TABLE OF CONTENTS

This publication represents joint effort b y

personnel of Flowserve Corporation .

Pump Requirements for the

Chemical Process Industries . . . . . . . . . . . . . . . . . . . . . . . 1
Head and Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 3
Sixth Edition
Centrifugal Pump Principles . . . . . . . . . . . . . . . . . . . . . . 29
(price $10.00)
t I3 4 Total Head Required by afPipingSystem . . . . . . . . . . . . 33
. I 9 5 Net Positive Suction Head . . . . . . . . . . . . . . . . . . . . . . . . 6 2
el$g Flow of Liquids in Pipes . . . . . . . . . . . . . . . . . . . . . . . . . 8 2
C. 1 9
7 The Affinity Laws Pertaining
TO OBTAIN A COPY OF THE PIPE FRICTION MANUAL
CONTACT: c. 1 9 t o Centrifugal Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
A.P. WHERRY & ASSOCIATES, INC . c.1 3 Viscosity
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96
712 LAKEWOOD CENTER. NORTH
14600 DETROIT AVENUE 1-3
CLEVELAND. OHIO 44 107 C !3 Specific Speed and Suction Specific Speed . . . . . . . . . . .113
c t3 1 0 Impeller Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .115
-311 Bearings and Bearing Life . . . . . . . . . . . . . . . . . . . . . . . .118
'12 Shaft Sealing Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 2 2
c 9
13 Shaft Deflection, Whip and Runout . . . . . . . . . . . . . . . 1. 2 7
, ' 9
C 914 Vibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. 3 0

Copyright O 1980 Flowserve Corporation . 6 915 Heat Generated in Pumping. . . . . . . . . . . . . . . . . . . . . . 133

.
Printed in the United States of America .
316 Self-priming Centrifugal Pumps . . . . . . . . . . . . . . . . . . . .136
All rights reserved . This book. or parts thereof.
may not be reproduced in any form
without permission of the publishers .
'-
,- 3
j17 Centrifugal Pump Trouble Shooting Guide . . . . . . . . . . .1 3 9
Suction Piping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. 4 9
i 3l8
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .154

.. v+**$*i*
 PUMP REQUIREMENTS FOR THE
CHEMICAL PROCESS INDUSTRIES
d : Pumps for the chemical process industries differ from those used in
6, ; other industries primarily in the materials from which they are made.
While cast iron, ductile iron, carbon steel, and alumimum or copper-
@ ; base alloys will handle a few chemical solutions, most chemical pumps
e ; are made of stainless steel, nickel-base alloys, or more exotic metals
@gv. such as titanium and zirconium. Pumps are also available in carbon,
glass, porcelain, rubber, lead and whole families of plastics, including
@, : phenolics, epoxies and fluorocarbons.
*, Each of these materials has been incorporated into pump designs for
just one reason: t o eliminate or reduce the destructive effect of the
*
'
chemical on the pump parts.
Since the type of corrosive liquid will determine which of these
@*
materials will be most suitable, a careful analysis of the chemical to
-4 be handled must first be made.
C+ ,. SOLUTION ANALYSIS
C,
t '.
w

* Most important in studying any solution will be a knowledge of its

''
C:

t'
'
4 constituents. This includes not only the major constituents,,'but the
minor ones as well, for in many instances the minor constituents will
be the more important. They can drastically alter corrosion rates,
3 and therefore, a detailed analysis is most critical.
t Closely allied and directly related t o the constituents, is the concent-
ration of each. Merely stating "concentrated", "dilute", or "trace
3 quantities" is basically meaningless because of the broad scope of inter-
e s pretation of these factors.
'-
c:
' For instance, some interpret "concentrated" as meaning having a
concentration greater than 50% by weight; whereas others interpret any
3 concentration above 5% as being "concentrated". Hence, it is always
&:. 9 desirable to cite the percentage by weight of each constituent in a given
solution. This eliminates multiple interpretation and permits a more
*' 3 accurate evaluation.
c* j
 I t is also recommended that the percentage by weight of any trace
* ,-I

Pump design is a very critical factor when solids are in the solution.
quantities be cited, even if this involves only parts per million. For @ -q
I t is not uncommon for a given alloy t o range from satisfactory t o
example, high-silicon iron might be completely suitable in a given
environment in the absence of fluorides. If, however, the same envir-
+ -, completely unsatisfactory in a given chemical application with hy-
draulic design being the only variable factor. Failure t o cite the
onment contained even a few tenths of one percent of fluoride, the & \ presence of solids is a common occurrence. This undoubtedly is the
high silicon iron could suffer catastrophic failure. t - b
cause of catastrophic erosion-corrosion failures experienced in many
pumps.
PROPERTIES OF THE SOLUTION The presence of air in a solution can be quite significiant. In s o m e
instances, it is the difference between success and failure, in t h a t air
Generalized terms such as "hot", "cold", or even "ambient" are can conceivably render a reducing solution into an oxidizing one.
ambiguous in that they can be interpreted in different ways. The Under these circumstances, an altogether different material may be
preferred terminology would be the maximum, minimum and normal required. A good example of this would be a self-priming, nickel-
operating temperature in degrees, either F or C. Chemical reactions, molybdenum-alloy pump for handling commercially pure hydrochloric
in general, increase in rate of activity approximately two times with acid. This alloy is excellent for the commerically pure form of t h e acid,
each increase of 18 O F. in temperature. Corrosion can be considered but any condition that can induce an even slightly oxidizing tendency
a chemical reaction, and with this is mind, the importance of temper- could render this same alloy completely unsuitable.
ature or temperature range is obvious.
A weather-exposed pump installation is a good illustration of the OTHER FACTORS IN SOLUTION ANALYSIS
ambiguity of the term "ambient", for there could be as much as
150 O F . difference between the extremes of temperature t o which When a pump is used for transferring or recirculating a solution,
the pump could be subjected. This is particularly important for t h e
there is a possible buildup of corrosive products or contaminants that
pump materials that are subject t o thermal shock in addition t o in-
can influence the service life of the pump. Such a buildup can have a
creased corrosion rates in high temperature environments.
beneficial or deleterious effect; and for this reason the possibility of
More often than not, little consideration is given t o the pH of process such buildup should be an integral part of evaluating solution char-
solutions. This can be a critical and well-controlled factor during pro- acteristics.
duction processing, and it can be equally revealing in evaluating sol- Inhibitors and accelerators can be intentionally o r unintentionally
ution characteristics for material selection. One reason the pH may be added t o the solution. Inhibitors reduce corrosivity whereas acceler-
overlooked is that generally it is obvious that the corrosive is acidic or ators increase it. Quite obviously, n o one would add an accelerator
alkaline. However, this is n o t always true, particularly on process sol- t o increase the corrosion rate on a piece of equipment; but the addition
utions that may have the pH adjusted so that it will always be either of a minor constituent, as a necessary part of a given process, may serve
alkaline or acidic. When this situation exists, the precise details should the same purpose. Thus, the importance of knowing whether such
be known so that a more thorough evaluation can be made. constituents are present.
I t is also quite important t o know when a solution is alternating Where purity of a product is of absolute importance, particular note
between acid and alkaline conditions, because this can have a pro- should be made of any element in the pump alloy that may cause con-
nounced effect on materials selection. Some materials, while entirely tamination problems, whether it be discoloration of product or solution
suitable for handling a given alkaline or acid solution, may not be suit- breakdown. In some environments, pickup of a few parts per billion of
able for handling the same solutions alternately. certain elements can create severe problems. This is particularly im-
Erosion-corrosion, velocity, and solids in suspension are closely allied portant in pumps, where velocity effects and the presence of solids can
in chemical process industries pump services. alter t h e end-results.
 Depending o n the process liquid, continuous or intermittent contact t 'i excellent information retrieval system is n o t available, it can be very
can affect service life. Intermittent duty, in some enviroments, can difficult t o obtain the desired information.
be more destructive than continuous duty. If the pump remains half t **
full of corrosives during periods of downtime, it may suffer accelerated t 3 TYPES OF PUMP CORROSION
corrosion at the liquid interface. Perhaps of equal importance is e -*
whether or not the pump is flushed and/or drained when not in service. The corrosion encountered in chemical pumps may appear t o be
tr 9 unique compared t o that found in the other process equipment. Never-
CORROSIVES AND MATERIALS . i theless, like other types of chemical process equipment, pumps will
experience only eight forms of corrosion. I t is not the intent here t o
Materials for pump applications can, in general, be divided into two describe in detail these forms of corrosion, b u t only t o provide a brief
very broad categories; metallic and nonmetallic. The metallic category description so that the various forms can be recognized when they
can be further subdivided into ferrous and nonferrous alloys, both of occur.
which have extensive application in the chemical industry. The non- 1. General, or uniform, corrosion is the most common type,
metallic~can be further subdivided into natural and synthetic rubbers, characterized by essentially the same rate of deterioration
plastics, ceramics, glass, carbon and graphite. over the entire wetted or exposed surface. General cor-
For a given application, a thorough evaluation of not only the rosion may be very slow or very rapid, but is of less concern
solution characteristics but also the materials available should be made than the other forms because of its predictability.
to ensure the most economical selection. 2. Concentration-cell, or crevice, corrosion is a localized form
resulting from small quantities of stagnant solution in areas
SOURCES OF DATA such as threads, gasket surfaces, holes, crevices, surface
deposits, and under bolt and rivet heads. When concent-
T o evalute materials for chemical pump services, various sources of ration-cell or crevice corrosion occurs, a difference in
data are available. The best is practical experience within one's own concentration of metal ions or oxygen exists in the stagnant
organization. It is not unusual, particularly in large organizations, t o area compared t o the main body of liquid. This causes an
have a materials group or corrosion group whose basic responsibilities electrical current t o flow between the two areas, resulting
are t o collect and compile corrosion data pertaining t o process equip- in severe localized attack in the stagnant area. Usually, this
ment in service at the company's various plants. These sources should form of corrosion does n o t occur in chemical pumps except
be consulted whenever a material evaluation program is made. perhaps under gaskets, or in designs where the factors
A second source of data is laboratory and pilot-plant experience. known t o contribute to concentration-cell corrosion have
Though it quite logically cannot provide as valuable and detailed been ignored.
information as actual plant experience, it can certainly be indicative 3. Pitting corrosion is the most insidious, destructive form of
and serve as an important guide. corrosion, and very difficult t o predict. It is extremely
The knowledge of suppliers can be a third source of information. localized. It is manifested by small or large holes (usually
Though suppliers cannot hope to provide comments on the specific small) and the weight loss due t o the pits will be only a
details of a given process and the constituents involved, they normally small percentage of the total weight of the equipment.
can provide assistance and corrosion coupons t o facilitate a decision. Chlorides in particular are notorious for inducing pitting
Technical journals, handbooks and periodicals are a fourth source of that can occur in practically all types of equipment. This
information. A wealth of information is contained in them, but if an form of corrosion, in some instances, can be closely allied
 1
*
*
to concentration-cell corrosion, as pits may initiate in the - 6. Galvanic corrosion occurs when dissimilar metals are in
same areas where such corrosion is manifested. Pitting is '-
contact, or otherwise electrically connected, in a corrosive
q
commonplace at areas other than stagnant ones; whereas, medium. Corrosion of the less noble metal is accelerated
concentration-cell corrosion is basically confined t o areas Q and the more corrosion resistant metal is decreased com-
of stagnation. 4 pared with their behavior when riot in contact. T h e further
Stress-corrosion cracking is localized failure caused by the apart that the metals or alloys are in the electromotive
combination of tensile stresses and a specific medium. 9
series, the greater the possibility of galvanic corrosion.
Undoubtedly, more research and development has been r When it is found necessary t o have t w o dissimilar metals
conducted o n this form of corrosion than any other. Never- in contact, caution should be exercised t o make certain that
9
theless, the exact mechanism of stress-corrosion cracking is t h e total surface area of the less resistant metal far exceeds
still not well understood. Fortunately, castings d u e t o their that of t h e more corrosion resistant material. This will tend
basic overdesign, seldom experience stress-corrosion crack-
ing. Corrosion fatigue, which can be classified as a form
*
-
t o prevent premature failure simply by providing a substant-
ially greater area of the more corrosion prone material. This
of stress-corrosion, is of concern in chemical pump shafts * form of corrosion is not common in chemical pumps, but
a
because of repeated cyclic stressing. Failures of this type 7 may be of some concern with accessory items t h a t may be
occur a t stress levels below the yield point, due t o the cyclic A
in contact with t h e pump parts, and are subjected t o the
4
application of the stress. pumped solution.
Intergranular corrosion is a selective form of corrosion at 3 7. Erosion-corrosion is characterized by accelerated attack
the alloy grain boundaries. I t is associated primarily with 9 resulting from the combination of corrosion and mechanical
stainless steels but can also occur with other alloy systems. wear. I t may involve solids in suspension and/or high
In stainless steels, it results from subjecting the material 3 velocity. I t is quite common with pumps where t h e erosive
t o heat in the 800 to 1,600 O F. range. Unless other alloy 3 effects prevent the formation of a passive surface o n alloys
adjustments are made, this form of corrosion can be pre- 9 that require passivity t o be corrosion resistant. T h e ideal
vented only by heat treating. It is easily detectable in material t o avoid erosion-corrosion in pumps would possess
castings, because the actual grains are quite large compared 3 the characteristics of corrosion resistance, strength, ductility
t o wrought material of the equivalent composition. In 3 and high hardness. Few materials possess such a com-
somes instances, uniform corrosion is misinterpreted as
intergranular corrosion because of the etched appearance 3 bination.
Cavitation is considered a special form of erosion-
of the surfaces exposed t o the enviroment. Even in ideally 3 corrosion that results from the collapse or implosion of gas
heat-treated stainless steels, very slight accelerated attack 5 bubbles against the metal surface in high-pressure regions.
can be noticed a t the grain boundaries because these areas The stresses created are high enough t o actually remove
are more reactive than the grains themselves. Caution is 3 metal from the surface and destroy passivity. T h e same
required t o avoid misinterpreting general and intergranular 9 material qualities mentioned above are desirable for cav-
corrosion. Stainless steel castings will never encounter
3 itation resistance, but an increase in suction pressure will
ir.tergranular corrosion if they are properly heat treated eliminate the problem in most cases.
after being exposed t o temperaures in the 8 0 0 t o 1,600 O F. 9 8. Selective-leaching corrosion involves removal of o n e element
range. 9 from a solid alloy in a corrosive medium. Specifically, it is
9
9
 typified by dezincification, dealuminumification, and polytetrafluoroethylene (PTFE) and fluorinated ethylenepropylene
graphitization. This form of attack is not common t o (FEP) are finding wide application. Where strength and chemical
chemical pumps, because the alloys in which it occurs resistance are needed, a variety of fiber-reinforced plastics ( F R P )
are not commonly used in heavy chemical applications. are available. Epoxy, polyester, and phenolic are three of t h e more
popular F R P materials. Polyvinyl chloride, polyethylene and poly-
TYPICAL MATERIALS OF CONSTRUCTION
propylene are also finding application. Plastics are gaining in popularity
because they offer the corrosion resistance of the more expensive
T h e most widely used metallic materials of construction for chemical metals a t a fraction of the cost. However, they have strength limit-
pumps are the stainless steels. Of the many available, the most popular ations and it is doubtful that plastics will ever completely replace
are the austenitic grades, such as the cast equivalents of T y p e 3 0 4 and metals.
Type 316, which possess superior corrosion properties compared t o the Ceramic or glass construction is avoided whenever possible, because
martensitic or ferritic grades. of the poor mechanical properties of these materials. However, for
The steels are used for a wide range of corrosive solutions. many extremely corrosive services at elevated temperatures, glass or
They are suitable for most mineral acids a t moderate temperatures and ceramic are the most suitable because of their extreme chemical inert-
concentrations. The notable exceptions are hydrochloric and hydro- ness.
fluoric acids. In general, the stainless steels are more suitable for Carbon or graphite construction is generally used for the same kinds
oxidizing than for reducing environments. Organic acids and neutral-to- of services as are ceramic o r glass. The primary reason for using carbon
alkaline salt solutions are also handled by stainless steel pumps. or graphite instead of glass o r ceramic is that the former are suitable for
Carbon steel, cast iron, and ductile cast iron are also frequently used services where H F o r alkalis are handled.
for the many mildly corrosive applications found in most plants.
For the more severe or critical services, the high-alloy stainless TYPES OF CHEMICAL PUMPS
steels such as Alloy 20 are frequently specified.
Nickel-base alloys, because of their relatively high cost, are generally T h e second step in selecting a chemical pump is based o n the char-
used only where n o iron-base alloy is suitable. This family of corrosion acteristics of the liquid and on the desired head and capacity. I t should
resistant materials includes: pure nickel, nickel-copper, nickel-chrom- be noted that not all types of pumps are available in every material of
ium, nickel-molybdenum, and nickel-chromium-molybdenum alloys. construction, and the final selection may depend on availability of
Aluminum, titanium and copper-base alloys such as bronze o r brass designs in the proper material.
are the most frequently used nonferrous metals for chemical pumps. Centrifugal pumps are used extensively in the chemical process
Zirconium has also found application in a few very special areas. industries because of their suitability for practically any service. They
Both natural rubber and synthetic rubber linings are used extensively are available in an almost unending array of corrosion resisting mater-
for abrasive and/or corrosive applications. Soft natural rubber generally ials. While not built in extremely large sizes, pumps with capacity
has the best abrasion resistance, but cannot be used a t as high a temper- ranges of 5,000 t o 6,000 gpm are commonplace. Heads range as high
ature as semi-hard natural rubber or the synthetic rubbers such as as 5 0 0 t o 6 0 0 feet, a t standard electric motor speeds. Centrifugal
Neoprene and butyl. In most cases, the hard rubbers and synthetic pumps are normally mounted in the horizontal position; b u t they may
rubbers also possess better chemical resistance. be also installed vertically, suspended into a tank, or hung in a pipeline.
Plastics are among the fastest growing families of pump materials. Disadvantages include reduced performance when handling viscous
A multitude of new plastics have been introduced in recent years. For liquids, and the tendency t o lose prime when comparatively small
the ultimate in chemical resistance, the fluorocarbon resins such as amounts of air o r vapor are present in the liquid.
 P *
t 3
Rotary pumps such as the gear, screw, deforming-vane, sliding-vane, e 2
axial-piston, and cam types are generally used for services requiring 5 0 0 @ 3
t o 1,000 psi discharge pressure. They are particularly adept a t pumping
liquids of high viscosity o r low vapor pressure. Their constant displace- t r
ment at a set speed makes them ideal for metering small quantities of k 4
liquid. Since they operate on the positive-displacement principle, they
are inherently self-priming. When built of materials that tend t o gall o r
c
seize on rubbing contact, the clearance between mating parts must be
increased, resulting in decreased efficiency. The gear, sliding-vane and 9
cam units are generally limited t o use o n clear, nonabrasive liquids.
Reciprocating pumps have, t o a great extent, been replaced by
centrifugal or rotary units, except for special applications. They are t @
still used extensively where their variable-speed and variable-stroke t $
capabilities are important process considerations. This characteristic,
in combination with their inherent ability t o handle volatile and very
viscous liquids, makes them particularly suitable for metering and L 2
injection systems where low capacity and h ~ g hheads are normal duty
parameters. These pumps are available for discharge pressures as high
as 50,000 psi. Disadvantages include comparatively high NPSH require- @L 3
ments, the vulnerability of available check-valve materials t o chemical +- j
liquids, and relatively poor performance where solids, abrasives o r dirt
are present in the process liquid. Most commercially available recip-
rocating pumps use multiple cylinders, i.e., duplex, triplex, o r quin- C 4.
tuplex, t o smooth the pulsating flow generated by the reciprocating L, 3
motion.
Diaphragm pumps are also classed as positive-displacement pumps
because the diaphragm acts as a limited-displacement piston. Pumping j
action is obtained when the diaphragm is forced into reciprocating FIGURE 1 1
motion by mechanical linkage, compressed air, or oil from a pulsating
em- 3
external source. This type of construction eliminates any connection t 9
between the liquid being pumped and the source of energy, and thereby t_ ,3
eliminates the possibility of leakage. This characteristic is of great
importance when handling toxic or very expensive liquids. Disadvan- t- .9 Regenerative-turbine pumps easily handle flow rates u p t o 1 0 0 gpm
tages include a limited selection of corrosion resistant diaphragm mater- and heads u p t o 7 0 0 feet. When used for chemical service, the internal
3
ials, limited head capacity range, and the necessity of using check valves clearances must be increased t o prevent rubbing contact, resulting in
in the suction and discharge nozzles. Construct~ondetails are shown in
E. decreased efficiency. These pumps are generally unsuitable for solid-
Figure 1.I. k. 9 liquid mixtures of any concentration.
Eu 3
10
C" 9
DESIGN CONSIDERATIONS FOR CHEMICAL PUMPS
Gasket materials must be selected t o resist the chemical being han-
dled. Compressed asbestos, lead, and certain synthetic rubbers have
Practically all the major wetted components of chemical pumps are been used extensively for corrosion services. In recent years, the
cast. Needless t o say, it is a fruitless venture to thoroughly evaluate the
fluorocarbon resins have come into widespread use, d u e to their almost
detailed characteristics of the pumped liquid and the material to be
universal corrosion resistance.
used if the component castings d o not satisfy the quality needed t o
The power end of pumps consists of the bearing housing, bearings,
provide good service life. This is probably of more concern in chemical
oil or grease seals, and the bearing lubrication system. This assembly
pump applications than in any other type of service, because loss of
is normally made of iron or steel components, and thus must be
~ r o d u c tand downtime can be extremely costly; and leakage can be
designed t o withstand the severe environment of the chemical plant.
very dangerous.
For example, when venting of the bearing housing is required, special
Of the several factors that determine whether or not a cetain material
means of preventing the entrance of water, chemical fumes, or dirt
can be used f o r a particular pump design, mechanical properties are the
must be incorporated into the vent designs.
most important. Materials may possess outstanding corrosion resistance
T h e bearing that controls axial movement of the shaft is usually
but may be completely impossible t o produce in the form of a chemical
selected to limit movement t o 0.002 in. o r less. End-play values above
pump because of limited mechanical properties. Hence, awareness of this limit have been found detrimental t o mechanical-seal operation.
these properties is essential for any material being considered in a
Water jacketing of the bearing housing may be necessary under
corrosion evalution program. Such a program gives a relatively good
certain conditions to maintain bearing temperatures below 225 " F, the
indication of whether or not a particular design may be available. Since
upper limit for standard bearings.
most materials are covered by ASTM or other specifications, such
Maintenance of a chemical pump in a corrosive environment can be
sources can be used for reference purposes. A table of mechanical
very costly and time consuming. I t can be divided into two categories,
properties and other characteristics of proprietary materials not in-
preventative and emergency. When evaluating materials and design
cluded in any standard specifications should be readily available from
the manufacturer of the material. factors, maintenance aspects should be high o n the priority list. T h e
Weldrnents or welded construction should impose n o limitation, ease and frequency of maintenance are very critical factors in any
providing the weldment is as good as, or better than, the base material. preventative-main tenance program.
Materials requiring heat treatment in order to achieve maximum cor-
rosion resistance must be heat treated after a welding operation, or STUFFING BOX DESIGN
other adjustments must be made, to make certain that corrosion
resistance has not be sacrificed. The area around the stuffing box probably causes more failures of
chemical pumps than all other parts combined. T h e problem of es-
Wall sections in pumps are generally increased over the mechanical
tablishing a seal between a rotating shaft and the stationary pump
design requirements, so that full pumping capability will be maintained
parts is one of the most intricate and vexing problems facing the pump
even after the loss of some material to the corrosive enviroment.
designer.
Pressure containing parts are also made thicker so they will remain L Packings of braided asbestos, lead, fluorocarbon resins, aluminum,
serviceable after a specified amount of corrosive deterioration. Areas graphite and many other materials, or combinations of these materials,
subject to high velocities, such as the cutwater of a centrifugal casing, have been used t o establish the seal. Inconsistent as it seems, a small
are further reinforced t o allow for the accelerated corrosion caused by amount of liquid must be allowed to seep through the packing t o
high velocities in the liquid. L : lubricate the surface between packing and the shaft. This leakage
rate is hard t o control, and the usual result is overtightening of the pregnated with grease for lubrication, but these lubricants break d o w n
packing t o stop the leak. The unfortunate consequence is the rapid a t temperatures above 250 O F. resulting in further temperature increases
scoring of the shaft surface, making it much harder t o adjust the because of friction. O n e of the less obvious results of this temperature
packing t o the proper compression. Recommendations as t o the type increase is corrosive attack o n the pump parts in the heat zone. Many
of packing t o be used for various chemical services should come from materials selected for the pumping temperature will be completely
the packing manufacturer. unsuitable in the presence of the corrosive at elevated stuffing b o x
Mechanical shaft seals are used extensively o n chemical pumps. Once temperatures. Another source of heat is the chemical solution itself.
again, the primary consideration is selection of the proper materials These liquids are often in the 3 0 0 F. range, and some go as high
f o r the type of corrosive being pumped. Stainless steels, ceramics, as 700 O I?.
graphite and fluorocarbon resins are used t o make the bulk of the seal T h e best answer t o the heat problem is removal of the heat by means
parts. Several large manufacturers of this equipment have very com- of a water jacket around the stuffing box. While heat conductivity is
plete files on seal designs for various chemical services. Typical seal rather low for most chemical pump materials, the stuffing box area
installations are shown in Figures 1.2 and 1.3. generally can be maintained in the 200 O F. t o 2 5 0 ° F. range. This
The operating temperature is one of the most important factors cooling is of further benefit in that it prevents the transfer of heat
affecting the stuffing box sealing medium. Most packings are im-

-
DOUBLE-INTERNAL SEAL WITH
TAPS IN FOLLOWER FLANGE

SINGLE-INTERNAL SEAL,

SINGLE-INTERNAL BALANCED SEAL,

FLEXIBLY MOUNTED SEAT
INTERNAL SEALS WITH VENT AND DRAIN

CLAMPED-IN SEAT WlTH

Flush directly
over seal facer
'
TYPICAL INSIDE OR OUTSIDE SEAL
FLUSHING OF SEAL FACES
FIGURE 1.2
FIGURE 1.3
 * ;
e s
along the shaft t o the bearing housing, thereby, eliminating other t ;resist the tendency t o wear under the packing or mechanical seal
problems around the bearings.
Stuffing box pressure varies with suction pressure, impeller design, e , parts. Further, it must be capable of withstanding the sudden temper-
ature changes often encountered.
and the degree of maintenance of close-fitting seal rings. Variations e Since it is n o t economically feasible t o make the entire pump shaft
in impeller design would include those using vertical or horizontal t, r of stainless alloys, and physically impossible t o make functional carbon,
seal rings in combination with balance ports, as opposed t o those usingc , glass, or plastic shafts, chemical pumps often have carbon steel shafts
back vanes or pump-out vanes. All impeller designs depend upon a with a protective coating or sleeve over the steel in the stuffing box
close-running clearance between the impeller and the stationary pump CL f area. Cylindrical sleeves are sometimes made so that they be removed
parts. This clearance must be kept as small as possible t o prevent , and replaced when they become worn. Other designs use sleeves that
excessive recirculation of the liquid, and resulting loss of efficiency. are permanently bonded t o the shaft t o minimize runout and deflec-
Unfortunately, most chemical pump materials tend t o seize when V tion.
subjected t o rubbing contact. Therefore, running clearances must be c i Another method of obtaining a hard surface in this region is t h e
increased considerably above those for other industries. ; welded overlay or spray coating of hard metals o n t o the base shaft.
At pressures above 1 0 0 psig, packing is generally unsatisfactory c . Materials applied by these techniques are generally lacking in corrosion
unless the stuffing box is very deep, and the operator is especially t resistance, and have not been widely accepted for severe chemical ser-
adept at maintaining the proper gland pressure o n t h e packing. Mech-e
anical seals incorporating a balance feature to relieve the high pressure
; vice. Ceramic materials applied by the plasma-spray technique possess
are the best means of sealing a t pressures above 1 0 0 psig. e a excellent corrosion resistance, but cannot achieve the complete density
required t o protect the underlying shaft.
The pump shaft can create additional stuffing box problems. Ob- @ 2 Composite shafts using carbon steel for t h e power end and a higher
viously, a shaft that is out-of-round or bent will form a larger hole j alloy for the wet end have been used extensively where the high alloy
in the packing, allowing liquid t o escape. Lack of static or hydraulic end has acceptable resistance. Since the t w o ends are joined by various
balance in the impeller produces a dynamic bend in the shaft, resulting 3 welding techniques, the combination of metals is limited t o those that
in the same condition. Undersize shafts, or those made of materials 9 can be easily welded. On such assemblies, t h e weld joint and t h e heat
that bend readily, will deflect from their true center in response t o affected zone must be outside the wetted area of the shaft.
radial thrust on the impeller. This action produces a secondary hole in Elimination of the stuffing box and its problems has been the object
the packing, and again allows the liquid t o escape. of several designs for chemical pumps other than the diaphragm pump
Mechanical seal performance is also impaired when the shaft is bent
or deflects during operation. Since the flexible member of the seal
t $ previously described.
Vertical submerged pumps use a sleeve-type bearing in t h e area
must adjust with each revolution of the shaft, excessive deflection 3 immediately above the impeller t o limit the flow of liquid u p the shaft.
value, the flexible seal member will be unable t o react with sufficient
speed t o keep the seal faces together, allowing leakage a t the mating
=
results in shortened seal life. If the deflection is of more than nominal 3 For chemical service, the problem of materials associated with this bear-
9 ing and its lubrication have been major disadvantages.
Canned pumps, wherein the motor windings are hermetically sealed
faces. An arbitrary limit of 0.002 in. eel "cans", also avoid the use of a stuffing box. The
h u m deflection o r runout of the sha quid is circulated through the motor section, lubricating
In the stuffing box region, the sh type bearings that support the rotating assembly. Disadvan-
resistance at least equal to, and preferably better than, that of wetted tages again center on the selection of bearing materials compatible
parts of the pump. In addition, this orrosive liquid; the lubrication of these bearings when hand-

16 17

I
 in strength a t higher pumping temperatures. Phenolic and epoxy parts
ling nonlubricating liquids; and the probability of the liquid paths
through the motor section becoming clogged when solid-liquid mixtures are subject t o gradual loss of dimensional integrity because of their
creep characteristics. The low tensile strength of the unfilled resins
are handled.
again dictates a design that will place these parts in compression,
DESIGNING WITH SPECIAL MATERIALS and eliminate bending stresses. Typical construction details are
shown in Figure 1.4.
I
As described earlier, a number of low-mechanical strength materials
have been used extensively in chemical pump construction. While
breakage problems are inherently associated with these materials,
their excellent corrosion resistance has allowed them t o remain com-
petitive with higher strength alloys. Of course, their low tensile
strength and brittleness make them sensitive t o tensile or bending
stresses, requiring special pump designs. The parts are held together by
outside clamping means, and braced t o prevent bending. T h e unit must
also be ~ r o t e c t e dfrom sudden temperature changes and from mech-
anical impact from outside sources.
Although produced by very few manufacturers, high silicon iron is
the most universally corrosion resistance metallic material available a t
an economic price. This resistance, coupled with a hardness of approx-
imately Brine11 520, provides an excellent material for handling abrasive
chemical slurries. T h e material's hardness however, precludes normal
machining operations, and the parts must be designed for machine
grinding. The hardness also eliminates the possibility of using drilled
or tapped holes for connecting piping to the pump parts. Therefore,
special designs are required for process piping, stuffing box lubrication, FIGURE 1.4
and drain connections.
Ceramics and glass are similar t o high silicon iron in regard to hard- Polytetrafluoroethylene and hexafluoropropylene possess excellent
ness, brittleness and susceptibility t o thermal or mechanical shock. corrosion resistance. These resins have been used for gaskets, packing,
Pump designs nlust, therefore, incorporate the same special consider- mechanical seal parts, and flexible piping connectors. Several pumps
ations. made of these materials have reached the market in recent years.
Glass linings or coats on iron or steel parts are sometimes used t o
Problems associated have centered around these materials' tendency t o
eliminate some of the undesirable characteristics of solid glass. While
cold flow under pressure, and their high coefficients of expansion
this usuage provides for connecting process piping, the dissimilar
compared t o the metallic components of the unit. Pumps may b e made
expansion characteristics of the two materials generate small cracks
in the glass, allowing corrosive attack. of heavy solid sections, as illustrated in Figure 1.4 or may use more
conventional metallic components lined with t h e fluorocarbon
Thermosetting and thermoplastic materials are used extensively in
materials as shown in Figure 1.5.
services where chlorides are present. Their primary disadvantage is loss
 C d
DIMENSIONS FOR HORIZONTAL CHEMICAL PROCESS PUMPS-TABLE 1.1
e 2

6% 4
9 AB
12%
6% 4

CHEMICAL PUMP STANDARDS

2% 11 4

Early in 1962, a committee of the Manufacturing Chemists 9

Association (MCA) reached an agreement with a special committee of
the Hydraulic Institute on a proposed standard for chemical process
pumps. This document was referred to as the American Voluntary
Standard (AVS), or MCA Standard. In 1971, it was accepted by the & 3
~ m e r i c a n~ a t i o n a lStandards Institute (ANSI), and issued as ANSI
Standard B123.1. In 1974, B123.1 Was redesignated as ANSI B73.1. ' s u r r a o n c o n n e c t , o n ma'v nave l a m e d boll holes D~mens~ons
tn Inches
Many pump manufacturers in the U. S. and a few in foreign countries C 9 It is the intent of this Standard to outline certain design features
are building pumps that meet these dimensional and design criteria. 3 that will minimize maintenance problems. The Standard states, for
It is the intent of the Standard that pumps of similar size from all instance, that the pump shaft should be sized so that
C: 3 maximum
sources of supply shall be interchangable with respect to mounting
shaft deflection, measured at the face of the stuffing box' when the
dimensions, size, and location of suction and discharge nozzles, input c: 9 pump is operating urider its most adverse conditions, will not exceed
shafts, base plates and foundation lists the pump
dimensions that have been standardized.
c. 3 0.002 in. The Standard does not specify shaft diameter, since impeller
, 5 diameter, shaft length, and provision for operating with liquids of high
specific gravities would determine the proper diameter. '
20 2; j 21

, . ".
 The Standard also states that the minimum bearing life, again under
the most adverse operating conditions, should not be less than two
e ' Chapter 2
t 2
years. Bearing size is t o be determined by the individual manufacturer HEAD AND PRESSURE
and will depend o n the load t o be carried. C
Additional specifications in the Standard include: hydrostatic test e J
T o understand the performance and application of centrifugal
pressure, shaft finish a t rubbing points, and packing space.
C .' pumps, it is essential that we grasp fully the meaning of t h e term
Other dimensional standards are in use in foreign countries o n both
horizontal and vertical pumps. In 1971, the International Organization C; $ "head" and its relationship t o pressure.
for Standardization (ISO) reached agreement on a set of dimensional
standards for horizontal, end-suction, centrifugal pumps. This HEAD
document, I S 0 2858, is in metric units, and describes a series of pumps
of slightly higher capacity range than described in B73.1. I t does n o t Head is a term for expressing pressure. Figure 2.1 shows t h e relation-
include design criteria as to minimum shaft deflection, minimum ship between head and pressure. T h e gauge in t h e pipe near t h e bottom
bearing life, o r other characteristics required t o reduce maintenance. of the tank in Figure 2.1 measures t h e pressure created by t h e weight
The British Standards Institution issued BS4082 in 1966, t o describe of t h e liquid above its centerline. We call the distance from t h e center-
a series of vertical inline centrifugal pumps. While dimensional inter- line of the gauge t o t h e surface of the liquid t h e "static head" above t h e
changeability was the primary reason for this standard, it also includes gauge. T h e relationship between a static head and t h e pressure it
requirements for hydrostatic testing of the pump parts. I t is made up creates is:
of two sections: Part 1 covers pumps wherein the suction and discharge h x sp. gr.
P= Equation 2.1
nozzles are in a horizontal line (the "I" type); and Part 2 covers pumps 2.31
wherein the nozzles are on the same side of the pump and parallel t o
each other (the "U" configuration).
2.31p Equation 2.2
h =
sp. gr.
REFERENCES
where :
1. "Corrosion Data Survey," National Assn. o f Corrosion Engineers,
Houston, 1 9 6 7 . h = head, in feet of liquid
2. Fontana, M. G. and Greene, N . D., "Corrosion Engineering," p = pressure, in pounds per square inch
McGraw-Hill, New York, 1 9 6 7 .
3. Lee, J. A , , "Materials o f Construction for Chemical Process Industries," sp. gr. = specific gravity of the liquid
McGraw-Hill, New York, 1 9 5 0 .
4. "Proceedings, Short Course o n Process Industry Corrosion," National
Assn. o f Corrosion Engineers, Houston, 1 9 6 0 . In Figure 2.1 the pressure is created by a static head of liquid. In
pum? applications we frequently encounter pressure that is not created
by static head. However, regardless of its source, any pressure can be
converted into units of equivalent head using Equation 2.2. (Table 3
in the Appendix t o this Manual contains means for converting other
pressure terms into equivalent heads). T h e conversion of all pressure
terms into units of equivalent head simplifies most pump calculations.
 PRESSURE & HEAD Under these conditions, the pressure gauge in the discharge line will
indicate a lower head than it did in Figure 2.1. This is because, in a
typical pound of liquid passing the gauge in Figure 2.2, some of t h e
----
- -
-----
----
--- --
-----
--7 100 feet
pressure energy, which was indicated as head o n the same gauge in
Figure 2.1 has now been converted into kinetic energy that the liquid
has acquired by virtue of its velocity and some of the pressure energy
-----
----- has been converted into heat by fluid friction in the piping ahead of the
-----
---- - - gauge.
-----
- --
---

FIGURE 2.1
iq - - - - - --
T h e head, or pressure energy, that has been converted into kinetic
energy is called "velocity head". It is not a "loss" of energy, but
instead a change in energy form. In fact, if we cut t h e velocity of the
HEAD AND ENERGY liquid in half, half of the kinetic energy would convert back t o pressure
energy, and the head indicated by the gauge would increase
One of the advantages of using "head, in feet of liquid" t o denote accordingly.
pressure is that it is equal to the foot-pounds of the pressure energy T h e head that is converted into heat by fluid friction is called the
available from each pound o f liquid. For instance, we can say that each "friction head" or "friction loss". From t h e hydraulic standpoint, it is
pound of liquid at gauge level in Figure 2.1 has 100 foot-pounds of a true loss, because heat energy is not hydraulically convertible back
pressure energy due t o the 100 foot head of liquid above it. This con- into pressure energy.
cept of head as an indication of energy, enables us t o consider the Going back t o Figures 2.1 and 2.2, we can calculate the amount of
change of pressure due t o flow, as the conversion of energy. head which was converted into kinetic energy (velocity head), if we
know the velocity of the liquid passing the gauge. The relationship is
VELOCITY HEAD AND FRICTION HEAD as follows:
Figure 2.2 shows an arrangement similar to Figure 2.1 except that v
the liquid is allowed t o flow out of the pipe at the bottom of t h e tank h, = - Equation 2.3
2g
and additional liquid is supplied t o the top of the tank t o keep the
static head constant. where:

h, = velocity head, in feet of liquid

V = velocity, in feet per second
g = acceleration due t o gravity, or 32.2 feet per second,
per second

We can also determine the friction loss but this is a more involved
calculation. I t is described in Chapter 4, "Total Head Required by a
Piping System".

FIGURE 2.2 ?
'
 c 3
e! 3
STATIC PRESSURE J abbreviation "psi" is used. Gauge pressure is generally measured by a
2 bourdon tube gauge or U-tube manometer with one side open t o
Sometimes pressure measured by a gauge, such as those in Figures atmosphere. Gauge pressure is, therefore, relative t o atmospheric
2.1 and 2.2, is called "static pressure". Thls is t o distinguish lt from pressure. T o be more precise, it is ihe amount by which t h e measured
"total pressure" which includes velocity head. T h e velocity head acts .$ pressure is greater than atmospheric pressure.
only in the direction of flow and, therefore, does not register on a For example, when steam pressure is sald t o be "100 psig" ( 1 0 0
gauge which is installed wlth its Inlet perpendicular t o flow. G- 4
pounds per square Inch, gauge) the steam pressure is 1 0 0 ~ o u n d sper
It is important that we note the difference between t h e terms "static square inch greater than atmospheric pressure. If the atmospheric
head" and "static pressure" and apply them properly. For example,
in Figure 2.1, there is a " static head" (difference in elevation) of 1 0 0
> pressure were 14.7 psia at the gauge, then the absolute pressure o f the

feet acting at the centerline of the gauge, and the gauge measures e ;
steam would be 1 0 0 plus 14.7, or 114.7 psia.
Vacuum is a special case of gauge pressure. It, too, is relative t o
a "static pressure" of 1 0 0 feet. On the other hand, in Figure 2.2, atmospheric pressure, but it is the amount by which the measured
the "static head" is still one hundred feet but the gauge measures a pressure is less than atmospheric pressure. In olher words, we could
"static pressure" which is less than 1 0 0 feet because some of the
t- ; say that "vacuum is negative gauge pressure': For example, a reading of
"static head" has been lost due t o fluid friction and some has been 2 "3.0 psi, vacuum" indicates that the pressure is 3.0 pounds per square
converted t o velocity head which does not register on a gauge which & 9 inch below atmospheric pressure or -3.0 PSI, gauge. So that if the
is installed wlth its inlet perpend~cularto flow. atmospheric pressure at the vacuum gauge were 14.7 psia, we could
calculate that 3.0 psi, vacuum is equivalent t o 14.7 minus 3.0, or 11.7
ABSOLUTE PRESSURE, GAUGE PRESSURE, VACUUM, AND k d psia.
DIFFERENTIAL PRESSURE The relationship between absolute, gauge, and vacuum pressures can
e 2
be defined by the following verbal equations.
Another aspect of pressure that we should understand is that all 9
pressures are measured with respect t o some basic pressure and the 3 Equation 2.4
resulting measurement can be identified accordingly. The pressure absolute pressure = atmospheric pressure + gauge pressure
measurement identifications most frequently encountered in pump e 2
applications are: "absolute pressure", "gauge pressure", "vacuum", C and
and "differential pressure".
Absolute pressure is usually expressed in "pounds per square inch,
es: a Equation 2.5
absolute", the abbreviation of which should always be "psia" not "psi". @t: a absolute pressure = atmospheric pressure - vacuum
Absolute pressure measurement is relative t o the complete absence of E j
pressure, as in a perfect vacuum. In other words, absolute pressure is
the amount by which the measured pressure exceeds a perfect vacuum. er 9
For example, when the atmospheric pressure is said t o be "14.7 psia" e= -9 From these equations, it is obvious that each of the above terms
(14.7 pounds per square inch, absolute), it is 14.7 pounds per square express a pressure differential; however, the term "differential pres-
9
inch greater than perfect vacuum. sure" is usually applied t o a difference in pressure between two points
Gauge pressure is commonly expressed in pounds per square inch, C 3 in a system. F o r example, if the pressure is 15 psig o n one side of a
gauge, which is usually abbreviated "psig", although sometimes the c valve and 5 psig on the other, then the differential pressure across the
valve is 15 - 5, or 1 0 psig.
c 5
26 Lt !9 27

e 5
 Chapter 3
CENTRIFUGAL PUMP PRINCIPLES

Absolute, gauge, vacuum, and differential pressure measurements A centrifugal pump adds to the pressure of liquid passing through it
can be converted into equivalent head terms, but it should be t, r by increasing that liquid's velocity. Figure 3.1, with views of a typical
remembered that the equivalent heads are relative t o the same basic centrifugal pump, shows how this happens.
pressures as were the original measurements. C The liquid enters the pump at the suction flange, A. At this point
C g the liquid velocity is essentially the same as in the pipe leading t o the
6 8 pump. From A, the liquid flows into the impeller eye, B, where it is
picked up by the impeller vanes, C. The vanes accelerate the liquid
in the direction of the impeller rotation so that as it leaves the impeller,

\
ROTATION
A- A
CENTRIFUGAL PUMP

FIGURE 3.1
the liquid velocity approaches that of the vane tips. The casing, D,
guides the liquid t o the discharge neck, E, which converts some of the ;lmn-.1z
l1 FLOWSERVE CORPORATION
W O N . OIUO I , SPECIAL
I
velocity energy t o pressure energy by slowing the liquid from the
casing velocity t o the discharge pipe velocity.

TOTAL HEAD

The sum of the gauge pressure head and the velocity head at the
discharge flange minus the sum of the corresponding heads at the
suction flange equals the energy (in foot-pounds) added per pound
of liquid pumped and is called the "total head" developed by the
pump. In equation form:

where:
H = total head
hgd = discharge gauge head
I
h ~ d= discharge velocity head e; g1
c.---II

hgs = suction gauge head

C
Serler
2 F-"-=

h,, = suction velocity head FIGURE 3.2

e 3
CENTRIFUGAL PUMP PERFORMANCE CHARACTERISTICS

The performance characteristics of a centrifugal pump are clearly

'
&; 9

9
The curve marked "bhp" shows the horsepower required by the
pump for any flow rate of the liquid with a specific gravity of 1.0. In
the example cited above, the pump requires 10.0 horsepower t o deliver
500 gpm of water. If liquid with a specific gravity of 1.5 were pumped
defined by its performance curves. Figure 3.2 shows the performance 3 instead of water, the horsepower required would be 1.5 times 10.0 or
curves of a typical centrifugal pump operating at 1 7 5 0 rpm. A similar 15.0 horsepower.
set of curves can be developed for any speed at which the pump is t= The third curve marked "EFF", shows the pump's efficiency, in
operated and, for most pumps, a set of curves is published for each of 3 percent, at any capacity. At 500 gpm, this pump is 63% efficient.
the common induction motor speeds. Pump efficiency is determined mathematically. It is equal t o t h e rate
In Figure 3.2, the curve marked "H" shows the total head developed at which the pump imparts energy to the liquid, divided by the rate a t
at any flow rate. For example, this pump develops a total head of 50 9 which the pump requires energy, the resulting value being multiplied
feet when pumping 500 gallons per minute. c by 100 t o put it in terms of percent. We determine the rate at which
the pump imparts energy, by converting the capacity (Q) from gallons Chapter 4
per minute info pounds per minute, and then multiplying by the total
head (H) t o get ft-lb per minute. The energy rate required by the pump TOTAL HEAD REQUIRED BY A PIPING SYSTEM
is its bhp, which we also convert t o ft-lb per minute. Mathematically,
this derivation is as follows: The basic purpose of any pump is t o move liquid from one point
t o another through a system of piping. Usually this system consists
Imparted E n e r ~ yRate of suction and discharge tanks, a pump, and interconnecting piping
EFF =
Required Energy Rate made u p of pipe, valves, elbows and tees. The energy required t o move
Q x 8.33 x sp. gr. x H a pound of liquid, at t h e desired flow rate, from the suction tank t o t h e
-
bhp x 33,000 x sp. gr. discharge tank is called "the total head of a system" or, more
commonly, "the system head". The pump has t o supply this energy.
Notice that the specific gravity appears in both the numerator In other words, the total head developed by the pump has t o equal
and the denominator of the expression and, therefore, cancels out, the total head required by the system.
t o give us the following equation for efficiency: Normally, the system head is divided into two parts t o simplify its
calculation. The head tending t o move liquid from the liquid level in
the suction tank t o the pump is called the "total suction head". The
HxQ x 100 Equation 3.2
head tending t o prevent flow of the liquid from the pump t o the
= 3960 x bhp at 1.0 sp. gr.
discharge tank is called t h e "total discharge head".
Both of these total heads can be further subdivided into three con-
In our example, we find that: 3 tributing heads. These are: static head, surface pressure and friction
head. They are discussed in the following paragraphs.
500x 50x100
EFF =
3960 x 1 0 STATIC HEAD
= 63%
Two useful characteristics of a centrifugal pump can be observed in The static head component of the total suction head is called the
Figure 3.2. First, that the head produced by the pump is limited t o "static suction head". It is defined as "the height of the liquid surface
that which is developed at shutoff (zero gpm). Therefore, even if there in the suction tank above the centerline of the pump". If t h e liquid
were a complete blockage somewhere in the discharge line of the pump, surface is below the centerline, the static suction head assumes a
the pressure on the equipment between the pump and the blockage negative value. Negative suction head is referred t o as "suction lift".
would n o t exceed the shutoff head of the pump. T h e static head component of the total discharge head is called t h e
Second, the horsepower decreases as the flow is throttled back t o "static discharge head". It can be defined as "the height of the highest
shutoff. This means that, in the event of a blockage, as described liquid surface in the discharge system above the centerline of the
above, the motor driving the pump would not be overloaded. pump". When the open end of the discharge piping is above the liquid
A third characteristic, not apparent in Figure 3.2, is that the level in the discharge tank, the open end is "the highest liquid surface".
centrifugal pump produces a steady flow. This, combined with the When the open end is submerged, the liquid level in the tank becomes
centrifugal pump's low starting torque, makes a very easy load for any the "highest liquid surface". When the "highest liquid surface" is
driving unit.
below the centerline of the pump's suction nozzle, the static discharge FRICTION HEAD
head assumes a negative value.
Figure 4.1 depicts the static suction head and static discharge head Friction head is the energy required (per pound of liquid pumped)
of a simple pumping system. t o overcome friction and turbulence losses that occur as the liquid
flows through the piping system. Suction friction head is the sum of
t h e energy losses occurring in the suction line. Discharge friction head
is the corresponding sum of energy losses in t h e discharge line. T h e
value of these losses varies with pipe length, diameter and internal
finish, with the quantity of flow, and with the type of valves and
fittings.
The Pipe Friction Manual, published by t h e Hydraulic Institute, gives
complete and detailed descriptions of how t o calculate the friction loss
STATIC DISCHARGE in pipe, valves, and fittings. T h e following paragraphs summarize t h e
HEAD
information given in the Pipe Friction Manual.

FRICTION LOSSES IN PIPES

Three methods of calculating the friction loss in pipes are

demonstrated in t h e Pipe Friction Manual, but the first method is the
one most generally used. Tables 1through 31 in t h e Manual list various
diameters of asphalt-dipped cast iron pipe as well as Schedule 40,
wrought iron or steel pipe. Flow rates are given in cubic feet per
minute and gallons per minute with the corresponding friction loss
FIGURE 4.1 in feet per 1 0 0 feet of pipe. When the pipe diameter and flow rate are
known, the friction factor can be read directly from these charts. Note
in Paragraph F - 5 of the Manual, the statement, "The values of the
SURFACE PRESSURE friction head ( h f ) in feet of fluid per 1 0 0 feet of pipe apply t o any
fluid having a kinematic viscosity, v = 0.00001216 square feet per
Pumping systems often begin or terminate in a tank which is under second (1.130 centistokes) which is the value for pure fresh water at'60°
a pressure other than atmospheric. This pressure o n the liquid surface F."
affects the system head. The surface pressure, converted t o feet of A further restriction is given in Paragraph F-7, "No allowance has
liquid, is called either t h e "suction surface pressure" or the "discharge been made for age, differences in diameter resulting from
surface pressure", dependent, of course, upon which surface it acts. manufacturing tolerances, or any abnormal conditions of interior
Surface pressure can be expressed in "feet of liquid, gauge" or "feet surface. Any factor of safety must be estimated from t h e local con-
of liquid, absolute" as described in Chapter 2. ditions and the requirements of each particular installation."
 The second set of tables, Figures 5 through 1 7 in the Manual, show
t s where:
friction loss moduli for the incompressible flow of viscous fluids. These t 0
graphs should be used in calculating the friction loss when the liquid = frictional resistance, in feet of fluid
C i hf
has a viscosity different from that of water. These charts, made up for V = average velocity, in feet per second, in a pipe
each size of steel or wrought iron pipe, show capacity in gallons per G # of corresponding diameter
minute, a range of viscosity values, and the friction modulus for 100 6 k g = acceleration of gravity, 32.2 feet per second,
feet of Schedule 4 0 pipe. They are read by going vertically up from the per second
desired capacity t o the proper viscosity line and then horizontally t o
G d
K = resistance coefficient for the valve or fitting,
the modulus scale on the left. The friction loss in feet of liquid is equal G 3 from the Pipe Friction Manual
to the modulus multiplied by 2.31. The restrictions given for the first
method, i.e., "no allowance for abnormal conditions of the interior
G a
surface" and "no allowance for age, differences in diameter, etc." also C: #
apply to these charts. e Values of V2/2g may be obtained from Tables 1through 3 1 in the
The third method of calculating pipe friction loss, the Darcy Pipe Friction Manual. Resistance coefficients for typical valves, elbows,
equation, is the most vksatile since it is applicable to any fluid (viscous
or non-viscous) in anjr pipe (brass, concrete, wood stave, new used, '
t 3
tees, and other pipe fittings are shown in Tables 32 (a) and 32 (b).
These values are averages and the exact friction loss or K factor
coated, etc.). It is also the most complicated, however, and should be
used only when the other methods do not apply. Use of the Darcy
equation is described and illustrated in Section 7 in the Pipe Friction
Manual. Reynolds numbers, pipe roughness factors, and friction factors
' published by the particular equipment manufacturer should be used if
available.
One other method of calculating the loss in pipe fittings and valves
3 is sometimes used. This method is called "the equivalent pipe length
for use in the Darcy equation are given in Figures 20, 21, 22, 23, and ;)?'method".,,~able 4 in the Appendix to this Manual specifies the lengths
24, in the Manual. This method must be used when the pipe diameters ofitraight pipe which are equivalent in flow resistance t o pipe fittings
are not standard, when the pipe material is other than those described of various types and sizes. The equivalent length of straight pipe for
in the first two methods, or when the surface roughness is other than
that of clean pipe.
e each fitting is added to the actual length of straight pipe and the losses
are then calculated on the basis on one piece of pipe.
At the present time, this method is the only one available for use in
FRICTION HEAD LOSS IN VALVES AND FITTINGS e 3 determining the friction loss in valves and fittings when handling a
viscous liquid. After the equivalent length of straight pipe has been
The loss of head in a valve or fitting may be estimated by the use of selected, the friction loss can be determined by use of the charts
this formula: for viscous liquids, Figures 5 through 1 7 of the Pipe Friction Manual.
t $ Losses incurred in passing through other auxiliary equipment such
Equation 4.1 t , as heat exchangers, filters, strainers, etc., should be obtained from the
equipment manufacturer.
e 3
 TOTAL DISCHARGE HEAD AND TOTAL SUCTION HEAD c' 3 SYSTEM HEAD
C b
The three sources of head discussed in the previous paragraphs
contribute to both the total discharge and total suction heads. These
GP * As stated earlier, the system head (or total head of the system) is
divided into the total discharge head and the total suction head t o
total heads can be calculated using the following equations: simplify the calculation. After determining these contributing heads,
i as shown in the previous section, the system head is evaluated using the
following equation:
hs = k, bps-hfs
+
Equation 4.2 Gf i
where: System head = total discharge head - total suction head

hs = total suction hea.d H = hd-h, Equation 4.4

h,, = suction static head
hps = suction surface pressure
It should be noted that the system head is a differential value;
hf, = suction friction head thus to solve equation 4.4 correctly, both hd and h, must be
expressed in the same units. In other words, both must be "gauge"
and
or both must be "absolute" values.
Equation 4.3 It should also be understood that, because the friction components
of the system head vary with flow rate, the total head calculated by
where: Equation 4.4 occurs at the specified rate of flow for which the friction
heads were determined.
hd = total discharge head
hSd = static discharge head CALCULATION EXAMPLES
hpd = discharge surface pressure
The examples which follow are selected t o illustrate the methods
hfd = discharge friction head for calculating total heads of typical pumping systems. T h e friction
loss values are obtained from the Pipe Friction Manual of the Hydraulic
Institute.
Obviously, the three contributing heads must have been converted
~reviouslyto feet of liquid before Equation 4.2 or Equation 4.3 can be e 3
is done using Equation 2.2 o r the conversions e j
Appendix of this Manual.
'
that the total head calculated by either of the e 3
above equations will be in "feet of liquid, gauge" or "feet of liquid.
- , e 5
absolute" depending o n which term was used t o express the surface
Dressure. e 3
Example 1 2. The suction tank is open, therefore the suction surface
pressure equals atmospheric pressure, or:
hps = 0 feet, gauge
3. The suction friction head is given as:
hfs = 4 feet at rated flow
4. The total suction head using equation 4.2 is:

hs = hss bps - hfs

+

= - 5 + 0 - 4
= -9 feet of liquid, gauge,at rated flow

Note that the total suction head is a gauge value because

the suction surface pressure is a gauge value.
FIGURE 4 2
3' b. Total discharge head calculation
Determine the total head for the above system at rated flow. The 3 1. The static discharge head is:
pump is to transfer 100 gallons per minute of heavy acid, 1.7 specific
gravity, water viscosity, from the suction tank to the discharge tank.
e 3 hsd = 100 feet
The liquid level in the suction tank is 5 feet below the centerline of the C 3 2. The discharge tank is also open to atmospheric pressure,
pump. The friction losses in the suction line are 4 feet at the rated e; 3 thus:
flow. The liquid level in the discharge tank is 100 feet above the
suction nozzle centerline. Friction losses in the discharge line are c 3 hpd = 0 feet, gauge
25 feet at the rated flow. G- 3 3. The discharge friction head is given as:
Solution
e 3 hfd = 25 feet at rated flow
c 3 4. The total discharge head per equation 4.3 is:
Divide the system into two sections with the pump as the dividing e 3
line.
e 3 h d = hsd + hpd + hfd
a. Total suction head calculation e u =
=
100 + o + 25
125 feet of liquid, gauge, at rated flow
1. The static suction head is negative because the liquid c 3
level in the suction tank is below the centerline of the c 3 ~ o t that
e the pump ~n Example 1 has to be pnmed
before ~twill operate See Chapter 16 Also note that, ln
pump. Thus:
hss = -5 feet e; 3 starting, the pump has to fill the dscharge prpe to ~ t s
hlghest level This means that the pump must be capa-
ble of producing a head that 1s at least equal to the
c 3 d~fference~n elevatron between the Ilqurd level ID the
suctlon tank and the hlghest point In thedischarge llne
c 3
c. Total system head calculation
~ i ' a square edge inlet, four feet of pipe, one gate valve, and one

From Equation 4.4 we find:

H = h, - h,
'II
C

C
I @
90° flanged elbow, all of which are 6 inches in diameter. The minimum
level in the vacuum receiver is 4 feet above the pump centerline. T h e
pressure o n top of the liquid in the vacuum receiver is 2 0 inches of
,,,y ".~llllm
= 1 2 5 - (-9)
= 1 3 4 feet of liquid at rated flow
Note that both the total suction head, h, , a n d the total
discharge head, h,, , must be in the same units, in this case 6I 3 Divide the system into two sections with the pump as the dividing
line.
"feet of liquid, gauge", t o solve Equation 4.4 correctly. &1 3
Example 2 a. Total suction head calculation

1. T h e suction side of the system shows a minimum static

head of 4 feet above suction centerline. Therefore, the
static suction head is:
el h,, = 4 feet

--
2. Using the conversion from Table 3 of the Appendix t o
this Manual, the suction surface pressure is:

4. yl
m
1 I
q 3
= -22.8 feet. -gauge
-
I- - 6 ; a 3
-. T- h- -p -q -i ~- -c-f-i-n- nf- -r-i -c-f-i-n-n- head
a - - ---,
h,
--Is ,
~ n i l a l qt h p qilm
---- -- ---
I

contributing friction losses in the suction line. Friction

FIGURE 4.3 c 9 loss in 6 " pipe a t 1000 gpm from Table 15 of the Pipe
Friction Manual, is 6.17 feet per hundred feet of pipe.
c 3
4 x 6.17
Determine the total head for the above system at rated flow. The e 9 In 4 feet of pipe the friction loss = -
100
pump is t o transfer 1 0 0 0 gpm of weak acid, specific gravity 0.9944, e 9
viscosity equal to water, from the vacuum receiver t o a storage tank.
The discharge line rises 50 feet vertically above t h e pump centerline C 3 0.3 feet =

and then runs 4 0 0 feet horizontally. There is o n e 90°flanged elbow 3 Friction loss coefficients (K factors) for t h e inlet, elbow
in this line. All of the piping is 6 inch, Schedule 40, steel pipe. T h e and valve can be added together and multiplied by the
pump takes its suction from the bottom of the vacuum receiver through
c 3 velocity head as follows:
c 5
42
c 9 43

6 9
 Fitting -
K From Table Friction loss in 6" pipe at 1000 gpm from Table 15, is
6" Square edge inlet 0.50 32(a) 6.17 feet per hundred feet of pipe.
6" 9 0 Flanged elbow 0.29 32(a) 450
6" Gate valve 0.11 32(b) In 450 feet of pipe the friction loss = 100 x 6.17
= 27.8 feet
Total coefficient, K = 0
-.90
Friction loss in 6 "elbow:
From Table 32 (a), K = 0.29
v2
From Table 15, the velocity head = - = 1.92 v2
2g From Table 15, -- = 1.92
2g
Friction loss = K
v2 2g
Total friction loss in fittings = K -
2g
= 0.90 x 1.92 = 0.6 feet
= 1.7 feet
The friction loss in the sudden enlargement at the end of
the discharge line is called the exit loss. In systems of this
Total friction loss on suction side is: type, where the area of the discharge tank is very large in
comparison to the area of the discharge pipe, the loss equals
h,, = 0.3 + 1.7 V2/2g, as shown in Table 32 (b).
= 2.0 feet at 1000 gpm v2
Friction loss at exit = -
2g
4. From Equation 4.2 the total suction head becomes: = 1.9 feet

h, = h,, + h P s - hfs
= 4 + (-22.8) - 2.0 The discharge friction head is the sum of the above losses,
that is:
= -20.8 feet, gauge at 1000 gpm hfd = 27.8 + 0.6 + 1.9
= 30.3 feet at 1000 gpm
b. Total discharge head calculation
4. From Equation 4.3 the total discharge head becomes:
1. Static discharge head = hsd = 50 feet
2. Discharge surface pressure = hpd = 0 feet, gauge hd = hsd + hpd + hfd
3. Discharge friction head = hfd = sum of following losses: = 50 + 0 + 30.3
= 80.3 feet, gaugeat 1000 gpm
 -.
c. Total system head calculation
GI
/ "' b. Total discharge head calculation

The total system head is calculated using Equation 4.4 G/ 1. The static discharge head hSd , will change from 5 0 feet
as follows: t o 40 feet, since the highest liquid surface in the discharge
H = h, - h_b
U
GI is now only 40 feet above the pump centerline. (This
= 80.3 - (-20.8) =I value is based o n the assumption that the vertical leg in
the discharge tank is full of liquid and that as this liquid
= 101.1 feet at 1000 gpm 6 1r falls it will tend to pull the liquid up and over the loop
T
Example 3
6 19 in the pipe line. This arrangement is normally called a
siphon leg).
Determine the total head for a system exactly like that of Example
2 except that the discharge line has a vertical leg inside the discharge
tank that extends downward a distance of 10 feet from where it enters 2. The discharge surface pressure is unchanged, that is:
the tank. There is an additional 1 2 feet of piping in this arrangement
and another 90° flanged elbow.
hpd = 0 feet, gauge

3. The friction loss in the discharge pipe will be increased

by the additional 1 2 feet of pipe and the additional elbow.

In 1 2 feet of pipe the friction loss = -

l2 x 6.17
100
20" Hg.
0.7 feet =
The friction loss in the additional elbow = 0.6 feet

4'
Therefore, the discharge friction head increases as
follows:
hfd = 30.3 + 0.7 + 0.6
u = 31.6 feet at 1000 gpm
FIGURE 4.4

Solution 4. By Equation 4.3 the total discharge head becomes:

a. Total suction head calculation hd = hsd + hpd hfd+

= 40 + 0 + 31.6
The total suction head will be the same as in Example 2, = 71.6 feet, gauge at 1000 gpm
that is:
hs = -20.8 feet, gauge
c. Total system head calculation e 3 b. Total discharge head calculation
C r
By Equation 4.4 the system head is: G P 1. The static discharge head, hSd ,will change from 40 feet
t o 38 feet. The static discharge head is measured from the
H = h, - h, 6 r centerline of the pump to the surface of the liquid in the
= 71.6 - (-20.8) 6 i
= 92.4 feet at 1000 gpm
;;;I
-.
r
discharge tank.

\
2. The discharge surface pressure in the open discharge tank
Example 4 & V remains at atmospheric pressure, that is,
6 3 hpd = o feet, gauge
Determine the total hezd for a system exactly like that of
Example 3, except that the vertical leg of the discharge line inside the GI 3 3. The discharge friction loss will be increased by the
tank extends down an additional 5 feet and discharges 3 feet below the C ' i additional 5 feet of pipe.
surface of the liquid. 5 x 6.17
In 5 feet of pipe, the friction loss = -
100
= 0.3 feet

e! 3 The discharge friction head increases as follows:

e8a hfd = 31.6 + 0.3
= 31.9 feet at 1000 gpm

A t --2 4. By Equation 4.3 the total discharge head becomes:

= 69.9 feet, gauge at 1000 gpm

c '3 c. Total head calculation
. e 3 And by Equation 4.4 the system head is:
FIGURE 4.5

Solution = 69.9 - (-20.8)

= 90.7 feet at 1000 gpm
-
a. Total suction head calculation Example 5
c 3
The total suction head will be the same as in the Example
2 and 3 that is: h, = -20.8 feet.
c 3 Determine the total head for the following system at the design flow
rate. The pump is to transfer 100 gpm of a viscous liquid from the
 =e I( Rl
-3
open suction tank to the pressurized discharge tank. The specific =I * 1. Because the minimum level in the suction tank is above
gravity of the liquid is 1.6 and the viscosity is 600 SSU at the pumping
temperature.
The liquid level in the suction tank varies from 4 t o 6 feet above the
pump centerline. The liquid level in the discharge tank varies from 30
t o 36 feet above the centerline of the pump, and the pressure above
=I
=I *
6 I 2.
the pump centerline, the minimum static suction head is
positive, that is:
h,, = + 4 feet
Because the suction tank is open to atmosphere.
the liquid is maintained at 3 0 psig. The pump will be primed initially Cl
v
f I h,, 0 feet, gauge

,
by using liquid from the discharge tank, and it will thereafter maintain =
its prime.
The suction piping is made up of 20 feet of 2 " Schedule 40 steel '
I I 3. The suction friction head is the sum of all the friction
pipe, three 2 " 9 0 ° flanged elbows and a 2 " gate valve. The discharge + 3 losses in the suction system which are found as follows:
line contains 1 0 0 feet of 2 " Schedule 40 steel pipe, two 2 " 900 flanged G I i T h e friction loss in the pipe entrance, elbows, valves,
elbows and a 2" flanged gate valve. c : @ and pipes, will be found by the equivalent length of new
straight pipe method. Equivalent lengths may be found in
' i Table 4 of the Appendix t o this Manual.

Equivalent length of new 2" pipe for:

1-2" reentrant inlet = 8.5 feet
3-2" flanged elbows = 3 x 3.1 = 9.3 feet
1-2" flanged gate valve = 2.6 feet
Total equivalent length = 20.4 feet
Actual length of 2" pipe = 20.0 feet
Total length of 2" pipe (equivalent plus actual) = 40.4 feet

Friction loss Modulus for 1 0 0 feet of 2" pipe at flow

rate of 100 gpm of liquid of 6 0 0 SSU viscosity = 19. See
FIGURE 4.6 e 3 Figure No. 1 0 in the Pipe Friction Manual
e a Loss in feet of liquid = Modulus x 2.31
Solution
e 3 = 1 9 x 2.31
Divide the system into two sections with the pump as the dividing e 3 = 44 feet per 1 0 0 feet of pipe
line. The minimum possible level in the suction tank and the maximum
possible level in the discharge tank must be used to determine the
e ;) The resulting suction friction head is:
head requirements since the pump may have t o operate under these e 3
conditions. e 9 hf,
--40.4 x 44
- 100
a. Total suction head calculation
e .a = 17.8 feet at 100 gpm
t 3
50 e 3
 'I'
= lI' =
e
4. From Equation 4.2 the total suction head is: Friction loss in pipe = 44 x 108.81100

hs = hss hps-hfs
+
=I@ = 47.9 feet

= 4 + 0-17.8 Friction loss a t pipe exit = 1.4 feet

= -13.8 feet, gauge at 100 gpm The discharge head is:

hfd = 47.9 + 1.4

'otal discharge head calculation = 49.3 feet

T h e discharge head is made u p of static head d u e to the By Equation 4.3 t h e total discharge head is:
4.
elevation of the discharge liquid surface, pressure head due t o
the 30 psig pressure on the liquid in the discharge tank, and
t h e friction losses encountered in delivering t h e liquid t o t h e hd = hsd + hpd + hfd
discharge tank. = 36 + 43.3 + 49.3
= 128.6 feet, gauge a t 100 gpm
1. The maximum static discharge head is:
4i 3
hsd = 36 feet 1
& a
c. System head calculation
2. By Equation 2.2, t h e discharge surface pressure is: t ' 3
30 psig x 2.31
I ;$ And by Equation 4.4 the system head is:
-
-
hpd
1.6 el 3 H = hd-hs
= 43.3 feet, gauge el 3
= 128.7 - (-13.8)
3. T h e discharge friction head is found using Table 4 in ' 3 = 142.5 feet
this Manual and Figure No. 10 in the Pipe Friction Manual, e 3
as follows.
b 3
Equivalent length of new 2" pipe for: e 3
2-2" flanged elbows = 2 x 3.1 = 6.2 feet !!$
1-2" flanged gate valve = 2.6 feet C 9
Total equivalent length = 8.8 feet
Actual length of 2" pipe = 100.0 feet e 3
Total length of 2" pipe (equivalent plus actual) = 108.8 feet c -9
e "9
cr 3
c 3
 * rr
c 3
Example 6
c $ absolute pressure (see Equation 2.4) and then converting
t o absolute head by Equation 2.2.
Determine the total head for this system. The pump is t o handle ' I,
300 gpm of a 1.3 specific gravity liquid. The viscosity of the liquid d
120 + 14.7) 2.31
is the same as that of water. The pressure a t the pump suction flange is G Ir +

30 psi, absolute. The discharge pressure is 120 psi, gauge. T h e pump 5 1 d = 243.4 feet, absolute
has a 3 inch suction and a 2 inch discharge. The atmospheric pressure
is 14.7 psia.
51 4
GI I 3
Note the 4 foot head correction t o the pump centerline.

S ' $ b. The discharge velocity head at 300 gpm is found in Table 9

6 I -,
1
of the Pipe Friction Manual t o be:
120 psig ph
c ' +
12.8 feet at 300 gpm
psia
e 3 hvd =

c 3 c. T h e suction gauge reading is in absolute terms so it needs

e 3 only t o be converted t o feet of liquid, absolute using
Equation 2.2.
e a
e 3 hgs = 30 x 2.31
c a
+

1.3
= 54.3 feet, absolute
FIGURE 4.7 G a
e 3 Note the 1 foot head correction t o the pump centerline.
C d
Solution d. T h e suction velocity head at 300 gpm is found in Table 11
e a of the Pipe Friction Manual t o be:
At the beginning of this chapter we stated that "the total head c a
developed by the pump has t o equal the total head required by the 3
system". In this example, we determine the total head required by h~~ = 2.6 feet a t 300 gpm
the system when we calculate the total head developed by the pump e 3
using Equation 3.1. Evaluating Equa e. The total head developed by the pump using Equation
procedure:

a. The discharge gauge head, corrected t o the centerline of H = (hgd hvd) - (hgs hvs)
+ +

the pump,in feet of liquid absolute is found by adding = (243.4 + 12.8) - (54.3 + 2.6)
the atmospheric pressure to the gauge reading t o get = 199.3 feet a t 300 gpm

54
 Solution

Determine the total head and plot the system curve for the following Divide the system into two parts with the pump as the dividing
system. This pump is used t o unload tank trucks of 1.2 specific gravity, line.
water viscosity, liquid. T h e liquid will be pumped into a process line
that is a t 1 0 psi, gauge. T h e desired capacity is 1 6 0 gpm. Priming will a. Total suction head calculation
be accomplished by an auxiliary means.
The suction line is made up of 2 swivel joints, 5-90° flanged elbows, 1. T h e pump must be capable of operating down t o the
1 flanged gate valve, a bell mouth inlet, and 15 feet of pipe. All of this minimum level in t h e tank truck under which condition
piping is 3 inch, Schedule 40, new steel pipe and all of the fittings are the static suction head is:
3 inch . T h e minimum level in the tank truck is 5 feet below the pump h,, = -5 feet
centerline.
T h e discharge piping is made up of 2 flanged gate valves, 3-900 screw- 2. T h e truck tank is open t o atmospheric pressure during
ed elbows, 1 flanged tee (flow is line t o line) and 5 0 feet of pipe. This unloading, therefore the suction surface pressure is:
piping is 2 inch, Schedule 4 0 steel pipe and all of t h e fittings are 2 inch.
bps = 0 feet, gauge
T h e process line is located 1 0 feet above the pump centerline.
ei 2 ".
2 Sinrn frirtinn
~ r r - u ~ v lncc
~ n
1 ~ 0 0xror;nc
v a ~ l c ..r;th
o W ~ L L L tC h
I I Cn fl-...
L ~ V W ..-en ..., ...:I1
LaLC, w e WIII

determine the suction friction loss for each of several

arbitrarily selected capacities, using Table 4 of the Ap-
pendix t o this Manual t o obtain equivalent lengths. T h e
two swivel joints will be treated as four long radius flanged
elbows.
I
lo' C 3 Equivalent length of new 3" pipe for:
l-Bell mouth inlet = 0.7 feet
5-3" flanged elbows = 22.0 feet
1-3" flanged gate valve = 2.8 feet
2-3" swivel joints = 13.6 feet
Total equivalent length = 39.1 feet
Actual length o f 3" pipe = 15.0 feet
FIGURE 4.8 Total length of 3" pipe (equivalent plus actual) = 54.1 feet
 *
G a
I
G v
-,
@ @
T h e friction loss at each capacity is equal t o the friction
'
-,
1
b. The total discharge head calculation
loss per 1 0 0 feet (as given in Table 11, of the Pipe Friction 1 C *
1 The static discharge head is:
Manual) multiplied by the total length of pipe divided by
100.
* -2
* hSd = 1 0 feet
d *
2. The discharge surface pressure is converted t o feet by
Capacity Friction Factor
(feet1100 feet)
Total Pipe Length Suction Friction
(feet)
Loss * a using Equation 2.2 as follows:
gp* hfs (feet)
40 0.44 54.1 0.2
V 3 -
- x 2.31
hpd sp. gr.
80 1.57 54.1 0.8 G 3
120 3.37 54.1 1.8 1 -
- 1 0 x 2.31
160 5.81 54.1 3.1 C * 1.2
200 8.90 54.1 4.8
c a =1 9 . 3 feet, gauge
c 3 3. The discharge friction loss will also be calculated for
4. The total suction head is calculated by solving Equation
3 each of the selected capacities in the same manner as the
4.2 for each of the flow rates selected above.
e 3 suction friction loss calculation.

h,= h,, + bps - hf, Equation 4.2 e+ 3

6 9 Equivalent length of new 2" pipe for:
2-2" flanged gate valves = 2 x 2.6 = 5.2 feet
C $ 3-2" screwed elbows = 3 x 8.5 = 25.5 feet
C -3 1-2" flanged tee = 1.8 feet
6 3 Total equivalent length of 2" pipe = 32.5 feet
Static Suction Suction Total Actual length of 2" pipe = 50.0 feet
Suction
Head
Surface
Pressure
Friction
Loss
Suction
Head
c 3 Total length of 2" pipe (equivalent plus actual) = 82.5 feet
Capacity "ss h ~ s hfs s
G- 3
(feet) (feet, gauge) (feet) (feet, gauge) ez 9
-
(g~m)
In addition t o the above, there is a friction loss at the
40 -5 0 0.2 -5.2 * 9 sudden enlargement where the discharge pipe enters the
80 -5 0 0.8 -5.8
process line. From Table 32 ( b ) of the Pipe Friction
120 -5 0 1.8 -6.8 &
k. 9
160 -5 0 3.1 -8.1 Manual, we see that this loss equals the velocity head
200 -5 0 4.8 -9.8 9 which is found in Table 9 of the Pipe Friction manual.
e 9
e 3
e 3
c; 9
 c. Total system head calculation

1. The total system head is determined for each of the

selected capacities using Equation 4.4.
H = hd - h, Equation 4.4
i ' 3
& ' a Total Total Total
Discharge
Total Pipe Friction Friction Friction Discharge Suction System
Capacity Pipe Friction Head Head Head
g Pm
Length Factor
Loss I n
Pipe
Loss A t
Enlargement
Head
hfS
& a Capacity hd "s H
(feet) (ft.1100 ft.) (feet) (feet)
(feet) I ""%
(gpm) (feet, gauge) (feet, gauge) (feet)
5; r
40 82.5 3.1 2.6 0.2 2.8 40 32.1 -5.2 37.3
80 82.5 11.4 9.4 0.9 10.3 C 1 3 80 39.6 -5.8 45.4
120
160
82.5
82.5
24.7 20.4 2.1 22.5 c : 3 120 51.7 -6.8 58.5
43.0 35.5 3.6 39.1 160 68.4 -8.1 76.5
200 82.5 66.3 54.7 5.7 60.4 200 89.7 -9.8 99.5
t ' 3
e 3
The calculated total system head, H , is plotted against
4. The total discharge head is calculated by solving
Equation 4.3 for each of the above flow rates.

hd = h ~ d + hpd + hfd Equation 4.3

Static Suction
Discharge Surface Friction Discharge
Head Pressure
Capacity hsd h ~ d
(gpmi (feet, gauge) (feet, gauge) (feet) (feet, gauge)
40 10 19.3
80 10 19.3
120 10 19.3
160 10 19.3
200 10 19.3

CAPACITY IN G.P.M.
FIGURE 4.9
 Chapter 5 G j
. AVAILABLE NPSH
NET POSITIVE SUCTION HEAD C; 1

6 8 A more precise definition of available NPSH is "the difference

The one hydraulic characteristic that is least understood by the
people who maintain, or operate centrifugal pumps is the Net Positive ,
- between the total suction head and the vapor pressure of the liquid,
in feet of liquid, a t the suction flange". We can measure the total
Suction Head. Most of the literature available on this subject is com- i r suction head of the pump and we can find vapor pressure from t h e
~ l e xin nature and tends t o confuse those who have had little exper-
; ;
ience in pump work. The purpose of this article is t o explain as simply
as possible the many factors affecting NPSH. Q

r
'
4
liquid temperature. T h e difference between these two values is t h e
available NPSH. T h e following equation is the mathematical expression
of the definition for available NPSH:

NPSH AND CAVITATION

5 a hsv = hsa - hvpa Equation 5.1
\
Perhaps it would be best t o start off by saying that we check the 6 w
NPSH conditions for each pump application t o determine whether c i where:
the liquid t o be pumped will vaporize inside the pump. Vaporization
within a pump is called "cavitation". Cavitation reduces a pump's c v
performance and may damage the pump. (2 2 = available net positive suction head, in feet of liquid
T o understand the occurence of cavitation, it is important t o remem- h,,
ber that a liquid will vaporize a t a comparatively low temperature if c, 3
hSa = total suction head, in feet of liquid, absolute
its pressure is reduced sufficiently. Water, for instance, will vaporize t 3
at 1 0 0 F . if it is exposed t o a vacuum of 2 8 inches of mercury. The h v p a= vapor pressure of liquid a t suction nozzle, in feet o f liquid,
pressure at which a liquid will vaporize is called its "vapor pressure".
e a absol~lte
Table 1 in the Appendix gives the vapor pressure of water between c;. 9
4 0 and 240 O F. c 3
NPSH CONDITIONS c a REQUIRED NPSH
:a
From the preceding paragraph, we see that a reduction in pressure The required NPSH can be defined as "the reduction in total head as
can cause a liquid t o vaporize if it is close t o its vapor pressure. The the liquid enters the pump". Figure 5.1 illustrates how this reduction
pressure on the liquid entering a centrifugal pump is reduced as it ' = a . in head takes place.
moves from the suction flange t o the point a t which it receives energy e In the pump shown in Figure 5.1, there may be an entrance loss at
from the impeller- O b v i o u s l ~ ,we must compare this reduced pressure the ,,,tion flange, A. The suction nozzle, B, even though very short,
j
t o the vapor pressure entering the pump t o determine whether the has a loss due t o friction. T h e turn that the liquid makes a t C in going
liquid will vaporize. This is what we do when we check the NPSH t 3 from axial flow in the suction nozzle t o outward radial flow in t h e
conditions of an application. We call the proximity of the liquid j impeller is accompanied by both friction loss and turbulence. T h e
t o its vapor pressure its "available NPSH" and the pressure reduction leading edges of the vanes act as obstructions in the path of t h e liquid
inside the pump, the "required NPSH". We compare the available NPSH e and an entrance loss occurs a t each vane edge. Any pre-rotation of the
t o the required NPSH. When the available NPSH is equal t o or greater ' 3 liquid as it enters the impeller changes the inlet angle and results in
than the required NPSH the pump will not cavitate. more turbulence. All of these losses occur before the liquid is acted
62 63
 upon positively by the impeller vanes. Once the liquid is in the im-
peller, with the vanes pushing from behind, the pressure starts t o in-
crease and eventually reaches the full discharge head. A theoretical
plot of liquid pressure in going through this section of the pump is
shown in Figure 5.1.
The pump manufacturer determines the required NPSH for each
pump by test procedures, and plots the results on the standard perform-
ance curves for that pump.

NPSH PROBLEMS

If the available NPSH is not greater than that required by the pump,
many serious problems can result. There will be a marked reduction
in head and capacity, or even a complete failure t o operate. Excessive
vibration can occur when sections of the impeller are handling vapor
and the other sections handling liquid. Probably the most serious
problem is pitting and erosion of the pump parts, resulting in reduced
tr 43 life. This is caused by the collapse of vapor bubbles as they pass t o
t 3 the regions of higher pressure. This cavitation phenomenon is usually
accompanied by excessive noise and vibration. As the vapor bubbles
t* 3 collapse, the adjacent walls are subjected t c a tremendous shock from
c 3 the inrush of liquid into the cavity left by the bubble. This shock
c a actually flakes off small bits of metal and the parts take on the appear-
ance of having been badly eroded. This erosion shows up not a t the
c a point of lowest pressure where the bubble is formed, but further
W
a:
3 = 4 downstream where the bubble collapses.
The energy expended in accelerating the liquid t o high velocity in
0)
W
t a
= filling the void left by the bubble is a loss, and causes the drop in head
W
a:
3
a associated with cavitation. The loss in capacity is the result of pumping
a mixture of vapor and liquid instead of liquid. Water, for example,
at 70 O F . increases in volume about 54,000 times when vaporized, and
0 thus even a slight amount of cavitation will reduce the capacity.
z A pump operating with insufficient available NPSH will often pump
J t J spurts of liquid. This is caused by the following chain of events. As
1 I I I I
t - 3 the pump is started, the liquid accelerates in the suction nozzle until
A B C
POINTS ALONG LIQUID PATH
D E
= 3 it reaches the capacity at which it is t o operate. As it accelerates, the
RELATIVE PRESSURES I N THE
ENTRANCE SECTION OF A PUMP
= friction losses increase and lower the absolute pressure until the liquid
3 flashes into vapor. As soon as this happens, the pumping action is
FIGURE 5.1 t 5 reduced, and the flow decreases. With the decreased flow, the losses
e '9
 i
are lower, the absolute pressure is higher, the liquid does not vaporize, hvpa= vapor pressure of the liquid, at the pumping temperature,
rr in feet of liquid, absolute .
and the pump starts to pump again. This increases the flow, reduces
the pressure, etc., until the whole cycle is repeated. This results in an 4
erratic flow rate with spurts of liquid being thrown from the discharge Note that the first three terms in Equation 5.2, equal the total
pipe. suction head, h,, , and if we replace the first three terms with h,, ,
d we get Equation 5.1 which is a mathematical definition of available
CALCULATING AVAILABLE NPSH OF A PIPING SYSTEM @ NPSH.
Each calculation of available NPSH for a piping system requires the
There are five typical pump installations for which the available five following steps:
NPSH should always be calculated. These are: 1 ) when the pump is
installed an appreciable height above the liquid level; 2) when the S t e p 1 : Determine the suction surface pressure, hpsa .
pump takes suction from a tank under vacuum; 3) when the liquid This is the pressure on the surface of the liquid in the suction tank.
has a high vapor pressure; 4) when the suction line is usually long; and When the suction tank is open, the suction surface pressure equals
5) when the pumping system is at an altitude considerably above sea atmospheric pressure. When the suction tank is closed the pressure
level (where the atmospheric pressure is reduced as shown in Table 2 in on the surface of the liquid must be measured. In either case, the
the Appendix). The available NPSH can be calculated by use of the pressure must be converted to feet of liquid, absolute, using Equation
following formula: 2.2 or one of the conversions from Table 3 of the Appendix t o this
Manual.
Equation 5.2 S t e p 2 : Determine the static suction head, h,, .
This is the height, in feet, of the liquid surface in the suction tank
where: above or below the pump centerline. When the liquid level is below the
pump centerline, the static suction head is a negative value.
h,, = available net positive suction head in feet of liquid Step 3: Determine the suction friction head, hf, .
This is the sum of all the friction losses in the suction line from its
hpsa = suction surface pressure,in feet of liquid, absolute, on the inlet t o the suction flange of the pump, at the specified flow rate. In
surface of the liquid from which the pump takes its the following examples, we calculate these losses following the pro-
suction. This will be the atmospheric pressure, in the case cedure described in Chapter 4 . The friction loss factors are from the
of an open tank, or the absolute pressure above the liquid Pipe Friction Manual of the Hydraulic Institute. Pipe friction per 100
in a closed tank. feet of pipe and velocity head are found in Tables 1 through 31 in that
Manual and K factors for pipe fittings are found in Tables 32 (a) and
h,, = static suction head, in feet of liquid. In other words, the 32 (b).
height, in feet, of the liquid surface in the suction tank Step 4 : Determine the vapor pressure, hvpa , of the liquid at the
above or below the pump centerline. (Positive if the liquid pumping temperature, in feet of liquid, absolute. If the liquid pumped
level is above the pump, negative if the liquid level is is water, the vapor pressure for any temperature between 40 O and
below the pump). 240 O F . is found in Table 1 of the Appendix t o this Manual. For
other liquid, the vapor pressure must be known and converted t o feet
hf, = friction head loss, in feet of liquid, between the liquid of liquid, using Equation 2.2 or one of the conversions from Table 3
surface in the suction tank and the suction flange of the of the Appendix.
Pump . Step 5 : Calculate the available NPSH from Equation 5.2 using the
values determined in steps 1 through 4. S t e p 3: Head loss per 100 feet = 17.4 from Table 2-PFM
In the following examples, the abbreviations "PFM" for "Pipe Velocity head = 1.42 from Table 2-PFM
Friction Manual" and "APP" for "Appendix" to this Manual, are used
hf (for the pipe) = 25 x 17.4/100 = 4.4 feet
to denote the source of data.
hf (for the valve) = 0.80 x 1.42 = 1.1 feet
Example 1
hf (for the elbow)= 0.37 x 1.42 = 0.5 feet
Suction lift.
At 100 gpm, h,, = 6.0 feet
S t e p 4 : hvpa = 0.6 feet from Table 1-APP
Step 5: hsv = hpsa + hss - hfs - hvpa Equation 5.2

= 12.3 feet of liquid at 100 gpm

If this pump were relocated at an elevation of 5,000 feet, the only

FIGURE 5.2 term, in equation 5.2, that would change isthe suction surface pressure
hpsa . Table 2-APP shows that the standard atmospheric pressure at
5,000 feet is 12.2 psia.
The above pump is handling 100 gpm of water at 60 O F. The 2.31 p Equation 2.2
Step 1 : h =
suction lift is 1 5 feet. The suction line consists of 25 feet of 2 inch sp. gr.
steel pipe with one 90 O flanged elbow and one 2 inch foot valve. The
barometric (atmospheric) pressure is 29.96 inches of mercury.
The available NPSH at 100 gpm is found as follows:
= 28.2 feet, absolute
inches of mercury x 1.13
Step 1 : h = from Table 3-APP
sp. gr.
S t e p 2: (As above)
Step 3: (As above)

= 33.9 feet, absolute

e. 3 Step 4 : (As above)
s t e p 5: hsv = 28.2 + (-15.0) - 6.0 - 0.6
t.. 3 = 6.6 feet at 100 gpm
S t e p 2: hss = -15.0 feet
e 3
Example 2 S t e p 4: h =
2.31~ Equation 2.2
sp. gr.
Suction lift and high specific gravity.

= 0.3 feet, absolute

S t e p 5: hsv = hpsa + hss - hfs - 'vpa Equation 5.2
= 19.4 + (-5.0) - 3.3 - 0.3
= 10.8 feet a t 100 gpm
Example 3

FIGURE 5.3 Suction lift and high vapor pressure.

The above pump is handling 100 gpm of acid with a specific gravity
of 1.7. The suction lift is 5 feet. The 2 inch suction line consists of 10
feet of pipe with one 90 O flanged elbow and one 2 inch foot valve.
The barometric pressure is 29.12 inches of mercury. The vapor
pressure of the acid at pumping temperature is 0.20 psia.
The available NPSH a t 100 gpm is found as follows:

S t e p 1: h =
inches of mercury
from Table 3-APP FIGURE 5.4
sp. gr.

The above pump is handling 500 gpm of gasoline with a specific

= 19.4 feet, absolute gravity of 0.73. The suction lift is 5 feet.
The vapor pressure a t pumping temperature is 11.5 psia. The suction
line consists of 10 feet of pipe, one 90 O flanged elbow and one flanged
S t e p 2: h,, = -5.0 feet gate valve. The atmospheric pressure is 14.7 psia.
S t e p 3: Head loss per 100 feet 17.4 from Table 2-PFM
= The available NPSH a t 500 gpm is found as follows:
Velocity head = 1.42 from Table 2-PFM
2.31 p Equation 2.2
S t e p 1: h =
hf (for the pipe) = 10 x 17.4/100 = 1.7 feet sp. gr.
hf (for the valve) = 0.80 x 1.42 = 1.1 feet
hf (for the elbow) = 0.37 x 1.42 = 0.5 feet

At 100 gpm, hf, = 3.3 feet = 46.5 feet, absolute

 Step 2: h,, = -5.0 feet
6 i
Example 4
Step 3: Head loss per 100 feet = 13 feet from Table 13-PFM G r
Velocity Head = 2.47 feet from Table 13-PFM i 8 Flooded suction and low suction surface pressure.
hf (for the pipe) = 1 0 x 13/100 = 1.3 feet
hf (for the valve) = 0.16 x 2.47 = 0.4 feet
hf (for the elbow)= 0.31 x 2.47 = 0.8 feet
hf (for the pipe inlet) = 0.05 x 2.47 =0.1 feet
At 500 gpm, hf,= 2.6 feet

Step 4 : h 2.31 p FIGURE 5.5

= Equation 2.2
sp. gr.
The above pump handles 1000 gpm of water at 100 O F. The pres-
sure on top of the liquid in the suction tank is 20 inches of mercury,
vacuum. The 6 inch suction line consists of 4 feet of pipe, one 9 0 O
= 36.4 feet, absolute
flanged elbow and one flanged gate valve. The barometric pressure
is 29.5 inches of mercury. The specific gravity and vapor pressure of
water at 100 O F. are given as 0.994 and 2.2 feet, absolute in Table
t 4
1-APP.
Equation 5.2 c -9
Step 1 : Equation 2.5 states:
= 46.5 + (-5.0) - 2.6 - 36.4
absolute pressure = atmospheric pressure - vacuum
= 2.5 feet at 500 gpm
h,,, = 29.5 - 20
In this example, high vapor pressure has produced a very low avail- e "a = 9.5 inches of mercury, absolute
able NPSH. A pump with an NPSH requirement equal to or less than !$ ") inches of mercury x 1.13 from Table 3-APP
the above, would be hard to find, since 2.47 feet of the available NPSH h =
sp. gr.
is in the form of velocity head. This leaves almost no head to accomo- Cr 3
date losses in the entrance to the pump. e 9 hspa = 9.5 x 1.13
0.994
= 10.8 feet. absolute

e 3
S t e p 2: h,, = + 4.0 feet
e ,:.
 Step 3: Head loss per 100 feet - u.1' ,I - '--
L Z U I-- l a u ~ el'a -""""
I I m-L'- rrlvl
5' 3 ~ h b_b _ = +5.0 feet
S t e 2:
17nl,.n~t..

hr
hi (for the
I,,-.,.A
c l V L I b Y llcall

(for the pipe)

- -
valve)
= 1 0'3 f...
- I.JA L L

=
=
OTable
~ 15-PFM
4 x 6.171100
0.11 x 1.92
=
=
0.2 feet
0.2 feet
6
i
ol
r
0
Step 3: Head loss per 100 feet = 32.4 feet from Table 13-PFM

hf
Velocity head =

(for the pipe)

6.32 feet from Table 13-PFM
= 1 0 x 3241100
n - n r n n
= 3.2 feet
-4
I
1.1 ~ e e t
0 ,

hf (for the valve) = V.I~ x o . 3 ~ =

hf (for the elbow) = 0.29 x 1.92 = 0.6 feet
hf (for the pipe inlet) = 0.5 x 1.92 = 1.0 feet ~3 hf (for the elbow) = 0.31 x 6.32 = 2.0 feet
r 3 hf (for the pipe inlet) = 0.5 x 6.32 = 3.2 feet
At 1000 gpm, hf, = 2.0 feet
@ 9 A4 "..-
~ n n I, = a K fnnt
Step 4: hvpa= 2.2 feet, absolute from Table 1-APP r 3
Step 5: h,, = hpsa + hss - hfs - hvpa Equation 5.2
6 i
= 10.6 feet at 1000 gpm
Example 5 Step 4: h v p a = hss = 11.2 feet absolute
c 3 Step 5: h,, = hpsa + hss - hfs - hvpa Equation 5.2
Flooded suction and boiling liquid.
e a = 11.2 + 5.0 - 9.5 - 11.2
3 = - ~ ~ f ~ ~ t ~ t n n n ~ n m

I .c -3 Two important observations can be made concerning the solution of

the above equation. First, that when pumping a boiling liquid, the
available NPSH equals the static suction head minus the suction friction
head (h,, - hfs) because the suction surface pressure and the vapor
FIGURE 5.6 pressure equalize one another. Secondly, we should recognize that a
negative NPSH is a physical impossibility because it indicates that the
friction losses exceed the head which is available to overcome them.
It is required that the above pump handle 800 gpm of water at f ;) From these observations, we can conclude that when handling a
160°F. The liquid in the suction tank is at its boiling point,that is t o boiling liquid, the static head must exceed the suction friction head by
say that the pressure on the liquid equals the vapor of the liquid. The
3 the amount of the NPSH required. If the pump in the above example
4 inch suction line consists of 1 0 feet of pipe with one 90 O flanged c 3 required 3 feet NPSH, we would have to raise the liquid level 7.5 feet
elbow and one flanged gate valve. : 3 (increasing the static suction head t o 12.5 feet) or reduce the friction
losses t o 2 feet to provide 3 feet of NPSH available to the pump.
r l
Step 1: hpsa = hvpa= 11.2 feet, absolute from Table 1-APP
= 3
 -.
Example 6 '6 /I 'r hf
hf
(for the pipe)
(for the valve)
=
=
(L + 12) x 8.91100 =
0.23 x 1.17 -
-
0.09L + 1.1feet
0.3 feet
Maximum suction lift. hf (for the elbow) = 0.33 x 1.17 - 0.4 feet
;I 6. hf (for the pipe inlet)= 0.05 x 1.17 - 0.1 feet
C l r
* I $ At 200 gpm, hf, = 0.09L + 1.9 feet
u
- hss I 4 1 4

2.31 p Equation 2.2

Step4:h =
sp. gr.
I
FIGURE 5.7

The above pump is handling 200 gpm of acid with a specific gravity
of 1.3, and a vapor pressure of 0.50 psia. The barometric pressure is = 0.9 feet, absolute
760 millimeters of merdury. The 3 inch suction line consists of 12 feet
of horizontal pipe, one 90 O flanged elbow, one flanged gate valve and
a vertical pipe whose length, L, has not yet been determined. A min-
imum suction inlet submergence of 2 feet is required. The pump
requires 10.0 feet of NPSH at 200 gpm.
Determine the maximum suction lift that this pump will support Step 5: Available NPSH, hSv , will equal the required NPSH (10.0
without cavitating. feet) when the suction lift is maximum.
hsv = hpsa h,, - hfs - hvpa
+
Equation 5.2
millimeters of mercury
Step 1: h = from Table 3-APP
22.4 x sp. gr. 10 = 26.1 + (2 - L ) - (0.09L + 1.9) - 0.9

= 26.1 feet, absolute 1.09L = 15.3 feet

L = 14.0 feet

Therefore, the maximum suction lift at 200 gpm is:

Step 2: h,, = - (L -2)
2-L
= h,, = - (L - 2)

Step 3: Head loss per 100 feet = 8.9 feet from Table 11-PFM = -(14-2)
Velocity head = 1.17 feet from Table 11-PFM = -12 feet
 Step 3: hvs = 1.4 feet at 100 gpm from Table 9-PFM
Step 4 :
i
5 1 3
Example 8

2.31 p Determine the NPSH of a 2 x 1%pump when handling 200 gpm of

h Equation 2.2
' I1 '
Q ~ Z water a t 90 'F. The barometer reading a t the test site is 29.0
=
sp. gr. clear
inches of mercury. The lowest suction gauge pressure that could be
hvpa = 2.31 x 0.20 obtained when the total head developed by the pump was within 3
1.7 v 3 percent of its normal total head, was 17.7 inches of mercury, vacuum.
i ~ d
=

Step 5 : hsv = hgs

0.3 feet, absolute
+ ha + hvs - h p a Equation 5.3
6
6
i1 J
step1:h =
inches of mercury x 1.13 from Table 3-APP
sp. gr.

= -9.8 + 19.3 + 1.4 - 0.3

= 10.6 feet at 100 gpm
= -20.1 feet, gauge

t ' d
inches of mercury
t cS Step 2: h = from Table 3-APP
TESTING FOR REQUIRED NPSH sp. gr.

T o test a pump for its NPSH requirement, we gradually reduce

the total suction head until the liquid just starts t o vaporize in the
impeller, causing a drop in the total head developed by the pump. A = 32.9 feet. absolute
three percent drop in total head is generally considered t o indicate in-
cipient cavitation. At this point, we record the suction gauge pres-
sure, the flow rate, the pumping temperature, the barometric pressure c 3
and the pump rpm. This information enables us t o solve equation 5.3 c 3 Step 3: h,,. - = 5.7 feet at 200 gpm from Table 9-PFM
t o determine the available NPSH. Step 4 : h v p a = 1.6 feet, absolute from Table 1-APP
When a pump is a t the point of hcipient cavitation, the NPSH avail- Step 5 : h,, = hgs + ha + hv, - 'vpa Equation 5.3
able at its suction flange has been reduced t o the degree that it just e 3
equals the NPSH required by the pump. In other words, solving equa-
tion 5.3 for this condition gives us both the NPSH available a i d t h e
NPSH reauired. e a = 16.9 feet
The example which follows illustrates the determination of one point e 3
o n a typical NPSH curve. A series of these points must be plotted e 1)
against capacity in order to describe the NPSH required by the pump at This value is then plotted as the required NPSH of t h e pump a t the
all capacities. 200 gpm capacity point.
c 1)
 Chapter 6
Equation 6.1
FLOW OF LIQUIDS IN PIPES
The Pipe Friction Manual, published by the Hydraulic Institute,
contains complete tables showing friction losses for most sizes of pipe.
I 6
These tables should be used whenever possible, but it may sometimes
be necessary t o estimate new flow conditions when system changes = 56 feet
are made. There are four formulae which can be used t o approxi
the effect of such changes on flow rates and pressures.
The first two deal with changes in the rate of flow in an existing
pipe line. This situation is often encountered in a customer's
when it becomes necessary t o change the capacity being
the system. Assuming that the pipe size is to remain unchanged, THE
FRICTION LOSS WILL VARY AS THE SQUARE O F THE CAPAC-
ITY RATIO. If the capacity is doubled the friction losses will quad- C ' 3
ruple.

-- - -
hf2 = hfl ($)2 Equation 6.1 ----- -
-
when:

D2 = D1 C 3 FIGURE 6 1

C 9
where: C
Condition
7 Pipe L i n e @
F
J' Gauge A Gauge B Loss
I As shown 300
hfl = original friction head loss c 3 2 Unchanged 500
hf2 = new friction head loss e . 3
The second relationship is a corollary t o the first. I t states that THE
e 3
Q1
Q2
Dl
=
=
=
original flow rate
new flow rate
original pipe size
= 3
CAPACITY VARIES AS THE SQUARE ROOT O F THE HEAD ON
THE LIQUID (again assuming the pipe size is unchanged). If the head
is doubled, the capacity will increase 1.41 times. If it is tripled, the
%3 capacity will Increase 1.73 tlrnes.
D2 = new pipe size
e s
As an example, assume 300 gpm flowing through a pipe line will suf- -9 Equation 6.2
fer 2 0 feet of friction head loss. Then 500 gpm through the same line
will lose:
t 9
e, 3
82 C 3
 when: The third and fourth rules apply to situations where the size of the
pipe line is to be changed. The pipe diameters may be increased t o
D2 = Dl lower the friction losses so that a smaller pump may be used or so that
more NPSH will be available; or it could be decreased so that the
where: friction losses will limit the capacity of a pump. The third rule applies
when the capacity is to remain the same, and states THE FRICTION
H~ = original head LOSS IS INVERSELY PROPORTIONAL TO THE FIFTH POWER O F
H2 = new head THE PIPE DIAMETER RATIO. In other words, the losses will increase
very quickly when the pipe diameter is decreased. If the pipe diameter
is reduced to lh of its former size, the friction losses will increase t o 3 2
For example, if a 160 foot head would deliver 300 gpm through a times their former value.
specified pipe line, a 100 foot head would deliver:
Equation 6.3
Equation 6.2
when:

= 237 gpm as shown in Figure 6.2 Let us assume a 3" diameter pipe can handle 300 gpm with a 20 foot
friction loss. The same flow rate through a 2" diameter pipe would
create :

Equation 6.3

= 152 foot loss

The same flow through a 4" line would create:

FIGURE 6.2

Condition Pipe Line Gauge A GPM f 3 \ I

= 5 foot loss
1 As shown 160 30 0 ",
2 Unchanged 100 237 ." .
L
This is shown in Figure 6.3
 Equation 6.4

= 109 gpm

and a 4" diameter pipe would deliver:

FIGURE 6.3

Condition GPM Pipe Dia. Gauge A Gauge B Loss

1 300 3 160 140 20
2 300 2 160 8 152 = 616 gpm
3 300 4 160 155 5
The final flow relationship applies when the head is constant and the
pipe diameter is to be changed. The head of water behind a dam, or
from a large storage tank, could be assumed to be constant. Under
these conditions, THE CAPACITY OF A PIPE WOULD VARY AS
THE 2.5 POWER OF THE DIAMETER RATIO. This seems logical,
since we found in rule three that the friction values increase and de-
crease very rapidly with changes in pipe diameter and in this case the
friction values are the only factor controlling the flow. Doubling
the pipe diameter will increase the flow 5.66 times its original value.
Tripling the pipe diameter would result in 15.6 times as much capacity.
FIGURE 6.4

Equation 6.4 Condition Pipe Dia. GPM

t .
l
1 160 3 300
when: e a 2 160 2 109
k. 3 3 160 4 616
H, = HI Notice that these characteristics may be used in sequence to deter-
k- 3
hf2 = hfl
As shown in Figure 6.4, assume that a 3" diameter discharge pipe
'a
C. j
mine all of the flow conditions. In rule two, for instance, after the new
capacity has been determined, the new friction loss value may be
obtained by using rule one.
delivers 300 gpm under a specified head. Under the same head, a 2" Note that these rules are based on the assumption that the liquid
pipe will then deliver:
e. 3 being handled has the same viscosity as water.
C 3
 1.The capacity is directly proportional t o the ratio of impeller
Chapter 7 diameters, that is:
THE AFFINITY LAWS Equation 7.4
PERTAINING TO CENTRIFUGAL PUMPS

Sometimes it is necessary to estimate the performance of a cent- where: D = impeller diameter

rifugal pump whose impeller diameter or speed may not be shown on a
standard curve. The approximate curves for a new impeller diameter or 2. The head is directly proportional t o the square of the ratio of
speed can be determined by means of the affinity.laws. impeller diameters, that is:
These laws state that when comparing the performance of a pump
at two different rotative speeds:

1.The capacity is directly proportional to the ratio of impeller

H2 = ,):(
Hl Equation 7.5

speeds, that is:

3. The horsepower is directly proportional t o the cube of the ratio
Equation 7.1 of impeller diameters, that is:

where: Q = capacity Equation 7.6

N = rpm
2. The head is directly proportional to the square of the ratio im- The laws relating t o impeller diameters are not as accurate as those
peller speeds, that is: relating to rotative speed. In order for them t o be equally accurate
Equation 7.2 it would be necessary for all the dimensions of the pump t o be changed
in proportion t o the impeller change. In most practical applications,
the impeller is trimmed to a smaller diameter, leaving t h e remaining
where: H = head dimensions unaffected. However, the results are sufficiently accurate
t o allow reasonable predictions of performance t o be made.
3. The horsepower is directly proportional to the cube of the ratio The affinity laws are used t o estimate new performance curves and
of impeller speeds, that is: t o correct test data sheets t o standard speed. The following examples
3 will illustrate their use:
BHP, = BHP (%) Equation 7.3
Nl
CHANGING IMPELLER DIAMETERS
where: BHP = horsepower
Assume the performance of a 3500 rpm pump with a 9 "diameter
The laws can also be restated as follows t o compare the performance impeller Dl is known (See Figure 7.1), and that it is desired t o find
of the two different impeller diameters in the same pump at the same the performance of an 8%'' diameter impeller (D,) at the same speed.
speed : This can be calculated from the 9" diameter curve, as shown in the
following steps:
FIGURE 7.1
 r* 9
--
r 3
-.
4. The horsepower of a pump varies as the cube of the ratio of * 3 9 " DIAMETER CURVE @ 3500 RPM
the impeller diameters. Two readings are necessary to obtain 3 BHP EFF. NPSH
G B HEAD
the cube ratio on a slide rule. Since the horsepower (BHPl) of 240 287 35.5 49.0 28.0
the 9" diameter impeller at 240 gpm is 35.5, set the indicator at r 2
35.5 on the "B" scale and read 31.6 under the hairline on the 3
"A" scale. Now set the indicator at 31.6 on the "C" scale and Q 3
read the new horsepower (BHP2) under the hairline on the "D"
scale which is 29.9 hp. Q d
6 a
D2
or BHP2 = BHP~(,) Equation 7.6 -
@ 3
6
8% " DIAMETER CURVE @ 3500 RPM
3 C J
= 35.5g) c d GPM HEAD BHP EFF. NPSH

= 29.9 hp 189 282 27.5

G &f lr l nnn nc -r)

5. The net positive suction head (NPSH) varies as the square of the
ratio of the impeller diameters in the same manner as the head
varies.

NPSH, = NPsH~ ($1 2

Equation 7.5
(Modified) Please note that the efficiency remains constant as the capacity,
head, and horsepower are stepped down. For example, on the 9" curve
the efficiency is 45% at 160 gpm and remains 45% at the new point on
the 8lh" curve. This can be checked by calculating the efficiency
using Equation 3.2:
= 25.0 feet QxHx100
EFF. =
3960 x bhp at 1.0 sp. gr.
Repeating the above steps for a numberof points taken away from
the original (9") curve will produce an equal number of points from
which the stepped down (8%") curve can be plotted. This procedure
is facilitated by making a tabulation as shown below:
 _- I
CHANGING ROTATIVE SPEED b ' r) 4. The horsepower of a pump varies as the cube of the ratio of im-
9' 3 peller speeds. Two readings are necessary t o obtain the cube ratio
Pump performance at a speed not shown on the standard curve can on the slide rule. Since the horsepower at 3500 rpm and 240 gpm
be estimated in the same manner. For example, to obtain a 9 diameter ' 3
is 35.5, set the hairline indicator at 35.5 on the "B" scale and
curve at 2900 rpm N2 from a 9" diameter curve at 3500 rpm (N1): ' 3 read 24.4 on the "A" scale. Now set the hairline indicator at 24.4
4. 3 on the "C" scale, read the new horsepower under the hairline of
the "D" scale which is 20.2 hp.
1.Set 3500 on the "C" scale of the slide rule over 2900 on the "D" k 3
scale. Once this has been done it will be unnecessary t o move the
slide throughout the rest of the operations.
Equation 7.3
2. Since capacity varies as the direct ratio of impeller speeds, set the
hairline indicator at a known capacity, say 240 gpm, on the "C"
scale and read the new capacity on the "D" scale.
The new capacity is 199 gpm.

Equation 7.1

5. The net positive suction head varies as the square of the ratio
of impeller speeds in the same manner as the head varies.

= 199 gpm

3. The head of a pump varies as the square of the ratio of the im-
NPSH2 = NPSHI (2) 2

Equation
(Modified)7.2
peller speeds. Since the head on 3500 rpm curve at 240 gpm is
287 feet, place the hairline indicator at 287 on the "B" scale and
read the new head on the "A" scale. The new head is 197 feet.

or H2 = HI (?) 2
Equation 7.2
= 19.2 feet

A tabulation can be made, as illustrated in the step down of di-

ameters, t o simplify the procedure.

= 197 feet
 I
4- I .) True o r Newtonian fluids are unaffected by the magnitude and kind of
Chapter 8 motion t o which they are subjected. Mineral oil and water are typical
wl2
VISCOSITY examples of this type of liquid.
@ l J
CHARACTERISTICS OF VISCOUS LIQUIDS @ 1 33
1
Dilatant fluids are those in which the viscosity increases as agitation is
increased. Liquids of this type can become practically solid within a
The viscosity of a liquid is defined as a measure of its resistance t o pump or a pipe line. Examples are candy compounds, clay slurries, and
'
I

flow. It may be considered as the internal friction resulting when one 3 similar heavily filled liquids.
layer of a fluid is made t o move in relationship t o another layer. The
so called "thick" liquids, such as automobile transmission lubricants,
molasses, varnish, etc., have high viscosities. "Thin" liquids, such as
water, gasoline and hydrochloric acid, have low viscosities. The per-
' '
6
J

3
Plastic fluids have a definite "yield value" which must be exceeded
before flow will start, after which the viscosity decreases with an
increase in agitation. Tomato catsup is a good example of this type of
formance of a centrifugal pump is adversely affected when the liquid I$ 3 material. Once the "yield value" has been exceeded the catsup will
being handled has a viscosity higher than that of water. Since flow is gush from t h e bottle.
associa'ted with shearing forces between sections of a liquid, more 6 3
energy is required t o pioduce flow in a viscous liquid than in a non- 6 3 Psuedo-plastic fluids also show decreasing viscosity with increasing
viscous liquid. This added energy is evidenced by an increase in horse- agitation, but d o not have a yield value. Many emulsions fall into this
; category.
power and a reduction in head, capacity, and efficiency of the pumping
unit. C d
Liquids that have high specific gravities are not necessarily highly Thixotropic fluids generally exhibit a decreasing viscosity with in-
viscous. Mercury, for instance, with its very high specific gravity (13.7)
has a viscosity of only 0.118 centistokes, whereas transmission lu-
bricants that are lighter than water, have viscosities of more than 500
'a
2 ;.)
creasing agitation, although the viscosity at any particular rate of
motion may depend upon the previous agitation of t h e liquid.
Examples are asphalts, glues, molasses, cellulose compounds, greases,
centistokes. These t w o properties of liquids are, therefore, entirely paints, soaps, starches, and tars.
independent of each other.
a
The viscosity of a liquid varies appreciably with changes in temper- :a Viscosity can be expressed in "absolute" or "kinematic" terms. T h e
ature, but very little with changes in pressure. A common example of 3 basic unit of absolute viscosity is the poise. T h e common unit for
this is maple syrup. When chilled, it is very stiff and hard t o pour. A expressing absolute viscosity is the centipoise, which is 1 / 1 0 0 of a
slight warming thins it, however, allowing it t o flow quite freely. For
this reason, the viscosity number must always be qualified by giving
r- poise. Water a t 68.4' F, has an absolute viscosity of 1.0 centipoise.
The basic unit of kinematic viscosity is the stoke. The common unit
the temperature a t which this viscosity was determined. In selecting
a centrifugal pump, it is imperative that the viscosity value used be in
= a for expressing kinematic viscosity is the centistoke, which is 1 / 1 0 0 of a
stoke. Kinematic viscosity is related t o absolute viscosity as follows:
agreement with the pumping temperature of the liquid. Inaccuracies in a
this respect could result in the selection of the wrong size pump and g Equation 8.1
motor.
*; 3
There are a five basic types of liquid as defined by their behavior Kinematic Viscosity - Absolute Viscosity (centipoise)
during agitation at a constant temperature. = a (centistokes) sp. gr.
t 3
96 t 3 97
Since the specific gravity of water at 68.4' F is very close to 1.0, it friction forces that will be encountered in the pump. This is especially
follows, from Equation 8.1, that the kinematic viscosity of water at true when testing solid-liquid mixtures. The solids tend to clog the
68.4' F is for all practical purposes, 1.0 centistokes. orifice of other types of viscosimeters, resulting in high viscosity
reading even though the liquid phase of the slurry may be non-viscous.
MEASURING VISCOSITY Tables 36 (a) through (f) of the Pipe Friction Manual list the vis-
cosities of many common liquids at several temperatures. Notice the
The viscosity of a liquid is determined by an instrument called a wide variation in viscosity possible with comparatively small variation
viscosimeter. There is an almost endless variety of these instruments, in temperature.
most of which are designed for specific liquids, specific viscosity ranges, The friction loss in pipe, valves, and fittings increases with viscosity,
or other specific purposes. The Saybolt viscosimeters are used in the as illustrated in Chapter 4.
United States more than any other type. Liquids of low to medium
viscosity are measured by the Saybolt Universal Viscosimeter, and PUMP PERFORMANCE WHEN HANDLING VISCOUS LIQUIDS
liquids of high viscosity are measured by the Saybolt Furol
Viscosimeter. A measured volume of liquid is allowed t o flow through The hydraulic losses that occur in a pump are due t o the viscosity
an orifice of specified proportions and the time of efflux noted. This of the liquid being handled. The standard performance curve is valid
time in seconds, is then called the SSU number (Seconds Saybolt only when the pump is handling liquid with a viscosity close t o or less
Universal) or SSF numbers (Seconds Saybolt Furol). These numbers than 1.0 centistoke.
have been so widely accepted that they are often used in place of, or in When handling a liquid with an appreciably higher viscosity, the
addition to, the actual viscosity in centistokes. Most of the charts pump will not perform as shown on the standard curve. The head,
published by the Hydraulic Institute show viscosities in both scales. capacity, and efficiency will be less, while the horsepower will increase.
The Irany, Zahn and Redwood viscosimeters operate on the same Unfortunately, there is no acceptable analytic method of establishing
principle. Readings from these instruments can be compared t o each the pump performance when the liquid has a viscosity other than that
other by means of conversion factors or charts such as shown on pages of water. A large number of experiments have been conducted as a
74 and 75 of the Pipe Friction Manual published by the Hydraulic means of developing an empirical method of predicting this perform-
Institute. ance. The data from these tests have been formulated into charts or
Other viscosimeters such as the Ostwald, Bingham, and Ubbelohde nomographs so that the pump performance can be estimated for liquids
operate in much the same manner, except that a capillary tube is used of practically any normal viscosity.
in place of the orifice. These instruments are used primarily on liquids
of low viscosity. VISCOSITY CORRECTION NOMOGRAPH
The Brookfield viscosimeter is the rotational type wherein a disc is
rotated in the liquid to be tested. The torque or drag imposed on the Figure 8.1 provides a means of estimating the viscous performance of
disc is measured by a calibrated spring and indicated on the dial directly a centrifugal pump when its water performance is known. The em-
in centipoise. Viscosimeters of this type are particularly useful in pirical correction factor values contained in this chart encompass the
measuring the viscosity of non-Newtonian liquids, since they will results of a great number of tests of variously sized pumps, operated
indicate whether the liquid is dilatant, plastic, psuedo-plastic or thixo- over a wide range of conditions. While experience has shown the curves
tropic. The rotating disc has approximately the same friction factor to be conservative, the viscous performance estimate that they provide,
acting upon it as does an impeller. For this reason, an instrument is only approximate. If exact viscous performance data are required,
of this type gives the most reliable viscosity reading in respect t o the performance tests should be made using the viscous liquid in question.
 Figure 8.1 is t o be used only for standard centrifugal pumps, not
for mixed or axial flow pumps. I t 'is limited t o Newtonian liquids.
Slurries, paper stock, dilatant or thixotropic liquids may produce
widely varying results. In addition, the chart is only t o be used where
there is adequate NPSH available because the exact effect of viscosity
on NPSH required is not predictable.
The following symbols and equations are employed in the applic-
ation of Figure 8.1 t o pump viscosity problems:

Qvis = CQ x Qw Equation 8.2

H i = ( 1 - CH x Qw/[Qw ] ) x Hw Equation 8 . 3
BHPvis = CH x sp gr x BHP, Equation 8.4

Qvis = viscous capacity, in gpm 6- J

HVIS 1 viscous total head, in feet c 3
B H P ~ ,= viscous brake horsepower 9 50- 0
0 o 0
o 0
o o o o o
T:
a~

0 zI l l 0
UI
0
N ?
0
UI
0
N
0
7 W m w
cf
= water (non-viscous) capacity, in gpm l 1 l l 1 1 1 1 1 1 I l 1 l 1
Qw I l l 1 1 I I 1111 l l l l l l l l l l l 3
2
!LL
[Q, ] = water capacity at best efficiency, in gpm
Hw = water total head, in feet G 9
BHP, = water brake horsepower
CQ = viscosity correction factor for capacity
c d
0
c' 9 0
CH = viscosity correction factor for total head
CHP = viscosity correction factor for horsepower c' 3
613
c la I
In
9
0
T -
V) a2
7

t II = @
CORRECTION OF WATER PERFORMANCE CURVES 0 0 0 0 0 0

TO VISCOUS PERFORMANCE CURVES I 1 I I I I I I 1 1 1,1,1,1//11,1

I I I I 1 U1I 1 1 1 1 0I ~ l l l l i i l 0l l 0 0
e;: zr 0 9
0
z 2 2 S z
The viscosity correction factors obtained from Figure 8.1 enable us
t o derive the viscous performance curves of a pump from its standard
I3 0 7

water performance curves. The procedure is as follows: e I

1 3 L
 Step 1 : Make up a worksheet as shown in Figure 8.2 wherein,
Step 3: On Figure 8.1, align a straight edge with the value of
each of the vertical columns represents the perform-
H,, on line A and the value of [Q,J on line B.
ance parameters at a given flow rate.
Mark line C where it intersects the straight edge. Now
align the straight edge with the mark on line C and
Step 2: At the best efficiency on the water performance the value of the viscosity on line D. Mark line E
curves, read the capacity, [Q,] , the total head, H,,
where it intersects the straight edge. Read and record,
and brake horsepower, BHP,. Record their values
in column 3 of the worksheet, the Horsepower
in the appropriate spaces in column 3 of the work-
Correction Factor, CHP. Now align the straight edge
sheet. with the mark on line E and point F. Mark line G
where it intersects the straight edge. Read and record,
in column 3 of the worksheet, the correction factors,
Cg,and CH .

Step 4: From the water performance curves, select three or

four more flow rates which are evenly spaced over the
range of the curve t o be plotted. Enter these values
in the remaining Q, spaces in the worksheet.

Step 5: From the water performance curves, read the total

heads H,, and the brake horsepower, BHP,, that
correspond to the flow rates selected in Step 4. Enter
these rates in the designated spaces on the worksheet.

Step 6 : Calculate the values of CH x Qw 1 [&,I and

record the results in the appropriate spaces on the
worksheet.

Step 7 : Use Equations 8.2, 8.3 and 8.4 to compute the

viscous capacities, Qvi, , heads, Hvi,, and viscous
brake horsepowers, BHPvi, . Enter the computed
values on the worksheet.
'"- 'a Step 8: Plot the viscous performance curves using the com-
t '3 puted viscous data from the worksheet.
'Brackets [ ] indicate Q w at best C"- '3
efficiency point The following example demonstrates the use of the above procedure.
FIGURE 8.2
r .a
Step 4: From Figure 8.3 we pick four additional flow rates 0, r 9 Using Equation 8.4 we calculate the viscous brake
400, 700, and 840 gpm and enter them as Q, , 3
horsepower as follows:
in columns 1, 2, 4, and 5, of the worksheet.

Step 5:
' 3 At point 1 BHPvis = 1.5 x 1.23 x 11.8 = 21.8
Again from Figure 8.3, we find the total heads (157, t
4 At point 2 BHPvis = 1.5 x 1.23 x 21.9 = 40.4
146, 117, and 88 feet) and water brake horsepowers
(11.8, 21.9, 29.7, and 33.2) corresponding .to the f 3 At point 3 BHPvis = 1.5 x 1.23 x 26.8 = 49.4
selected flow rates and enter them o n the worksheet @
3 At point 4 BHPVi, = 1.5 x 1.23 x 29.7 = 54.8
as Hw and BHP, . At point 5 BHPvis = 1.5 x 1.23 x 33.2 61.3
' 3
=

Step 6: Calculate CH x Qw/[Q, ] C' 3 These values are then entered on the worksheet.
.-
Atpoint1 C H x Q w/[Q,] = 0.14~01585=
' 13
0 ,
At point 2 CH x Q, / [ Q, ] = 0.14 x 4001585 = 0.10 3
At point 3 CH x Q, / 1Q, ] = 0.14 x 5851585 = 0.14 3
Atpoint4 CHxQ,lIQw] = 0.14~700/585=0.17
At point 5 CH x Qw / [ Qw ] = 0.14 x 8401585 = 0.20 IS
c a
Step 7 : Using Equation 8.2 we calculate the viscous capacities 3
as follows:
At point 1 Qvjs = 0.68 x 0 = 0 gpm
= d
F' 3
At point 2 QVis = 0.68 x 400 = 272 gpm *-

At point 3 Qvis = 0.68 x 585 = 398 gpm

* 3
At point 4 Qvis = 0.68 x 700 476 gpm d
= -
At point 5 Q,,, = 0.68 x 840 = 571 gpm 6 3
f a
Using Equation 8.3 we calculate the viscous head as C 3
follows:
c' p a
At point 1 Hvis = ( 1 - 0 ) x 157 =I57 feet r", ' 3
At point 2 Hvis = ( 1- 0.10) x 146 =I32 feet t, -
At point 3 Hvis = ( 1- 0.14) x 132 =I14 feet
= , a
At p o i n t 4 H v i s = (1-0.17)x117 = 9 7 f e e t
At point 5 Hvis = ( 1 - 0.20) x 88 = 70 feet
-
A ,
3
e ,;r 'Brackets [ ] indicate Q w at best
efficiency point
FIGURE 8.4
 ' j 9
6 18
S t e p 8: Data from the worksheet, Figure 8.4, is then plotted *. 1 3
as shown by the dotted line curves in Figure 8.5.

SELECTION OF A PUMP FOR VISCOUS CONDITIONS

T h e most common use for Figure 8 . 1 is in selecting a pump t o satisfy

viscous requirements. T o d o so presents an obvious problem, in that
t o obtain correction factors from the chart we need t o know the best
efficiency of the pump which is t o be used. T o circumvent this pro-
blem, we follow a procedure that obtains a preliminary set of viscosity
factors which are admittedly incorrect but, nevertheless are sufficiently
accurate t o permit us t o select a pump that will satisfy the viscous
requirements. Once having selected the basic pump, we can obtain true
factors from Figure 8.1 which allow us t o select the correct impeller
diameter.
The procedure employs the following equations which are transposed
from Equation 8.2 and Equation 8.3:

Qw
-
--
Qvis Equation 8.2a
CQ
Hvis
Hw = ( 1 - c ~ x Qw/[Qwl) Equation 8.3a

The procedure is as follows:

Step 1 : In Figure 8.1 align Hvis on the A line with Qvis

(or 3 0 gpm when Qvis is less than 30) o n the B line.
Determine the preliminary viscosity correction
factors, CHPl , CQ andCH l c ~ r r e ~ p ~ n d tion these
g
conditions and the specified viscosity. When this
procedure indicates a horsepower correction factor
greater than 1.6, assume CHPl = 1.6, CQ =0.58
and CHI = 0.18.

Step 2: Using Equation 8.2a, determine the preliminary

water capacity, Qwl , corresponding t o Qvis .
 3-
-
S t e p 3: Using Equation 8.3a and assuming Q, / [Q, ] = 1.0 v
Example 2
determine the preliminary water, head, Hwl , cor- o 9
responding t o Hvls . select a pump t o move 272 gpm of viscous liquid against a total
head of 1 3 0 feet. The viscos~tyof the liquid is 6 0 0 0 SSU and its
Step 4: Select a pump that satisfies Q, and Hwl . 6 3 speclflc gravity 1s 1.23. Determine the viscous brake horsepower at
Read the head and capacity, Hw and [Q,] ,a t the the operating point and at the end o f the curve.
best efficiency point of the impeller curve that is 9
nearest to, but falls below, Hwl a t Qwl . k
- 'I$ ~olutlon
S t e p 5: @ 3
From Figure 8.1, obtain a secondary set of correction From page 74 of the Plpe Friction Manual we find that 6 0 0 0 SSU =
factors, C Q 2 and CH 2 , corresponding t o H, , --- a 1 3 2 0 centistckes.
and [Q,] determined in step 4. * 9
6- 3 Step 1 From Figure 8 . 1 at 1 3 0 feet, 272 gpm, and 1 3 2 0
Step 6: Using CQ from step 5 and Equation 8.2a, compute centistokes we find:
the secondary water capacity, Qw2 , corresponding 3
t o Qvis - c ' 3 cHPl= 1.60
3 CHI = 0.18
Step 7: Using [Q,] from step 4, CH from step 5 and
Qw2 from step 6 in Equation 8.3a, compute a sec- 3 C~ 1
= 0.58
ondary water head, Hw2 , corresponding t o HVi, .
G d
G & Step 2 From Equation 8.2a:
S t e p 8: Select t h e impeller diameter that satifies Hwg and
Qw . If the diameter is close t o that selected in C 3
step 4 , we can consider the pump selection t o be c- = -272
' i #
Qw I 0.58
complete and proceed t o step 9. If the diameter
differs by more than H inch from that selected by 3 = 4 6 9 gpm
step 4, we must repeat steps 4 through 8 , substi-
tuting the values of H, and Q w 2 from step 8
f 3
for Hwl and Qw in step 4: 3 Step 3 From Equation 8.3a, assuming Qw /[ Qw 1 = 1.0
et: 'a
S t e p 9: Use the procedure previously described for plotting 130
the viscous performance curves or calculating the c i) Hw I - ( 1 - 0.18 x 1)
viscous brake horsepower of the selected pump and ' i)
the impeller .
= i)
= 1 5 9 feet

The following example illustrates the above procedure.

Step 4 We find that a 4 x 3 - 1 3 pump operating at 1 7 5 0
et ;a rpm will satisfy Qw and Hw 1 . From the water
performance curves of thls pump shown in Figure
 '1'
C" !
4
9
8.3, we select the 12lh" diameter impeller as being W i 3
nearest to, but falling below, H, at Q, . We %i 9 Chapter 9
read the best efficiency head and capacity of the SPECIFIC SPEED AND SUCTION SPECIFIC SPEED
12%" impeller curve as being.: - 3
H, = 132 feet '
h
3 Specific speed is a dimensionless index number which is used t o
relate the hydraulic performance of a centrifugal pump to the shape
[&,I = 585 gpm J
and physical proportions of its impeller. An indication of how impeller
h.
Step 5: From Figure 8.1, we find the secondary correction shape and proportions vary over the normal range of specific speeds
factors corresponding to 132 feet and 585 gpm t o be: 6 9 can be seen in Figure 9.1.
+

CHp2 = 1.5 3
?

Cg2 = 0.68 3 SPECIFIC SPEED

C5
- 7 1 I I T I I I I I I ( I l l I
CH2 = 0.14 *+ 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0

6 3
0
In
0 0 0 0 0
w r-como Y
0
"Y
0
0
cu
0
*
0
s
0 0 0 0 0 0 0
5: g e z g $ 0
In
Y

Step 6: Using Equation 8.2a:

- 272
Qw2 - 0.68
P
= 4nn onm 3
RADIAL-VANE
-
--
.
FRANCIS-VANE
. -. -... r",.,
MIA~V-~LUW AAIML-r~wn
,,,&I

Step 7: From 8.3a, using [Q, 1 from step 4 we find that: FIGURE 9.1
E 3
w = 130 E 3
lLW
( 1- 0.14 x 4001585)
aP The numerical value of specific speed is found from the following
= 144 feet r
3 equation:
=
*r

Step 8: Looking again at Figure 8.3 we see that the 121h" miP
-
~ n l h a1
li',-.~~~t;nn

curve satisfies Q, and HW2;therefore, the 4 x 3 - 13

-pump- with a 12%" impeller is the correct selection 3
for this application. 3 whew-

Step 9 : This step has already been completed in Example 1 ,

t a N, = specific speed
since we are dealing with the same pump ( 4 x 3 - 13) 3 N = rotative speed, in revolutions
.-.. per ..minute . - I

and the viscosities are equal (6000 SSU - 1320 '", ' a Q = capacity, at best efficiency, in gallons per minute
.. .. . . .-.. --.. :-..,.ll-"
centistokes). From the worksheet of Example 1 , a H =
. I

total head developed by maxlmurn alarnewr urlpeuel

1: .L...-

we see that the viscous brake horsepower at the at best efficiency, in feet
design point (where Qvi, = 272 gpm) is 40.4 and 3
at the end of the curve Goint 5) it is61.3.
F- 'a
 I t can be shown by applying the affinity laws (See Chapter 7) t o the "1
terms of the above equation that the specific speed of a given impeller 9 Chapter 10
is constant regardless of its rotative speed. R . 4 IMPELLER LOADING
Suction specific speed is a dimensionless rating number which I

indicates the relative ability of centrifugal pumps to operate under 3 In addition to the obvious torsional reaction, centrifugal pump
conditions of low available net positive suction head (See Chapter 5). 3 impellers are also subjected to appreciable radial and axial loading.
The equation for suction specific speed is: F

3
RADIAL LOADING
Equation 9.2
At the flow rate of best efficiency, in a well designed volute pump,
where: the pressure developed by the impeller is more or less uniform about
its entire circumference. The resultant radial force on the impeller
S = suction specific speed is essentially zero. However, a t all other flow rates, the pressure is not
N = rotative speed, in revolutions per minute uniform and this pressure variation, acting on the projected impeller
Q' = capacity, at best efficiency, in gallons per minute area, produces a resultant radial force o n the impeller. At low flow
h,, = net positive suction head required by maximum rates, the pressure distribution in a volute type casing is such that
diameter impeller at best efficiency, in feet the impeller surfaces closest t o the discharge are acted upon by high
pressures and those on the other side of the cutwater are acted upon
Depending on impeller design, suction specific speeds will vary in by comparatively low pressures. The resulting unbalanced force acts
numerical value from below 4,000 to above 11,000 with the higher towards the shaft at a point approximately 240 degrees from the cut-
values indicating lower net positive suction head requirements. water, as shown in Figure 10.1.

At capacities greater than that of best efficiency, the resultant

force acts towards the shaft at a point directly opposite that shown
in Figure 10.1.
 I
F

3
In a "circular" type casing, the resultant radial load is minimum at " ,1 3 OPEN CLOSED
- I
shutoff and increases to maximum near the best efficiency point. 3
Consequently, a pump which has been selected to operate near its , 3 w 0 50
best efficiency point will experience much lighter radial loads if its - IILL
L ~
casing is the volute type rather than the circular type. ' 3 o w
sz
' 9 040
CALCULATING RADIAL LOADS -- moo
+-I
' 3 U W
- '
K
The radial load on an impeller in a volute type casing can be cal- 3
culated by the use of the empirical formula: - -. 2:
030

b 3 L:
Equation 10.1
-. *cU
t L301
o 0.20

P=KQ x K x H x sp gr x Dz x B, " 9 $?
+3
-I-'
where:
2.31
' 3 2
SY
o,,
r 3 iig
Kg = I-($,)" n = 0.7 + 2.6 (Ns - 500) : 9 -
UJ
3000 o 00
: 3
where: iP SPECIFIC SPEED N s

P = resultant radial force, in pounds : 3 FIGURE 10.2

AXIAL LOADING
Kg = capacity factor C 3
K = radial thrust factor at shutoff, from Figure 10.2 3 Single suction impellers, like those used in Durco pumps, are subject
H = total head at Q gpm, in feet to an axial thrust which is the net result of three forces. One force
3
Dz = O.D. of impeller, in inches P
is caused by the pressure behind the impeller tending to push it towards
s
the suction nozzle. A second force is the result of pressure on the
BB width of impeller at O.D., in inches
Q
=
= capacity, in gpm, at which radial thrust is t o be calculated
+
I
a other side of the impeller tending t o push it toward the motor. The
third is caused by the suction pressure which acts on the cross-sectional
capacity, in gpm, at best efficiency of pump
1 3 area of the shaft, as if it were a piston in the stuffing box. The pres-
Q, =
specific speed of pump 5 3 sures which act on the impeller area are created by the impeller. They
Ns =
t, 3 are directly proportional t o specific gravity of the liquid pumped.
A quick examination of the above equations, reveals that the radial The outboard bearing must be able to absorb this axial thrust and
load is greatest at shutoff, where Q is and Kg = 1.0. As Q in- C- 3 keep the rotating element from moving axially. The life of this bearing
creases, ~g and the radial load decrease to zero at the best efficiency
3 is, courw, dependent upon the axial load, and can be greatly in-
point where Q = Q, As Q increases to valtles greater than Q, , creased by designing the impeller so that the net axial force is very
KQ and the radial load increase as negative values. The negative sign e 3 low. The formulae for calculating the net axial thrust vary with each
indicates that the force is now in the opposite direction as shown in e j iolpeller style,
Figure 10.1.
116
k 3
117
&
*I 3
 Chapter 11 RADIAL LOAD
BEARINGS AND BEARING LIFE
Radial thrust acting on the impeller creates radial loading on both
Two of the most important parts in a centrifugal pump are the bearings. The magnitude of the radial load at each bearing can be
bearings. This may seem t o be a point for discussion, but consider determined by the use of the following equations:
the vital role that they play in pump operation. They must allow P
the shaft to rotate with practically negligible friction, so that the RI

horsepower requirements will be kept t o a minimum. They hold the

rotating element in its proper position relative to the stationary parts
of the pump, both radially and axially, so that rubbing cannot occur.
They must be able to absorb the forces that are transmitted t o them
from the impeller and give trouble free service for long periods of
time.
The present-day ball bearing is a piece of precision equipment, man-
FIGURE 11.1
ufactured t o extremely close tolerances. It is made, assembled, cleaned,
lubricated, and packaged in a dust free factory where temperature
and humidity are rigidly controlled. When properly mounted, lub- Equation 11.1
ricated, and protected from dirt, a ball bearing will carry radial and
axial loads at high rotative speeds for years.
In order t o obtain the maximum service from a bearing, the shaft and
housing must be machined to the same exacting tolerances that are where:
used in making the bearing. Locating shoulders must be a t right angles
to the shaft centerline so that the bearing will be squared with the R = radial load a t bearings 1and 2, in pounds
shaft. The housing bores must be in almost perfect alignment t o P = radial thrust on impeller, in pounds, from Equation 10.1
insure that the bearing will not be forced t o operate in a twisted pos-
a&s = dimensions shown
ition. The lubricant, be it grease or oil, must be of the right viscosity,
and it must be present in the proper quantity. The seals that retain the AXIAL LOAD
lubricant in the housing must also be capable of excluding dirt and
water from the housing. All of the parts which surround the bearings In any two bearing system, one of the bearings must be fixed axially,
must be free from dirt at the time of assembly. The bearings them- while the other is free t o slide. This arrangement allows the shaft t o
selves should not be unpacked until the pump is ready for assembly, expand or contract without imposing axial load on bearings, and yet
and then should be installed and covered as soon as possible. definitely locates one end of the shaft relative t o the stationary parts
BEARING LOAD ' -
of the pump.
In Durco pumps, the outboard bearing (the closest one t o the
Assuming that all of the above conditions are satisfied, the life of 3 coupling), is fixed The inboard bearing is free t o slide within
a bearing is dependent upon the load that it must carry and the speed c the housing bore to accommodate thermal and contraction
of operation. The Loads on pump bearings are imposed by the radial
of the shaft.
and axial forces acting on the impeller as described in Chapter 10. 3
Since the outboard bearing is fixed in the housing, it must carry the
'"- 3
118 119
axial thrust in addition t o radial thrust. The axial thrust is considered The "rating life" of a bearing varies inversely as the cube of the
t o be acting along the centerline of the shaft, and therefore, is the same applied load. Obviously, if the bearing supports one of the smaller
at the outboard bearing as it is at the impeller. The radial and axial impellers, or operates a t a lower speed, or if the flow rate more nearly
loads combine t o create a resultant angular load a t the outboard bear- approaches the best efficiency flow, the "application rating life" will
ing. be markedly greater than two years. For example, if the application
flow rate is 50% of the flow of the best efficiency, the radial load will
BEARING LIFE
be about 3/4 of that calculated for shutoff and the "application rating
life" will be more than twice the "design rating life".
A bearing is capable of operating for a specific number of revolutions
under a specified load, before the first evidence of failure appears. BEARING LUBRICATION
Naturally, at higher loads this number decreases, and at lighter loads
it increases. Remarks o n bearings would be incomplete without comments o n
A bearing is said to have reached the end of its life when it develops lubrication. Either grease or oil lubrication is satisfactory s o long as
excessive noise, excessive heat, or has worn internally t o the point it is done properly.
where it n o longer will hold the shaft in its correct position. Ex- As a general rule, grease is preferred when:
amination at this time will disclose small pits in the raceway or ball
surfaces where particles of metal have been removed. These pits are 1.Temperatures are n o t excessive, usually n o t over 200' F.
formed when the surface metal fails from repeated flexural stress. The 2. Speed does not exceed bearing manufacturer's recommended
metal has a definite ability t o resist deflection, but when the load limit for grease lubricated bearings.
applied t o it exceeds this ability, fatigue failure occurs. High unit 3. Extra protection is required from dirt, fumes and other con-
pressures are applied when the bearings are under-sized, when they taminates.
operate a t unusually high temperatures, when they operate in a twisted 4. Prolonged periods of operation without maintenance are expected.
position, or are otherwise improperly installed.
The life of an individual bearing in a specific application cannot be Oil is preferred when
foretold because of unavoidable variations in bearing materials and
manufacturing. However, it is possible t o predict by statistical analysis 1.Operating temperatures are consistently high.
that 90% of a large number of specific size and type of bearing in a 2. Speed exceeds bearing manufacturer's limit for grease lubricated
specific application will last longer than a certain time. That certain bearings.
time is called the "rating life" of that size and type of bearing in that 3. Dirt conditions are not excessive and oil tight seals can be utilized.
specific application. T h e "rating life" is calculated by following the 4. Determination of the lubrication conditions is more easily
procedure outlined in ANSI Standard B3.15-1972 or the equivalent detected.
procedure usually included in each bearing manufacturer's catalog. 5. Bearing design does not lend itself t o grease lubrication.
Durcopump bearings were selected to provide a "design rating life"
of at least two years (365 days per year, 24 hours per day) when Regardless of the type of lubricant, proper lubrication is essential
supporting a shaft which is driving the largest impeller that it is design- if the bearing is to live up t o its predicted life. Oil levels must be
ed t o accommodate. The "design rating life" is based also o n the maintained, contamination must be prevented, regreasing must be done
additional assumptions that the maximum impeller is rotating a t its properly. T o insure optimum lubrication, it is important that the
maximum allowable speed against a closed discharge valve, both of Durco pump user follow explicitly the instructions in the Operation
which conditions maximize radial loads. Manual which he receives with his pump.
 .-
.--I I Y J
" 9 PACKING GLAND
Chapter 12
@ 3
SHAFT SEALING DEVICES
@ J
I t is highly desirable that centrifugal pumps be provided with a ' 9
suitable shaft sealing arrangement so that the pumped liquid will not $
leak t o the atmosphere. Sealing problems confronting centrifugal pump
manufacturers are many because:
@ 3
1.The seal is mounted on a rotating shaft which may .-
n o t be running concentric, or may be worn. 3
b 3
2. The pressure differential may vary greatly. 6 3
3. Abrasive materials are often encountered. 4 3
SHAFT SEAL CAGE
$ 4
4. Temperature changes are usually experienced. FIGURE 12.1
C 3
5. Corrosive materials are handled frequently. 5' 3
3 The lubrication port provides a means of introducing additional
PACKING liquid to help in the above functions, and the seal cage proviaes a
,-
means of distributing the additional liquid around the periphery of the
A typical pump stuffing box is shown in Figure 12. 1. The packing
usually consists of fibrous material that is first woven into rope and
v 3 shaft. "
In pumps handling corrosives, the rope packing is usually ma.de trom
I
.. . . ..
then cut and molded into packing rings of rectangular cross-section.
," . . ,----. . . ".. .
r
asbestos, tetrafluoroethylene (Yl'P'h), o r graphlte tlber. 'l'he choice
These rings
- are made sufficiently soft so that pressure from the packing- between these three is made on the basis of cost versus corrosion
gland forces them to fill the space between the stuffing box and the
shaft.
P
r- 3 resistance.
The three major virtues of a packed stuffing box are:
I
The packing is usually impregnated with a viscous liquid which:
1.I t is relatively inexpensive in initial cost.
1.Provides lubrication between the packing and the
shaft. 2. It rarely is the cause of an unscheduled shutdown
of the pump since its sealing efficiency usually
2. Seals off the microscopic spaces between the deteriorates slowly.
packing fibers.
3 . T o compensate for wear, the packing can be
3. Helps t o distribute heat developed by friction adjusted or completely replaced without dis-
between the packing and the shaft. assembly of the pump.
 The two major shortcomings of a packed stuffing box in a chemical parts 1 and 6 , held in contact by a compression spring, part 8. Ring
pump are: 1 is stationary and ring 6 rotates with the shaft. There are three sealing
points in the typical mechanical seal installation. One exists between
1.Leakage ring 6 and the shaft, 5. Another is between ring 1 and the end of the
stuffing box. The third is normal t o the pump axis, between the mating
2. The cost of packing maintenance. faces of rings 1 and 6.
The ability of rope packing to minimize leakage depends on its The seal between ring 6 and the shaft is usually accomplished by
ability to maintain a close clearance between itself and the rotating one of the following means:
shaft. Any radial movement of the shaft, due to runout, deflection,
whip or bearing looseness, tends to open up the clearance and increase 1. A PTFE sealing element, part 7, held in compression
the leakage along the shaft. Every pump shaft exhibits these forms of by the spring, part 8.
radial movment to some degree, but in chemical pumps they tend to be
accentuated by the severity of the applications. 2. An elastomeric O-ring in place of the PTFE element.
In addition to radial movement, thermal expansion and contraction
3. A bellows of PTFE, elastomer, or metal.
of the packing tend to open up the clearance along the shaft. Since
the stuffing box temperature can change appreciably with changes in
pumping conditions, this adds to the leakage problem. A vital requirement of the seal at this point is to permit ring 6 , to
Futhermore, even in the easiest of applications, it is usually necessary move relative to the shaft. This enables ring 6 t o remain aligned with
to allow some leakage t o occur in order to remove the heat that is stationary ring 1 , regardless of the radial and axial movement of the
generated by friction between the packing and the shaft. This leakage shaft on which it is mounted and with which it rotates. The relative
may amount to only a few drops per minute, but in a chemical pump movement between ring 6 , and the shaft can result in loss of the seal
this may be unacceptable due to the value of the product lost or at this point due to abrasion between the sealing surfaces or failure
the safety hazard it may create. of the bellows in flexure. This is why a pump with minimum radial
While it is true that a packed stuffing box can be adjusted to min- and axial shaft movement helps to reduce seal maintenance cost.
imize leakage due to corrosion, wear, and the conditions described A flat gasket, part 3, usually effects the seal between ring 1 and the
end of the stuffing box. Any gasket material that is resistant to the
above, the cost of labor required to make the adjustment can lead t o
excessively high maintenance costs if the pump has to have frequent liquid to be sealed is satisfactory. This sealing point is rarely the cause
attention. of seal maintenance.
The rotating seal between ring 1 and 6 is achieved by lapping both
mating surfaces until they are extremely flat and smooth. When
MECHANICAL SEALS
these surfaces are pressed together by the spring, part 8 , only a mirco-
scopically thin film of liquid exists between them, and the leakage of
In most applications, a mechanical seal can be selected t o overcome
liquid through this gap is imperceptible.
the shortcomings of rope packing; however, the mechanical seal will
usually be higher in initial cost and presents the possiblity of sudden In conventional seals, the thin film of liquid between the seal faces
is the key t o successful performance of the rotating seal, If the liquid
failure, requiring shutdown of the pump for seal replacement.
film has adequate lubricating properties, the rotating seal will be
A typical mechanical seal is shown in Figure 12.2. Though the
maintained almost indefinitely. On the other hand, if the film of liquid
details of construction may vary from one make of seal to another,
contains solids, or if it heats up and loses its load carrying capacity, or
the basic components are similar. Each has two mating seal rings,
if it crystallizes when the pump is shutdown, thereby cementing the
 Chapter 13
SHAFT DEFLECTION, WHIP AND RUNOUT

As explained in the previous chapter, radial movement of a pump

\ I shaft reduces the performance and life of its stuffing box seal. Radial
1 STATIONARY SEAT-ANTI-FRICTIONAL MATERIAL movement of the shaft occurs in three forms: deflection, whip and
SUITABLE FOR WEARING
SURFACE. runout.
2 GLAND PLATE These forms of radial movement are of vital concern t o the pump
3 GASKET user. This concern is recognized by ANSI B73.1, the Standard for
4 STUFFING BOX
5 ROTATING SHAFT centrifugal chemical pumps. I t limits the radial deflection and runout
6 SEAL RING - ANTI-FRICTIONAL MATERIAL of the pump shaft to a maximum of 0.002 inches at the face of the
SUITABLE FOR WEARING stuffing box, and calls for impeller balancing t o minimize shaft whip.
SURFACE.
7 SHAFT PACKING-USUALLY TEFLON OR
" 0 RlNG SHAFT DEFLECTION
8 SPRING
9 RETAINER
FIGURE 12.2 A shaft deflects, or is pushed from its normal position, because of
unbalanced radial loads on the impeller which are described in
seal faces together, then the seal will fail rapidly. Most of the variations Chapter 10. While at rest, the shaft is straight and is centered in the
in seal installations are designed t o insure a continuous liquid film, with stuffing box. When pumping, the unbalanced radial loads make the
optimum lubricating qualities. shaft bend and take up a new position in the stuffing box.
I t is beyond the scope of this book t o go into all the variations from
the standard mechanical seal installations shown in Figure 12.2. They
are well documented in the literature of the manufacturers of mech-
anical seals. However, the selector of a mechanical seal for a specific
application should keep in mind three factors when choosing from
FIGURE 13.1
these variations:
P-
3 Figure 13.1 shows a shaft in its normal position and also in an
1.That the seal be made of materials that will resist
the deleterious effects of the liquid being sealed. ' a exaggerated deflected position. Please note that the shaft revolves on
its own centerline even when it is deflected, because the load is
constant in direction and magnitude. In other words, it maintains its
2. That the condition of the pump be such that it L
bent position as long as the speed of rotation and the quantity being
will not contribute to premature seal failure due to 3
I..
pumped are not changed. Since the shaft remains in this same position
excessive radial and axial movement of the shaft.
as long as the capacity is not changed, and since it is revolving on its
;5 own centerline, the stuffing box packing can be adjusted t o allow a

-
3. That the seal installation insures the continuance minimum of leakage. When the pump is stopped o r the rate of flow
of a liquid lubricating film between the sealing t z,
drastically changed, the packing may have t o be readjusted to
faces. t - compensate f o r the change in deflection.
 CALCULATING SHAFT DEFLECTION
SHAFT WHIP
The shaft deflection is calculated by treating the shaft as a cantilever
beam and using an expanded version of the beam formula. The The impeller end of a shaft that is whipping revolves in such a
equations for calculating the deflection of a shaft with three major manner as t o generate a cone shape. The direction that it is moved
diameters in the overhung section are: from its true centerline changes 180' for each 180° turn of the shaft.
Shaft whip makes it almost impossible to have a stuffing box function

FIGURE 13.3

properly because the packing is pushed out t o the diameter of the

cone being generated. This leaves a hole which is larger than the
shaft diameter and excessive leakage will occur. Usually the force
that is causing whip is coming from within the rotating element. F o r
example, an impeller head that was not properly balanced would cause
FIGURE 13.2 whip since the heavy side of the impeller would always be on the same
side of the shaft.
It is entirely possible t o have a combination of shaft deflection and
shaft whip present at the same time. In this case the cone generated by
the whipping shaft is moved t o one side by the amount that the shaft
deflects.
where:
SHAFT RUNOUT
Y = shaft deflection, in inches, X inches from impeller
Shaft runout is another term that is often used in discussing pump
centerline
P = radial load, in pounds, from Equation 10.1 shafts. It is the amount that the shaft sleeve section is out of true when
the shaft is in the pump. Shaft runout is measured by fastening a dial
E = modulus of elasticity of the shaft, in psi per inch,
per inch indicator t o a stationary part of the pump so that its contact point
a, b, c , & s = shaft dimensions, in inches, as shown indicates the radial movement of the shaft surface as the shaft is rotated
d,, d b , d,, & d, = maximum shaft diameters, in inches, in sections, a , b, slowly. A runout check is made on every pump after it is assembled.
This inspection will disclose any out of roundness of the shaft, any
c, & s, eccentricity between the shaft and shaft sleeve, and any permanent
I,, Ib, I,, & Is = moments of inertia, in inches4 , corresponding to
bend in the shaft. Any of these defects would produce an oversize
diameters d, ,d b , d,, and d, .
X = distance, in inches, from centerline of the impeller to hole through the packing and make the stuffing box hard t o seal. This
point at which Y is to be calculated check will not disclose any shaft whip that would be caused by impeller
unbalance since the runout check is made by turning the shaft by hand.
 '- I '
'-I ' TABLE 14.1
Chapter 14 @ l a Maximum Acceptable Vibration
% l a
VIBRATION C, i3 Pump Speed Displacement *
When we speak of vibration in a centrifugal pump, we usually refer Mils Peak-to-Peak
% I d RPM
to vibration in a radial or axial direction. This type of movement and 3600 1.25
the force accompanying it tend to damage the bearings and other parts 6-- 1 J 1800 2.50
of the -pump.
- .-. 1200 3.75
It is customary to measure the vibration of a pump as close as
possible to each bearing, in two radial directions (vertical and
horizontal on a horizontal pump) and in the axial direction. The
vibration detecting device may measure the acceleration, velocity, or
-
6 , *Measured erpend~culart o the shaft axls a t p o ~ n ot n the bearing
housing a 8 a c e n t t o each bearing.

One means to extend the usefulness of a simple vibration reading is

the displacement of the surface which it contacts. Regardless of the
parameter measured, most instruments can display all three. Cen- 6 13 to establish a program of measuring and recording the readings from a
given pump at regularly scheduled intervals. The change in vibration,
trifugal pump vibration is usually expressed in terms of peak-to-peak 6 13 from one interval to the next is at least as important as the magnitude
displacement.
Vibration may originate within a pump and motor assembly or be
transmitted to it from some outside source. Table 14.2 lists the com-
'1 3
of the vibration. Thus, an increasing vibration level, even though it may
be relatively low in value, indicates an approaching problem; whereas,
an unchanging level, even though relatively high, gives evidence of a
mon causes of vibration.
stable, and probably satisfactory, operatiqg condition.
Regardless of its origin, if the vibration is large enough, some parts
If the vibration of a pump exceeds the generally accepted limit noted
of the pump will be damaged; and they, in turn, will generate additional
vibration. For this reason, vibration, in addition to indicating forces - in Table 14.1 , or if successive measurements show an increase in the
vibration level, it is advantageous to identify the source of vibration
and movement that may damage a pump, can also be evidence that the
pump has been damaged. ' I J3
I"

I
to facilitate corrective maintenance. One approach t o identification is
through vibration analysis. A study of vibration analysis is beyond the
unfortunately, a simple broad band vibration measurement does not
positively establish that the pump will be damaged by the vibration, or I a scope of this article, but basically, it involves determining the frequency
vibration level occurs. The significance of the
at which the
that damage has already curre red. However, there are some generally 3 frequency with regard to the source of vibration is shown in Table 14.2.
accepted vibration limits for centrifugal pumps which indicate that
more detailed vibration analysis or corrective maintenance is required.
-
c a When there is a dominant frequency with no apparent relationship
to pump rpm, the source of the vibration is probably external to the
These limits are shown in Table 14.1.
-a 3
pump. T~ reduce this form of vibration, the source must be located and
suppressed, or the pump must be isolated from it by means of flexible
connectors in the piping and/or vibration isolating mounts under the
C-
a base.

-
e.
3
t 3
t
b
- a 131
L-" ?,
 3
- I

3
TABLE 14.2
Relationship Between Frequency and Cause of Vibration @- 9 Chapter 15
3
HEAT GENERATED IN PUMPING
@ J
Frequency Amplitude Cause Remarks
C 3 In centrifugal pumps, some of the input energy is transformed by
I x RPM Largest in radial Unbalance fluid friction into heat, which increases the temperature of the liquid
b
direction. Proport- being pumped. The temperature increase is dependent upon the flow
ional t o unbalance.
)
e
3 rate o f liquid moving through the pump. At normal flow rates, this
I x RPM Axial direction Misalignment of Easily recognized
r temperature rise is negligible; while under shutoff (zero flow)
normally vibration50%or coupling or bearings by large axial vib-
+ conditions, the temperature will continue t o rise, resulting in eventual
more of radial and bent shaft ration. Excessive 3 damage t o the pump. I t is obvious that some minimum flow rate must
flange loading can @ 3 be maintained to prevent this from occurring. In order t o determine

Very high. Unsteady Bad anti-friction

contribute t o mis-
alignment.

Largest high-fre-
F

w
6
'
d
the minimum flow rate, the maximum allowable temperature must be
known.
T h e maximum allowable temperature rise is established by three
Several bearings. quency vibration 6 3 basic considerations:
times RPM near the bad
bearing. 3 1.The temperature at which the pumped product will
2 x RPM Mechanical loose- Check grouting
G d be adversely affected.
ness. and bedplate 6
rC 3 2. T h e temperature at which the properties of the
bolting.
"
F"
3 pumped product are changed so as to adversely
1, 2, 3, & Erratic or pulsing. Bad drive belts. Use strobe light " 3 affect the action of the pump (e. g. vaporization,
polymerization, etc.)
4 x RPM of t o freeze faulty b=-

be1ts. belt. " 3

c 3 3. The maximum temperature which various
1or2x Disappears when Electrical 3600 or 7200 cps P components of the pump and/or piping system
synchron- power i s turned for 60 cycle 3
ous fre-
quency
off. current
= 3
(e. g. mechanical seals, packing, gaskets, etc.)
can accommodate.

No. of Hydraulic forces Rarely a cause of

=" 3 *
.
In most applications, 10' F is an acceptable temperature rise;
impeller
vanes x
RPM
serious vibration
= 3
however, when NPSH is critical, the temperature rise should be limited
to5"F.orless.
-
&
3 The minimum flow rate which must be maintained may be calculated
I Extremely
high,
Cavitation -
b-
3 as follows:
&= 5 x BHPo x Chp Equation 15.1
random. to, AT x sp h t
55,
132 $ 3
where: 5 x BHP, x C , ,
Q=
Q = minimum flow rate, in gpm AT x s p h t
BHP, = the non-viscous performance curve horsepower at shutoff
we find:
sp ht = specific heat of liquid
5 x 11.8 x 1.5
AT = maximum allowable temperature rise, degrees F. Q=
10 x 0.85
= viscous horsepower correction factor. (See Chapter 8,
Ch P
Figure 8.1). = 10.4 gpm
T o find the temperature rise resulting from a known flow rate,
the Equation 15.1 can be written as follows:

5 x BHP, x Ch,
' 3 Example 2 Calculating Temperature Rise
AT = Equation 15.2 3
Q x sp h t Given the pump from the previous example, what temperature rise
1
would result from a flow rate of 2 gpm ?
Example 1 Calculating Minimum Flow Rate '- 3
Solution
Given a 4 x 3 -13 pump a t 1750 rpm (See Figure 8.3), handling
'- 3
a liquid with a 1.23 specific gravity, a 0.85 specific heat, and a viscosity
of 1625 centipoise, what is the minimum allowable flow rate for the
-
*
r
3
3
Having determined shutoff horsepower and the viscous efficiency
correction factor, substitute into Equation 15.2 as follows:
12lh" impeller if the temperature rise is to be kept below 10' F.? ;3
(See Figure 8.1. ). 5 x BHP, x Chp
3 AT =

Solution
- Q x spht

1. From Figure 8.3, read the non-viscous shutoff 3

horsepower for the 12lh" impeller equals 11.8. d
Irr -
2. Following the procedure outlined in Example 1of
' 3
r
Chapter 8, (which calls for the same pump, 3 5IINIMUM FLOW BY-PASS
operating under the same conditions as the *- 3
example) find the viscous horsepower correction
factor, Chp = 1.5. '3
5
3
In situations where the required flow rate is less than the minimum
;illowable flow, a by-pass line can be installed which returns a portion
o f the pumped liquid t o the suction reservoir.
3. Substituting into Equation 15.1. * 4 The sum of the required and by-pass flows should be equal t o or
f
-
"- -
i

-
,
w
greater than the calculated minimum allowable flow. The by-pass
flow can be controlled with a throttling valve or an orifice t o
arconlplish this purpose.
 " 1 J
-- I
k e ' 9
'" After establishing prime, the pump performs exactly as any other
Chapter 16 centrifugal pump, with the exception that it can prime itself auto-
SELF-PRIMING CENTRIFUGAL PUMPS
@ a
rnatically if it becomes air bound. For example: if the liquid level
I
3 drops below the end of the suction pipe, the suction piping will agaill
A standard centrifugal pump will NOT move liquid unless it is
primed. A pump is said to be "primed" when the pump casing and the
, 3 fill with air and pumping action will cease. As the level builds up
suction piping are completely filled with liquid. Units that are located ' agaln and covers the end of the pipe, the priming cycle will be repeated
and the pump will resume normal operation.
below the liquid level of the suction tank can be primed by merely r -
3 It should be remembered that these pumps are primarily designed t o
opening the suction and discharge valves, thereby allowing liquid to a
convey a liquid and are, therefore, not efficient air handling machines.
1-
flow into the pump by gravity. Standard centrifugal pumps that are - For this reason, particular attention should be given t o the pump
located above the suction level must be primed by some auxiliary 1 3 installation t o limit the amount of air that the pump will be required
means such as a vaccum pump, or an ejector. These means of priming 3 to evacuate. The suction lift and the volume of the suction line should
chemical pumps are generally unsatisfactory because of the corrosive
nature of the liquids being handled. Foot valves are sometimes used on ' be kept t o a minimum, and all the joints should be made air tight.
In addition, any pressure build up at the discharge of the pump will
the lower end of the suction line so that it is possible t o fill the pump 9 - 3 reduce its suction lift. Therefore, the discharge line should not include
and piping with liquid from an outside source; but this method demands I-
an elevated tank to store the priming liquid. All of these systems 3 check valves, closed valves, or any other obstruction that would block
the flow of air and cause pressure to build.
require someone to be present to operate the auxiliary equipment 3 The stuffing box area must be sealed tightly against the entrance of
initially, and to reprime the pump should it air-bind during operation. ,- outside air. hi^ is accomplished by use of or a
The self-priming pump is one answer to the above problems. I t
'
- mechanical seal. During the priming cycle particularly, the stuffing box
primes itself, it will reprime if it becomes air bound during operation, a
is under vacuum and air that enters from the outside must be removed
and does not require constant attention during operation. '" 3 in exactly the same manner as that from the suction piping. This
A well designed self-priming pump must be capable of efficiently -
removing the air from the suction piping. Removal of this air creates ' - 3 increases the priming time; and may, if the leak is large, prevent priming
altogether. Where possible, it is a good idea to provide a flushing line
a ~ a r t i a lvacuum in the suction line, allowing atmospheric pressure d ,,to the stuffing box t o prevent the leakage of air into the pump at that
t o force the liquid from the suction tank up into the pump, thereby point.
establishing prime. Such a pump must be capable of forming a seal When large volumes of air have to be evacuated from the suction
above the impeller so that atmospheric pressure cannot work back "
+

.t
piping, the priming time may become so extended that the liquid re-
through the discharge line to fill the vaccum. Finally, a self-priming 3 circulated within the pump becomes hot enough to vaporize. When
pump must function efficiently as a standard centrifugal pump after ,. 3 this happens the priming action ceases and the priming chamber has t o
it has attained prime and established flow. For this reason, the self- be refilled with liquid.
priming feature must not detract excessively from the pumping ability
3 The
of the unit. - lift that can be tolerated depends upon the specific
,. 3 gravity and the vapor pressure of the liquid being handled. Heavy
There are many different self-priming centrifugal pump designs, but,
in essence, they all function in the same manner. That is, they all - a llquids will not be forced as high into the suction line by atmospheric
pressure as will light liquids. Therefore, the maximum lift is lower for
retain a certain amount of ''priming liquid" when they are shut-down T 3 these heavy liquids. ~ i vapor ~ hpressure liquids will tend to vaporize as
Or lose prime, and the priming liquid in such a way
as to entrain air at their inlet side and to release it at their discharge
a they reach the areas of greatest vacuum, and the rnavimum lift for the=
side. - a
D, llquids is therefore reduced. Both of these factors are taken into
account in the calculation of the available NPSH of the system. It is
136 k 3 137
 '- I '
more accurate t o use the available NPSH to establish the maximum lift @
1a Chapter 17

'
than to use an arbitrary figure. In fact, every self-priming pump (I
installation should be thoroughly checked for available NPSH before
il
1 CENTRIFUGAL PUMP TROUBLE SHOOTING GUIDE
the pump is selected.
The self-priming pump has a definite place in the pump industry, I PROBLEM 1: PUMP NOT REACHING DESIGN FLOW RATE
but care must be taken t o insure that the pump is not misused nor v- 1. Probable Cause:
expected t o operate beyond its capabilities. - l3 Insufficient NPSH. (Noise may or may not be present).
13
Recommended Remedy:
Recalculate NPSH available. I t must be greater than NPSH
required by pump at desired flow. If not:
1.Refer t o Chapter 18.
2. Re-design suction piping, holding number of elbows and
number of planes to a minimum to avoid adverse fluid
rotation as it approaches the impeller.

Probable Cause:
3 System head greater than anticipated.
c.
' 3 Recommended Remedy:

. 3
" Reduce system head by increasing pipe size andlor reducing
number of fittings. Increase impeller diameter .
" 3
w
- @ C. Probable Cause:
Plugged impeller, suction line, or casing: product fibrous or
contains large size solids.

Recommended Remedy:
F o r fibrous material:
1. Reduce length of fiber when possible.
2. Reduce consistency when possible.
3. Consider oversized pump.

For solids:
- 1. If solid's size exceeds the maximum allowable sphere
published for the pump, a larger pump will be required.
&
--
-
w
w
2. Refer to: A. J. Stepanoff, Pumps and Blowers, Two-Phase
Flow, John Wiley & Sons, Inc., New York, 1965.
- 3 139
 - I. Probable Cause:
D. Probable Cause:
" 3 Impeller clearance too large.
Entrained air-leak from atmosphere on suction side.
' 3
Recommended Remedy : *- a Recommended Remedy:
Reset impeller.
1.Check suction line gaskets and threads for tightness. - 3
2. Properly re-pack stuffing box.
3. If vortex formation is observed in suction tank, install '-. 3
vortex breaker. '*-. 3
4. Check for minimum submergence, see Chapter 18.
' 3 PROBLEM 2: PUMP NOT REACHING DESIGN HEAD (TDH)
E. Probable Cause: A. Probable Cause:
En trained gas from process. Check all items listed under "Pumps not reaching design flow
rate".
Recommended Remedy:
Process generated gases may require a larger pump. Recommended Remedy:
Refer to remedies listed under Problem 1.
F. Probable Cause:
Speed too low.

Recommended Remedy: PROBLEM 3: NO DISCHARGE OR FLOW - PUMP RUNNING

Check motor speed against design speed.
A. Probable Cause:

G. Probable Cause:
Direction of rotation wrong.
'3 Not properly primed.

Recommended Remedy:
3
Recommended Remedy:
Reverse any two of three leads on a three phase motor.
'3 Repeat priming operation, recheck instructions.

3 B. Probable Cause:
Suction lift too high.
H. Probable Cause: -. 3
e- Recommended Remedy:
Impeller too small. -
3
A

Recommended Remedy: -
1.

C*
a
w
4
1.Rearrange piping.
2. Increase suction head if possible.
3. Determine if larger impeller would be better.
Replace with proper diameter impeller.
"5
- w 4. Select new pump to handle higher suction lift.
 ',-.I *
Probable Cause:
j P
) PROBLEM 4: PUMP OPERATES FOR SHORT PERIOD, THEN
Impeller, suction line, or casing is plugged. LOSES PRIME
a
Recommended Remedy: 1) A. Probable Cause:
For fibrous materials: Insufficient NPSH.
1.Reduce length of fiber when possible. , 3
2. Reduce consistency when possible. ' 3 Recommended Remedy:
3. Consider oversized pump.
a Recalculate NPSH available. It must be greater than NPSH
required by pump at desired flow. If not:
For solids: , it 1. Refer to Chapter 18.
1.If solid's size exceeds the published maximum allowable
sphere for the pump, a larger pump will be required. 1 3 2. Re-design suction piping, holding the number of elbows and
2. Refer to: A. J. Stepanoff, Pumps and Blowers, Two-Phase number of planes to a minimum to avoid adverse fluid
3 rotation as it approaches the impeller.
Flow, John Wiley & Sons, Inc., New York, 1965.
3
D. Probable Cause:
'
P-1
- 13
' B. Probable Cause:
Entrained air - leak from atmosphere on suction side.
Direction of rotation wrong.

Recommended Remedy:
" I* Recommended Remedy:
Refer to recommended remedy under Problem 1-D.
Reverse any two of three leads on a three phase motor.

E. Probable Cause:
Entrained air - leak from atmosphere on suction side. h
'I*
1 * PROBLEM 5: EXCESSIVE NOISE - WET END

Recommended Remedy: [e A. Probable Cause:

Refer to recommended remedy under Problem 1-D.
't 1133 L
Cavitation - insufficient NPSH available.

Recommended Remedy:
........................................
r- i
Refer to recommended remedy under Problem 1-A.
e
*,
- '3
P
- a
*- 3
k d
"
'k 3

143
 - I3
C
-. I d
B. Probable Cause:
Abnormal fluid rotation due t o complex suction piping.
*I$ Recommended Remedy:
@ l a 1.Work with clean tools in clean surroundings.
Recommended Remedy: 19 2. Remove all outside dirt from housing before exposing
bearings.
Re-design suction piping holding number of elbows and number C,
of planes t o a minimum to avoid adverse fluid rotation as it I
' 3. Handle with clean dry hands.
4. Treat a used bearing as carefully as a new one.
approaches the impeller. 6- ! a I 5. Use clean solvents and flushing oil.
6. Protect disassembled bearing from dirt and moisture.
7. Keep bearings wrapped in paper or clean cloth while not in
C. Probable Cause: use.
Impeller rubbing. "
.--.
3 8. Clean inside of housing before replacing bearings.
9. Check oil lip seals and replace as required.
Recommended Remedy: 10. Check all plugs and tapped openings t o make sure that they
are tight.
1.Check and re-set impeller clearance.
2. Check outboard bearing assembly for axial end play.
r
e3
. 3 c . Probable Cause:
........................................ #-

" 3 Brinelling of bearing identified by indentations on the ball races,

t- usually caused by incorrectly applied forces in assembling the
"
L
3 bearing or by shock loading such as hitting the bearing o r shaft
PROBLEM 6 : EXCESSIVE NOISE - POWER END " 3 with a hammer.
Ir

A. Probable Cause: "e 3 Recommended Remedy:

Overloaded bearing which is indicated by flaking.or spalling of the 3 When mounting the bearing on a shaft use a proper size ring and
bearing raceways. )c
apply the pressure against the inner ring. When mounting in a
3 housing, press against the outer ring. Be sure when mounting a
+
Recommended Remedy:
Check t o be sure that actual operating conditions d o not exceed
the maximum allowable suction and the maximum
-
P'

t,
3
-
bearing to apply the mounting pressure slowly and evenly.

allowable specific gravity. 3 D. Probable Cause:

=
~r
3 False Brinelling of bearing identified again by either axial or
circumferential identations usually caused by vibration of the balls
B. Probable Cause: 3 between the races in a stationary bearing.
Bearing contamination appearing on the raceways as a scoring,
b
- 3
Recommended Remedy:
pitting, scratching, or rusting caused by adverse environment and
entrance of abrasive waste materials from atmosphere.
3 1.Correct the source of vibration.
k
3 2. Where bearings are oil lubricated and employed in units
that may be out of service for extended periods, the shaft
0
a should be turned over periodically t o relubricate all bearing
- 6 surfaces at intervals of one t o three months.
E. Probable Cause:
2. Check all wiring, insulation and rotor windings t o be sure
Thrust overload on bearing identified by flaking ball path o n one that they are sound and all connections are properly made.
side of the outer race o r in the case of maximum capacity 3. Where pumps are belt driven, consider the elimination of
bearings, may appear as a spalling of the races in the vicinity of t h e static charges by proper grounding o r consider belt material
loading slot. These thrust failures are caused by improper that is less generative.
mounting of the bearing or excessive thrust loads.

Recommended Remedy: Probable Cause:

1. Follow correct mounting procedures for bearings. Bearing damage due t o improper lubrication, identified by one or
2. Check pump literature t o be sure that actual conditions d o more of the following:
not exceed the maximum allowable specific gravity o r 1. Abnormal bearing temperature rise.
suction pressure limitations of the particular pump. 2. A stiff cracked grease appearance.
3. A brown o r bluish discoloration of the bearing races.
4. Failure of the ball retainer.
F. Probable Cause:
Misalignment identified by fracture of ball retainer or a wide ball Recommended Remedy:
path on the inner race and a narrower cocked ball path on the 1. Be sure that the lubricant is clean.
outer race. Misalignment is caused by poor mounting practices 2. Be sure proper amount of lubricant is used. T h e constant
o r improperly machined shafts. For example, shaft shoulders not level oiler supplied with Durco pumps will maintain the
square with the shaft centerline, initial shaft bowed due t o proper oil level if it is installed and operating properly.
grinding inaccuracies, or possibly a bent shaft due t o improper In the case of grease lubricated bearings, be sure that there
handling. is space adjacent t o the bearing into which it can rid itself of
excess lubricant, otherwise the bearing may overheat and fail
Recommended Remedy: prematurely.
Check housing bore dimensionally t o be sure that both bores are 3. Be sure the proper grade of lubricant is used. Generally
true with each other. Check shaft to be sure that shoulders are speaking for oil lubrication, use an SAE 3 0 weight oil. F o r
square with shaft centerline and that all shoulder steps are con- grease lubrication a general purpose ball bearing grease with
centric and straight. an operating temperature of -15OF. t o 250°F. is satisfactory.
F o r more specific advice or if bearing temperatures exceed
200°F. consult the local Flowserve Office.
G. Probable Cause:
Bearing damaged by electric arcing identified as electroetching of
both inner and outer ring as a pitting or cratering. Electrical
arcing is caused by a static electrical charge emanating from belt
drives, electrical leakage or short circuting.

Recommended Remedy:
1.Where current shunting through the bearing cannot be
corrected, a shunt in the form of a slip ring assembly should
be incorporated.
PROBLEM 7 : EXCESSIVE LEAKAGE AT STUFFING
BOX-PACKED PUMP Chapter 18
SUCTION PIPING
A. Probable Cause:
Loose packing gland. The majority of hydraulic problems encountered in pumping systems
originate in the suction piping. For this reason, it is important to
Recommended Remedy: provide the best possible suction piping layout. One should adhere
Adjust packing gland while pump is running. to the following piping recommendations whenever possible t o min-
imize NPSH problems and t o maximize pump performance.
B. Probable Cause:
1. Pipe size must be large enough to carry the volume; if not, the
Packing excessively worn or improperly lubricated. high velocity will increase the friction losses. This may be the
case when a larger pump replaces a smaller one, without change
Recommended Remedy: in the suction piping.
Replace worn packing and check for proper lubrication. If the
packing is being lubricated by the pumpage itself, it is mandatory 2. Piping configuration must be such that the liquid is properly led
that there be no solids in the pumpage. t o the pump.

C. Probable Cause: EXAMPLES

Shaft bent (resulting in a whipping action). a) Ideal piping configuration with a minimum of 1 0
diameters of straight pipe between the source and the
Recommended Remedy:
pump suction. (Figure 18.1)
Replace bent shaft.

Probable Cause: - - - --
D. -- - --
--- -.
Shaft or shaft sleeve severely scored.
-- -
Recommended Remedy: -- SOURCE
- 10
Replace shaft or shaft sleeve. ~IAMETERS-

FIGURE 18.1
 " ' 4
a

9
Elevated source (Figure 18.2). Perferred piping to be in 9 Running two pumps from the same source, Figure 18.5, is
one plane, again with a minimum of 1 0 diameters of c)
straight pipe between the elbow and the pump suction. '- 9 the best piping configuration.
' 3
- - -

q- DIAMETERS -.
- -
FIGURE 18.2
FIGURE 18.5

If the elbow bolts directly to the pump's suction, problems

A "Y" fitting (Figure 18.6) should be used in place of a
may occur. (Figure 18.3) The liquid tends to follow the
3 tee, if branching is necessary.
longer radius creating a void or even reverse flow along the ,- 3
shorter radius, thus starving part of the impeller.
' 3

-- -
-
-
CI--

A--- h

- -.* -
-
- \\
- --
--

FIGURE 18.6
FIGURE 18.3
' 9
L
Tee header (Figure 18.7) causes the flow t o become
Avoid situations where there are many elbows in many " 3 turbulent. If the pump suction is close to the tee outlet,
planes. (Figure 18.4)
the turbulent liquid will enter the impeller.

FIGURE 18 4
FIGURE 18.7

150

Bart".:
 .jl
lr

--
'* 1 9
d) Straight o r elbow reducers mounted horizontally also v : s
-- b) Suction openings should be surrounded by sufficient water
permit the formation of voids or air pockets (Figure 18.8). s o the approaching flow lines remain uniform and
Placing a eccentric reducer in the proper manner will
' 3 turbulence does not occur. (Figure 18.12)
eliminate this problem. (Figure 18.9) @- 3
0
'- 3
SUCTION -...

e*,". a
'- 9
C- 3 FIGURE 18.12

FIGURE 18.8 FIGURE 18.9 '-3 The suction line must be deep enough t o prevent forma-
' 3 tion of whirlpools which may allow air t o flow down the
' 3
a-
funnel t o the suction intake. Figure 18.13 shows min-
imum submergence t o avoid surface distortions.
3. Suction inlet plakement, rather than pipe contour, can lead t o ' 3
-.
air pockets due t o entrainment.
-
* 3 4
V)

a) Free-falling discharge drags air bubbles into the reservoir.

(Figures 18.10 & 18.11) If the discharge lines are t o o
-3
3
16
12
close t o the suction inlet, air bubbles are pulled in with the >-
fluid. Solution: move pipes farther apart, or install a ,
3
"' C 8
0
0
baffle arrangement. " 3
L '>w
"3 W
a_
:3 a
z S=MINIMUM SUBMERGENCE, FT.
2
3 L2 FIGURE 18.13
a V)

' 3
( 1 ) Harold Woodhouse, Stevens Institute of Technology, "The
t 3 Suction Side of Centifugal Pumps" Power, May, 1966, Pg. 71-73
3
FIGURE 18.10 FIGURE 18.1 1
:3
=-a
- r
h

" 5 153
 Table 4 - Chemical Elements
*-I
v, ' ,J
Element Symbol
Atomic
Number Element Symbol
Atomic
Number ' I1I ,-*
.
' Table 5 - Conversion Charts
actinium Ac 89 mercury Hg 80 1 ACCELERATION OF GRAVITY (at sea level) (g)
alumin(i)um
americium
Al
Am
13
95
molybdenum
neodymium
Mo
Nd
42
60 1
1, 3 LatIIude. 0 10 20 30 40 45 50 60 70 80 90

antimony Sb 51 neon Ne
argon Ar(A) 18 neptunium NP 93 1 ' s 9780 9782 9786 9793 9802 9807 9811 9819 9826 9831 9832
arsenic
astatine
As
At
33
85
nickel
niobium
Ni 28 I 'ICTE Correcton for ali~ludeabove sea level - 0 003 rnis2 p e r 1000 rn
- 0 003 f t l s e c 2 p e r 1000 R
barium Ba 56 (columbium) 41
7
berkelium
beryllium
Bk
Be
97
4
nitrogen
nobelium 102 I3 8
-
bismuth Bi 83 osmium 76 ' 3 C.h U.S. Customary Units S.I. Units Other Metric Units
boron
bromine
cadmium
calcium
B
Br
Cd
Ca
5
35
48
20
oxygen
palladium
phosphorus
platinum
4
15
8
6
-*
loot
' 3
rod
.
mile naut. meter centimeter kilometer

0 0833 002778 -
- (mi.) -
mile
1 (m)
0 0254
(cm)
2 54 - (km)

californium Cf 98 plutonium
carbon C 6 polonium
cerium Ce 58 potassium
cesium Cs 55 praseodymium
chlorine C1 17 promethium
chromium Cr 24 protactinium
cobalt Co 27 radium
copper Cu 29 radon
curium Cm 96 rhenium
dysprosium DY 66 rhodium PHYSICAL DIMENSIONS - Area
einsteinium EdE) 99 rubidium
erbium Er 68 ruthenium
U.S. Customary Units S.I. Unil Other Metric Units
square
1
europium Eu 63 samarium u*
quan
1001
square
vard
square rrrtular quan squam square hrlomeier squam
mlle ml meter mtlhmelsr renttme1m Ihtlan) Morn,er
fermium Fm 100 scandium in I (V~I acre (ml') (elf m111 (m2) (mrn') (tmq (hm') (tm')
fluorine F 9 selenium ~ 1 5 9 4 x 1 024910x10"
~ 12732x10~ 64517r104 64517 64517
francium Fr 87 silver 2 2 9 6 x 1 0 ~ 35871r106 - 0 0929 92 900 929
gadolinium Gd 64 silicon 0000207 03228 x 10' - 08361 08361 x 10' 08361 xl0'
1 0001563 - 4 047x lo3 -
gallium Ga 31 sulfur 640 1 2 5889 r lo6 -
germanium Ge 32 strontium 5 1 1 1 2 ~ 1 0 ' ~1 05067r10~ 05067x10~ fl5ffi7rtflS
gold Au 79 tantalum
hafnium Hf 72 technetium
hclium He 2 tellurium
holmium Ho 67 terbium
hydrogen H 1 thallium 81 c 3 PHYSICAL DIMENSlOKS - Volume
indium In 49 thorium sn
, - U.S. Customary Units S.I. Unit Other ~ e f i n i t
iodine I 53 thulium 69
iridium Ir 77 tin .sn- 3 cubic lnch US gallon Imp. gallon cubic loot barrel* 1 cubic rneler I liter
iron
krypton
lanthanum
Fe
Kr
La
26
36
57
titanium
tungsten
uranium
74
I I . (US ual)
0 00433
(Imp gal)
0 00360
(Ha)
0 000579 ""' 1
0 000103 ("'
0 00001639 / 001639

lawrencium Lr 103 vanadium 23

54 -
'
lead Pb 82 xenon
lithium
lutetium
magnesium
Li
Lu
Mg
71
3

12
ytterbium
yttrium
zinc
70
39
30 -
+
r
9702
67 024 264 17
34 97
220
5615
35 31 6 2898
0 15897 158 97
1000
manganese
mendelevium
Mn
Md(Mv)
25
101
zirconium
t.' * ' t . ',ade custom one barrel petroleum 0111s equal to 42 US gal
 VELOCITY
DENSITY
U.S. Customarv Units S.I. Unit Other Metric Units
U.S. Customatv Units S.I. Unit Other Metric Unit 1
meters per centimeters meters kilometers
leet per min feet per sec miles per hr sec per sec per min per hr kilograms
(f~m) (ips) (mPh) (mls) (cmlsec) (mlmin) (kmhr) pounds per pounds per pounds per per cubic
US gallon cubic loot cubic inch meter grams per milliliter
1 0.01 667 0.01 136 0.00508 0.5080 0.3048 0.01829
(IbNS gal) (IbW) (lbhn.') (k0Im') (O/ml)
60 1 0.681 8 0.3048 30.48 18.29 1.097 1 7.481 0.004329 119.8 0.1198
88 1.467 1 0.447 44.70 26.82 1.609
0.1337 1 0.0005787 16.02 0.01 602
196.85 3.281 2.237 1 100 60.0 3.60
231 1728 1 27 680 27.68
1.9685 0.03281 0.02237 0.010 1 0.6 0.036 0.008345 0.06243 3 . 6 1 3 ~l o 3 1 0.001
3.281 0.05468 0.03728 0.01667 1.667 1 0.06
8.345 62.43 0.0361 3 1 000 1
54.68 0.9113 0.6214 0.2778 27.76 16.67 1
NOiE: Density expressed in grams per milliliter = specific gravity - grams per cubic centtmeter
NOTE: 1 knot = 1.15155 mph or 1.8532 kmlhr.
Angular Velocity: 1 revolution per minute (rpm) = 0.1047 radians per second PRESSURE OR HEAD
1 radian per second = 9.55 revolutions per minute U.S. Cuslomatv Units S.I. Unit Other Metric Units I
pounds per feet of
CAPACITY square inch water (60°F) inch ol newtons per
U.S. Customary Units S.I. Unit Other Metric Units 1 atmospheres (psi) in mercury (60°F) square meter kilopascal bar
Millions 1 14.096 33.93 30.005 101326 101.31 1.0132
01 US cubic cubic 0 06885 1 2.389 2.042 6895 6.895 0.06895
gallons US gallons Imp. gallons cubic leet 'barrels meters liters meters 0 02947 0.4331 1 0.8844 2987 2.987 0.02987
per day per min per min per sec per hr per sec per min per hr
0 03333 0.4898 131 1 3376 3.376 0.0376
(MGO) (gpm) (iopm) (cts) (bph) (m'ls) (Ilm) (m3hr)
0 000009669 0.00145 0.000335 0.000296 1 0.001 0.00001
1 694.4 578 1.547 992 0.0438 2628 157.72
0 009869 0.145 0.335 0.295 1000 1 0.01
0.00144 1 0.8327 0.002228 1.4286 63.08 x 1O6 3.785 0.2271
0 9869 14.50 33.5 29.5 100000 100 1
0.00173 1.2009 1 0.00268 1.7156 75.77 x 1U6 4.5454 0.2728
0.64636 448.86 373.8 1 641.23 2.832 x 102 1699 101.952 ENERGY, WORK, HEAT
0.00100 0.7000 0.5829 0.00156 1 44.166 x 106 2.6495 0.1590 U.S. Customarv Units S.I. Unit Other Metric Units I
22.824 15 852 13 188 35.316 22 643 1 60 000 3 600 loot pounds British thermal units horsepower-hrs
0.000380 0.2642 0.2198 0.000586 0.3774 16.67x106 1 0.0600 (It-lb) (014 (hp-hr)
0.00634 4.4028 3.666 0.00981 6.2898 2.777 x 1 0 V 6 . 6 6 7 1 1 0.001285 -

'78 2 1 0003930

I
'By trade custom, one barrel petroleum oil is equal to 42 US gal
gpm =m W
x Iblhr when w -density. lblcu It -
0i3i6
3 088
2544
0.0009478
0.003968
1
-
-
MASS. WEIGHT AND FORCE - 341 2 1.341
U.S. Customary Units (avoirdupois)
1 S.I. Unit Other Metric Units I1 POWER
I ounce
(02)
pound
(Ib)
short
ion
long
ton
I
1 gram
(9)
mt:c

loot-pounds
U.S. Customary Units
British thermal \
S.I. Unit Other Metric Units

per sec.
(It-lblsec)
horsepower
(hp)
0.00182 0.001 285
was
(w)
1.356
I kip;Hs
0.001356
metric
horsepower
0.001 84
0.7068 1.014
-78.2 1.415 1.434
,3 7376 1.341 x 10' 9.478 x 1O4 0001 0.00136
-37 6 1.341 0.9478 1000 1.360
542 5 0.9863 0.6971
' 2 T E For Viscosity Conversions see Tables llC-3 & llC-4.
 Table 6 ." -
-iction Loss for Water in Feet Per 100 Feet of Piw *!B 4 INCH
h-, .-- -
f-feet
6 INCH
hf-feet
Table Values for Schedule 40 Steel Pipe @
-
3 Discharge
GPM
V2/2g
feet
per-c100
UI pipe
feet
-:--
u r 111 m ~ r
per 100 feet
~
-f a
.
.
;.
.
m pnpr
v
1 INCH 1112 INCH
3 I0 0.000987 0.01017 20 0.000767 0.00487
V2/2g
feet
hf-feet
per 100 feet
of pipe
Discharge
GPM
V2/2g
feet
hf-feet
per 100 feet
of pipe
9
-
--
3 :i 0.00395
0.00888
0.0344
0.0702
40
60
0.00307
0.00690
0.0164
0.0337
0.0158 0.1 18 80 0.0123 0.0564
0.00214
0.00857
0.114
0.379
2
4
0.00154
0.00618
0.0492
0.164
,'
$
8
3 50
N)
0.0247 0.176 100 0.0192 0.0843
0.0355 0.245 140 0.0376 0.155
0.01927 0.772 6 0.01 39 0.333 0.246
.-. 70 0.0484 0.325 180 0.062 1
0.0343
0.0535
1.295
1.93
8
10
0.0247
0.0386
0.558
0.829
'" 3 80 0.0632 0.4 15 220 0.0927 0.357
90 0.0800 0.5 15 260 0.130 0.487
0.077 1 2.68 14 0.0756 1.53 9 100 0.0987 0.624 300 0.172 0.637

"-
0.137 4.54 18 0.125 2.42 ,. 140 0.193 1.165 400 0.307 1.09
0.2 14 6.86 22 0.187 3.52
1x0 0.320 1.86 500 0.479 1.66
0.420 12.8 26 0.261 4.81 770 0.478 2.72 600 0.690 2.34
0.694
1.036
20.6
30.2
30
34
0.347
0.446
6.26
7.92
"
- 3 760 0.667 3.74 700 0.939 3.13

1.45 41.6 38 0.577 9.78 3 300

400
0.888
1.58
4.89
8.47
800
900
1.23
1.55
4.03
5.05
1.93 54.6 44 0.747 12.9 - 500 2.47 13.0 1000 1.92 6.17
2.48 69.4 50 0.965 16.4 I.' 3 600 3.55 18.6 1200 2.76 8.76
3.09
3.78
4.53
86.0
104.5
124
60
70
80
1.39
1.89
2.47
23.2
31.3
40.5
'1 9 i:;: 4.84
6.32
25.0
32.4
1400
1800
3.76
6.21
11.8
19.4

5.35 146 100 3.86 62.2

.-- '- -3
7.7 1

2 INC
209 130 6.52

INC
103
". 3
re#
8 INCH
hf-feet
10 INCH
hf-feet
I
hf-feet hf -feet Discharge ~2/2g per 100 feet v2/2g per 100 feet
Discharge V2/2g :r 100 ft Discha] 2/2g :r 100 feet G (;PM feet of pipe feet of pipe
GPM feet of pipe GPM 'eet of ~ i ~ e 3