of the torque variation of the driving and of the driven

SAFETY, APPLICATION, AND SERVICE FACTORS AS APPLIED TO SHAFT COUPLINGS

Aoolication

71

Rubber

Urethane

Agitators

1.0

1.5

Agitator, heavy liquid .... ............................ 2

Vane Blowers

1.0

1.5

Compressor, centrifugal ................... ......... 1.25

Agitator, light liquid ................................... 1.25

.

Car pullers

1.5

2.0

Compressor, reciprocating ... ... ....

Lobe-type compressors

1.5

2.0

Fan, blower small ....................................... 1.25

Crane drives

1.5

2.0

Fan, blower heavy ...................................... 1.75

Cooling tower fans

1.5

2.5

Generator .................................................... 1.5

Paper mill chippers

2.5

3.5

Propeller ..................................................... 2

Banbury mixers

2.5

3.0

Wood and plastic machinery ...................... 1.5

Figure 16. Service Factors Vs Element Material. Notes: Service

factors for couplings using urethanes are larger than the ones that
use rubber, as urethanes absorb more energy (therefore become
hotter) when subjected to flexing, as caused by misalignment or
torque fluctuations.

...

.......... 1.75

Figure 17. Published Service Factors of an European Disk-Pack

Coupling Line. Notes: No service factor of one is listed! The
published ratings are therefore meaningless, as they cannot be
used for any application.
Coupling manufacturers should establish an "industry stan
dard" that covers safety and service factors. Without such a
standard, users' associations should consider mandating minimum
safety factors, and request proof that these factors are met.

machines. Most coupling manufacturers list two service factors,

one for the driver and one for the driven, factors which are
additive. The sum of the two factors becomes the new service
factor. It is important to remember that:
When the torque consumed is constant in time, the service
factor for the driven machine is one.

Service factors for smooth service should always be unity (1),

as the only purpose of service factors is to compare rough service
to smooth service.

When the torque generated is constant in time, the service

factor for the driving machine is zero.

Manufacturers should avoid cataloguing application factors,

as such factors are strictly users' tools.

Example:

An indiscriminate, across-the-board value for experience

(application) factors would not serve either purpose of such
factors: to allow for future upgrading, and to numerically show
users' confidence level in a coupling type, or in a manufacturer.
However, a standardized experience factor (such as the one recom
mended by API 671) should be used whenever there is no previous
experience with a particular type of equipment or coupling.

A reciprocating compressor is driven by a four-cylinder gaso

line engine. The compressor's service factor is two, while the
engine's service factor is 0.5. The selection of the coupling should
be made using a correction factor of 2 + 0.5 2.5.
A survey of service factors will reveal that recommendations of
coupling manufacturers from many nations have the smallest
factor larger than one for even the best possible conditions. This
practice indicates that the coupling ratings are effectively mean
ingless. The service factor tabulation of an European coupling
manufacturer is shown Figure 17.
=

CONCLUSIONS AND RECOMMENDATIONS

Coupling users must be knowledgeable about the process of
coupling selection, in order to avoid the pitfalls of meaningless or
misleading ratings, and the possibility of ending up with an
undersized coupling. To this end, the following recommendations
are made:
Users should question the veracity of coupling catalogs that
contain ratings which cannot be used even under the best conditions.

ACKNOWLEDGEMENTS
The authors used data from published catalogs of a large number
of coupling manufacturers in the preparation of the paper. Actual
data will be made available on request.
The examples herein were chosen from manufacturers in three
continents only to show that the practices described are universal
in nature. Even though tables and figures use data from specific
catalogs, which may be "American," "Japanese," or "European,"
the type of data shown can be found in many other catalogs. The
examples given are not intended to single out a particular manufac
turer or country.

72

PROCEEDINGS OF THE TWENTY-SECOND TURBOMACHINERY SYMPOSIUM