EBOOKS A Simple Guide to Understanding

WATTERSON
FOR THE Compressors
ENGINEERING ENGINEERING TECHNOLOGY COLLECTION
James M. Watterson
LIBRARY
Create your own This straightforward guide to compressors seeks to
Customized Content unveil a lot of myths surrounding compressors. In this
Bundle—the more book, we will be looking at most types of compres-
books you buy, sors, including the centrifugal compressors, the air
the greater your compressors, and of course the most troublesome of
discount! all compressors, the reciprocating compressors.
Having a compressor with minimal operating A Simple Guide
to Understanding
THE CONTENT problems does not only depend on the selection of
• Manufacturing the right type and size for your job. Detailed speci-

A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Engineering fications of all auxiliary equipment and operating

Compressors
• Mechanical
conditions, as well as keeping constant vigilance
& Chemical
Engineering
over the engineering and installation is imperative.
• Materials Science The Simple Guide will explain in a simple yet defini-
& Engineering tive manner which compressor type is best used for
• Civil & which job and what it can produce.
Environmental
Engineering
James M. Watterson has a PhD in mechanical en-
• Advanced Energy
Technologies
gineering and is currently serving as independent
training consultant with JMW, United Arab Emirates,
THE TERMS since January 2003, and as part-time senior lecturer/
• Perpetual access for lead trainer–commissioning and process start-up in
a one time fee the oil and gas refineries with ADNOC since February
• No subscriptions or 2011. He trains groups of maintenance, operations,
access fees instrumentation, and planning engineers. Previously
• Unlimited
he was managing director at JW Management and
concurrent usage
Marketing Consultancy from September 1985 to
• Downloadable PDFs
• Free MARC records October 2000.

For further information, James M. Watterson

a free trial, or to order,
contact: 
[email protected]

90000

9 781947 083707
A SIMPLE GUIDE TO
UNDERSTANDING
COMPRESSORS
A SIMPLE GUIDE TO
UNDERSTANDING
COMPRESSORS

DR. JAMES M. WATTERSON

A Simple Guide to Understanding Compressors

Copyright © Momentum Press®, LLC, 2018.

All rights reserved. No part of this publication may be reproduced, stored

in a retrieval system, or transmitted in any form or by any means—­
electronic, mechanical, photocopy, recording, or any other except for brief
quotations, not to exceed 250 words, without the prior permission of the
publisher.

First published in 2018 by

Momentum Press®, LLC
222 East 46th Street, New York, NY 10017
www.momentumpress.net

ISBN-13: 978-1-94708-370-7 (print)

ISBN-13: 978-1-94708-371-4 (e-book)

Momentum Press Engineering Technology Collection

Cover and interior design by S4Carlisle Publishing Service Private Ltd.

Chennai, India

First edition: 2018

10 9 8 7 6 5 4 3 2 1

Printed in the United States of America

 Abstract

This straightforward guide to compressors seeks to unveil a lot of myths

surrounding compressors. In this book, we will be looking at most types of
compressors, including the centrifugal compressors, the air compressors,
and of course the most troublesome of all compressors, the reciprocating
compressors.
Having a compressor with minimal operating problems does not only
depend on the selection of the right type and size for your job. Detailed
specifications of all auxiliary equipment and operating conditions, as
well as keeping constant vigilance over the engineering and installation is
­imperative. The Simple Guide will explain in a simple yet definitive manner
which compressor type is best used for which job and what it can produce.
Dr. Watterson is not a pure academic and often works at the “Coal
Face” during plant shutdowns and turnarounds, and commissioning. His
experience spans across many continents and varying industries. Com-
pressors is one of the subjects he has trained with many refineries across
the Middle East and Africa.
He has also published “diagnostic skills” and aid to troubleshooting
with compressors, pumps, and turbines.

KEYWORDS

API piping plans, compressor efficiency, compressors, failure modes,

piston rods, reciprocating compressors, rotary-lobe compressors, rotary-
screw compressors, safe operations, surge control, troubleshooting
 Contents

Acknowledgments ix
Chapter 1 Understanding Compressors 1
Chapter 2 Compressor Efficiency 53
Chapter 3 Basics of Surge Control
for Centrifugal Compressors 91
Chapter 4 Lubrication and Compressors 147
About the Author 155
Index 157
 Acknowledgments

I would like to thank a number of people who helped me make this book
possible, most of all my wife, who supplied endless cups of coffee at all
times. The names of other people are too numerous to mention, so I simply
say many many thanks to them.
 CHAPTER 1

Understanding
Compressors

Having a compressor with minimal operating problems depends not

only on the selection of the right type and size for the job. Detailed
specifications of all auxiliary equipment and operating conditions, as
well as keeping a constant vigilance over the engineering and installa-
tion phases, are also essential.
2  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

A compressor is frequently an expensive device, a critical item in a

process, and a rather complicated piece of equipment to specify and pur-
chase because of the many alternatives open to the engineer. Faced with
this wide choice, on what basis can a selection be made so that the chosen
compressor will do the intended job?
Since each type of compressor has specific characteristics, these must
first be understood. Then, the many factors that enter into the final deci-
sion to purchase a compressor and auxiliaries are considered.

COMPRESSOR TYPES

There are two basic mechanical methods of increasing the pressure of gas:
reducing its volume and increasing its velocity so that the velocity energy
may be converted into pressure.
Positive-displacement machines that increase the pressure by reducing
the volume are as follows:

• Reciprocating compressors, which have a piston moving within a

cylinder (Figure 1.1).
• Rotary-screw compressors, in which gas is squeezed between two
rotating intermeshed helices and the casing in which they are housed.
• Rotary-lobe compressors, through which gas is pushed by inter-
meshing lobes.

LEARN HOW A ROTARY-SCREW COMPRESSOR

WORKS!

On this page I will tell you about the pros and cons of the rotary-screw
compressor, what to look for when buying one, and common breakdowns
that might occur during the operation.
The rotary-screw compressor uses two rotors (helical screws) to com-
press the air. There’s a “female” rotor and a “male” rotor. The rotors are of
different shape, but fit each other exactly.
When the rotors start turning, air will get sucked in on one side and
get “trapped” between the rotors. Since the rotors are continuously turn-
ing, the air gets pushed to the other end of the rotors (the “pressure side”)
and new fresh air gets sucked in.
Because this is a continuous process, this kind of compressor doesn’t
make a lot of noise; it runs quietly and smoothly.
Compared to the piston-type reciprocating compressor, the rotary-
screw compressor is much more expensive, but it will use a lot less energy
over the years, resulting in a lower overall cost.
 Understanding Compressors  •   3

Compressor element (oil-free type). (Photo courtesy of Atlas Copco.)

OIL-FREE OR LUBRICATED COMPRESSORS

The rotary-screw compressor is available as oil-injected and oil-free ver-

sions. The basic principle is the same (the rotors “push” the air to one
side), but they are quite different machines.
Oil-inject models are by far the most common oil-injected screw-type
compressors. When you need a lot of air in your workshop or factory, this
type of compressor is usually the best choice!
Oil-free models are used for specific special applications. I’ve mostly
seen them on big factories like oil/gas or chemical refineries, big food fac-
tories, or other places where the compressed air must be 100 percent oil free
(otherwise it could contaminate the food, product, or chemical process).

OIL-INJECTED ROTARY-SCREW COMPRESSORS

How do they work? As its name implies, oil is injected in the compressor
element (where the two rotors turn), during the compression of the air.
What we end up with is a mixture of oil and air under pressure (commonly
about 7 bar).
In a special oil separator, the oil is separated from the air. Most of
the oil is removed by the centrifugal force, the remaining less percentage
of oil is separated by the separator (filter) element (it looks just like a big
4  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

air filter). The separator element should be renewed every 2,000 running
hours or so (depending on the manufacturer/model).
The oil is cooled and fed back to the compressor element to do its job
again. The compressed air, now without the oil, is directed to the pressure
outlet of the compressor, usually through an after-cooler (the air gets very
hot when it’s compressed).
There is no special oil pump to do all this; the oil flows by the pressure
differences inside the compressor.

Example of a small rotary-screw compressor. (Photo courtesy of Atlas Copco.)

Pros:
• Quiet operation
• High volume of air, steady flow
• Low energy cost

Cons:
• Expensive compared to piston compressors
• More suitable for continuous operation only

OIL-FREE ROTARY-SCREW COMPRESSORS

The basic workings are the same as that of the oil-injected screw compres-
sor, only this time, there’s no oil, only air! Because of this, the rotors used
 Understanding Compressors  •   5

are of superior quality with very little space between them. They do not
touch each other though; otherwise, they would wear down too quickly.
Because there’s no oil injected during compression, the compression
is usually done in two stages, because if we would compress the air in one
go from 1 to 7 bars, it would get really, really hot.
Stage one compresses the air to a few bars (say 3.5 bars). The air will
be very hot at this time, so it flows through an inter-cooler first before
entering the second stage. Stage two will compress the air further from 3.5
bar to the end level, mostly 7 bar.
Normally, the two stages will be built on one gearbox, with one elec-
tromotor driving them at the same time.

Pros:
• 100 percent oil-free air

Cons:
• More expensive than the oil-injected type
• Servicing/repairing more difficult and more expensive than the ser-
vicing/repairing of the oil-injected type
• More noisy than oil-injected compressors

An example of a big rotary-screw compressor (old-fashioned oil-free type).

(Photo courtesy of Atlas Copco.)
6  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

FAILURES

Rotary-screw compressors can fail in a number of ways. The most com-

mon problem is oil in the compressed air. Most of the times this happens
because the oil separator is not doing its job properly. The chances are that
the separator element is saturated with oil (read: you didn’t service the
compressor in time! There’s a fixed time limit [running hours] to change
the element!).
Another problem often encountered is water in the compressed air.
Since the compressor takes in a huge amount of air (with water vapor)
and compresses it to seven times a smaller volume, a lot of water will be
produced. Normally this water is drained using an electronic or mechan-
ical automatic drain. If this drain is broken, the water will stay in your
compressed air and fill up your air receiver and piping.
If the problem persists, it will most probably be a defective pressure
switch (which will start/stop, load/unload the compressor), or a defective
inlet valve (which opens and closes the air inlet of the compressor). If it’s
closed, the compressor is running in an “unloaded” condition and won’t
supply any air.

ROTARY-SCREW AIR COMPRESSORS

Rotary-screw compressors are another variety of compressors available to

industrial and DIY compressed air users.

CONCEPT

If you have ever used an electric drill in a piece of wood or metal, you may
have noticed that the chips or spirals of metal follow the contour of the
flutes up and out of the hole (most of them, at least).
A similar phenomenon occurs inside the rotary-screw compressor
housing.
At the wide end of the screw (sometimes there is one screw operating
against a housing, sometimes more than one), an inlet valve allows free
air into the screw housing when there is a demand. Free air flows into
the housing from the outside as there is a partial vacuum formed inside the
rotary-screw housing as the screw(s) rotate.
Inside the screw housing are the screws in a bath of oil. The oil is
there to provide a viscous, flowing, sealing method to help trap the air in
the rotary-screw flutes.
 Understanding Compressors  •   7

The air–oil mixture in the screw housing moves along the flutes from
the wide end of the screw toward the narrow end, pulling a vacuum behind,
thus drawing more air into the screw housing.
As the air–oil blend is pulled along the flutes of the screw, the space
in which the air is contained gets smaller and smaller. The diameter of
the screw is larger at the inlet end and smaller at the discharge end, thus
compressing the air. The amount of air trapped in the screw flutes does not
change as the air is moved along the narrowing path, but the volume that
air is in gets steadily smaller, thus compressing the air.
Manufacturers of rotary-screw compressors have their own ideas
of what constitutes the ideal geometry of the screw within their air
compressor.
Rotary-screw compressors may have just one screw (also sometimes
known as augers) or maybe two or more. Single-screw compressors func-
tion the same way as multiple-screw units, with the air being compressed
between the housing of the screw compartment and the screw ­itself, rather
than between two or more screws.
8  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

The following drawing will give you an idea of how the rotary-screw
concept works with two screws. The actual guts of the rotary-screw com-
pressor will vary depending on the designs of the company that manu-
factured that particular compressor. The drawing shows two screws. They
would be housed inside the screw compartment of the compressor, in a
bath of oil.
At the narrow end there would be an outlet valve, which feeds the com-
pressed air–oil mixture from the screw compartment and into a separator.
The separator has the job of removing as much oil from the com-
pressed air as possible, and then to release that compressed air into the
compressor receiver or into the plant main air lines.

FEATURES/COMMENTS

Although rotary-screw compressors are available in lower horsepower rat-

ings, it would appear that it is in the 20 to 25 horsepower and higher range
that many industrial compressor applications tend to move toward using a
rotary-screw compressor solution instead of other types of compressors.
One major manufacturer states that the rotary-screw air compres-
sor has become the most popular source of compressed air for industrial
applications.
That may be because of the need for a compressor with a continuous
duty cycle. Some rotary-screw compressor manufacturers claim a duty cycle
of 24/7/365, which is pretty incredible for any electromechanical device.
Rotary-screw compressors are available with a direct motor-to-screw
drive; others are belt driven. Each has its benefits and its own draw-
backs, the details of which are best obtained from the actual compressor
manufacturer.

Less Maintenance

The perception, one that is claimed by some manufacturers, is that

rotary-screw units have the least maintenance issues of all types of air
compressors and are touted as being the easiest to maintain in terms of
both regular maintenance and unscheduled downtime.

Reputed for Lower Cost

When you move up into the higher horsepower units, rotary-screw units
are reputed for their lower cost over a comparably sized reciprocating
 Understanding Compressors  •   9

compressor, and further, they boast lower operating costs than either vane
or reciprocating units.

Oil Carryover

Some manufacturers suggest their oil carryover from the compressor to

the compressed air supply of the plant is calculated in parts per million
per day, rather than the ounces or more of oil that can enter the plant-air
stream from older reciprocating models and some well-used vane models.

Lower Operating Noise

Other firms suggest that their rotary-screw units boast a very low oper-
ating decibel rating, and claim noise output levels far below other types
of compressors, an important issue to be considered for the benefit of
workers in the area. It is our experience that the lower operating sound
levels may not eliminate the need for a soundproof housing, unless the
compressor is well equipped with sound-deadening cladding.

Know Your Air Requirements

As noted earlier in this chapter, when you are moving toward selecting
a rotary-screw (or any compressor type), you first need to know how
much air you’ll need in cubic feet per minute (CFM) at the psi you need
for your plant, your tools, and all ancillary equipment for now and for
the future. You’ll want to build in excess volume available, as one sta-
tistic we’ve seen says, on average, more than 10 percent of all compres-
sor capacity is lost through leaks, despite the best efforts of the plant to
reduce wasted air.
Once you got the compressor size figured out (link; I hope the in-
formation here will help), you will want to review the up-front cost of
compressors from a host of manufacturers. Check their mean time,
between-failure rates, their parts and service costs, the life expectancy of
the unit with the duty cycle you will require, and the particular operating
costs.
Compressing air is expensive, and one compressor might provide
lower up-front capital costs, yet end up being far more expensive in the
longer term due to higher operating costs.
All factors having been considered, and certainly this is claimed by many
of the manufacturers of the rotary-screw type of compressor, the rotary-screw
compressor may surface as your best choice for industrial application.
10  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

• Sliding-vane compressors, where an eccentric cam (into which

sealing vanes slide) rotates inside a housing.
• Liquid-piston type, in which a partially liquid-flooded case creates
the equivalent of sliding vanes.
• Diaphragm compressors, in which a flexible diaphragm is pulsed
inside a concave housing.

The two types of compressors that convert velocity into pressure are:

• Radial-flow compressors, generally called "centrifugal compressors”

• Axial-flow compressors, known as "axial compressors”

In centrifugal compressors, the gas enters the eye of the impeller, and
the rotative force moves the fluid to the rim of each wheel or stage. Diffus-
ers convert the velocity head into pressure, and return passages are then
used to lead the gas to the compressor discharge or to the next impeller
stage.
In axial compressors, flow occurs through a series of alternating
rotating and stationary blades, and in a direction basically parallel to the
compressor shaft. Each passage through the rotating blades increases the
velocity of the fluid, and each passage through the stationary diffuser
blades converts the velocity head into a pressure head.

SELECTION CONSIDERATIONS

Not all types of machines are made in all pressure–volume ranges.

Figure 1.1 indicates, in a very general manner, the capacities of recip-
rocating, centrifugal, rotary-screw, and axial compressors available. The
more common usage is indicated by the deeper shading. Although this
figure does not indicate the theoretical or engineering limits of any design
(the limits are continually being expanded), it may be used as a guide to
current technology.
Since sealing systems for axial compressors are not as versatile as for
other types, normally only those gases whose leakage to the atmosphere
can be tolerated should be handled by this type of machine.
Rotary-lobe, sliding-vane, liquid-piston, and diaphragm compres-
sors have relatively small capacities and, typically, atmospheric-pressure
suction. Of these four types, the rotary-lobe compressor can deliver
the most gas, as its maximum suction volume is about 30,000 actual
cfm (acfm). A maximum discharge pressure of about 40 psia can be
 Understanding Compressors  •   11

Figure 1.1.  A process reciprocating compressor operates from vacuum to over

36,000 psi

attained. However, rotary-lobe compressors are most competitive

at capacities of 17,500 acfm or less and discharge pressures of about
22 psia.
Maximum inlet capacities of sliding-vane units are about 3,000 acfm,
or double this amount if a duplex compressor is used. The latter consists of
two compressors attached to a single drive. Maximum discharge pressures
of standard machines are about 65 psia in a single stage and 140 psia in
two stages.
The liquid-piston compressor has a maximum capacity of about
10,000 acfm and can deliver this amount of air (or gas) at about 30 psia.
Volumes of 300 acfm or less can be compressed to about 115 psia.
The three foregoing compressor types can produce moderate to high
vacuum, particularly in multiple stages.
Diaphragm compressors have much smaller volumetric capaci-
ties, with maximum flows ranging from 40 to perhaps 200 acfm. These
machines, however, can develop pressures up to 40,000 psi.
Before selecting a compressor type, one must decide how many
machines will be needed to handle the process load. In former years, recip-
rocating machines were used for almost all process applications. Since
the compressor capacity was low, large plants would require trains of
machines. As machine reliability and capacity increased, the tendency to
install two machines started, each with 55 percent or 60 percent capacity,
perhaps with a third unit as a spare.
The spare unit ensured operation at full capacity, but at an increased
compressor cost of about 50 percent. If the spare compressor were omit-
ted, but two half-size machines installed, one could still be reasonably sure
12  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 1.2.  The pressure–volume operating range (discharge pressure vs. feed
volume) of various compressors

of continued operation at all times. This was particularly important when

the process included equipment that could not be shut down frequently,
such as furnaces.
Later, to take advantage of larger machine capability, several services
were placed on the same frame.
Today, the situation is somewhat different, as more centrifugal com-
pressors are being used (Figure 1.3). For one thing, the downtime of a
rotating equipment generally is appreciably less than that of a recipro-
cating equipment. Therefore, in many instances a single centrifugal com-
pressor may be satisfactory. However, it must be recognized that when
a compressor is down, it will usually take longer to repair or overhaul a
centrifugal unit than a reciprocating one—unless a complete spare rotor
is available.
Also, the pricing structure of centrifugal compressors is quite differ-
ent from that of reciprocating ones. As a first very rough approximation,
one may assume that halving the size of a reciprocating compressor will
halve its cost. Yet, halving the size of a small centrifugal compressor may
only decrease its cost 20 percent, and halving the size of a large machine
may only reduce its cost 30 percent.
 Understanding Compressors  •   13

Figure 1.3.  A train of three barrel-type centrifugal compressors in a

1,400-ton/day ammonia plant

Furthermore, because of their flat operating characteristics, the run-

ning of centrifugal compressors in parallel may result in surging unless
very careful attention is given to avoiding unstable operation. Therefore, in
many process applications for which one centrifugal compressor will have
adequate capacity, an installed spare is not provided. In these instances, a
complete spare rotor may be bought.
The choice between reciprocating and centrifugal compressors is not
always simple, particularly for high-head, medium-capacity service such
as gasfield repressuring. If several reciprocating compressors are used,
each can be multiple staged to develop the desired head. The shutting
down of one machine would merely cause a decrease in plant output. But
if several centrifugal compressors were used in series, the failure of one
would stop the entire operation.

OPERATING CHARACTERISTICS

A positive-displacement compressor is characterized by a pressure rise–

volume curve that is almost vertical. (It is not completely vertical because
14  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

there is mechanical clearance, and slip and leakage from the discharge to
the suction; the slip increases as the compression ratio rises.) The com-
pressor delivers its gas against any pressure head up to the limit of its
mechanical strength and drive capacity. Capacity is almost directly pro-
portional to speed.
The characteristics of a centrifugal compressor are apprecia-
bly different. Generally, the pressure rise–volume curve is quite flat
(Figure 1.4a). (It may be somewhat steeper if a heavier gas is being com-
pressed.) A small change in the compression ratio produces a marked
effect on the compressor output. As the discharge pressure increases, the
flow ­decreases, and if the flow decreases too much, the machine will start
to surge.
Surging occurs when the velocity of gas leaving an impeller wheel is
too low to move the fluid through the machine. With no gas leaving the
impeller, the discharge pressure may drop. Should this occur, the machine
will again start to compress gas, and the cycle will be ­repeated. Such
intermittent operation may severely damage a compressor. The charac-
teristic curve can be modified by the installation of adjustable inlet guide
vanes (Figure 1.4b). These are most effective on machines having a few
stages. Adjustable diffuser vanes have been used on some machines.
In some installations, process requirements may dictate that the com-
pressor be run at the far right of the characteristic curve, where it is very
steep. Operating in this area requires careful control and is accomplished
at some penalty of compressor efficiency.
The volumetric capacity of a centrifugal compressor is almost directly
related to its speed; its developed head, to the square of the speed. (The
horsepower requirements are thus related to the cube of the speed.) The
efficiency of centrifugal compressors is lower than that of reciprocating
machines by perhaps 5 to 20 percent.
These characteristics establish the sensitivity of the compressor to
variations in flow conditions. For example, a change in the density of the
fluid being compressed will have little effect on either the volume of gas
pumped or the discharge pressure developed by a reciprocating machine,
although one would have to be sure that no component parts of the com-
pressor were being mechanically overstressed. Any variation in the density
of a gas being compressed will result in a proportionate change in the
weight of gas pumped.
On the other hand, because the head developed by a centrifugal com-
pressor depends only on the velocity developed, a change in gas density
will be directly reflected by a proportionate change in the developed dis-
charge pressure. However, at a given density, if the discharge pressure can
 Understanding Compressors  •   15

be permitted to change slightly, one can obtain large variations in volumet-

ric gas flow through the compressor.
The axial compressor has a very steep characteristic curve (Figure 1.4c).
The unit’s surge capacity is thus close to its operating capacity. However,
by providing a method of adjusting the angle of stator blades and inlet
guide vanes, a greater operating range can be obtained (Figure 1.4d).
Generally, the efficiency of an axial compressor exceeds that of
a multistage centrifugal machine by perhaps 5 to 10 percent. The ax-
ial compressor does not contain diaphragms that expand radially as
the compressed gas gets hot. This mechanical factor, combined with
higher efficiency, leads to greater freedom from temperature limits and
permits a higher compression ratio per case than do centrifugal units.

SPEED CONSIDERATIONS

The type of mechanical drive (including gears) that is used may influence
the choice of compressor.
Compressor and drive speeds are very pertinent if one wishes to avoid
gearing. The accompanying table 1.1 provides speed ranges of the most com-
mon types of compressors and drives. There are specially designed units, how-
ever, that do not fall within the ranges listed. One of these, for example, is a
carbon dioxide compressor with a suction volume of approximately 50 acfm at
the last wheel, which rotates at 25,000 rpm and delivers gas at 5,000 psi. The
tip speed of this compressor’s impeller is approximately 650 ft/sec. The com-
pressor itself is directly driven by a specially designed 1,000-hp steam turbine.

COST COMPARISONS

In very general terms, at low pressures and large flows, the purchase cost of a
reciprocating compressor may be estimated to be perhaps twice that of a cen-
trifugal machine of the same capacity (Figures 1.5–1.7). The cost differences
narrow as pressure increases or actual flow decreases. At high pressures and
low flows, costs may be quite close to each other. A reciprocating compres-
sor will need a more massive foundation, more protection from the envi-
ronment, and a more careful piping design to avoid vibration and pulsation.
On the same rough basis, one may estimate the costs of rotary-screw
and axial compressors to be about the same or less than that of centrifugal
units. In their most suitable applications, the costs of the screw and of the
axial compressors may be considerably lower.
16  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Table 1.1.  Speed Range for Compressors and Drives

Usual Speed
Compressor Types Range, Rpm. Remarks
Large reciprocating 300-600 Some even 1,000 to
compressors 1,500 rpm.
Small reciprocating air and 1,000-1,500
refrigeration machines
Rotary-screw 3,000-10,000
Process centrifugal units 3,000-12,000 Some large-
horsepower
machines up to
17,000 rpm.
Special, small-volume high- 30,000-50,000
head air centrifugals
Axial compressors 3,000-6,000 Some up to
16,000 rpm.
Large internal-combustion 300-600
engines and reciprocating
gas expanders
Small rotary and radial 3,000-8,000
engines
Mechanical-drive gas 10,000 or less Small gas-turbine
turbines and centrifugal compressor drives
expanders (over 1,000 hp.) have operated at up
to 50,000 rpm.
Mechanical-drive, back- 16,000 or less Condensing turbines
pressure steam turbines have lower
(3,000 to 40,000 hp.) maximum speeds.
Electric motors 3,600 or less

COMBINATIONS OF COMPRESSORS

At times, combining compressors may be worth considering. For example,

in compressing to a very high pressure, it may be possible to use a centrif-
ugal machine or a rotary-screw machine for a lower pressure and then pipe
the gas to a reciprocating unit. In some instances, axial and centrifugal
 Understanding Compressors  •   17

Figure 1.4.  Operating characteristic curves for centrifugal and axial compressors

Figure 1.5.  The cost of centrifugal compressors according to the inlet volume
and head output (head is feet of gas)

impellers may be placed on the same shaft. In addition, one might also
resort to placing axial- and centrifugal-compressor cases in a common
drive train.
As an alternative to an axial compressor, three or four single-stage
centrifugal compressors may be connected by a gear train to a single drive.
With the gas cooled after each stage of compression—and gears designed
to permit each stage to be run at its optimum speed—the efficiency of
18  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 1.6.  The base price of reciprocating and centrifugal compressors

according to horsepower unit

these centrifugals is comparable to that of an axial compressor, while their

operating characteristics are those of a centrifugal machine. Extensive
gearing, however, is a distinct disadvantage.
Additional significant characteristics of each type of compressor will
be discussed in a forthcoming article.

DRIVING MECHANISMS

Smaller units are usually electric-motor driven (direct or belt); for

medium to large units, a wide choice of drives are available. These include
motors (synchronous, induction, low or high speed); steam turbines
(back-pressure, condensing or controlled extraction); internal-combustion
engines (integral or direct connected); gas turbines (single or double shaft);
and expanders.
The selection of a drive depends to some extent on the compres-
sor service, but more important are the overall energy balance, energy
 Understanding Compressors  •   19

Figure 1.7.  The effect of operating pressure on the cost of centrifugal and
reciprocating compressors

utilization and availability, and heat-rejection methods. Within the limits

imposed by these criteria, the selection should stress a drive system that is
simple, dependable, and straightforward. The compressor is the reason for
the drive, not the other way around.
The drives of internal-combustion engines and steam turbines can
ordinarily be operated over a fairly large range of speeds. This may not be
the case, however, if gas-turbine or electric-motor drives are used.
Let us first consider gas turbines, almost all of which have axial-type
air compressors for pressure–air supply to the compressors. For single-shaft
units (air compressor, gas turbine, and driven unit on one shaft), the speed
range is most often determined by the steep performance curve of the
axial compressor rather than by the much flatter curve of the centrifugal
process compressor. Double-shaft machines permit constant speed for the
axial air compressor and variable speed for the process c­ ompressor. The
selection of sizes, speeds, and horsepower outputs of commercially avail-
able gas turbines is limited. Very often, a wide freedom of choice is not
available as to single- or double-shaft units.
Motor drives usually have a constant speed: In a limited number of
cases, variable-speed couplings, wound-rotor or multipole (PAM) motors
may be used. Large motors may be of the synchronous or induction kind.
For a unit driven at above-synchronous speeds (3,600 rpm for 60 Hz),
the choice should be based on the total cost of the motor and the speed
increaser. Thus, a 1,800-rpm induction motor and its speed increaser may
cost less (including operating costs) than a 1,200-rpm synchronous motor
with its speed increaser.
20  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Constant-speed centrifugals in process plants tend to operate at a high

enough average load so that the economic rewards of power-factor correc-
tion obtained by the use of a synchronous motor are minor.
Fossil-fuel drives are used when initial and operating costs are more
attractive than those of steam or motor drives, sufficient electric power is
not available, and electric or steam sources are not reliable. In this last case,
the entire system must be carefully specified to ensure that minor items
such as cooling-water pumps, pressure switches, control air, etc. are inde-
pendent of any source of power, not as dependable as the compressor drive.
Internal-combustion engines are usually turbo-supercharged and may
be two or four cycle, integral with or separate from the compressor. The
type of engine can normally be selected on the basis of drive features,
including accessories and costs (purchase order, installation, fuel con-
sumption, spare parts, maintenance) independent of the compressor. It is
best, however, to include the drive as part of the compressor system.
The gear mechanical rating, including the American Gear Manufac-
turers Association (AGMA) service factor, should be selected so that the
gear rating does not become the limiting factor in the compressor and
drive train. Steam-turbine drives combined with a gear (with the turbine
at a lower speed than the compressor) are sometimes lower in cost than
higher-speed turbines. The policy of the user and his insurance carrier on
warehouse spares for gears affects the choice, since the gears and addi-
tional couplings increase the probability of outage.

MARKETING OF COMPRESSORS

Compressor systems and their drives range from small through m ­ edium
to large. Marketing methods range from those suitable for catalog items
handled by distributors to engineered systems oriented to specific
market areas such as chemical processing, gas distribution, petroleum
refining, and electric-power generation. Engineers holding discussions
with equipment suppliers, manufacturer representatives, or suppliers
of package or skid-mounted units should recognize these elements of
supplier organizations; such knowledge will aid in establishing the
scope and the detail most useful to include in purchase requisitions and
specifications.
Catalog items require little more than hardware description, as per-
formance is specified in published information. Engineered systems, on
the other hand, require the definition of performance requirements for
the overall compressor system. Hardware definitions also are needed to
establish the quality level of the system and its components, and to define
 Understanding Compressors  •   21

the number and type of auxiliary elements such as oil coolers, governor,
pumps, etc. Other hardware items include controls, heat-rejection sys-
tems, drives, gears, piping, ducts, and electrical wiring (cable, conduit,
trays, etc.).
The environment (indoor or outdoor) should also be made clear so
that due allowances can be made for access for construction and mainte-
nance, sound control and isolation, and area electrical classification.

SYSTEM SPECIFICATIONS

The form of a purchase specification should be the one most familiar to,
and most commonly used within, the issuing organization; typical forms
are given in API standards 617 and 618. Here we shall consider the content
of the specification.
Performance criteria must be carefully defined for the end use that
the compressor must have within the overall system. Following should be
included in the specification:

• The range of mass and volumetric flow as influenced by variations

in the inlet temperature, pressure, molecular weight, gas composi-
tion (vapor loading, compressibility factor, etc.), discharge pres-
sure, temperature, and flow of cooling fluids (water, air, etc.)
• Startup, standby, and shutdown conditions of the compressor and
of the entire system
• Mention of even traces of vapors, liquid droplets, dusts, or gases
that may be minor items for the chemistry of the process but may
cause fouling, gunking, seal problems, etc., either by themselves or
when mixed with lubricants or sealing fluids (items such as these
may appreciably influence the choice of the compressor type)
• Range of ambient temperature
• Altitude
• Area electrical classification
• Applicable codes and standards from such organizations as the
Tubular Exchanger Manufacturers Association (TEMA) and the
American Society of Mechanical Engineers (ASME)

Purchase specifications must define quality requirements for auxil-

iary equipment such as seals, piping systems (material and arrangement),
type and quality of control elements and systems, level of redundancy,
and shop testing (if any). Checklists for such items may be prepared
based on the available knowledge within an organization, as well as on
22  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

accepted references on the inspection of completed installations, such

as Chapter X of the API “Guide for Inspection of Refinery Equipment.”
Purchase specifications—or those prepared by the customer for use by
an engineer-constructor—should not limit the bidders from using their
knowledge and experience.
Required controls cover a very wide range of supply. The compressor
specification should include all elements that directly measure and control
any part of the compressor system. This includes local panels, receivers
from external inputs, and any items to provide outputs to external devices.
Devices for volume control as such—or those used to control mass flow
and provide anti-surge control—must be carefully defined as to which ele-
ments are supplied as part of the compressor system and which elements
are external. Thus, such items as inlet guide vanes are best included in the
compressor system, while anti-surge and recycle devices are usually best
considered as external to the system.
An increasing area of interest for controls consists of the types of
diagnostic devices used to measure, indicate, alarm, and record vibration
(velocity and displacement), axial movement, bearing temperatures, and
drive-motor-copper temperatures. Axial-movement and motor-copper
temperature indicators are best used for both alarm and shutdown. Other
instruments are most suitable for alarm only and as trend indicators. The
compressor supplier is in the best position to select the points of pickup
and recommend types of pickup and readout devices. Controls for units to
be attended only by remote or occasional local surveillance require very
careful attention.
Job cost and completion time is improved with proper use of
shop-assembled units. Typical packages, including skid-mounted units,
comprise refrigeration (chilled water and low-temperature brines) and
complete instrument and plant-air units. Purchase specifications should
therefore call for or permit packaged units to be offered where feasible.
The units should be such that they need merely to be set on simple founda-
tions, and have the power, cooling-water, and supply and discharge piping
connected.

FACTORS THAT MAY BE OVERLOOKED

Perhaps the most important single factor to determine is whether a pro-

posed piece of machinery has been used in a similar service and what its
history has been. This by no means implies that one should never install
a newly designed compressor or use an older design for a new applica-
tion. The first to use a new design may enjoy an advantageous position
 Understanding Compressors  •   23

before competitors follow suit. Furthermore, if the decision to begin the

construction of a plant cannot be delayed, it may be a matter of either
installing something new or installing a proven design that one realizes
may soon be outmoded. Features of a design that have not been proven
should be reviewed in detail, and perhaps consideration should be given to
courses of action to be taken if unexpected difficulties occur. Attempting
to foresee possible failures and developing corrective courses of action
is very time consuming, complex, and usually not very rewarding. But
when trouble does occur, such planning may more than compensate for
the effort spent.
If a compressor and its drive are investigated separately, one should
not overlook the direction of rotation (which may not always be change-
able) and its effect on gear requirements. One supplier should be given the
responsibility of completing a combined torsional and vibration analysis
of the entire system. An agreement as to who will do this work should be
reached as soon as possible, so that any required design changes can be
made with minimum difficulty.
Similarly, it is advisable to have one supplier assume the responsibil-
ity for collecting and correlating design data pertaining to noise emission
and for making the final recommendation for noise suppression. The best
procedure is to select one supplier and have him assume responsibility for
the overall unit.
When reviewing vendors’ proposals, there are several items, in addi-
tion to price and energy requirements, that should not be overlooked.
One may question what spares the supplier will generally have in his
shop. If the purchaser does not buy a spare rotor, how long will it take
to obtain a new one in the case of an emergency? Is there another com-
pany that might share the cost of a standby spare? Such sharing has not
been generally accepted because even though costs are reduced, the risk of
extended outage is increased. Nevertheless, this course of action may be
worth considering.
When a centrifugal compressor is bought, one should check the close-
ness of the operating point to the surge point. On low-capacity, high-head
wheels, these points may be quite near each other. It may not be possible
to reduce speed very much without resorting to bypass control or to the
installation of suction or discharge valves, etc. Critical speeds should also
be reviewed to be sure they are far enough away from any desired operating
speed, particularly if operation at reduced capacities may be considered.
Allowable noise levels are a function of frequency as well as intensity,
and levels must be lower at higher pitches. The amount of noise generated
depends on the type of compressor, its horsepower rating, compression
ratio, speed, etc. Silencers or acoustic coverings may be used to reduce the
24  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

emitted noise to levels acceptable by the user and by regulatory author-

ities. Methods of estimating compressor noise levels, and the effect of
various kinds of silencers, are detailed in the literature.
Drive and compressor characteristics must be evaluated. Operating at
the maximum continuous speed of the centrifugal compressors, which is
gen­erally 5 percent greater than the speed at the compressor rated point,
may call for a power expenditure 15 percent greater than the rated horse-
power. If the compressor is driven by a turbine, increasing the latter’s
capacity by 15 percent may impose a significant increase in the cost of
the drive and its auxiliary facilities, as well as a penalty of poorer effi-
ciency when running the drive at the rated compressor speed. Normally,
the permissible maximum mechanical speeds of the drive and the driven
unit should be the same.
It may also be good to review guarantees. In general, the following
applies (API Std. 617, 618) unless other representations are made: For
centrifugal compressors operating at a constant speed, the capacity is
guaranteed; the head may vary within +5 percent and −0 percent of that
specified; the horsepower (when corrected to the specified head-capacity
conditions) may vary by not more than 4 percent of the stated horsepower.
For centrifugal compressors operating at a variable speed (i.e., gas- or
steam-turbine drive in most instances), the capacity and the head are guar-
anteed but the speed is not. Horsepower may vary ±4 percent.
For reciprocating compressors, one may specify a guaranteed capacity
with no-negative tolerance, as well as a guaranteed maximum horsepower
and a specified speed. However, process industries frequently accept a
manufacturer’s capacity guarantee of ±2 to 3 percent rather than pay more
for the no-negative tolerance. When motor drives are to be used, obviously
one must review compressor-speed guarantees, much more carefully than
when turbine drives are involved.

INSTALLATION ENGINEERING

After the compressor unit is selected and a purchase order issued and
accepted, the next steps require continued vigilance. This is not simple
be­cause many more people in the engineering and supplier organizations
now become involved. Follow these guidelines to prevent certain items
from being neglected:

1. For all but simple catalog units, prepare a process and instrumen-
tation diagram or an engineering flow diagram for the complete
compressor system.
 Understanding Compressors  •   25

2. Establish the layout requirements including those determined by

operator assignment, that is, the number of operators assigned to
the compressor during normal operations, during startup only, etc.
Provide terminals for remote control if an operator will not be in
attendance at all times.
3. If a large, complex compressor is involved, hold meetings with the
supplier’s engineering group to establish schedules for the submis-
sion, review, and approval of the supplier’s engineering data and
drawings, and tentative plans for use of his servicemen during
installation and startup.
4. Review the compressor manufacturer’s drawings and those of
his suppliers to ensure that quality and performance criteria are
being met.
5. Review the torsional-vibration analysis and lateral critical studies
completed by the compressor and drive supplier to make certain
that no contemplated operating condition will cause the machine to
operate at a hazardous speed.
6. Review unpriced supplier orders. (Priced orders would not be made
available and are unnecessary.)
7. Review control plans, including startup, normal operation, sched-
uled and forced shutdowns, protective and safety devices for alarm
and shutdown, and the duties to be assigned to the operators.
8. Submit to the supplier, for his comments and review, design bases
and installation drawings for foundations, piping, and pipe sup-
ports. Such information should include the calculated forces and
moments (hot and cold) exerted by the piping on the equipment
flanges. Guidelines for allowable values are established by the com-
pressor and turbine suppliers on their outline drawings.
9. Review requirements for shop and field pressure and performance
testing. In most applications, test procedures established by the
supplier are sufficient. Establish procedures for shop erection and
match marking of the prefabricated pipe to be furnished by the
compressor supplier. Establish what shop tests are to be witnessed.
10. Establish requirements for such items as operating and maintenance
access (cranes, monorails, etc.); noise control within buildings and
other enclosures; protection from fumes and dust; winterizing; and
piping systems, including drains, vents, and access for field flush-
ing and cleaning.
11. Provide methods so that the installed dimensional accuracy of piping
right at the compressor is high, and thus compatible with the level
of dimensional exactness required by the machinery. Neglect of this
may require field changes in the piping arrangement to secure and
26  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

maintain the acceptable compressor alignment. Arrange the major

piping so that supports can be taken from concrete substructures
rather than from elevated steel structures. This is most important in
reciprocating compressors because it helps in attenuating vibration.
12. Make provisions for shop inspection during fabrication and assem-
bly, as well as during shop testing.
13. Obtain copies of expediting and inspection reports. Monitor deliv-
ery schedule.

VIGILANCE DURING INSTALLATION

Job pressures during installation, run-in, and startup create many hazards
to the achievement of quality results from the compressor system. To min-
imize the hazards, follow these guidelines:

Manufacturer’s Representatives—Err on the side of too much partic-

ipation in the field of representatives of the supplier’s service orga-
nization and of his major sub-suppliers (turbines, gears, motors).
Most suppliers include a specific number of days for such repre-
sentatives in the original proposal price (and thus in the purchase
order), with a per diem rate beyond this limit. Do not save these
“free” days for when troubles may be encountered. Use the days to
avoid trouble. Do not hesitate to obtain such service beyond the limit
of days included in the purchase order for such items as machine
setting and grouting, alignment, initial run-in, and actual startup and
demonstration.
Field Checking—During field checking of all kinds, not only con-
struction drawings but also flowsheets, operating manuals, etc. must
be consulted.
Foundations and Superstructures—Check these vital elements of the
system, using supplier drawings and installation recommendations,
as well as the engineer’s construction or working drawings. Dimen-
sional accuracy and quality of construction are both vital.
Materials of Construction—Check to see whether mill certificates
have been received and whether they are acceptable.
Piping, Ductwork, and Supports—Ascertain accuracy and structural
adequacy, so that excessive loads are not imposed on equipment
flanges. Also make sure that provisions are made for controlled
movement due to thermal expansion; proper line slopes are main-
tained; noise and vibration are attenuated; and resonant conditions
 Understanding Compressors  •   27

are avoided. Conduct hydrostatic and leak tests, remove all tempo-
rary blanks, and install rugged line strainers.
Flush, degrease, mechanically clean, and, when pertinent, clean with
chemicals. Chemical cleaning (pickling) of carbon steel piping for
such services as lubricating oil should also be done. Chemical clean-
ing of field-assembled systems should be considered very carefully
because the results can be very hazardous if any of the chemical
solution enters the compressor during startup operations. Proper
drainage and venting provisions are essential for effective and safe
removal of cleaning fluids. Chemical cleaning, moreover, does not
prevent trouble from mud, stones, welding rods, and slag.
Small piping systems such as sealing, venting, drain, and control con-
nections must be checked for continuity and completeness. Fre-
quently, these systems are not given enough attention. This kind of
piping (usually field run) must be arranged so as to permit access
to the compressor, as well as for lubrication and maintenance. The
supports must be sturdy to avoid leakage or rupture from vibration
during normal and upset operating conditions.
Instrumentation and Controls—For the proper functioning of
instruments and controls: (1) conduct completeness and conti-
nuity checks—hydraulic, pneumatic, electrical; (2) commission,
field-calibrate, and establish set points; (3) check for accessibility
for operator use and vision, and for adjustment and maintenance;
(4) check mounting locations, and methods to minimize vibration
pickup—avoid mounting directly on the compressor, light plat-
forms, or hand rails; (5) check the supply of instrument air and con-
trol power for adequacy and reliability.
Electrical Power and Lighting—For power, test for continuity, elec-
trical-insulation soundness, proper grounding, and settings of relays
(including the correctness of thermal overload trip devices), tight-
ness and quality of all connections, sealing of fittings, and use of
flexible connections for equipment and instruments.
As for lighting, check to see that it is adequate for those areas fre-
quently missed by general lighting such as control stations, instru-
ments, and lubrication points.
Thermal Insulation and Painting—Check to see that proper insula-
tion will: (1) avoid hazard to personnel; (2) prevent thermal shock to
the piping, compressor, and turbine from rain; (3) prevent fire from
occurring as a result of oil spillage on a hot metal surface.
Avoid “overcompleteness” such as (1) paint on valve stems, instru-
ments, etc.; (2) unnecessary insulation on flanges and flange bolting;
28  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

(3) external insulation on internally insulated brick or refractory piping

and ducts; (4) insulation restraint on expansion joints; and (5) insula-
tion that will limit the freedom of the required pipe movement.

DYNAMIC COMPRESSORS

INTRODUCTION

There are two forms of energy in any system. One is called potential energy
and the other is called kinetic energy. For example, compressed gas in a
static state exerts its pressure in all directions, as shown in Figure 1.8. When
the outlet valve is opened, the gas flows out at very high velocity. Depend-
ing on the flow rate, the pressure in the cylinder drops down. In this case,
the pressure energy is converted into kinetic energy. This kinetic energy is
capable of doing work such as driving a pneumatic wrench, h­ ammer, etc.
The higher the pressure, the higher will be the velocity and hence the
kinetic energy of the gas leaving the system.
A dynamic compressor adds energy to gas in the same manner that
an electric fan does. Consider a fan in operation and note the following
points:

1. It is the rotating blades of the fan that force the air to move.
2. Air that is at rest tends to remain at rest.
3. As the fan blades start turning, they push on the air. The stationary
air resists the push of the blades.
4. As the air resists the blades, the molecules of the air are brought
closer together.
5. When the air molecules are Pressure Pressure and
compressed, the volume of the energy Kinetic energy

air decreases.
6. As the volume of the air
decreases, its pressure increases.
7. The blades of the fan over-
come the resistance of the air
and thrust the air forward.
8. The faster the blades turn, the
faster the air is pushed.
9. The fan, by doing work on the
air, actually increases the pres-
sure and velocity of the air. Figure 1.8.  Two forms of energy
 Understanding Compressors  •   29

10. When velocity and pressure are added to a gas, its total energy
increases.
11. A dynamic compressor increases the total gas energy by adding
pressure and velocity to the gas.
12. The total energy of a gas leaving a compressor is greater than the
total energy of the gas entering the compressor.
13. The energy that a gas gains in a compressor is due to the work
done on it.

CENTRIFUGAL COMPRESSORS

If a body is set in motion, it tends to continue in motion unless some force

acts in the opposite direction to stop it. If there is no gravity pull, nor any
obstacle to deflect it, a body in motion travels in a straight line.
Suppose a ball attached to a string is set in motion, as shown in
Figure 1.9. Assume that there is no gravity and that the string has no effect
on the ball. The ball moves in a straight line.
Suppose the string is fastened to a fixed pivot point and then the ball
is set in motion, as shown in Figure 1.10. At first, the ball moves in the
direction of motion. When the string becomes taut, it deflects the ball.
Because of the deflection, the ball actually travels in an arc, or a circle.
Assuming it has enough energy, the ball continues to move in a series of
arcs. At each instant of its travel, the physical tendency of the ball is to
travel in a straight line. But instead, the ball travels in a circle because it is
held or deflected by the string.
Understanding Compressors

Figure 1.9.  Motion of a ball in straight line

30  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Pivot point

Figure 1.10.  Centrifugal motion of a ball

The string actually applies centripetal (pulling-in-toward-the center)

force, causing the path of the ball to change or curve. If the string breaks,
the ball flies out in a straight line. Any object traveling in a circle is kept
in that path of travel by a force called centripetal force. The force holding
the ball in the circle of motion, that is, from the ball to the pivot point is
the centripetal force.
To hold the ball in its position and path, an opposite force is required.
That force is called the centrifugal force. If the centripetal force is elimi-
nated, the object then moves in a straight line.
The force pulling an object in a circular path toward the center is cen-
tripetal force. The centrifugal tendency of the object is its tendency to pull
away from the center of rotation, or to pull against the centripetal force.
The centrifugal tendency acts in the direction opposite to the centripetal
force.
The centrifugal tendency is actually not a force but is the result of
the tendency of the object to move in a straight line while being pulled
toward a center of rotation by the centripetal force. Assume a ball bear-
ing is placed close to the center of a disk that has blades, as shown in
the figure below. As the disk begins to move, one of the blades forces
the ball bearing to move. The ball bearing tends to travel in a straight
path. The drawing shows the actual path of the ball bearing as the disk
rotates.
 Understanding Compressors  •   31

When the disk rotates, the bearing is forced away from the center of
the disk, as shown in the figure below.
32  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Let us consider two points A and B located on the disk. Point A is at

the tip of the disk, while point B is closer to the center of the disk.
When this disk starts rotating, as shown in the figure above, point A
covers a larger distance than point B. When the disk is rotating, point A
moves faster than point B. Anything that is being carried along by the rota-
tion of the disk has a greater velocity when it is near the outer rim of the
disk. If anything being carried along by the rotation of the disk also travels
outward from the center to the outer rim, it gains velocity.
The velocity of point A is proportional to the RPM (revolution per
minute), or the rotating speed of the disk. The energy picked up by the
material at point A is given by

KE = m × v2/2gc

If D is the diameter of the disk in meters and N is the RPM of the

disk, the longitudinal velocity is equal to Π × D × N/60 m/sec (i.e.,
3.28 × Π × D × N/60 = 0.17181 × D × N m/sec. The kinetic energy
gained will be m × v2/2gc)
To achieve this kinetic energy, work has to be done on the disk or impeller.
The figure given below is a compressor impeller. An impeller is made
of two plates separated by blades.

When the impeller begins to rotate, the blades force the air in the
impeller to move. Air molecules tend to travel in a straight line. Because
there is no centripetal force, the rotation forces the air molecules outward
from the center, or eye, of the impeller. As the air molecules move out-
ward, they gain velocity, or speed. The air also tends to oppose the push of
the blades, so the pressure of the air is increased. The impeller adds both
pressure and velocity to the air.
 Understanding Compressors  •   33

The tendency of air or gas to move outward from the center of a rotating
impeller is the centrifugal tendency. A compressor that uses centrifugal ten-
dency to impart pressure and velocity to a gas is called a centrifugal compressor.
The part of the centrifugal compressor that moves the gas is the
impeller. As the impeller rotates, it moves the gas toward its outer rim. As
the gas moves toward the outer rim of the impeller, its velocity increases.
This increase in velocity away from the eye creates a low-pressure
area at the eye. This low-pressure area causes a suction, which allows more
gas to enter. The impeller does work on the gas. The work is converted into
the energy that the gas gains, which is in the form of both pressure and
velocity. When the gas is at the tips of the impeller blades, it is at maxi-
mum velocity. As the gas leaves the impeller, it is thrust into a passageway
called the diffuser (refer figures given below). When the gas enters the
diffuser, the impeller is not acting directly on the gas.

The radius of the diffuser is larger than the radius of the impeller. Due
to the larger radius, the flow path of the gas through the diffuser is in a
larger spiral. Since the flow path is longer and there is no direct action by
34  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

the impeller blades, the velocity of the gas decreases. As the velocity of the
gas decreases, its pressure increases.
The diffuser converts the velocity of the gas to increased gas pressure.
Gas passes from the diffuser into the volute. In the volute, the conversion
from velocity to pressure continues.
Gas passes from the diffuser into the volute as shown below (single
stage/last stage of a multistage compressor).
In the volute, the conversion from velocity to pressure continues. In a cen-
trifugal compressor, work is done on a gas to impart both pressure and velocity.

A centrifugal compressor, by doing work on gas, imparts both pres-

sure and velocity to the gas. Then, the velocity of the gas is converted into
pressure within the compressor. Look at the compressor below.

• It has four separate impellers.

• Each impeller and diffuser makes a stage.
• This is a four-stage centrifugal compressor.
 Understanding Compressors  •   35

As the gas leaves the first impeller, it gains some velocity and pres-
sure. The increased velocity is partially converted into pressure in the
diffuser.
As the gas leaves the diffuser, it enters the return passage, which
guides it into the eye of the next impeller. When the gas enters the eye of
the second impeller, it has greater pressure than when it entered the eye of
the first impeller. Each impeller adds to the total energy of the gas. It may
be noted that the velocity added by the impeller is converted into pres-
sure energy within the diffuser. When the gas leaves the compressor, its
pressure is higher than the inlet pressure. The work done by a compressor
is the total energy added to the gas through impellers. A gas leaving the
compressor has added energy in the form of pressure and temperature.

AXIAL COMPRESSORS

Motion along the axis of a shaft is called axial motion. This takes place in
a straight line. A compressor in which the gas moves parallel to the axis of
its shaft is called an axial compressor. An axial compressor has stator and
rotor blades, as shown below.

The rotor blades are attached to the shaft and rotate with the rotary
motion of the shaft. The stator blades are attached to the casing, as shown
in the above figure.
The arrangement of the blades is such that there is a set of stator
blades between each two sets of rotor blades, as shown in the figure
below.
36  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

The rotor blades behave in the same manner as the blades of a fan.
As they rotate, they force the gas to move. The rotor blades impart both
pressure and velocity to the gas.
The rotor blades force the gas into the stator blades.
As the gas is thrust into the stator blades, the openings between the
blades act as diffusers and reduce the velocity of the gas. With the decrease
in the velocity, the pressure of the gas increases. The stator blades guide
the gas into the next set of rotor blades. The gas entering the second set
of rotor blades has a slightly higher pressure. Thus, each set of stator and
rotor blades increases the gas pressure.
In axial compressor, the pressure increase of a gas is achieved by
using many sets of stator and rotor blades. The blades in an axial compres-
sor are not of the same size. The blades get gradually smaller toward the
discharge end of the compressor, as shown below.

Suction Stator Blades discharge

Rotor
Blades

Casing

As the gas flows through an axial compressor, it occupies less vol-

ume successively in its flow path. Thus, the gas pressure increases. The
 Understanding Compressors  •   37

flow of gas in an axial compressor is somewhat linear and in the direction

of its axis.

REVIEW

Two forms of gas energy are pressure and velocity. Energy cannot be cre-
ated or destroyed, but it is convertible from one form to another. By doing
work on a gas, the compressor adds energy to the gas. The total energy of a
flowing gas is a function of its pressure, velocity, and temperature. Where
the gas velocity reduces, pressure increases (Bernoulli’s Theorem).

RATIO OF COMPRESSION

While the difference between suction and discharge pressures denotes the
work done on a gas system, the ratio of absolute discharge pressure to absolute
suction pressure is known as compression ratio. When a gas is compressed,
part of the energy input or work done is converted into heat and friction losses.
The ratio of compression, R, is the relation between the absolute dis-
charge pressure and the absolute suction pressure. If P2 is the discharge pres-
sure and P1 is the suction pressure, then the compression ratio R = P2/P1.
This means the compression ratio denotes how many times the dis-
charge pressure is greater than the suction pressure. In determining com-
pression ratio, only absolute pressures must be used. To get absolute
pressure, add the atmospheric pressure to gage pressure.
For example, compressor discharge pressure = 300 psig: Absolute
discharge pressure = 300 + 14.7 = 314.7 psia.

Example:

Air is compressed to 100 psig using an air compressor. What is the com-
pression ratio of the compressor?
Air enters at atmospheric pressure. Therefore, P1 = 14.7 psia
Discharge pressure = 100 psig = 100 + 14.7 = 114.7 psia
Therefore, compression ratio R = 114.7/14.7 = 7.81

Example (metric units):

Air is compressed to 6 kg/cm2g using an air compressor. What is the com-

pression ratio of the compressor?
Air enters at atmospheric pressure. Therefore, P1 = 1.033 kg/cm2a
Discharge pressure = 6 kg/cm2g. Absolute pressure = 6 + 1.033 = 7.033
38  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Compression ratio R = 7.033/1.033 = 6.81

Note how the pressures have been converted into absolute pressures.

COMPRESSOR CAPACITY

The capacity of a compressor is the volume of gas that it handles in a given

period of time. For example, CFM indicates the volume of the gas handled
by the compressor in 1 minute. The flow rate of a gas in CFM in a pipe line
depends on the velocity of the gas and the diameter of the flow path. For
the same velocity, the rate in CFM is higher for a larger diameter passage.

Example:

Air passes through two pipe lines, one with 6 in. diameter and another with
8 in. diameter. The velocity of air in both the pipes is 500 ft/min. What is
the flow rate of air through these two pipe lines?
Denote the diameter of the 6 in. pipe as d1 and that of the 8 in. pipe
by d2.

a. Area of cross-section of the 6 in. pipe = Π × d12/4 = (22/7) × (6/12)2/4

= 0.19643 ft2
b. Volume of air flow =
 velocity × cross-section area = 500 × 19643
= 98.2 cfm
c. Area of cross-section of the 8 in. pipe = Π × d22/4 = (22/7) × (8/12)2/4
= 0.34921 ft2
d. Volume of air flow = velocity × cross-section area = 500 × 0.34921
= 174.6 cfm

Result: Air flow rate in 6 in. and 8 in. diameter pipes for the velocity
of 500 ft/min are 98.2 and 174.6 cfm, respectively.

If the gas velocity is greater, then the pressure at the discharge section
is lower.
During compression, the volume of gas entering the compressor is
greater than the gas leaving the discharge. The ACFM is measured at the
suction of the compressor.

HEAD OF COMPRESSION

For compressing a gas to higher pressures, certain amount of work has

to be done on the gas. The work done on a gas may be expressed as foot
pounds or kilogram meters.
 Understanding Compressors  •   39

When 1 lb of gas is lifted or moved to a distance of 1 ft, then the work

done by the compressor is 1 ft·lb. Similarly, when 10 lb of gas is moved to
a distance of 1 ft, then the work done by the compressor is 10 ft·lb. When
1 lb of gas is lifted/moved to a distance of 10 feet, then also the work done
by the compressor is 10 ft·lb.
The head developed by a compressor is the distance or height to which a
column of gas can be moved at the average density of gas. This is an important
factor in calculating the head developed by a compressor. For each pound of
gas the compressor raises the head, a corresponding amount of work in foot
pounds has to be done on the gas. As shown above, if the head increases, the
number of foot pounds of work to be done on the gas per pound also increases.
Pressure may also be converted to head, as shown in the example below.

Example:

1,000 cfm of air is compressed by a compressor to a pressure of 6 kg/cm2g

pressure. What is the work required to be done by the compressor on the
air and what is the head developed?

I. Air inlet pressure = atmospheric pressure = 14.7 psia

II. Air outlet pressure = 6 × 14.22 + 14.7 psia = 100 psia
III. Air density at atmospheric pressure = 28.84/359 lb/ft3 = .0803 lb/ft3
IV. Weight flow rate of air = 1,000 × 0.0803 lb/min = 80.3
V. Air density at 100 psia = 28.84/50.19 = 0.5746 lb/ft3
VI. Average density of air = (0.0803 + 0.5746)/2 = 0.3274 lb/ft3

Differential pressure developed by the compressor = 6 × 14.22 psi

= 85.32 psia
= 12,286.08 lb/ft2
Dividing the above value by the average air density, we get
= 12,286.08/0.3277 ft
= 37,491.8 ft
Work done on 1,000 cfm air = 80.3 × 37,491.8 ft·lb/min
This is the work to be done by the compressor on 1,000 cfm of air.

RPM AND HORSEPOWER

For all rotating machines, RPM is an important parameter. The impeller of

a centrifugal compressor has to rotate (revolve) to move the gas.
As the RPM of the impeller increases, the velocity of the gas also
increases.
40  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Gas velocity is proportional to the RPM of the impeller, as shown

below.

D = diameter of impeller in feet

RPM of impeller = N revolutions/

minute

RPM of impeller = N revolutions/minute

Gas velocity at the tip of the impeller = Π × D × N ft/minute

Gas velocity
Diameter RPM (feet/min)
5 8,000 1,25,714
5 10,000 1,57,142
5 12,000 1,88,571

For any given RPM, a set amount of work in foot pounds or kilogram
meter is done on the gas per unit weight. Since this velocity is converted into
head or pressure, for a constant RPM, the head developed is fairly constant.
Whether the gas is heavier or lighter, the work done per pound of gas
is the same for the same RPM.
At a given RPM, the ACFM of gas the compressor compresses will be
constant. But the weight of the gas compressed will be more, in the case of
a heavier gas. When more weight of gas is compressed, the work done on
a heavier gas will be more.
When the rate of work done on a gas increases, the horsepower
­required to compress a heavier gas also increases.

Example:

Vol flow Density Flow Work

Gas (cfm) RPM (lb/ft3) (lb/min) (ft·lb/min)
Hydrogen 1,000 8000 0.0056 5.60 5.60
Ammonia 1,000 8000 0.0473 4.73 4.73
Air 1,000 8000 0.0803 8.03 8.03
 Understanding Compressors  •   41

RPM AND BREAK HORSEPOWER

The ratio of compression, defined as the ratio of the discharge absolute

pressure to the absolute suction pressure, is an indicator of the amount
of pressure that the compressor adds to a gas or gas mixture. At any
particular RPM, a dynamic compressor adds certain head (pressure) to
the gas. The total head/pressure developed depends on the compressor
design, its RPM, and the amount of gas flow at suction. At a fixed RPM
and inlet CFM of gas, the head developed by a centrifugal compressor
is the same irrespective of the weight of gas. The head developed by a
centrifugal compressor does not depend on the density of the gas and it
is possible to convert the feet of head into psi (pounds per square inch)
and vice versa.
The density of a gas does not affect the head developed, but affects the
discharge pressure of the compressor.

Example:

Two identical compressors handle 200 cfm of air and hydrogen at 12,000
rpm and as per the compressor characteristics; the differential head
­developed is 20,000 feet at discharge conditions. What will be the dis-
charge pressure under these conditions? (Densities at discharge condi-
tions for air and hydrogen are 0.65 and 0.15 lb/ft3, respectively.)

Case 1. Discharge head for air = 20,000 ft: Density = 0.65 lb/ft3
Discharge pressure = head × density = 20,000 × 0.65 lb/ft2
= 13,000 lb/ft2
= 90.30 lb/in.2 = 6.35 kg/cm2 g

Case 2. Discharge head for hydrogen = 20,000 ft: Density = 0.15 lb/ft3
Discharge pressure = head x density = 20,000 × 0.15 lb/ft2
= 3,000 lb/ft2
= 20.83 lb/in.2 = 1.46 kg/cm2g

From the above example, it is clear that for the same CFM flow, RPM
and the head developed the discharge pressure is high for a high-density
gas than for a low-density gas.
BHP refers to the break horsepower that is required by the com-
pressor shaft to achieve the desired compression ratio. Because the
gas density fluctuates very often in industrial systems, a centrifugal/
dynamic compressor tends to change power consumption while in
operation.
42  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

The ratio of compression, R, is the absolute discharge pressure divided

by the absolute suction pressure. R is an indicator of the amount of pres-
sure that the compressor adds to the gas.
At a given RPM, a dynamic compressor adds a certain head to the gas.
The total head added depends on:

• The design of the compressor

• The amount of gas flow and
• The operating RPM (speed)

As RPM increases, the total head developed by the compressor

increases.
At a fixed RPM and CFM, the compressor attains approximately the
same feet of head, regardless of the weight of the gas handled.
The head developed by the compressor does not depend on the den-
sity of the gas being handled.
The feet of head or the meter of head can be converted into PSI or
kg/cm2 equivalent.
Similarly, PSI or kg/cm2 can be converted into feet or meters,
respectively.
When a compressor at a given RPM is handling a heavier gas, the
work it does over a pound of gas is the same as the work done on a pound
of lighter gas.
Head represents the amount of foot pounds or kilogram meter of work
done per unit weight.
A compressor at a given RPM handles two different gases. The gas
that requires a larger volume per unit weight is the lighter gas.

SURGING

When a compressor is connected to a large system, which needs a large

amount of gas for the process, the gas intake will vary with the process
requirement. During startup, the gas demand will be very high. Hence the
compressor will run at its full capacity.
As the process proceeds, the demand for gas will start reducing. If
the system does not use the gas as quickly as the compressor delivers, the
system pressure will increase.
With the increase in the system pressure, the resistance to discharge
of the compressor increases.
Due to this, the compressor capacity reduces.
 Understanding Compressors  •   43

When the head required maintaining the flow increases above the
maximum head of the compressor, the gas flow stops.
Under this condition, the pressure within the compressor becomes
less than the system pressure. This results in the gas flow from the system
to the compressor, called “flow reversal.”
When some quantity of gas has gone to the compressor section, the
system pressure will drop.
When the system pressure becomes less than the maximum head of
the compressor, the compressor starts delivering the gas to the system.
The compressor operates at a lower capacity and higher head/pressure.
If the system pressure system builds up again to a high pressure, the
flow reversal will repeat. The cycle continues.
The rapid flow of gas back and forth the compressor is called surging.
Surging occurs when the compressor operates below the minimum
capacity.
The rapid reversals of surging set up severe vibrations in the compres-
sor and piping, which may cause damage to the compressor.
A compressor goes into surging because the flow of gas drops down
below the minimum stable limit.
Most compressors are protected against surging by anti-surge control
methods as explained below.
Refer to the figure shown below.
44  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Y axis shows the percentage of rated head and x axis shows the per-
centage of rated capacity of the compressor. A point higher on the graph
shows higher head. The compressor capacity varies between 40 to 110
percent. The curve shows, as the flow increases, head decreases.
The compressor is discharging into a system that requires 100 percent
of its rated head. According to the graph, the compressor is operating at
100 percent of its capacity.
Suppose the discharge system does not use as much gas as the com-
pressor delivers the pressure at the discharge end of the compressor
increases as may be seen from the graph.
When the gas flow reduces to 90 percent, the head developed increases
to 102 percent of the rated head. As the gas discharged from the compres-
sor reduces, the pressure developed by the compressor increases. When
the flow reduces to 50 percent of the rated flow/capacity, the compressor
no longer delivers the gas to the system and virtually the gas flow stops.
When the compressor pressure becomes lower than the system pres-
sure, the gas will flow from the system to the compressor (flow reversal).
When this condition is reached, the system pressure will start drop-
ping down, and the compressor will discharge gas to the system again.
This repeated process of forward and reverse flow is called “surging.”

INTERPRETATION OF CURVES

Performance curves show the limits of the compressor, which are the surge
point on the curve and the normal capacity limit on which the compressor
must operate.
Compression is controlled by making permissible changes in pres-
sure, flow, and temperature to keep the compressor from surging.

SURGE CONTROL

A centrifugal compressor has continuously falling pressure flow charac-

teristics. A family of parallel curves represents a compressor operating at
different speeds.
For a compressor operating at constant speed, the pressure increases
as the flow decreases and vice versa.
If the reduction in flow continues, at one point, the compressor
operation becomes unstable and a momentary flow reversal takes place.
This leads to a very rapid pulsating flow backward and forward through
the compressor internals, resulting in severe vibration and consequent
damage.
 Understanding Compressors  •   45

This phenomenon, known as “surge,” occurs at a surge point at a par-

ticular flow and a particular speed.
The surge point shifts to the right as the speed increases, as shown in
the figure below.

Discharge

Speed
N1

N2

N3

Stone
Surge wall
line

A B C A2 A1

Suction flow

Suppose the compressor is running at RPM N3 and the inlet gas flow
rate is A1. When the flow rate is reduced due to lower system intake, the flow
rate drops to A2. When the flow drops down further to the level A, on RPM
N3, the surge limit is reached. This results in a drastic drop in the gas flow
and the pressure leading to flow reversal from the system to the compressor.
The line joining the surge points of curves at different speeds is the
compressor surge line. To avoid surge, compressors are always operated at
flows greater than the surge point, that is, at a reasonable distance to the
right of the surge line.
As the flow increases, a point is reached at the right-hand extreme
of the compressor curve, after which the flow cannot increase, no matter
what changes are made on the outlet side. The point is known as stone-
wall for the compressor. No adverse effect occurs at stonewall conditions.
The operating regime of a centrifugal compressor lies between two limits:
surge limit and stonewall.
46  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

SURGE CONTROL TYPES

RPM Control

In this method, when the gas flow rate starts coming down, the RPM
(speed) of the compressor is increased by sending the signal to the speed
control governor. This will increase the steam rate and the RPM for the
same flow rate of gas.

Recycling/Pressure Control

In this method, the discharge pressure is controlled by a spill-back mecha-

nism, which recycles the excess gas back to the suction, as shown in the figure
below. In this control system, the gas passing through the compressor is com-
pressed to the required pressure and passes to the system. The pressure con-
trol valve located at the discharge section will maintain the header pressure.
When the pressure starts increasing, the spill-back control valve will
open and recycle the excess gas to the suction, without affecting the sys-
tem pressure.

Recycle Pressure
control valve

Gas inlet

Gas
discharge

Compressor
 Understanding Compressors  •   47

TROUBLESHOOTING COMPRESSOR PROBLEMS

Normal problems associated with compressors may be broadly classified

into two categories: mechanical problems and operational problems.
In industrial practice, centrifugal compressors are usually run by
steam turbines. Motor-driven centrifugal compressors are also used in
small-capacity compressors. The problem associated with the compres-
sor may not necessarily be due to the compressor alone, but may be due
to the accessories also. Hence, for troubleshooting compressor problems,
related accessories must also be checked. A checklist for the compressor
and accessories is given below.

Mechanical Problems

Mechanical problems associated with the compressors are:

• Excessive vibrations of bearings

• Axial displacement of compressor rotor
• Bearing running too hot
• Bearing running rough
• Pulsating pressure
• Too low discharge pressure
• Casing temperature running high
• Excessive noise in running

Excessive Vibration of Bearings

All high-capacity compressors are provided with a trip mechanism when
the vibration goes above a stipulated level. Vibration increase is a gradual
process. If the vibration shows an increasing tendency, check all the bear-
ing temperatures, gas flow rate, pressure developed, etc.
Normally, vibration will start only if some misalignment/roughness of
the bearing has developed. Vibration may also occur when the compressor
surges.
If the compressor surging is frequent, the bearing alignment and/
or the condition of the bearing may have deteriorated, resulting in
­excessive vibration. When the vibration starts increasing, the compres-
sor must be taken out of service and inspected for bearing conditions
and alignment.
48  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Axial Displacement of the Rotor

All centrifugal compressors are provided with an axial displacement
trip system. Up to 0.5 to 1.0 mL of displacement is tolerated in certain
machines. When the displacement exceeds this limit, the compressor will
trip automatically. Axial displacement takes place from the high-pressure
end to the suction side (low pressure). This displacement is prevented by
a thrust bearing, provided in the suction end. If the thrust bearing is dam-
aged due to surging, critical speed, etc., the displacement of the rotor will
tend to move toward the suction end. When the displacement exceeds a
certain limit, the rotor blades will tend to contact the diffuser. Due to the
high speed of the machine and the friction heat generated, there will be
total damage to the compressor if it fails to trip on the stipulated displace-
ment limit. Frequent compressor surging of the compressor increases the
axial displacement.

Bearing Running Too Hot

Most compressors are provided with bearing temperature indicators or
local bearing thermometers. When the temperature of the bearing shows
an increasing tendency, the immediate step is to check for proper bearing
lubrication. Inadequate and/or improper lubrication may lead to high bear-
ing temperatures.
An experienced person will be able to even foresee the conse-
quences of high bearing temperature. From a historic log data of the
compressor, it is possible to determine whether or not the compressor is
running or not.
The following table is a historic reading of a compressor, which gives
the most important readings of a compressor. The bold lines indicate the
design/stipulated values of the compressor operating parameters.

Bearing Bearing Displace- Vibration Vibration

1 2 ment Bearing Bearing
Month Temp °C Temp °C mL 1 mL 2 mL
Maximum 100 100 0.05 1.5 1.5
limit
1 90 82 0.01 0.028 0.01
3 105 85 0.01 0.250 0.05
5 112 87 0.02 0.580 0.07
7 121 89 0.034 1.05 0.09
 Understanding Compressors  •   49

Compressor RPM 15,000/Full Load

It may be noted from the above data that bearing 1 has a problem,
which is reflected in the vibration levels. Another indirect method of bear-
ing metal loss is to collect the lube oil sample from the abnormal section
and analyze for the presence of metals. If the lube oil shows the presence
of any metal, matching with that of the bearing, it may be assumed that
the bearing thickness is deteriorating and may fail at certain point of time.

Surging in Compressors
Besides flow variations, changes in the molecular weight of the gas being
compressed may also lead to surging, as shown below. The process condi-
tions linked to the compressor play an important role in compressor surges.

Example:

The process refers to a gas compressor, which compresses a mixture of

methane, ethane, propane, and propylene coming from a cracking process.
The gas mixture is compressed to recover heavy hydrocarbons from the
mixture by condensation in a condenser and to recover the dry gas for
further processing. The compressor is located as shown.

PC

Wet gas

Reflux drum

Compressor

Feed

The following figure is the characteristic curve of the gas compres-

sor shown above.
50  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

DischPr.PSI

200 B
C A
180

160

140

120

100

ACFM 10,000 15,000 20,000

Observations: Point “A” is the operating point

Points “B” and “C” are surge points

OPERATING PARAMETERS

Gas flow: 17,000 acfm

Molecular weight of gas: 34
Suction pressure: 20 psig
Discharge pressure: 200 psig

The compressor develops 180 psig differential pressure under normal

operating conditions as given above. The discharge pressure is fixed by
the absorber pressure. Suction pressure is controlled by the PC, as shown
above.
The compressor handles 17,000 acfm of gas, of which 10,000 acfm
goes to the absorber, while 7,000 acfm is recycled.
When the drum pressure starts rising, the PC will start closing. This
will result in low gas inflow and the operating point will move to the left,
toward the surge point.
When the molecular weight starts decreasing, the differential pressure
will also start decreasing, and at certain point of time, the compressor will
start surging.
Hence, it is necessary to check the molecular weight of the gas regu-
larly to avert compressor surging and damage to the compressor.
 Understanding Compressors  •   51

Similarly, lube oil analysis data for:

Flash point
Water content
Impurities and wear particles
Viscosity and viscosity index
Total acidity may be used to diagnose the compressor-related problems
for remedial action.
 CHAPTER 2

Compressor Efficiency
DEFINITION MAKES A DIFFERENCE
54  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Compressor efficiency ratings can be as misleading as interest rate

figures. To be certain of getting the most for your money when buying
compressors, you must be sure that efficiency comparisons are made
on the same basis.
The compression of air and other gases consumes a significant
amount of power in the chemical process industry. In the production of
industrial gases, such as oxygen, nitrogen, and helium, or in the rapidly
growing field of natural-gas liquefaction, compression power accounts for
more than 80 percent of total energy requirements. Clearly, compressor
efficiency has a direct impact on product costs.
This article compares isothermal and adiabatic efficiencies, shows
the effect of intercooling between stages of compression, and provides
convenient ways for obtaining preliminary figures on compression power.

ISOTHERMAL COMPRESSION

The basic equation used to calculate the work of gas compression in a steady-
flow process (and flow through a reciprocating compressor may be consid-
ered steady, despite the pulsations of the individual compression strokes) is:

w = ∫1 v dp (1)

where w is the work required per lb-mole, v the molal volume, and
p the pressure.
For preliminary calculations, the ideal-gas equation of state may be
used to relate pressure, volume, and temperature:

Pv = RT (2)

In fact, this equation will yield quite accurate results at pressures

­below 10 atm and at temperatures well above the critical temperature.
 Compressor Efficiency  •   55

If the gas could be compressed without friction and isothermally, the

work required to raise the pressure from p1 to p2 is:

wiso = RT1 ln ( p2 /p1 ) (3)

In this equation, R is the gas constant, 1.986 Btu/(lb-mole) (°R); T1

the initial temperature, °R; and work has the units Btu/lb-mole.
The isothermal efficiency of a compressor is defined as the ratio of
work calculated by Eq. (3) to the actual work required, or:

ηiso = wiso /wa (4)

In many cases, the numerical value of isothermal efficiency may

appear to be low, for example, 65 percent, and yet the actual work
required may be less than that of a compressor with an adiabatic effi-
ciency of 80 percent. We shall see how this apparent discrepancy comes
about.

ADIABATIC COMPRESSION

An ideal gas compressed adiabatically and without friction requires work

according to the expression:

k −1
 
 k   p2  k 
wad = RT1   − 1 (5)
 k − 1   p1  

where k is the ratio of specific heats, Cp/Cv, and all the other terms are as
defined before.
Adiabatic efficiency is defined similarly to isothermal efficiency:

ηad = wad /wa (6)

It is clear that the two definitions of compression efficiency vary only

in the standard of comparison, that is, ideal isothermal work versus ideal
adiabatic work. Figure 2.1 shows the variation in the ideal adiabatic work
as a function of the pressure ratio, p2/p1, and the specific heat ratio, k. In
all cases, this work is greater than the isothermal work shown by the lowest
curve. The influence of the specific heat ratio is pronounced. Monatomic
gases, such as helium and argon, have k = 1.66, and therefore, the adiabatic
56  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 2.1.  Adiabatic efficiency is greater than

isothermal work

work for the compression of such gases is significantly greater than the
isothermal work. Diatomic gases, such as nitrogen, oxygen, hydrogen,
and air, have k = 1.4. More complex gaseous molecules have a lower
value of k, such as 1.32 for methane.
As an example, consider the compression of air al 14.0 psia and
80°F to 56 psia. From the appropriate equation, or from Figure 2.1
with k = 1.40 and p2/p1 =. 4.0, we find wiso = 1,487 Btu/lb-mole and
wad = 1,824 Btu/lb-mole.
If the actual work required by a compressor is 2,280 Btu/lb-mole,
the compressor efficiency may be expressed as ηiso = (1,487/2,280)
(100) = 65 percent and ηiso = (1,824/2,280) (100) = 80 percent. Both
of these efficiency definitions are equally valid, although the adiabatic
efficiency is usually used when no effort is made to cool the gas during or
between stages.

EQUAL-WORK EFFICIENCY RATIO

We have seen in the preceding example that for the same actual work of
compression, the adiabatic and isothermal efficiencies vary by a factor
 Compressor Efficiency  •   57

of 1.23. From Eq. (3–6), the equal actual work of compression, wa,
­results when:

k −1
 
 k   p2  k − 1
 
 k − 1   p1 
ηad
=  (7)
ηiso ln( p2 /p1 )

The equal-work efficiency ratio given by Eq. (7) is plotted as a function

of pressure ratio in Figure 2.2. With the same conditions as in the earlier
example, k = 1.4 and p2/p1 = 4.0, the equal-work efficiency ratio is 1.23.
To illustrate how the equal-work efficiency ratio may be used, consider
the problem of evaluating compressor bids. Again, we shall compress air
from 14.0 psia and 80°F to 56 psia. If one vendor guarantees 80 percent
adiabatic efficiency and another 70 percent isothermal efficiency, the quo-
tation of the latter must indicate a lower power requirement, despite the
lower efficiency. We found that the equal-work efficiency ratio for these
conditions is 1.23. The ratio of quoted efficiencies is 1.14. Whenever the
actual ratio is less than the equal-work ratio, the compressor whose effi-
ciency is based on the isothermal efficiency will require less power. Above
the equal-work curve, the adiabatic efficiency machine requires less power.

Figure 2.2.  Ratios for the actual work of compression

58  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

MULTISTAGE COMPRESSION WITH

INTERCOOLING

Whenever the overall pressure ratio exceeds 4, consideration must be

given to multistage compression with intercooling. This is necessary to
keep the temperature of the compressed gas at a safe value, for example,
365°F for air compressors with hydrocarbon lubricants as well as to reduce
the overall power requirement.
Ideal intercooling would reduce the temperature between stages to
the suction temperature. The effect of staging with intercooling is shown
in Figure 2.3. The ideal work of compression in steady flow, given by
Eq. (1), is represented by the areas to the left of the curves. The iso-
thermal curve represents the minimum work requirement, whereas the
adiabatic curve encloses a much larger area. The intermediate dashed
curve represents two-stage compression with intercooling to the suction
temperature, 80°F.
Now let us consider a two-stage compressor, compressing air from 14.0
psia and 80°F to 140 psia, with intercooling to 80°F. Should the efficiency
of a compressor be based on the ideal isothermal or the ideal adiabatic com-
pression, as shown by the solid curves in Figure 2.3? Actually, we should
refer to stage efficiency but clearly keep the overall efficiency in mind.

Figure 2.3.  Dashed curve represents two-stage adiabatic compression with

intercooling to the suction temperature; solid curves depict adiabatic and
isothermal compression of gases
 Compressor Efficiency  •   59

Because, in this example, the overall pressure ratio is 10.0, we may

compute the ideal work of compression as: wiso = 2,470 Btu/lb-mole and
wad = 3,480 Btu/lb-mole.
These numbers may also be found from Figure 2.1. For each stage
of compression, with a pressure ratio of 3.16, the ideal adiabatic work
of compression is 1,460 Btu/lb-mole, or a total, for two stages, of 2,920
Btu/lb-mole. Using a realistic adiabatic stage efficiency of 80 percent, the
actual work of compression is 2,920/0.80 = 3,650 Btu/lb-mole. We now
have three equally accurate, but quite confusing, compressor efficiencies:
Adiabatic compressor efficiency: (3,480/3,650) (100) = 95%
Adiabatic stage efficiency: (2,920/3,650) (100) = 80%
Isothermal compressor efficiency: (2,470/3,650) (100) = 68%
Efficiency is a word with many meanings. We have seen that a higher
numerical value of efficiency based on one definition may actually rep-
resent poorer performance than a lower numerical value based on another
definition. The equations and graphs presented in this discussion are
­designed to refresh practicing engineers on these points. There is no doubt,
however, that for a given compression job, the compressor that requires the
least horsepower is the most efficient.

HOW TO ACHIEVE ONLINE AVAILABILITY OF CENTRIFUGAL

COMPRESSORS?

Here is practical information that will ensure the best selection and
proper maintenance of centrifugal compressors and their drivers,
60  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

bearings, seals, couplings, speed gears, and lubrication and control

systems.
Centrifugal compressors are an integral part of the chemical process
industries (CPI). They are used extensively because of their smooth oper-
ation, large tolerance to process fluctuations, and higher reliability than
other types of compressors.
Centrifugal compressors range in size from pressure ratios of 1.3:1
per stage to as high as 12:1 on experimental models. We will limit our dis-
cussion to compressors having pressure ratios below 3.5:1 because these
are the ones used extensively in the CPI.
A proper selection of these compressors involves making complex
decisions, since the successful operation of many CPI plants depends on
the smooth and efficient functioning of such units. To ensure the best se-
lection and proper maintenance of centrifugal compressors, the engineer
must have a wide knowledge of many engineering disciplines.

SPECIFICATION CRITERIA

Detailed compressor specifications can vary from customer to customer.

Some provide only the basic information, such as pressure, flow rate, type of
gas, driver, and site conditions. Others present a lengthy document detailing
types of bearings, rotor response, lubrication system, acceptable tolerance on
performance, etc. The latter requires that the engineer be very conversant with
compressors and their total support systems. A starting point for the specifi-
cations is to be found in the publications of the American Petroleum Institute
(API) for turbomachinery. Some of the applicable ones are listed in Table 2.1.
Specifications for the API standards are written by user engineers,
with the input of manufacturers and engineering contractors. Thus, the
standards represent a wealth of experience and are a very good base from
which to start turbomachinery specifications.
Many decisions, regardless of details contained in specifications,
have to be made in advance by the engineer. Some of these decisions may
involve company philosophy about various units; others could be strictly
job oriented. In writing a job specification, we must have an understanding
of the major problem areas. These will now be covered.

COMPRESSOR LAYOUT

The general topography of the plant must be known so that the proper site
can be established. Whether the unit will be grade mounted or mezzanine
 Compressor Efficiency  •   61

Table 2.1.  Sources for compressor specifications

Standard Title
API 611 General Purpose Steam Turbine for Refinery Services
API 612 Special Purpose Steam Turbine for Refinery Services
API 613 High-Speed Special Purpose Gear Units for Refinery
Services
API 614 Lube and Seal Oil Services
API 616 Combustion Gas Turbines for General Refinery Services
API 617 Centrifugal Compressors for General Refinery Services
API 670 Noncontacting Vibrating and Axial Position Monitoring
Systems

Available from: American Petroleum Institute, 2101 L St., N.W.,

Washington, DC 20037.

mounted is important in determining the foundation characteristics.

Enough space should exist for the ducting, so that the inlet conditions to
each stage allow the flow to enter without large distortions of velocity and
pressure.
Accessibility requirements should be kept in mind so that repair and
maintenance work on the unit can be performed with relative ease. The
location of the oil system for the unit is also an important aspect. It is
advisable to locate the oil reservoir away from the base plate, with the
bottom sloped toward the low drain point. Enough space should be pro-
vided so that the return-oil lines can enter the reservoir away from the
oil-pump suction. This would greatly reduce the disturbance of the pump
suction, and also help in keeping the reservoir-retention time to around
10 minute.

COMPRESSOR ENVIRONMENT

The environment in which a machine operates is as important as any other

factor. In many cases, this detail is often overlooked, or described in a
phrase such as “extreme cold climate.” The vendor needs to know much
more. He must know what extreme cold means (usually below −25°F) and
what the transition weather is.
Weather creates many problems—though few precautions are taken
­because operating problems are either not recognized or are glossed over.
62  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

All cold-weather applications need to have some icing protection, especially

in the air-inlet supplies. Fuel-supply ventilation, pneumatic controls, and ac-
tuators need to have some degree of deicing. Many types of deicing systems
exist. The two most common use exhaust gases or compressor bleed-air.
Tropical climates present their own problems, such as excessive
corrosion, high moisture content, and high ambient temperatures that
­increase both the horsepower required and the required cooling capacity
of the lube-oil systems. Desert locations require special filtration systems
to prevent the erosion of the blades and special sealing on joints to prevent
micron-sized sand particles from entering the lube system, etc.
Machine usage is an environmental consideration because intermit-
tent operation can be the most severe kind of service. The aggressiveness
of the gas to be handled and its temperature determine the materials of
construction to be used.

COMPRESSOR TYPES AND ARRANGEMENTS

It is not always obvious what type of compressor is needed for an applica-

tion. Of the many kinds, some of the more significant are the centrifugal,
axial, rotary, and reciprocating. Figure 2.4 will aid in the selection of a
­compressor. For very high flows and low pressure ratios, an axial-flow
compressor would be best. Axial-flow compressors usually have a higher effi-
ciency but a smaller operating region than a centrifugal machine. Centrifugal

Figure 2.4.  Compressor capacities

 Compressor Efficiency  •   63

compressors operate most efficiently at medium flow rates and high-pressure

ratios. Rotary and reciprocating compressors (positive-displacement
­machines) are best used for low flow rates and high-pressure ratios.
The general configuration for compressors and their drive trains must
fit the location, environment, and type of compressor. A decision must be
made as to whether the units are to operate in a series or in parallel, and
it requires a knowledge of the necessary discharge pressure and flow. In
most cases, a number of casings are connected together to form a “com-
pressor train.” This is nothing more than connecting various compressors
in a series. The limit to the number of casings so connected is due to the
rotor dynamics of the coupled train.
A decision must also be made as to what type of mounting is desirable
for a given arrangement. Most turbomachinery is mounted on structural-steel
plat­forms, referred to as base plates or skids. These platforms are then
­positioned on a mass of concrete at the job site by installing them on sole
plates or through direct grouting. Platforms should be considered part of the
foundations, with great care exercised in their design. Insufficient rigidity
can allow the rotating machinery to excite these platforms.

DRIVER SELECTION

The three main types of drives for centrifugal compressors are (1) steam
turbines, (2) gas turbines, and (3) electric motors. The decision of which
drive is best is not always easy. Selection depends on many factors, such
as location, process, and unit size.
For remote locations, gas turbines are mostly used, due to their low
maintenance and the ability to prepackage the units. Their light weight
makes them a must for offshore platforms. For chemical plants, steam
turbines are widely used, due to the needs for process steam. In this man-
ner, CPI plants can use energy more efficiently. For smaller flows, electric
motors drive the compressor, usually through speed-increasing gears.
Typical ranges for the various drives are shown in Figure 2.5, from
which we note that the higher the flow, the lower the speed. At high flows,
the compressor diameter must be large; therefore, the speed must be
­reduced to maintain the same stress levels in the machinery.

COMPRESSOR DESIGN AND CONFIGURATION

To properly design a centrifugal compressor, we must know the operating

conditions—the type of gas, and its pressure, temperature, and molecular
64  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 2.5.  Different compressor speeds

weight. We must also know its corrosive properties so that proper met-
allurgical selection can be made. Gas fluctuations due to process insta-
bilities must be pinpointed so that the compressor can operate without
surging.
Centrifugal compressors for industrial applications have relatively
low pressure ratios per stage. This is necessary so that the compressors
can have a wide operating range, and stress levels can be kept at a mini-
mum. Due to the low pressure ratios for each stage, a single machine may
have a number of stages in one “barrel” to achieve the desired overall
pressure ratio. Figure 2.6 shows some of the many configurations. Some
of the factors to be considered when selecting a configuration to meet
plant needs are:

1. Intercooling between stages can considerably reduce the power

consumed.
2. Back-to-back impellers allow for a balanced rotor thrust and mini-
mize overloading the thrust bearings.
3. Cold inlet or hot discharge at the middle of the case reduces oil-seal
and lubrication problems.
4. Single inlet or single discharge reduces external piping problems.
5. Balance planes that are easily accessible in the field can appreciably
reduce the field-balancing time.
6. Balance piston with no external leakage will greatly reduce the
wear on the thrust bearings.
 Compressor Efficiency  •   65

Figure 2.6.  Centrifugal Compressor Configurations

7. Hot and cold sections of the case that are adjacent to each other will
reduce thermal gradients and thus reduce case distortion.
8. Horizontally split casings are easier to open for inspection than
­vertically split ones, reducing maintenance time.
66  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

9. Overhung rotors present an easier alignment problem because

shaft-end alignment is necessary only at the coupling between the
compressor and the driver.
10. Smaller, higher-pressure compressors that do the same job will
reduce foundation problems but will have greatly reduced opera-
tional range.

IMPELLERS AND THEIR FABRICATION

Centrifugal-compressor impellers are either shrouded or unshrouded.

The blading for the impellers can have one of three configurations—most
commonly radial, followed by backward-curved, and, in some rare cases,
forward-curved. Backward-curved blades have the largest operating range
and are usually higher in efficiency. The radial-bladed impeller produces
a higher head but has a smaller operating range and a slightly lower effi-
ciency than the one with backward-curved blades.
Open, shrouded impellers that are mainly used in single-stage applica-
tions are made by investment-casting techniques or by three-dimensional
milling. Such impellers are used, in most cases, for the high-pressure-ratio
stages. In the CPI, the shrouded impeller is the most common.
Figure 2.7 shows several fabrication techniques. The most common
type of construction is seen in Figure 2.7a and 2.7b, where the blades are fil-
let-welded to the hub and shroud. In Figure 2.7b, the welds are full penetra-
tion. The disadvantage of this type of construction is the obstruction to the
aerodynamic passage. In Figure 2.7c, the blades are partially machined with
the covers and then butt-welded down the middle. For backward-lean-angle
blades, this technique has not been very successful, and there has been dif-
ficulty in achieving a smooth contour around the leading edge.
Figure 2.7d illustrates a slot-welding technique and is used where the
blade-passage height is too small, or the backward-lean angle too high, to
permit conventional fillet welding. In Figure 2.7e, an electron-beam tech-
nique is used to weld on the shroud or the hub. This technique is still in its
infancy and work needs to be done to perfect it. Its major disadvantage is
that electron-beam welds should preferably be stressed in tension, but for
the configuration of Figure 2.7e, they are in shear. The configurations of
Figure 2.7g through 2.7j use rivets. Where the rivet heads protrude into the
passage, aerodynamic performance is reduced.
Materials for fabricating these impellers are usually low-alloy steels,
such as AISI 4140 or AISI 4340. AISI 4140 is satisfactory for most appli-
cations; AISI 4340 is used for larger impellers requiring higher strengths.
For corrosive gases, AISI 410 stainless steel (about 12 percent chromium)
 Compressor Efficiency  •   67

Figure 2.7.  Fabrication types for compressor blades

is used. Monel K-500 is employed in halogen-gas atmospheres and in

oxygen compressors be­
­ cause of its resistance to sparking. Titanium
­impellers have been applied to chlorine service. Aluminum-alloy impel-
lers have been used in great numbers—especially at lower temperatures
(below 300°F). With new devel­opments in aluminum alloys, this range
is increasing. Aluminum and titanium are sometimes selected because of
their low density. This can cause a shift in the critical speed of the rotor,
which may be advantageous.

ROTOR DYNAMICS

Rotor movement and its effect on the performance of the entire unit
is the most important aspect of a centrifugal-compressor design. Most
compressors for the CPI are built in accordance with API 617 spec-
ifications. The natural frequency of the rotor should not occur in the
variable-speed range of the compressor, so as not to excite any of these
frequencies. In newer high-speed compressors, many operate above
68  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

their first critical speed. Shafts that operate above the first critical are
said to be “flexible shafts.”
It is desirable that the first critical not be around half the design speed.
Otherwise, a problem known as “oil whirl” may be induced. Oil whirl is a
major cause of instability in turbomachines. It may occur in journal bear-
ings or in the seals in which the shaft and the stationary seal are separated
by a film of fluid.
Many types of whirling motions occur in turbomachinery and are
generated by various factors, such as:

• Hydrodynamic or “oil whip” or half-frequency whirl is caused by

the instability of the fluid film in the bearing.
• Aerodynamic whirl is usually induced by tip clearances on com-
pressors and turbines. It cannot be eliminated on balancing.
• Dry-friction whirl is usually initiated by rubbing caused by a large
unbalanced force.
• Torque-induced whirl occurs in large flexible rotors under large
power levels.
• Gyroscopic-induced whirl may be caused by bowed shafts and
skewed disks.
• Asymmetric whirl is caused by the difference in shaft stiffness and
is most violent near the first critical.

Balancing is a major problem for the newer flexible rotors that oper-
ate, in many cases, above the first critical and, often, above the second or
third critical. Figure 2.8 shows the various modes that the rotor shaft
­undergoes as it passes through these criticals. High-speed balancing of
these rotors is sometimes a must for a smooth operation. This inevitably
means field balancing, because there are only a few test-stands that can
balance these rotors at design speed.
Figure 2.9 is a typical rotor-response curve for a four-stage rotor.
Here, the rotor is operating above the first critical, but the steepness of
the curve near the design point is a cause for concern. Modification of
the rotor and a change in bearing stiffness will move the slope from the
design point.

BEARINGS FOR HIGH-SPEED MACHINES

Journal and thrust bearings are among the most important components to
assure the maintenance-free running of high-speed turbomachines. Bear-
ings in these machines range from simple journal bearings and flat thrust
 Compressor Efficiency  •   69

Figure 2.8.  Rotor shapes & speeds

bearings to multiwedge designs for both thrust and journal bearings. Some
of the many factors that enter into the selection of such bearings are:

• Speed range of shaft

• Maximum misalignment that can be tolerated by the shaft
• Critical-speed analysis, and influence of bearing stiffness on this
analysis
• Loading of the compressor impellers
• Oil temperature and viscosity
• Foundation stiffness
• Axial movement that can be tolerated
• Type of lubrication system and its contamination
• Maximum vibration levels that can be tolerated

All rotating machines vibrate when operating, but the failure of the
bearings is mainly due to their inability to resist cyclic stresses. The level of
vibration that a unit can tolerate is shown in the severity charts (Figure 2.10).
70  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 2.9.  Rotor speeds & response curves

These charts are modified by many users to reflect the critical values for
their machines.

JOURNAL BEARINGS

The journal bearing for turbomachinery has a fluid film that carries the
load. Film thickness in most applications ranges from 0.0003 in. for gases
to 0.008 in. for hydrostatic oil-lubricated bearings.
In this chapter, we will discuss only journal bearings in which a positive
supply of lubricant is fed to the bearing at all times. Figure 2.11 shows a num-
ber of such bearings. The circumferential-grooved bearing normally has the
oil groove at half the bearing length. This provides better cooling, but reduces
load capacity by dividing the bearing into two parts. The cylindrical bearing,
used in turbines, has a split construction with two axial oil-feed grooves at the
split. The pressure-dam bearing is used where bearing stability is required.
The most common bearing is the tilting-pad type, whose most important
feature is self-alignment when the bearing is used with spherical pivots. This
bearing offers the greatest increase in fatigue life because of these advantages:

1. Self-aligning for optimum alignment and minimum limit

2. Thermal conductivity backing material to dissipate heat developed
in the oil film
 Compressor Efficiency  •   71

Figure 2.10.  Vibration levels for compressors

3. Thin babbitt layer, centrifugally cast to a uniform thickness of about

0.005 in. (thick babbitts greatly reduce bearing life; babbitt thickness
of about 0.01 in. reduces the bearing life by more than half)
72  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 2.11.  Bearing types

4. Oil-film thickness can be varied by changing the number of pads,

directing the load onto or in between the pads, or changing the axial
length of the pad. Oil-film thickness is critical when making bear-
ing stiffness calculations.
 Compressor Efficiency  •   73

THRUST BEARINGS

The most important function of a thrust bearing is to resist the unbalanced

force developed in the working fluid of the machine and to maintain the
rotor in its position within the prescribed limits. A complete analysis of
the thrust load must be conducted. Compressors with back-to-back rotors
greatly reduce the load on thrust bearings.
Figure 2.12 shows a number of different types of such bearings.
When properly designed, the tapered-land thrust bearing (Figure 2.12a)
can take and support a load equal to that of a tilting-pad thrust bearing.
With perfect alignment, it can match the load of even a self-equalizing
tilting-pad thrust bearing. Figure 2.12b is a nonequalizing tilting-pad

Figure 2.12.  Trust bearings Vs unbalanced forces

74  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

thrust bearing that pivots on the back of the pad along a radial line.
Figure 2.12c is a nonequalizing tilting-pad bearing whose pads are sup-
ported on spherical-pivot points. Since this allows the pads to pivot in
any direction, alignment is not as serious a problem as in the other two
types. Figure 2.12d is the Kingsbury-type self-equalizing thrust bearing.
This bearing virtually eliminates the problem of misalignment. The major
drawback is that standard designs require more axial space than do a
nonequalizing type.

MISALIGNMENT

The amount of misalignment that can be tolerated in high-speed tur-

bomachines depends on the types of journals and thrust bearings used.
Figure 2.13 shows the misalignment for such bearings. In a journal bear-
ing, mis­alignment will cause the shaft to contact the end of the bearing.
Thus, journal length is a criterion in the amount of misalignment that a
bearing can tolerate. Obviously, a shorter length bearing can tolerate more
misalignment. In a thrust bearing, misalignment will cause the loading
of one segment of the thrust-bearing arc and the unloading of the oppo-
site segment. This effect is more pronounced at higher loads and with
less-flexible bearings.
To adjust for misalignment other than by using tilting-pad bearings,
various techniques must be employed. The simplest and most common is
the so-called the cold alignment (also known as base alignment) method.
Once this is done, hot alignment is carried out. Hot-alignment techniques
measure the changes when the unit is operational and temperature growth
is stabilized, so that accurate alignment data are obtained. There are many
ways to perform these tests. The recommended procedure is to do cold
alignment by using the “reverse-indicator graphical plotting” and the hot
alignment by using an optical alignment technique.
Reverse-indicator graphical plotting is normally obtained when the
unit is cold. This is done by first laying out the desired hot operating
line on a graph paper (Figure 2.14). This line shows the desired equilib-
rium operating conditions. Then, the desired cold position of the shaft is
plotted, based on the thermal-growth information that must be supplied
by the manufacturer. Actual positions of the shaft in the field are taken.
This information is plotted, the difference computed, and shims added
to the supports. (The major assumptions here are that the computation
measures actual vertical growth, and that growth in the horizontal di-
rection is zero.) This procedure is then repeated, after hot checking the
alignment.
 Compressor Efficiency  •   75

Figure 2.13.  Bearing misalignment

For hot checking, an optical alignment procedure is recommended.

The heart of this technique is the optical instrument with built-in optical
micrometers to measure the displacement from a precise and referenced
line of sight. After the initial cold alignment, the train is ready for startup.
Optical reference points are established at each end and each side of each
unit of the train, and as close to the couplings as possible. On most units,
this is at the bearing housing.
76
Figure 2.14.  Alignment techniques
 Compressor Efficiency  •   77

A jig transit is readied, a line of sight established, and readings taken

at each reference point in the compressor train in the vertical plane. For
the horizontal plane, a somewhat similar procedure is followed, and data
are recorded for reference points. The train is then started up and operated
at design or near-design conditions, and temperatures are allowed to stabi-
lize. Then, another set of optical readings are taken. In this manner, when
a comparison is made with the cold readings, actual thermal growth can
be calculated. Thus, the actual thermal growth can be plotted on the graph,
and new corrections can be computed.
The technique outlined here is used on new machines. On old
machines, a reverse technique can be applied. First, a hot-alignment
check is made, then a cold optical check, followed by the mechanical
reverse-indicator readings. The information is plotted, and realignment
measurements are taken.
The outline of the technique given here is simple, but in practice one
must develop this skill. Some of the major problems encountered in align-
ment are due to pipe strain. This is caused by the piping being offset from
the intake or exhaust nozzles from a few thousandths of an inch to sev-
eral inches. Many engineers cannot understand how a “small” pipe is able
to move a large, heavy piece of machinery. Tension on pipe hangers can
change the vibration level considerably.
Another contributor to the alignment problem is the gear casing.
Thermal growth in the new fabricated cases has been unpredictable. In
some instances, gear cases rise with a twist to create another problem that
is very difficult to correct. Short couplings also present a problem and
magnify misalignment. Some users are now specifying that the coupling
spacer will be at least 18 in. long.
The preceding discussion has presented a general outline of align-
ment techniques. The alignment of high-speed machinery must be
accurate; otherwise, major problems will arise. Sometimes, to prevent
major shutdowns to correct the misalignment, heaters are added to
one or the other leg of the unit to align them while running. This tech-
nique is not a cure but rather a temporary relief when shutdown is
impossible.

COMPRESSOR SEALS

The internal seals that prevent leakage around the impellers are usually
labyrinth type, as shown in Figure 2.15. They have a series of circum-
ferential knife edges that are positioned close to the rotating impeller. If
damaged by rubbing, erosion, or corrosion, these knife edges will lose
78
Figure 2.15.  Function of labyrinth seals- explanation of abc shows the leakage paths
 Compressor Efficiency  •   79

their effectiveness. In some cases, the knife edges are machined onto the
rotating part, while a sleeve of soft material is positioned on the station-
ary part. The rotating part then cuts a groove into the stationary sleeve,
reducing leakage considerably and thus increasing the compressor effi-
ciency. The stationary sleeve can be manufactured in babbitt-lined steel,
ceramic, compressed steel-fiber material, etc., while the knives are made
of high-quality steel.
Shaft seals are usually mechanical-contact or liquid-film types, as
seen in Figure 2.16. In some cases where a small amount of leakage
can be tolerated, labyrinth seals are used. The oil- or liquid-film seal
consists of two stationary bushings that surround the shaft with a clear-
ance of a few thousandths of an inch. Oil at a nominal rate of 10 gpm
is introduced between the bushings at a positive pressure higher than
that of the process gas and leaks in both directions along the shaft. The
oil is retained in the seal housing by “O” rings. To limit the inward oil
leakage, the differential pressure across the inner bushing is only a few
pounds per square inch. The inner-leakage rate varies from 1 to 4 gph,
depending on the size of the seal, but is independent of the gas pressure
being contained.
This leakage is collected in a chamber that is usually separated from
the gas stream by a labyrinth seal. Overflow of oil from the leakage cham-
ber and its subsequent entry into the compressor is the biggest problem.
Mechanical-contact seals have two major elements (Figure 2.17).
These are the oil-to-process gas seal, or carbon ring, and the oil-to-
uncontaminated-seal-oil-drain seal, or breakdown bushing. This seal can
maintain a lower inner-leakage ratio with higher oil-to-gas differential
pressure.
In operation, the seal-oil pressure is maintained at about 25 to 50
psia over the process-gas pressure against which the seal is sealing.
High-pressure oil enters the seal cavity, completely filling it. Some of
the oil (ranging from 2 to 8 gph) is forced across the carbon-seal face
and flows out the contaminated-oil drain. The mechanical seal’s great
advantage over the oil-film seal is that it has a minimum effect on rotor
dynamics. On the other hand, when the oil-film bushings lose their
free-floating feature, they can upset the stability of the rotor when oper-
ating at high speeds.

BALANCING

Balancing of high-speed rotating machinery is very crucial in the smooth

running of the machine. Machines designed to operate higher than their
80  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 2.16.  Liquid film seal function

first critical speed need even more care. Presently, in most vendors’
shops, each wheel is balanced separately and then shrink-fitted to the
shaft. During this process, shaft runout must be closely watched, since
large changes could indicate that the impeller is not installed square to
the shaft. When all rotors are fitted on the shaft and the final balancing
is done, correction is usually made to the last impeller installed if an
unbalance exists.
 Compressor Efficiency  •   81

Figure 2.17.  Function of contact seals

The best technique for high-speed rotors is to balance them at their

rated speed, and not in low-speed machines. This is not always possible in
the shop, and so is often done in the field. To achieve best results, balanc-
ing should be performed at all rotating planes. This is done with the influ-
ence-coefficient technique. Here, the effect of unbalance on each plane,
and its influence on the other planes, is noted. Once the influence coeffi-
cients of each plane are known, balancing can be performed at each plane
and a smooth rotor obtained. Figure 2.18 shows rotor-response curves
­before and after balancing.

COUPLINGS

Flexible gear couplings have been the most widely used. In many new
machines, they are being replaced by disk-type couplings. Gear-type
82  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 2.18.  Balancing techniques before & after

couplings consist of a hub gear and a sleeve gear. In most cases, male
teeth are integral with the hub, but some couplings have male teeth
integral with the sleeve. Most of these couplings are mounted on a
tapered-shaft end. In many instances, a key and keyway are provided to
transmit torque; in others, the coupling is shrink-fitted on the tapered
shaft.
Lubrication of the gear coupling is accomplished by a continu-
ous or batch-packed technique. With continuous lubrication, nozzles
spray oil into the coupling’s teeth. The oil is recirculated through the
machine’s lubricant system, which dissipates the heat generated in the
coupling, and thus maintains a relatively constant temperature. With
batch-packed or seal lubrication, a recommended grease or oil is sealed
in the coupling, which is changed as necessary or during scheduled
shutdown. The major advantages of the batch-packed technique are
that the lubricant is the best available for the application and that the
­lubricant does not get contaminated. For high-temperature applications,
continuous lubrication is recommended. Such lubrication requires a
good filtration system; otherwise, a centrifuging of contaminants in the
gear teeth may occur.
The disk-type or metal-flexible coupling, shown in Figure 2.19, con-
sists of two hubs rigidly mounted by interference fit or by flange bolting
to the driven and driving shafts. Flexible elements, one attached to each
hub, compensate for misalignment. They are connected by a spacer (usu-
ally tubular) to span the gap between shafts. The major problem of this
coupling is its potential to vibrate when excited at its resonance frequency.
 Compressor Efficiency  •   83

Figure 2.19.  Flexible couplings

The major advantage is that it needs no lubrication system and can tolerate
a higher degree of misalignment.

LUBE-OIL SYSTEMS

API Standard 614 covers in detail the minimum requirements for lubrica-
tion systems, oil-type shaft-sealing systems, and control-oil supply sys-
tems for special-purpose applications.
The base system consists of a reservoir that should be separately
placed from the base plate. The working capacity of the oil tank should
be at least 5 minute, based on normal flow. The reservoir-retention time
should be 10 minute, based on normal flow and the total volume below the
minimum operating level.
Arrangements for heating the oil should also be made. If an immer-
sion heater is used, maximum watt density should be 15 W/in.2. If steam
heating is used, the heating element should be external to the reservoir.
The oil system should be equipped with a main oil pump, a standby
pump, and, for critical machinery, an emergency pump. Power sources for
the main and standby pumps should be different. For example, if steam is
used to power the main pump, the standby pump should be electrical. The
emergency pump is also usually electrical, but will be driven by either a
d.c. supply or a completely separate a.c. supply.
Twin oil coolers should be provided and piped in parallel, using
a single multiport transfer valve to direct the flow to the coolers. The
84  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

coolers should carry water on the tubeside and oil on the shellside. Oil-
side pressure should be greater than the waterside pressure. Twin full-
flow filters should be located downstream of the oil coolers. Do not pipe
the filters with separate inlet and outlet block valves. This could cause
loss of oil flow from the possible human error of blocking the flow during
filter switching.
Oil for turbomachinery should be of correct viscosity and formu-
lated with the required chemical additives to prevent rusting, resist
oxidation and sludging, be noncorrosive to machine parts, resist foam-
ing, and separate rapidly from water. The oil should be checked peri-
odically to detect changes in viscosity, pH and neutralization number,
and precipitation.

HIGH-SPEED GEARS

In many compressor applications, high-speed gears are used between the

drive (especially, electric motors) and the compressor. The application of
gear drives to large trains involving high speeds and power levels has never
been an easy task. Currently, high-hardness gearing is used, with tooth
loads in the range of 1,500 to 2,000 lb/in. of face, at pitchline velocities
of 20,000 to 30,000 ft/min. Normally, the gear drives used in such an
­application are of the single-helical or double-helical type, with rotors
carried in sleeve-type bearings. The double-helical gear is the first choice
for high-reliability applications.
Proper alignment of gears to the shafts is very important, and hot
checks on alignment are usually a must for high-speed gearing. When pos-
sible, gears should be run-in on initial startups. Speed and load should be
increased percentage wise. Lube-oil pressure and temperature should be
closely observed, and adjustments made to the lube system, as required.
Oil pressure must be maintained at all costs. High-speed gearing must be
treated with care.

CONTROL SYSTEMS

The controls for most compressor trains consist of two major systems: one
for the lubrication system and one for the compressor. For the lubrication
system, minimum alarms are: low oil pressure, low oil-pressure trip (at
some point lower than the alarm point), low oil level in the reservoir, high
oil-filter differential pressure, high thrust-bearing metal temperature, and
 Compressor Efficiency  •   85

high oil temperature. Each pressure and temperature-sensing switch should

be in a separate housing. The switch type should be single pole, double
throw, and furnished as open (deenergized) to alarm and close ­(energize)
to trip. Pressure switches for alarms should be installed, with a “T” con-
nection for a pressure gage and bleeder valve to test the alarms. Tempera-
tures should be monitored in the oil piping to and from the coolers, and
at the outlet of each radial and thrust bearing. Bearing-metal temperature
should also be measured, since problems will show up much faster in the
metal temperature than in the oil temperature. Figure 2.20 shows that the
change in metal temperature is much higher than the change in oil tem-
perature as load is increased.
Pressure gages should be provided at the discharge of the pumps,
bearing header, control-oil line, and seal-oil line. Each atmospheric oil-
drain line should be equipped with steel nondestructive bull’s-eye flow
indicators, positioned for viewing through the sides.
For the compressor, a complete instrumentation package should be
provided, as shown in Figure 2.21. In this manner, a full analysis can be
made for both vibration and performance characteristics. For vibration
analysis, both proximity probes and accelerometers have been suggested.
Figure 2.22 shows the ranges for which the various probes are applicable.
The reason for two different transducers is that proximity probes show
only the movement of the shaft at the location of the probes. The probes do
not pick up high-frequency problems such as blade-passing or gear-mesh
frequencies.

Figure 2.20.  Temperature changes

86  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Turbine

➊ ➋
2 Radial probes 2 Radial probes
1 Key phasor 1 Accelerometer
2 Thrust probes 1 Journal-bearing temperature
1 Accelerometer 1 Steam pressure
1 Journal-bearing temperature 1 Steam temperature
2 Thrust-bearing temperature
1 Steam pressure
1 Steam temperature
Low case
➌ ➍
2 Radial probes 2 Radial probes
1 Accelerometer 1 Accelerometer
1 Journal-bearing temperature 2 Thrust probes
2 Process pressure 1 Journal-bearing temperature
2 Process temperature 2 Thrust-bearing temperature
1 Process pressure
1 Process temperature
Gear box
➎ ➏
1 Accelerometer 1 Journal-bearing temperature
1 Journal-bearing temperature

➐ ➑
1 Journal-bearing temperature 1 Journal-bearing temperature
1 Accelerometer
1 Key phasor
High case
➒ ➓
2 Radial probes 2 Radial probes
1 Accelerometer 1 Accelerometer
1 Journal-bearing temperature 2 Thrust probes
2 Process pressure 1 Journal-bearing temperature
2 Process temperature 2 Thrust-bearing temperature
1 Process pressure
1 Process temperature

Monitoring probes required for compressor train

Figure 2.21.  Monitoring probes

 Compressor Efficiency  •   87

Figure 2.22.  Vibration analysis

Figure 2.23 is a real-time plot that shows the blade-passing

frequencies, dominating as the compressor approaches surge. High
subharmonic frequencies indicate oil whirl, and second harmonics
indicate some mis­alignment. In some cases, catastrophic failure has
occurred even though no indication was given by the proximity probes.
When properly used, proximity probes are excellent in locating sub-
harmonic and misalignment problems. Thus, by using a combination
of proximity probes and accelerometers, a complete picture of rotor
dynamics is obtained.
Surge-control instruments on the market are basically static devices.
These measure the pressure rise across the compressor, and sense the
flow rate from pressure drop across an orifice. The pressure rise and
flow rate are then compared with values programmed into the device,
and when they exceed these values, the signal is given for a bypass valve
to open.
The biggest drawback of this system is that it does not provide for
any changes in the molecular weight of the gas or the degradation of the
compressor itself. Presently, programs are under way that measure the
dynamic change in the pressure head in the boundary layer. When flow
reversal is sensed, the bypass valve is opened. This program is still being
evaluated.
88
Figure 2.23.  Vibration Spike Types
 Compressor Efficiency  •   89

Until now, monitoring systems have consisted of no more than vibra-

tion monitors. With the advent of real-time analyzers and minicomputers, a
whole new era is opening in monitoring and diagnostic systems. These new
systems indicate problem areas, give expected life of various components,
and schedule maintenance. They correlate both process and mechanical
parameters to fully diagnose the compressor train and run it at its most
efficient point. These systems use a performance matrix in combination
with trending data and real-time analysis to properly project and diagnose
problems. A good monitoring and diagnostic system can save thousands of
dollars in downtime and energy.
 CHAPTER 3

Basics of Surge
Control for Centrifugal
Compressors
92  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Surging is an unstable operating condition that causes erratic com-

pressor performance. Here is a basic anti-surge control scheme for
centrifugal compressors, along with suggestions on how to cope with
variations in operating conditions and how to handle multi-compressor
arrangements.
Compressors, like other major processing equipment, should be con-
trolled as effectively as possible to ensure efficient plant operation. In
devising a control scheme, the designer must consider how to avoid an
unstable operating condition known as surging.
Simply stated, surging occurs when the compressor throughput is re-
duced to a point sufficiently below design conditions so that erratic per-
formance results. The job of the anti-surge system, then, is to detect the
potential upset and automatically compensate for it by maintaining a flow
through the compressor in excess of the surge condition.
Before getting into the details of how this system works, we must
first understand the phenomenon of surging from the standpoint of
compressor-system variables.

SURGING CHARACTERISTICS

A typical family of characteristic curves for a variable-speed centrifugal

compressor is shown in Figure 3.1. These curves depict the adiabatic head
as a function of actual inlet-volume flow, with a characteristic curve for
each rotational speed.
 BASICS OF SURGE CONTROL  •   93

Figure 3.1.  Characteristic curves for a typical

centrifugal compressor are based on Eq. (3)

If we assume that the compressor is operating at Point A on the 100

percent-speed curve, the inlet flow is Q and the head is L. Now if the
external load resistance gradually increases while the speed remains con-
stant, the flow decreases and the operating point will move to the left along
the 100 percent-speed characteristic curve. When it reaches Point B, the
flow decreases to Q’, and the head increases to L’, the maximum head the
compressor can produce at this speed.
At this point, the characteristic curve is practically flat and the oper-
ation of the compressor becomes unstable. This condition is called surg-
ing, and it appears as rapid pulsation in the flow and discharge pressure,
producing high-frequency reversals in the axial thrust on the compressor
shaft. In some machines, surging can become severe enough to cause
­mechanical damage. To avoid this danger, a control system must be devised
to prevent the compressor from operating in this unstable area.

BASIC CONTROL THEORY

In all centrifugal machines (both pumps and compressors), the head pro-
duced is proportional to the square of the speed, while the flow rate is
linearly proportional to the speed. These relationships are described in the
following equations:

Q = CN (1)
94  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

L = CN 2 (2)

Substituting for N yields:

L = CQ 2 (3)

This equation is used to calculate the parabolic surge curve usually

provided by the compressor manufacturer.
Since there is no way to measure the adiabatic head directly, we must
find another quantity to use in its place. The one most easily related to
head is compression ratio, Rc.

R c = P2 / P1 (4)

Equation (5) is used to relate the compression ratio to the adiabatic

head:

Rc = (1 + Lmφ/1545T1Z1)1/φ (5)

where

φ = (K − 1)/K

When the compressor handles gas of only one composition, and when
the suction temperature is nearly constant, the quantities m, φ, T1, and Z1
are constant, and Eq. (5) reduces to:

Rc = (1 + CL)−1/φ (6)

Although Eq. (6) shows that the relation between Rc and L is not linear,
when this equation is plotted for most of the gases commonly encountered
(e.g., air, natural gas, etc.), the curves approach straight lines over the nor-
mal operating ranges. The quantity (Rc − 1) can then be substituted for L
without producing significant distortion of the surge curves.
Equation (3) shows that the relation between the adiabatic head and
Q2 is linear. Figure 3.2 is a plot of L and the corresponding values of (Rc
− 1) versus Q2 for the same range of heads for natural gas to show the
magnitude of the error introduced by substituting (Rc − 1) for L in Eq.
(3). At greater heads, this error increases and may require an adjustment
in the system to correct it, but in most cases, the error is small enough to
be neglected.
 BASICS OF SURGE CONTROL  •   95

This permits us to write the equation:

R c − 1=CQ 2 (7)

The differential pressure across the compressor is:

ΔP = P2 − P1 (8)

From Eq. (4),

P2 = P1Rc

Then:

ΔP = P1Rc − P1

ΔP = P1(Rc − 1) (9)

Rc − 1 = ΔP/P1 (10)

Substituting Eq. (10) into Eq. (7) provides:

ΔP/P1 = CQ2 (11)

Nomenclature
c Constant (not necessarily the same in all equations)
h Differential pressure across primary in. of water flow device,
K Specific heat ratio
L Adiabatic head, ft.-lb./lb.
m Molecular weight
N Compressor speed, rpm.
P Pressure, psia.
Q Actual inlet-volume flow, cfm.
Rc Compression ratio
T Absolute temperature, °R.
V Specific volume, cu. ft./lb.
W Mass flowrate, lb./min.
z Supercompressibtlity factor
Greek letters
Δ Differential
φ Ratio of (K - 1) to K
Subscripts
D Design
1 Suction
2 Discharge
96  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 3.2.  Error resulting from replacing the

adiabatic head with the com­pression-ratio factor
is insignificant

To measure Q, we must have a primary device, such as an orifice or

a venturi tube, in the compressor suction line. If we let h represent the
differential pressure across this primary device in inches of water, we can
calculate the mass flow, W, to the compressor from the equation:

W = C hP1 / T1 (12)

V = C (T1 / P1 ) (13)

Q = WV (14)

Substituting Eqs. (12) and (13) in Eq. (14) produces:

Q = C hP1 / T1 × T1 / P1
(15)
Q = C hT1 / P1

If, as before, we assume the temperature to be constant, then:

Q = C h / P1
Q 2 = Ch / P (16)
 BASICS OF SURGE CONTROL  •   97

Figure 3.3.  Control line is generally displaced about 10

percent of Q, or 20 percent of ΔP, from the surge curve

Substituting Eq. (16) in Eq. (11) yields:

∆P / P1 = C ( h / P1 )
∆P = Ch (17)

Equation (17) is used to calculate the surge curve, which will be our
basic design equation. It verifies two important points:

1. There is a linear relationship between ΔP and h, as seen in Figure 3.3

and
2. The surge curve is not affected by variations in suction pressure, P1.

Figure 3.3 also shows the control line, displaced to the right of the
surge curve. Moving the control line to the right reduces the possibility of
reaching surge conditions in case a rapid decrease in flow causes an over-
shoot to the left of the control line. However, if the control line is placed
too far to the right, gas will be bypassed unnecessarily and power wasted.
Consequently, the location of the control line represents a compro-
mise based on the actual operating conditions of each individual system.
As a general rule, the control line is displaced about 10 percent of flow,
98  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

or 20 percent of differential pressure. However, there are cases where the

compressor must operate close to the surge curve, and it is necessary to
reduce this margin.

EFFECT OF TEMPERATURE CHANGES

In deriving Eq. (17), we assumed that the compressor inlet temperature

was constant. Since this is not always true in actual practice, it is necessary
to investigate the effect of changes in this variable on the performance of
the control system.
In order to determine the change in ΔP resulting from a change in
T1, we start with Eq. (5), which expresses the relationship between com-
pression ratio (Rc) and temperature. If it is assumed that the head and gas
composition are constant, the factors L, m, and φ in the equation will be
constant. The factor Z1 may be somewhat affected by the temperature, but
for this part of the analysis, it will be considered constant.
Equation (5) reduces to:

Rc = (1 + C/T1)1/φ (18)

Figure 3.4 shows a plot of this equation for two gases, air and natural
gas, that have widely different molecular weights and specific heat ratios.
The curves show the magnitude of the change in the compression ratio for
a 120°F. change in the inlet temperature in a compressor producing a head
of 40,000 ft·lb/lb.
From Eq. (9), ΔP = P1 (Rc − 1), it is evident that for any particular
value of suction pressure (P1), ΔP will vary directly with Rc − 1, and the
shape of the curves for ΔP versus T1 is the same as that for Rc versus T1.
Now, we will consider the effect of changes in the differential pres-
sure across the orifice, h, resulting from variations in suction temperature.
Starting with Eq. (15),

Q = C hT1 / P1
Q 2 = C ( hT1 / P1 ) (19)
h = C (Q 2 P1 / T1 )

Then, for specific values of Q and P1

h = C/T1 (20)
 BASICS OF SURGE CONTROL  •   99

The curves in Figure 3.4 also show that both ΔP1 and h vary inversely
with the absolute inlet temperature, T1. If these temperature effects were
identical, both sides of Eq. (17) would be changed equally and the sys-
tem would be exactly self-compensating. However, this is not the case—­
although both variables change in the same direction, the magnitudes are
different, with the result that there is a shift in the slope of the surge curve.
In most cases, this problem can be solved simply by placing the con-
trol line to the right of the surge curve for the highest temperature and
accepting the fact that, when minimum temperature conditions exist, some
gas may be bypassed unnecessarily.
If temperature variations are too great, or if the compressor must
operate close to the surge curve, it will be necessary to provide tempera-
ture compensation in the control system. The instrumentation required to
perform this function will be described later.

MOLECULAR-WEIGHT CHANGES

In most processes, each compressor handles the same gas all the time. Typ-
ical examples are air, refrigerant, and natural-gas pipeline compressors. In
some applications, however, the compressor may be called on to handle
gases of varying composition.
Since a variation in composition almost always results in a change in
molecular weight, it is important to determine the effect of this change on
the surge curve. Since the surge curve is plotted with ΔP versus h, we must
determine how each of these variables is affected.
Referring back to Eq. (5):

Rc = (1 + Lmφ / 1,545T1 Z1 )1/ φ

for fixed values of L, T1, and Z1, this equation reduces to:

Rc = (1 + Cmφ )1/ φ (21)

This equation indicates that, as with temperature, there is a nonlinear

relation between molecular weight, m, and differential pressure across the
compressor, ΔP. This is further complicated by the fact that a change in
molecular weight is accompanied by a change in the specific heat ratio, K,
and consequently the factor φ.
In some cases, such as the lighter saturated hydrocarbons (i.e., meth-
ane, ethane, and propane, and mixtures of these gases), there is a definite
100  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

relationship between m and K, as shown in Figure 3.5. This permits an

accurate calculation of the change in Rc and ΔP for a given change in m.
However, in some processes, the compressor may be handling two or
more gases where there is no correlation between m and K. An example of
this would be a methane compressor that is also periodically used to purge
the system with nitrogen. Under such a condition, each value of Rc and ΔP
must be calculated individually to determine the magnitude and direction
of the change.
To compute the effect of changes in m on the value of h, we must
include m in Eq. (15), which then becomes:

Q = C hT1 / P1m (22)

h = CQ 2 P1m / T1 (23)

Then, for specific values of Q, P1, and T1,

h = Cm (24)

By comparing Eqs. (21) and (24), it is apparent that the system will
not be exactly self-compensating for changes in molecular weight. For
lighter hydrocarbons, changes in m produce changes in the same direction
in both ΔP and h, but in different magnitudes. Figure 3.6 shows how the
surge line shifts in slope for changes in m from 16 to 20.

Figure 3.4.  Curves for both compression ratio, Rc,

and h vary inversely with gas inlet temperature
 BASICS OF SURGE CONTROL  •   101

Figure 3.5.  Definite relationship exists between

molecular weight and specific heat ratio for
lighter hydrocarbons

Figure 3.6. Increasing m shifts the surge curve to

the left
102  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

It is not possible to generalize, however, concerning the over-

all effect of molecular-weight changes because of the influence of the
specific-heat-ratio factor. As mentioned earlier, when two or more unre-
lated gases are handled by the same compressor, the slope of the surge
line must be calculated for each condition. And if there is too great a
divergence to be covered by a single control line, provision must be made
to adjust the slope as required. This is usually done manually.

Figure 3.7.  A basic scheme to provide anti-surge control

Figure 3.8.  “Batch” controller provides anti-windup protection

 BASICS OF SURGE CONTROL  •   103

INSTRUMENTATION

The instrumentation required to use Eq. (17) is shown in Figure 3.7. This
equipment, of course, may be either pneumatic or electronic. A high-range
differential-pressure transmitter measures ΔP and sends an output signal
that becomes the measurement signal to the surge controller. The flow
transmitter is connected across a primary device such as an orifice or a
venturi in the compressor suction line, and its output signal, h, is fed to a
ratio station. There it is multiplied by the constant C and becomes the set
point of the surge controller.
The surge controller should have, in addition to proportional and reset
functions, an anti-windup function, sometimes referred to as the “batch”
feature.
To understand the need for a “batch” feature, consider Figure 3.8.
­Under normal conditions, the compressor operates in an area some distance
from the control line. This results in an offset between the measurement
and the set point of the controller. In a standard proportional-plus-reset
controller, this causes the output signal to wind up to either its high or low
limit.
In this condition, the proportional band and the operating point will
be on opposite sides of the set point or control line, and no control action
will be obtained until the measured operating point reaches the control
line. If the measurement approaches the control line rapidly, it will over-
shoot before the controller can unwind, and the compressor may surge.
The anti-windup or “batch” function is arranged so that when the con-
troller output reaches its limit, the reset loading is adjusted to shift the
proportional band to the same side of the control line as the measurement,
as indicated in Figure 3.8. Then, if it approaches the control line rapidly,
the measurement enters the proportional band, and the control action starts
before it reaches the con­trol line; overshoot is thus reduced or eliminated.
If the measurement change is slow—that is, no faster than the reset rate—
the controller will perform as a standard instrument.

INSTRUMENT RANGES

The basic anti-surge control system shown in Figure 3.7 includes two dif-
ferential-pressure transmitters—one measuring the differential pressure
across the primary flow device, h, and the other measuring the differential
pressure across the compressor, ΔP.
If flow measurement is used only for anti-surge control, the primary
device can be designed for optimum conditions. The preferred design for
maximum flow is about 25 to 30 percent above the maximum volume flow
104  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

indicated by the surge curve. To keep pressure loss as low as possible, the
maximum differential pressure, h, should be as small as practical.
If flow measurement is used in the normal operation of the compres-
sor, the primary device will probably be designed for a considerably higher
maximum flow. As a result, the value of h at the surge line will be a much
smaller percent of full scale and may require an undesirably high setting of
the ratio station. Under these circumstances, it is better to install a second
transmitter with a lower range for use only with the anti-surge system.
Calculations required to establish the range of the ΔP transmitter d­ epend
on the variables used by the compressor manufacturer to plot the surge curve.
When the curve is plotted with discharge pressure versus flow for a given
suction pressure, a simple subtraction produces the maximum ΔP at surge.
If the surge curve is plotted using compression ratio versus flow, the max-
imum value of operating point (OP) can be obtained from Eq. (9). However, if
the manufacturer has used adiabatic head versus flow to draw the surge curve,
calculations become a little more involved, and it is necessary to use Eq. (5) to
calculate the compression ratio and then compute ΔP from Eq. (9).
Having determined the maximum ΔP, the transmitter range should
be set at some reasonable amount above this value (e.g., 20–25 percent).
In order to plot the surge curve in the form shown in Figure 3.3, which
is expressed by Eq. (17), about four points should be selected from the man-
ufacturer’s surge curve, and the corresponding values of ΔP and h computed.
The ΔP values can be calculated in the same manner as the range of
the ΔP transmitter is determined. To compute the corresponding values of
h, we use Eq. (15):

Q = C hT1 / P1
(25)
h = CQ 2 P1 / T1

Using the subscript D to indicate the design conditions for the pri-
mary flow device, the equation for full-scale flow is:

H D = CQ 2D PD / TD (26)

and dividing Eq. (25) by Eq. (26) yields:

h = ( hD )Q 2 P1TD / Q 2D PD T1 (27)

If the Rowing pressure and temperature are the same as the design
conditions, Eq. (27) reduces to:

h = hD (Q / QD )2 (28)
 BASICS OF SURGE CONTROL  •   105

Points plotted using ΔP and h will usually fall very close to a straight
line. The control line in Figure 3.3 is located by adding 5 to 10 percent
to the flow values and calculating the corresponding values of h. This
provides the necessary safety margin to keep the compressor out of the
surge area.
Some compressors with very high compression ratios have parabolic
surge curves in the lower range that straighten out, or even bend to the
right, at the upper end. These curves, when replotted as ΔP versus h, ob-
viously do not produce a straight line. However, in almost every case, a
control line can be drawn that will provide adequate protection over the
normal operating range of the machine. If this control line does not pass
through the origin, it is necessary to add bias between the ratio station and
the controller.

RATIO SETTING

With the control line established, as in Figure 3.3, the slope of the line (the
factor C in Eq. 17) represents the setting of the ratio station. Rearranging
Eq. (17) produces:

C =ΔP/h (29)

In calculating C, values of ΔP and h must be expressed in terms of the

percentage of full scale of their respective transmitters. This calculation
can best be shown by an example. If we assume, as in Figure 3.3, that the
transmitters have the following ranges:

ΔP = 0–500 psi

h = 0–50 in. of water

and if we select an arbitrary point, A, on the control line where:

∆P = 300 psi
h = 22 in. of water
then: ∆P = (300 / 500)100 = 60%
and : h = (22 / 50)100 = 44%
Using these values in Eq. (29) gives:
C = 60 / 44 = 1.36
and the equation becomes:
∆P = 1.36 h
106  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

AUTOMATIC TEMPERATURE COMPENSATION

As we mentioned earlier, the slope of the surge curve changes with varia-
tions in the compressor inlet temperature. When these changes in the inlet
temperature are large and occur frequently, and when the compressor must
operate near the surge curve, it may be wise to provide automatic tempera-
ture compensation in the anti-surge control system.
This can be done by replacing the manually adjusted ratio station in
Figure 3.7 with an analog dividing computer, as shown in Figure 3.9. By
properly scaling the computer, the equation:

ΔP = Ch/T1 (30)

can be solved. This equation indicates that the temperature correction is

inversely proportional to the first power of the absolute temperature.
Although, in theory, this is not precisely correct, the error introduced by
this assumption is negligible.
When the largest temperature changes occur over long periods
of time (e.g., seasonally), changes in the slope of the control line can
be made manually to keep the system as simple as possible (as in
Figure 3.7).

Figure 3.9.  Control arrangement to provide

automatic temperature compensation uses an
analog dividing computer
 BASICS OF SURGE CONTROL  •   107

INLET-GUIDE VANES

Constant-speed centrifugal and axial compressors are frequently equipped

with adjustable inlet-guide vanes to control the output flow. In addition,
the axial compressor may have adjustable stator blades. Besides influenc-
ing the compressor output, moving these vanes also changes the slope of
the surge curve. In other words, there is a surge line for each vane position,
just as there is for each inlet temperature.
Since the magnitude of this change in the slope of the surge line is dif-
ferent for each compressor design, there is no convenient way to express
it mathematically. However, it is generally true that moving the vanes in
the counter-rotation direction will reduce the slope of the surge line (i.e., it
moves the surge line to the right). It is also generally true that the change
in slope is not a linear function of the vane position.
There are several solutions to this problem. If the change in slope is rea-
sonably small, it may be possible to use the surge line for the maximum count-
er-rotation vane position and set the control line accordingly, as in Figure 3.3.
This places the control line on the safe side of the surge line for all vane
positions. But it also provides an unnecessarily wide safety margin when the
vanes are moved in the pre-rotation direction. This could result in bypassing
gas at times when it is not required. This drawback must be weighed against
the advantage of using the simple control system shown in Figure 3.7.
When vanes are positioned from a manual station, and the position is
not frequently changed, it may be satisfactory to use the basic control sys-
tem and adjust the setting of the ratio station manually. This can be done
by providing the operator with a table showing the correct ratio setting for
each vane position.
When frequent changes in the vane position produce large variations in
the slope of the surge line (as when they are operated by a controller), it is usu-
ally desirable to make the changes in the ratio setting automatically. A system
like the one for automatic temperature compensation (Figure 3.9) can be used.
The signal from the controller, t, the vane operator is a measure of the vane
position and, instead of the temperature measurement, can be fed into the ana-
log computer to adjust the ratio setting. This system produces a linear relation
between the ratio setting and the vane position that is not precisely correct.
However, the errors introduced are usually small enough to be neglected.

DISCHARGE-LINE FLOW MEASUREMENT

The anti-surge control system based on Eq. (17) uses a primary

flow-measuring device in the compressor suction line. This system offers
108  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

the very important advantage of being self-compensating for changes in

suction pressure and, at least partially, for variations in suction tempera-
ture. These benefits are not realized when the flow measurement device is
located in the discharge line.
In some installations, because of the piping configuration or the size
of the line, it is not feasible to put the primary flow device in the suction
line. This may also be true when the compressor operates with a very low
suction pressure and any additional pressure drop in the line cannot be
tolerated. It may be possible to resolve the problem by using the inlet eye
of the compressor as an orifice. However, it must first be established that
this measurement is a usable one (i.e., noise free, repeatable, and represen-
tative). It also requires individual calibration.
If this cannot be done, it is necessary to install the primary flow device
in the discharge line and modify the anti-surge control system accordingly.
This modification consists of adding analog computers to calculate the
value of h for use in Eq. (17). The instrumentation to perform this compu-
tation is shown schematically in Figure 3.10.
Under steady-state conditions, the weight and flow (in lb/min or in
scfm) are the same at both suction and discharge, that is,

W1 −  W2 (31)

Using Eq. (12) to substitute for W provides:

C1 h1 P1 / T1 = C2 h2 P2 / T2

If the primary flow devices are designed for the same maximum values
of weight flow, and if we make

h1 P1 / T1 = h2 P2 / T2

then

h1 = h2 ( P2T1  / PT
1 2) (32)

With these relative maximum values of h1 and h2, the same relation
will hold true at any point on the scale and we can use Eq. (32) to calculate
h1 under all conditions. The surge curve and control line can be plotted in
the conventional manner, and the setting of the ratio station can be com-
puted by the method described earlier.
 BASICS OF SURGE CONTROL  •   109

Although the inlet temperature may vary considerably, in some com-

pressor applications, the ratio T2/T1 will be nearly constant (remember
these are absolute temperatures).
If the temperature ratio is assumed constant, Eq. (32) reduces to:

h2 = Ch2P2 /P1 (33)

This greatly simplifies the control system by eliminating the tempera-

ture transmitters as well as two analog computers (i.e., the one to divide T2
by T1 and the one to multiply the temperature ratio by the pressure ratio).
When the compressor operates at a constant speed, the instrumenta-
tion can be further simplified. Under this condition, the compression ratio
is constant at the surge point, so Eq. (33) becomes:

h1 = Ch2 (34)

This makes it possible to eliminate two more analog computers from

Figure 3.10 and use the standard basic control system, by replacing h1 with
the corrected value of h2.

COMPRESSORS IN A SERIES

When two or more compressors are connected in a series, using the con-
trol system shown in Figure 3.7 on each individual machine will provide
the best protection against surge. This, of course, requires a recirculation
control valve for each compressor.
In some designs, two compressors in a series are driven by the same
prime mover, and consequently, they both always run at the same speed.
The machines are sometimes considered as a single unit: The manufac-
turer may provide one surge curve for the combination and the installation
usually includes only one recirculation control valve. For surge control,
the system shown in Figure 3.7 is used.
When compressors are independently driven at variable speed, but
have a single recirculation control valve, the surge is controlled by the
pneumatic system shown in Figure 3.11. Each compressor is equipped
with its own ratio station and controller, but the two are combined into
an auto-selector system to enable either controller to operate the common
recirculation valve as required.
If, at times, either compressor operates singly, the three-way valves in
the controller output lines should be installed to ensure a positive air supply
110  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

to the selector relay of the controller in operation. If both compressors will

always operate together, the three-way valves may be eliminated altogether.

COMPRESSORS IN PARALLEL

When centrifugal compressors are operated in parallel, the problem of prop-

erly dividing the load is always present. The solution is particularly difficult
when the compressors have dissimilar characteristics. But even when the
characteristics are identical, some means must be provided to ensure that
the desired load distribution is obtained. Controls to accomplish this will
not be discussed here, except as they affect the anti-surge control system.
Figure 3.12 shows the effect of inadequate control. With a differential
pressure of 80 psi and both compressors running at 100 percent speed, the
load is equally divided and the flow through each machine is Q1.
If, however, for some reason such as dissimilar governor characteris-
tics, one compressor is operating at only 98 percent speed and the other is
at 102 percent, the flows will be Q2 and Q3, respectively. If the difference
in speeds is large enough, the flow through one machine will decrease
sufficiently to cause surging.
From Figure 3.12, it can be seen that the flatter the characteristic
curves, the more difficult the problem becomes. For this reason, centrif-
ugal compressors require more precise methods of load distribution than
axial-flow compressors, which have much steeper curves. This suscepti-
bility to surge makes it even more essential to provide positive protection
for centrifugal machines during parallel operation.
The control arrangement shown in Figure 3.13(a) allows maximum
­operating flexibility with minimum supervision. It consists of a standard surge
control system for each compressor, which provides complete protection at
all times, whether the machines are operating singly or together. It is even
possible to start a compressor with the surge controller in automatic while
the other com­pressor is running. As the newly started machine approaches its
operating speed, the surge controller holds the recirculation valve open just
far enough to bring the compressor up along the control line and keep it from
surging. And this system can be used for any number of machines.
Figure 3.13(b) shows an alternate surge control system that requires
less equipment. It uses only one ratio station and controller, one OP trans-
mitter, and a single recirculation control valve. When both compressors
are operating, the low selector relay chooses the lower flow signal and
sends it to the ratio station. If only one machine is running, its flow signal
is connected to the ratio station by a manual selector switch. This selection
can also be per­formed automatically with relay logic, if desired.
 BASICS OF SURGE CONTROL  •   111

Figure 3.10.  A schematic diagram shows the

instrumentation required when the primary flow-
measuring device is located in the centrifugal-
compressor discharge line

Figure 3.11.  The system for controlling

compressors in a series
112  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 3.12.  Load distribution for parallel

compressors can be seriously affected if they run
at different speeds

Figure 3.13.  Alternate control arrangements for centrifugal compressors

operating in parallel

The cost savings resulting from the use of this system, however, are
not obtained without some sacrifice. For example, each compressor must
be equipped with a manual recirculation valve for starting. With one ma-
chine operating, the recirculation control valve cannot be used for starting
the second compressor, so the operation must be performed manually and
without automatic surge protection.
 BASICS OF SURGE CONTROL  •   113

IMPROVED SURGE CONTROL FOR CENTRIFUGAL

COMPRESSORS

Surge may occur without being detected by conventional control sys-

tems, cutting compressor and process efficiency and causing compres-
sor damage. Excess flow recirculation or blowoff can help avoid some
surges, but energy costs are high. Here is a control system that copes
with all surges without incurring such penalties.
The operation of a centrifugal compressor can become unstable due to
changes in many conditions such as flow rate, pressure, and the molecular
weight of gas. This causes rapid pulsations in flow, called surge. No sys-
tem is immune to sudden upsets at one time or another.
We have surge-tested air compressors, gas compressors, centrifugal
compressors, and axial compressors, at high and low pressures. The am-
plitude of the flow drop and the frequency of surge cycles vary. However,
there are certain common results: The pressure change is less than the flow
change in every case that we have recorded; flow usually drops extremely
fast just before surge, and then always drops precipitously during surge, at
least in the authors’ experience, a time of 0.05 second from set point flow
to reverse flow being common.
The speed of some surges is such that conventional control systems
cannot detect them, let alone react to them. So, often even when records
indicate that no surge has taken place, stripping down a compressor will
reveal surge damage, ranging from changes in clearances, which exact a
penalty in compressor efficiency, to destruction of parts.
Unstable operation—whether it is detected or not—affects the opera-
tion of the process that relies on the unit.
Typically, surge is prevented by recirculating some of the flow or by
blowing off the excess flow. Constant recirculation of flow—often at 30
to 40 percent of the amount needed by the process—is common. For in-
stance, this is found frequently in units used to compress chlorine and wet
gases, among others. Obviously, this excess flow costs energy. And it does
not necessarily avert all surges.

IMPROVED SURGE CONTROL

Here, we will present an improved surge-control strategy and instrumen-

tation system. This system copes with even very rapid upsets, minimizing
disturbances to process operations. The amount of recirculation is reduced
sharply, saving significant amounts of energy. Furthermore, the system
lends itself to actual field calculation of the surge limit.
114  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

With conventional closed-loop control schemes, once surge begins,

the compressor control system oscillates, and the only way to stop this
is by manual override. Here, if surge occurs, the compressor will auto-
matically be brought out of it during the first surge cycle, and the control
system will reset itself to prevent such oscillations from recurring.
Conventional systems can be set into oscillation by operator errors;
such errors are minimized by this system. Also, fluctuations in the pres-
sure and flow rate of the compressed gas are kept to a minimum, result-
ing in relatively minor disturbances to the process. (The reasons for these
­advantages will be explained later.)
In order to develop this control system, we will first relate surge to
compressor operation, then determine where surge begins and where to
set the control system to prevent it. Next, we will discuss instrumenta-
tion, control strategies, and calibration of the equipment. Compressors in a
­series and in parallel will not be discussed here, due to limitations of space.

SURGE CHARACTERISTICS

Figure 3.14 shows characteristic curves for a typical single-stage centrifugal

compressor. Each rotational speed (N1, N2, N3, N4) has its own characteristic
curve.
If the compressor is operating at speed N1 at Point A (mass flow rate WA)
and the flow is decreased, surge will occur at WB. The pressure has increased
from PD,A. to PD,B. Point B lies on the surge limit line, a naturally occurring

Figure 3.14.  Surge characteristics

 BASICS OF SURGE CONTROL  •   115

line that is peculiar to each compressor geometry. The area to the left of this
line is the surge zone, where there are oscillations of flow and pressure.
The line to the right of the surge limit line is the surge control line,
which is an artificial line set by an engineer. These lines are the boundaries
of the recirculation zone, where recirculation or blowoff is used to prevent
surge. To the right of the control line is the safe-operating zone, where
recirculation is considered unnecessary. In order to develop the control
strategy, we must first develop improved equations for the control and
limit lines.

SURGE LIMIT LINE

Here, we will formulate an equation for the surge limit line that does not
assume constant inlet temperature and molecular weight. Typical equa-
tions assume that these parameters are constant. Corrections must be
­applied when conditions change.
The head of a centrifugal compressor is a function of angular velocity,
speed of rotation, inlet volumetric flow rate, and impeller diameter. Using
a form of the kinetic energy of the compressor, given by Davis and Cor-
ripio, and performing a dimensional analysis:

gc H p  Q 
= f  3 3 , Ma, Re (35)
N 2 D2  ND 

At high velocities at which centrifugal compressors operate, the vari-

ation in Reynolds number, Re, with the velocity of the gas is negligible.
The variation in Mach number, Ma, with velocity is also negligible. Mach
number can be assumed constant as long as the gas velocity does not
­approach the speed of sound. We will assume that operation is well below
this speed in the region close to surge.
For a given compressor geometry, under these conditions, if Hp/N2 is
plotted against Q8/N, then the performance curves for different speeds will
reduce to one curve, called the universal performance curve (Figure 3.15).
The surge limit line is reduced to a single point, the surge limit. Since
the surge limit is fixed for a particular geometry, the values of this point
are constant:

Hp
= C1 (36)
N2

Q3
= C2 (37)
N
116  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Thus, the curves of the surge limit line follow the fan law, which states
that the inlet volume is proportional to speed, and the head is proportional
to the square of speed. The surge limit may be defined by either Eqs. (36)
or (37). These equations apply not only to single-stage centrifugals but
also to multistage units without intercoolers.
While Eqs. (36) and (37) can be used to set the surge limit (and the
surge limit line), this is impractical since HP and Qs depend on a measure-
ment of molecular weight.
Variations in molecular weight cannot be continuously measured at
the high speeds present in compressors and, therefore, must be eliminated
from the equation of the surge line. Also, HP and Qs depend on variations
in the inlet temperature and pressure. We will eliminate molecular weight
and temperature from the equation of the surge limit line. Eqs. (36) and
(37) are combined to eliminate N:

C3 H P = Qs2
(38)

Polytropic head is given by:

C4 Z av T3 ( Rcσ − 1)
Hp = (39)
M σ

Substituting HP in Eq. (38):

C5 Z av T3 ( Rcσ − 1)
= Qs2 (40)
M σ
The volumetric flow rate equals the mass flow rate divided by density:

W C6W
Qs = = (41)
density ( Ps M / Z s Ts )

Substituting Qs in Eq. (40):

( Rcσ − 1) W 2 Z82T8
C7 Z av = (42)
σ P82 M

Using an equation for gas flow across an orifice in the suction of the
compressor:

∆Por , s P8 M
W = C8 (43)
T8 Z8
 BASICS OF SURGE CONTROL  •   117

Equation (42) becomes:

( Rcσ − 1) ∆Por , s P8 MZ82T8

C9 Z av = (44)
σ Z8T8 P82 M
Or

Z av ( Rcσ − 1)
C9 P8 = ∆Por , s (45)
Z8 σ

Equation (45) may be written for the discharge side of the compres-
sor. Since the mass flow rate at any instant in the compressor is the same
at suction and discharge, Eq. (43) can be written as:

∆Por ,8 P8 M ∆Por ,d Pd M
W = C8 = C8 (46)
T8 Z8 Td Z d

Or

Pd T8 Z8
∆Por ,8 = ∆Por ,d (47)
P8 Td Z d

Figure 3.15.  Performance curves

118  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 3.16.  Surge control

Nomenclature
ba, bb, bn Distance between surge control line and surge limit line
(n= 1,2,3). in. H2O
Cn Constant (n= 1,2,3,. . .). Units vary from equation to
equation
D Impeller diameter, ft
d1 Bias added by summing device in No. 5, Fig. 7, psi
d2 Bias added by summing device in No. 7, Fig. 7, psi
gc Acceleration of gravity, ft/s2
Hp Polytropic head, (ft-lb)/lb
k Ratio of specific heat at constant pressure to specific heat
at constant volume
k1 Slope of surge control line, dimensionless
 BASICS OF SURGE CONTROL  •   119

M Molecular weight
Ma Mach. number
N Speed of rotation, rpm
ΔPc Differential pressure across compressor, psi
ΔPor,s’ΔPor,d Orifice pressure at suction; discharge psi
Ps, Pd Suction pressure; discharge pressure, psi uction;
discharge psi pressure, psi
Qs Volumetric flowrate in suction, actual cfm
Rc Pressure ratio across compressor
Re Reynolds number
Ts , Td Temperature at suction; discharge, R
W Mass flowrate, scfm
Zav Average compressibility = (Zs + Zd)/2
Greek litters
σ k −1
kη p

ηp Polytropic efficiency

Equation (45) becomes:

Z av Z d Td P82 ( Rcσ − 1)
C10 = ∆Por ,d (48)
Z82T8 Pd σ

If we assume that charges in compressibility are small, then Zav = Zd

=Zs. The term (Rcs − 1)/s can be approximated by a linear function
(Figure 3.16). If Rc is less than 2.5, the approximation C11(Rc − 1) can be
used. Equation (45) and (48) become:

C12 ( Pd − P8 ) = C12 ∆Pc = ∆Por , s (49)

P8Td PT
C13 ( Pd − P8 ) = C13 8 d = ∆Por ,d (50)
Pd T8 Pd T8

However, White arrived at the same result by assuming that the inlet
temperature and molecular weight were constant; he later offered correc-
tions for these. Here, we have shown that such corrections are unnecessary.
120  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

As in White’s case, here the surge limit line is unaffected by variations in

suction pressure.
If Rc is greater than 2.5, (Rcs − 1)/s may be approximated by C14(ΔPc
− 1) + a and the surge limit line becomes:

C15 ∆Pc + a2 P8 = ∆Por , s

(51)

When changes in inlet pressure are negligible:

C15 ∆Pc + b1 = Por , s

(52)

where b1 is a constant.
Note that the above equations apply to centrifugal compressors with-
out intercoolers. With intercoolers, it is necessary to consider each section
as a separate compressor or to use an approximation that considers the unit
as a whole.
Eqs. (49), (51), and (52) are recommended for calculating the surge
limit line, as they contain the fewest number of variables, all of which can
be easily measured. Sometimes, the surge limit line is obtained from Eq.
(37). When this is done, it is necessary to neglect variations in molecular
weight, and this can lead to inaccuracies, even on air compressors, since
the humidity of ambient air varies.

Figure 3.17.  Detection of surge

 BASICS OF SURGE CONTROL  •   121

Figure 3.18.  Explanation of flow drops

SURGE CONTROL LINE

The three common forms of the surge control line are shown in
Figure 3.14. The optimal position of this line is parallel to the surge limit
line (Figure 3.14a). To minimize recirculation, the surge control line
should be set as close to the surge limit line as possible. Setting the control
line with a slope less than that of the limit line (Figure 3.14b) can lead to
excess recirculation at high pressures, and surge at low pressures during
stopping and startup. The third method is to select a minimum safe vol-
umetric flow and set a vertical control line (Figure 3.14c). This can lead
to excess recirculation at low pressures and surge at high pressures. Many
systems measure the flow in the discharge without correcting for suction
conditions. This gives maximum recirculation with minimum surge pro-
tection and is not recommended.
If it is desired to keep the control line parallel to the limit line, how
close together can the two lines be? This depends upon how accurately the
surge control line is set and how well it accounts for changes in inlet tem-
perature and pressure, and molecular weight. Also critical to the location
of the control line is the effectiveness of the anti-surge system in handling
upsets. This effectiveness depends on the control strategy chosen and the
nature of the anti-surge system, its transmitters, controller, and anti-surge
valve.
122  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

To set the surge control line equidistant from the limit line, Eq. (49)
can be used. The control line is displaced to the right by some fixed
amount, ba:

C12 ∆Pc + ba = ∆Por , s (53)

Equation (53) applies when Rc is less than 2.5. For values of 2.5 and
greater, Eq. (51) may be used. The control line is displaced by some fixed
amount, bb:

C15 ∆Pc + a2 Ps + bb = ∆Por , s (54)

INSTRUMENTATION AND VALVES

How fast must the anti-surge system be to detect the onset of surge and
effectively stop it?
It is often thought that a very rapid response is not necessary; also,
many believe that protection against large upsets is unnecessary. However,
upsets that are both fast and large are not rare, and these can be caused by
factors such as jammed check valve, operator errors, and shutdowns of
process equipment triggered rapidly by protective systems.
We have found that surge cycles can occur faster than is thought pos-
sible and that often surge goes undetected. This is because conventional
pneumatic controllers are too slow to detect this phenomenon. Only fast
transmitters can cope with the high speeds of compressor transients. We
recommend a diffused-silicon electronic transmitter. We also recommend
pressure-differential transmitters with a lag not greater than 0.035 second
as necessary to maintain surge control and to calibrate the surge limit line.
Also, recorders (used for calibrating and testing the anti-surge system)
must have a chart speed not less than 25 mm/s.
Figure 3.17 compares the response of a diffused-silicon transmitter
with that of a pneumatic device. Pressure differential was measured across
a venturi. At about 3.2 second, surge began, and the pneumatic transmitter
was too slow to detect flow fluctuations.
The slowness of most transmitters and recorders explains why records
will show no surge, yet upon inspection, compressors reveal changed
clearances or damaged impellers, seals, or bearings. Such undetected
surges will gradually alter the compressor, resulting in decreased effi-
ciency. Some electronic transmitters are dampened and are too slow for
use here; others are not stable at high speeds.
 BASICS OF SURGE CONTROL  •   123

Figure 3.18 shows how quickly surge can begin. Tests were made
on a centrifugal air compressor that supplied a blast furnace. Here, surge
­occurred at about 14 second. Occasionally, there are pre-surge oscillations
(incipient surge). Flow drops quickly before surge occurs. As surge begins,
flow drops precipitously—typically from a set point level to a reverse con-
dition in about 0.05 second, regardless of the pressure or compressor speed
(see Figure 3.17).

ANTI-SURGE VALVE AND CONTROLLER

For proper surge control, the anti-surge valve must be large and fast acting.
It should be capable of recirculating 100 percent of the design flow rate.
When flow is decreasing and operation is moving close to surge, it
is necessary to effect a quick increase in flow. The rate at which flow can
be increased depends on the response of the valve and its size. The larger
the valve, the greater the effect of its opening. The valve should be able to
recirculate the entire flow rate because sometimes complete blockage of
the system occurs, and all of the flow must be recirculated.
Field tests show that the full stroke of the anti-surge valve should be
from 0.5 to 1.5 second. New valves can meet this requirement; existing
valves can be speeded up with boosters. Boosters increase the time of
­response by increasing the pneumatic signal for the valve actuator.
Why should the transmitter have a maximum response time of 0.035
second, while the anti-surge valve has a maximum time of 1:5 second? The

Figure 3.19.  Basic scheme for isolated anti-surge and process control can
result in surge (a). Integrated process and anitisurge control system—two
devices protect against surge (b)
124  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 3.20.   Anti-surge control

surge control line is set close to the surge limit line—typically, the distance
between them is set at 15 percent of design flow. Thus, a partial valve
stroke will be enough to stop the movement toward surge; further opening
of the valve will restore the operating point to the surge control line.
The controller must be fast as well. Pneumatic controllers are too slow
to reliably prevent surge. Digital controllers should be used with caution.
Since these devices look in sequence at each control loop, the scanning time
may not be short enough to detect the transients typical of compressor surge
cycles. Microprocessors must have a scanning time less than 0.1 second.

CONTROL: ISOLATED ANTI-SURGE LOOP

First, we will look at the anti-surge control system without considering

its interaction with the process controller. Then we will integrate the two
 BASICS OF SURGE CONTROL  •   125

systems. In developing our control system, two typical systems will be

studied first: a proportional-plus-reset controller and a relay (on–off) con-
troller. Consider an electronic analog anti-surge controller that has a pro-
portional-plus-reset response, with an anti-windup device (Figure 3.19a).
The anti-windup device is necessary due to the nature of the proportional
and reset functions. Normally, the compressor operates in an area some
distance from the control line, resulting in an offset between the measure-
ment and the set point of the controller. As a result, the output signal winds
up to its high or low limit.
The proportional band and the operating point will be on opposite
sides of the set point or control line; control will not be affected until the
measured operat­ing point reaches the control line. Surge can result if the
measurement approaches the control line quickly, since overshoot will oc-
cur before the controller can unwind.
Anti-windup adjusts the reset loading to shift the proportional band to
the same side of the control line that the measurement is on when the con-
troller reaches its output limit. Then, if the control line is approached rap-
idly, the measurement enters the proportional band and the control starts
before the value reaches the control line. Overshoot is reduced.
Derivative control is not recommended, for it can open the anti-surge
valve far from the surge control line and can cause system oscillations.
Rapid oscillations in flow, even in the safe-operating zone, can cause the
valve to open because of the nature of the derivative response.
The response speed of this controller depends on the width of the pro-
portional band and the reset time. These parameters influence the stability
of the system. Decreasing the proportional band or increasing the reset
time increases the speed of the controller’s response, but past a certain
point, the system stability will be disturbed—all closed-loop control sys-
tems have a stability limit.
The speed of an anti-surge controller’s response is limited mainly by
the inertia of the compressor and its networks, the transmitters, and the
anti-surge valve and its actuator.
The slower these elements are, the slower the controller must be set.
The inertia of the compressor and its piping network cannot be changed,
but the inertia of the transmitter and the valve and its actuator can be de-
creased by a proper selection of faster elements.
Once fast transmitters and an anti-surge valve are selected, the pro-
portional band level and reset time are set, based on the parameters of the
compressor and its network. These parameters include the volume of the
system, the length of the pipes, and the inertia of the compressor. This lim-
itation is common to all closed control loops with feedback. Thus, modu-
lated control responds well to slow upsets and give good control. However,
126  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

the valve cannot be opened quickly for fast upsets, due to limitations of
system stability.
Now consider using a relay for control. The relay is part of an
open-loop system that opens the anti-surge valve to a preestablished level
after the compressor’s operating point reaches the surge control line. The
output of a relay device can change from minimum to maximum at any
speed, without upsetting the system’s stability. This is because (1) the relay
is an on–off device and there is no limit as to how fast the device can be
turned on and off and (2) the system is open loop and oscillations cannot
be set up in the control system.
However, the relay system has certain disadvantages for anti-surge
controllers. The preadjusted output level may be either lower or higher
than that required for protection. If it is lower, the compressor’s operat-
ing point will cross the surge limit; if it is higher, the compressor will
be operated with large amounts of recirculation or blowing off when the
anti-surge valve is opened. Also, the operator can be unsure as to when to
reset the relay without endangering the compressor.
Comparing the closed-loop system to the open-loop system has led
to a patented method that combines the advantages of both, while limiting
their shortcomings.
Figure 3.20 shows the controller with inputs for Eq. (52), one form of
the surge control line. The input corresponding to ΔPc passes through a scaler
(No. 1) and summing unit (No. 2) that transform it to the required form:

k1 ∆Pc + b1 (55)

where k1 is the slope of the surge limit line and b1 is the normal distance
between the surge limit line and the surge control line.
The value of k1ΔPc + b1 is compared with the value of ΔPor,s by the
proportional-plus-reset controller (No. 3). If ΔPor,s is greater, then the con-
troller’s output is set at zero.
When ΔPor,s is less than the signal given by Eq. (55), the controller begins
producing a signal that increases until k1ΔPc + b1 becomes equal to ΔPor,s .
As a result, the compressor’s operating point will be restored to the
surge control line. This system will work well if disturbances are small and
slow, but cannot open the anti-surge valve quickly enough to prevent surge
when disturbances are large or fast.
To handle large disturbances, the output from the summing device
(No. 2) is decreased by a fixed amount, d1, which is typically set at one
half the value of b1. This addition is done by the summing device in ele-
ment 5. The comparator in element 5 compares the value of ΔPor,s with
that of k1ΔPc + b1 − d1. When ΔPor,s is smaller, a signal is sent to the curve
 BASICS OF SURGE CONTROL  •   127

generator. The output of the curve generator follows the shape shown
in Figure 3.20; it increases nearly instantaneously, then exponentially
­decreases to zero. This signal goes to a summer (No. 4), overriding the
signal from the ­proportional-plus-reset controller.
Thus, for slow upsets, the valve will be opened slowly by the action
of the proportional-plus-reset controller; for fast upsets, No. 5 will quickly
open the valve, overriding the slow response of the controller.
The output from No. 5 decreases exponentially, allowing the propor-
tional-plus-reset element to take over.
Thus, the operating point is restored to the surge control line as in any
other anti-surge closed loop. After the output of No. 5 decreases to zero,
No. 6 automatically resets the override system.
Field tests have shown that, if the surge control line is set far enough
from the surge limit line, then the combination of the p­ roportional-plus-reset
response with the relay override signal is sufficient to prevent surge.
However, if the surge control line is set too close to the limit line,
the operating point of the compressor can cross the limit line before the
override signal appears: As a result, surge may begin, causing an almost
instantaneous flow decrease. Under such circumstances, the deviation of
the compressor’s operating point from the surge control line increases.
Here, it is necessary to move the surge control line to the right—the
task of No. 7. This element, like No. 5, subtracts a fixed amount (d2) from
the signal from No. 2 and compares the sum with ΔPor,s. When the sum is
greater than ΔPor,s, a preselected bias, b2, is added:

k1ΔPc + b1 + b2 = ΔPor,s (56)

Bias b2 moves the surge control line to the right. These three elements
(3, 5, and 7) can stop surge during the first cycle and keep the operating
point at a safe distance from the surge limit line, even if gas consumption
decreases to zero. This system protects the compressor against changing
operating conditions, not just at design conditions. The reset pushbutton
(No. 8) restores the surge control line to its original position.
Test results on this system are shown in Figure 3.21. Here, rather than
using a butterfly valve, the compressor was controlled by changing its
speed. The speed set point is a record of how the process controller changes
the set point of the speed governor. The action of the anti-surge system is
seen in the graph of the anti-surge controller output. The horizontal part
corresponds to the anti-surge valves being closed; the negatively sloped
line corresponds to the opening of the valve by the proportional-plus-reset
controller (No. 3 in Figure 3.20); and the vertical drop corresponds to the
action of the override controller (No. 5 in Figure 3.20).
128  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Figure 3.21.  Anti surge control effects

 BASICS OF SURGE CONTROL  •   129

Figure 3.22.  Anti surge control effects

130  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

INTERACTION WITH PROCESS CONTROL

A process control system is shown in Figure 3.19a. This consists of a pres-

sure transmitter, a pressure controller, and a butterfly valve with an actuator.
By closing the butterfly valve, the process controller forces the com-
pressor’s operating point to move toward the surge. If the point crosses the
surge control line, the anti-surge controller will open the anti-surge valve.
With both control systems, the deviation in the compressor’s oper-
ating point from the surge control line depends on the speed of response
of the anti-surge controller and of the process controller, and on the dead
times of the anti-surge valve and its actuator.

Figure 3.23.  Anti surge control effects

 BASICS OF SURGE CONTROL  •   131

Figure 3.24.  Effects of decrease in specific power

levels

If the process controller is faster than the anti-surge controller, then

this difference may interfere with surge control. The dead time of the
anti-surge valve can further add to this problem. Until the output of the
anti-surge controller overcomes the dead time, the process controller may
push the compressor’s operating point over the surge limit line.
This interaction between an isolated surge control system and the pro-
cess controller occurs not only with the control of discharge pressure but
also with the control of suction pressure and flow.

INTEGRATED SYSTEM

System performance can be improved by changing from an isolated anti-surge

loop to an integrated process control and protective system (Figure 3.19b).
In this system, the butterfly valve is controlled by the output of the
summing device. This applies equally to a compressor that is under vari-
able speed control. The output of this device depends on the output of
the anti-surge controller as well as on that of the process controller. The
operating point will, after reaching the surge control line, follow this line
rather than the line of constant pressure.
With integrated process control and anti-surge protection, the effect of
the dead time of the anti-surge valve will be minimized, since two devices
132  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

are used to protect the compressor from surge—the anti-surge valve and
the butterfly valve. Test results of the integrated system are shown in
Figure 3.22, with a variable-speed compressor.

SETTING AND TESTING EQUIPMENT

The surge limit line can be located precisely, since the control system
stops surge during its first cycle.
To calibrate the surge limit line, the slope of the surge control line,
k1, is set at zero to avoid the response by the proportional-plus-reset con-
troller. The distance between the surge limit and surge control lines, b1,
is set at 15 percent of the flow rate, and b2 is set at 75 percent. This is
because the control line is vertical and a value of b1 = 15 percent will
induce surge. The distance between the surge control line and the control
line of the override anti-surge relay element, d1, is set at zero, so that
override will start immediately. The level of the override signal opening
the anti-surge valve must be set at 100 percent; the time of its exponential
decrease must be not less than 3 minutes. This ensures that the valve is
kept open.
If the discharge line of the compressor is closed completely, surge
will be induced. At the beginning of the first surge cycle, the override
anti-surge relay element opens the anti-surge valve at maximum speed.
At the same time, the comparator (No. 7; Figure 3.20) moves the surge
control line to the right, making the distance between the surge control
and surge limit lines b1 + b2 equal to 90 percent, which is safely out of the
surge zone. This will keep the anti-surge valve open, and surge will be
stopped. The outputs of the transmitters, ΔPc and ΔPor,s, will be obtained at
the moment surge occurs.
The ratio ΔPor,s/ΔPc yields k1. If the control line is not a straight line,
this test can be repeated to generate different values of k.
Setting d1 depends on the speed of response of the ­proportional-
plus-integral part of the anti-surge controller. If the discharge line is closed
slowly, the override anti-surge relay should not operate. Experience indi-
cates that d1 may be between 2 percent and 5 percent of ΔPor,s.
The distance b1 between the surge limit and the surge control line is
selected so that the severest disturbances (e.g., the fastest possible com-
plete closing of the discharge line) will not cause surge. This distance can
be narrow, because the improved anti-surge controller operates at such a
high speed. In most cases, b1 is set between 5 percent and 10 percent of
ΔPor,s. Careful selection of k1 and b1 will reduce the energy expense for any
compressor that operates with recirculation or blowoff.
 BASICS OF SURGE CONTROL  •   133

EXAMPLE OF ENERGY SAVINGS

A 4,000 hp air compressor was being operated at constant 40 percent

blowoff (see Figure 3.23). The specific power required by this machine is
shown in Figure 3.24. The flow rate was 9,400 scfm at 310 psi.
The unit had a pneumatic surge control system that was supplied by
the manufacturer. In addition to constant blowoff, damage was frequent—
the machine had to be rebuilt twice during the year. The control system
described here was installed at a cost of about $15,000. About half the cost
was for instruments, the other half for installation. After installation, the
anti-surge valve was closed and recirculation was kept to a minimum. A
valve booster was not needed.
Assuming operation on a yearly basis, the savings in energy costs are
calculated, with electricity costs of $0.024/kWh:

kW h
Savings = 0.746 × 8,760
hp yr
hp $0.024
× (0.4225 − 0.3825) × × 9, 400scfm
scfm kWh
= $59,000 / yr

CAN YOU RERATE YOUR CENTRIFUGAL

COMPRESSOR?

Do you have a centrifugal compressor that you would like to use

above its capacity or pressure rating? Here is how to determine if it is
possible.
Compressor users often want to obtain increased production, which
can take the form of increased capacity, or increased pressure rise, or both.
One way to obtain an increase in production lies in the purchase of
new equipment; however, one can often rerate (i.e., rebuild) an existing
134  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

compressor to provide the required increase. Even if the compressor

­requires a new rotor and all internal stationary hardware, one can usually
still salvage the compressor casing.
Most often, compressors are rerated to obtain greater throughput,
with only a minor change in pressure levels. Under these conditions, one
can probably reuse a significant amount of existing hardware.
One problem always appears when considering a rerate. No way exists
to change the casing size or the bearing centerline distance. One cannot
change the nozzle locations, and only a limited number of impellers will
fit in any section of the compressor.
When he designs new equipment, the manufacturer starts with a
“clean sheet of paper” and can do almost anything to satisfy the require-
ments of the process. However, the manufacturer must forego this luxury
when rebuilding an existing unit.
When considering a rerate, the project engineer must first study the
process and determine the flow required to handle the scheduled produc-
tion increase. What pressure levels will he require of the compressor? How
will the compressor’s operation affect the overall process? He must de-
termine the answers to these and other questions before considering the
compressor for possible rerate.
A frustrating thing can happen to a project engineer. He designs the
process and determines the new required capacity and associated pressure
levels at which he expects his compressor to operate, only to learn from
the manufacturer that the compressor cannot meet intended requirements.
He then heads back to the drawing board and attempts to answer the per-
tinent questions until he arrives at a set of numbers compatible with both
the compressor and the process.
The project engineer could often save a great deal of time if he under-
stood the rerate capabilities of his compressor. If he could figure out for
himself what his compressor could deliver, he could determine the proba-
bility of reaching his scheduled production increase, without communicat-
ing with the manufacturer, until he derived a feasible set of numbers. He
might save days, or even months, depending on the extent of the process
change and its relation relative to the compressor casing. He would know
better than to try to put “10 pounds of coal into a 5-pound bag.”
Actually, the engineer need possess only a minimal amount of compres-
sor knowledge and spend a minimal amount of calculation time in order to
determine the rerate feasibility of a compressor. He must consider the following:

1. Capacity: Will the size of the nozzle accept the projected increase
in flow?
 BASICS OF SURGE CONTROL  •   135

2. Horsepower: Will the motor, motor gear, turbine, or turbine gear

handle the increased horsepower?
3. Pressure: Can the casing accommodate the intended pressure
­levels? Can the compressor aerodynamically meet these levels?
4. Speed: Will the compressor handle the required speed within
the API (American Petroleum Institute) critical-speed limita-
tions?

We will now review these four considerations, presenting them, as

above, in the order of their importance.

RERATE FEASIBILITY: CAPACITY

The most important consideration for capacity lies in nozzle sizes. One
must determine whether the nozzle size will pass the required volume flow
with a reasonable pressure drop. An existing compressor possesses fixed
nozzle sizes—the geometric size of the nozzle sets the maximum feasible
volume flow possible. Hence, we can reduce the analysis to the consider-
ation of inlet velocities.
Although inlet-velocity limits vary with conditions, a good rule of
thumb sets the limit at a maximum of 140 ft/s for air and lighter gases.
Owing to the inverse proportionality between the inlet-velocity limit
and the square root of the molecular weight of the gas, one should gen-
erally limit heavier hydrocarbons such as propane (mol. wt. = 44.06) to
approximately 100 ft/s.
Equation (57) determines the actual inlet velocity:

Va = 3.06Q / D2 (57)

Since velocity limitations set the maximum capacity, the maximum

velocity of 140 ft/s can limit one when rerating a compressor. With geo-
metrically sound inlet piping, the engineer may find higher inlet velocities
possible.
Manufacturers often furnish compressors with inlet nozzles other
than the main inlet. They design some compressors with side-load nozzles,
some with interstage cooling nozzles, and some with a combination of
both. Use the above inlet-velocity criteria for all inlet nozzles, and check
all inlet nozzles for gas velocities.
Once the design engineer decides that the compressor can pass the
desired capacity, he can investigate the effect of the increase on the horse-
power required.
136  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

RERATE FEASIBILITY: HORSEPOWER

The project engineer must include power requirements as the second ma-
jor consideration in a rerate-feasibility analysis. This holds especially true
in the case of motor-driven compressors, since customers usually do not
purchase oversized motors along with the original equipment. However,
one can usually rebuild gears and turbines to provide greater power ca-
pacity. Motor drives may require buying a new motor, which may lead, in
some cases, to foundation problems.
The compressor will require power approximately proportional to the
increase in weight flow desired. This means that an increase in weight flow
of 20 percent will dictate an increase in horsepower of at least 20 percent
(or 1.2 times the original power). Moreover, the manufacturer will usually
suggest at least an additional 10 percent horsepower availability to take
care of overload should this occur. Therefore, a 20 percent weight-flow
increase will result in an increase of approximately 32 percent in power
requirements (1.2 + 10%(1.2) = 1.32), should the project engineer wish to
design a bit conservatively (see Figure 3.25). Note also that in the case of a
motor-gear drive, general practice requires the allowance of an additional
2 percent horsepower for gear losses.
If the driver cannot deliver the required horsepower to the compressor,
the engineer must either buy a new driver or relax the scheduled production
increase until he can reduce the required weight flow to within the capabili-
ties of the driver. In cases where critical power requirements prevail, driver
capabilities formulate the starting point of a rerate-feasibility analysis.

RERATE FEASIBILITY: PRESSURE

The project engineer must next consider the hydrostatic test pressure. The
compressor maker, at the time of manufacture, hydrostatically tests the
casing to 150 percent of the maximum expected operating pressure. Com-
pressor aerodynamics may set the maximum operating pressure, or the
process may set it by the existence of a relief valve in the system.
If the expected pressure levels of the new process exceed the nameplate
maximum pressure, one should make a new hydrostatic test. CAUTION:
Sometimes, in a compressor made up of two or more sections (as with inter-
stage cooling, side loads, or extractions), the test engineer will test the com-
pressor sectionally, with the inlet section at a lower pressure than the outlet
section. In such a case, the project engineer must review both sections for
the possibility of exceeding the maximum, sectional, safe-operating pressure.
 BASICS OF SURGE CONTROL  •   137

He must then investigate the compressor’s aerodynamic ability to de-

liver the required head. He can more easily do so in terms of polytropic
head—the amount of work done by the compressor (which differs from
the input work by the polytropic efficiency). Equation (58) relates the pol-
ytropic head to the desired pressure ratio:

H P = ZRT [ n / ( n − 1)][ rp( n − 1) / n − 1] (58)

Although the polytropic efficiency can vary from 68 percent on small

impellers to 83 percent on ideal impellers, for a feasibility analysis one can
assume a polytropic efficiency of 70 percent. If the engineer analyzes the
role of the polytropic efficiency in Eq. (58), he will find that the error in the
“n/(n − 1)” multiplier tends to counterbalance the error in the “(n − 1)/n”
exponent, thus considerably reducing error in the calculated polytropic head.
Equation (58) can give both the original polytropic head and the re-
rate polytropic head requirement. With the original rated speed known, the
“Fan Law,” Eq. (59), determines the approximate rerate speed:

N rerate = N original ( H Prerate / H Poriginal )1/ 2 (59)

The speed calculated from Eq. (59) may turn out too high for safe
operation (see the considerations of rotational speed discussed in the next
section). However, the manufacturer can sometimes obtain increased poly-
tropic head by adding an impeller, or replacing the existing impellers with
ones of lower backward lean. The manufacturer will ultimately have to
determine the feasibility of obtaining the required head.
The project engineer can make a quick estimate of the maximum head
capability of his compressor. If the compressor contains a “blank” stage
(stage space with no impeller), Eq. (60) will show the approximate maxi-
mum head capability with no change in speed:

HPmax = HPoriginal [( a + b ) / a ] (60)

One can then apply the “Fan Law” along with Eq. (60) to determine
the approximate polytropic head capability of the compressor. This pro-
cedure should yield a polytropic head within 10 percent of the maximum
compressor capability.
The above procedure for estimating the required polytropic head, al-
though presented for straight through units, will also work for units with
side loads, extractions, or interstage cooling, provided one uses the sum of
the head requirements for each section as the total head.
138  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Determine the head for Section 1, based on the operating parameters

of Section 1 (i.e., the inlet temperature, inlet pressure, discharge pressure,
etc.); the head for Section 2, based on the operating parameters for Sec-
tion 2; and so on. The sum of all sectional head requirements yields the
total head requirement for the rerate. One can determine the total head
requirement for the original rated point in a similar manner and derive the
approximate rerate speed from Eq. (59) as before.
One important effect of an increase in head requirement remains to be
analyzed, that is, the head effect on the power requirement. Horsepower is
directly proportional to polytropic head:

WHp
GHP = (61)
33,000η P

As can be seen, an increase of 20 percent in the required polytropic

head will raise the power requirement 20 percent if the efficiency
remains constant. For example, if an increase of 20 percent in the
weight flow is coupled with one of 20 percent in the polytropic head,
the power requirement will rise by 44 percent (1.20 × 1.20 = 1.44).
To this figure, an additional 10 percent horsepower should be added
for overload. A 2 percent gear loss should also be included, where
appropriate.

Figure 3.25.  A centrifugal compressor: typical performance

 BASICS OF SURGE CONTROL  •   139

RERATE FEASIBILITY: SPEED

The final consideration is that of rotational speed. Two major criteria must
be satisfied:

1. The rotational speed must be slow enough to not overstress the

impellers.
2. The rotational speed should fall within certain limits of the com-
pressor’s first and second critical speeds.

The first criterion above is best viewed from the standpoint of me-
chanical tip-speed limitations. The mechanical tip speed can be deter-
mined by:

u = πdN/720 (62)

Mechanical tip-speed limits vary with manufacture, size, impeller

material, and type of construction. A typical limit on mechanical tip speed
might be 900 ft/s. When using the “Fan Law” of Eq. (59), the rotational
speed that results in a tip speed of 900 ft/s for the largest impeller should
be considered the upper limit (providing that this speed falls within API
critical-speed limitations.
The project engineer will not know what effect the rebuild will have
on the compressor critical speeds when conducting the feasibility analy-
sis, except possibly through experience with such effects. Since the bear-
ing centerline distance is fixed and most rebuilds can be accomplished
without a change of bearings, it can generally be assumed that the critical
speeds will not be affected (at least, as a first approximation).

SAMPLE PROBLEM

For the purpose of discussion, let us consider a “straight through” cen-

trifugal compressor on a dry-air process, having the following nameplate
data: inlet capacity = 11,000 ICFM; inlet temperature = 90°F; rated inlet
pressure = 14.5 psia; rated discharge pressure = 55 psia; rated power
input = 1,700 hp; rated speed = 8,100 rpm; maximum continuous
speed = 8,500 rpm; first critical speed = 4,800 rpm; rated molecular
weight = 28.97; K = Cn/Cv = 1.4; maximum discharge pressure = 65 psia.
An investigation of the data files on the compressor reveals that the
second critical speed is equal to 10,800 rpm and that the smallest wheel
140  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

diameter is 22 in. Also the cross-sectional drawing of the compressor indi-

cates that the inlet nozzle diameter is 20 in.
The desired rerate is an increase in the inlet capacity to 12,300 ICFM
and an increase in the discharge pressure to 60 psia, with all other inlet
conditions unchanged:

1. Compute the inlet velocity, based on the new inlet-volume flow

using Eq. (57):

Q  12, 300 
Va = 3.06 2
= 3.06  2 
= 94 ft/s
D  (20) 

Since this is an acceptable inlet velocity, the proposed capacity is

feasible.
2. Since the rated inlet conditions have not changed, the increase in
weight flow will be proportional to the increase in volume flow, and
therefore, the power requirement due to the change in volume flow
will increase by the same proportion:

GHPrerate W Qrerate 12, 300

= rerate = = 1.12
GHPoriginal Woriginal Qoriginal 11,000
GHPrerate = 1.12GHPoriginal = 1.12(1.700) = 1, 910 hp

Note that, up to this point, the driver will have to be capable of:

(1.1)(1,910) = 2,100 hp

plus 2 percent excess horsepower if a gear is involved.

3. Since the nameplate maximum discharge pressure is 65 psia, the
compressor will not have to be hydrostatically retested, providing
that the process will not allow the value of 65 psia to be exceeded.
4. The approximate polytropic head can now be calculated for both the
original and the rerate condition from Eq. (58):
Original:

n / ( n − 1) = [ K / ( K − 1)]η P = [(1.4 / 1.4 − 1)](0.76) = 2.66

H p = ZRT ( n / n − 1)[( P2 / P1 )( n − 1) / n − 1]
= (1.0)(1,545 / 28.97)(550)(2.66)[(55 / 14.5)1/ 2.66 − 1]
= 50,700
Rerate:
Hp = (1.0)(1,545 / 28.97)(550)(2.66)[(60 / 14.5)1 / 2.66 − 1]
= 55,000
 BASICS OF SURGE CONTROL  •   141

The new required speed can be determined from the “Fan Law,”
Eq. (59):

N rerate =N original H p rerate /H poriginal

= 8,100 55,000 / 50,700 = 8, 440 rpm

The smallest wheel diameter is 22 in.; therefore, from Eq. (62):

π dN (π )(22)(8, 440)
u= = = 810 ft/s
720 720

The new rotational speed results in a satisfactory mechanical tip

speed. API states that the second critical speed must be 20 percent
above the highest operating speed. Assuming that the new rerate
speed is the highest for the new process, the second critical speed
must be at least:

(1.2)(8,440) = 10,130 rpm

The second critical speed (10,800 rpm) is higher than that re-
quired and therefore the rotational speed is feasible.
5. The total increase in gas horsepower can now be determined. The
new horsepower will be proportional to the increase in the poly-
tropic head and weight flow (in this case, volume flow):

GHPrerate  H prerate   Qrerate 

=
GHPoriginal  H p   Qoriginal 
original  

 55,000   12, 300 

= = 1.21
 50,700   11,030 
GHPrerate = (1.21)(1,700) = 2,060 hp

The driver must therfore be capable of:

(1.1)(2,060)=2,270 hp

plus 2 percent if a gear is involved.

Since the inlet velocity, maximum operating pressure, and
­required rotational speed are within satisfactory limits, the rerate
is feasible.
142  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Nomenclature
a Number of impellers in original rotor
b Number of “blank” stages in original configuration
d Tip dia. of largest impeller, in.
D Inside dia. of nozzle flange, in.
GHP Gas horsepower
Hp Polytropic head, ft-lbf/lbm
K Ratio of specific heats, dimensionless
N Rotational speed, rpm
n/(n – 1) [K/(K–1)]ηP
Q Flange volume flow, ft3/min, based on rated inlet
conditions (i.e., inlet pressure, temperature,
compressibility factor and molecular wt
r
P Pressure ratio, Pdischarge/Pinlet
R Gas constant, 1,545/mol. wt.
T Inlet temperature, °R
u Mechanical tip speed, ft/s
Va Actual inlet velocity, ft/s
W Weight flow, lbm/min
Z Average compressibility factor
ηP Polytropic efficiency

SUMMARY

We have analyzed the four major considerations of a rerate-feasibility

analysis: (1) capacity, (2) horsepower, (3) pressure, and (4) speed. The
nozzles must be large enough to pass the required flow—or, to state this
another way, the maximum capacity will be limited by the size of the
inlet nozzle. The driver must be capable of delivering the required power.
The casing must be capable of handling the pressure levels, both mechan-
ically and aerodynamically. The required speed must be within certain
mechanical limitations and should meet the API critical-speed require-
ments. Once the project engineer is satisfied that the above criteria can be
met, a request for a quotation can be forwarded to the manufacturer with
greater confidence.
 BASICS OF SURGE CONTROL  •   143

EASY WAY TO GET COMPRESSION

TEMPERATURES

An engineer or operator can quickly find the temperature of gas discharg-

ing from a compressor with this nomograph. All he needs to know is the
suction pressure and temperature, the discharge pressure, and the ratio of
specific heats for the gas.
The basic equation for compressor discharge temperature is:

( n − 1) / n
T2 = T1 ( P2 − P1 )

where: T2 = absolute discharge temperature, °F + 460

T1 = absolute suction temperature, °F + 460
p2 = discharge pressure, psia.
P1 = suction pressure, psia.
n = ratio of specific heat at constant pressure to specific heat at con-
stant volume
p2/p1 = compression ratio
The ratio of specific heats, n, can be calculated, read from tables, or
estimated from the chart.
If the compressor suction or discharge pressure is beyond the range in
the nomograph, the nomograph can still be used by calculating the com-
pression ratio and taking that as the starting point. It should be noted that
the absolute values of temperature and pressure in the equation have been
converted to °F and psig.

Example:

A hydrocarbon gas whose molecular weight has been given as 21 is fed to

a compressor at 60 psig and 70°F, then compressed to 350 psig. What is its
temperature leaving the compressor? Read n = 1.26 on the plot of molec-
ular weight versus n. Connect 60 psig on the P1 scale with 350 on the P2
scale (note that these pressures are psig on the scale and not psia) and read
the compression ratio, p2/p1, as 4.88. Connect this with 1.26 on the n scale,
note the intersection on the pivot line, and align this intersection with 70°F
on the T1 scale to read 275°F on the T2 scale.
144  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS
BASICS OF SURGE CONTROL  •   145
 CHAPTER 4

Lubrication and
Compressors
148  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

LUBRICATING AIR COMPRESSORS

Choosing the proper lubricant for plant air compressors pays off in
terms of long life and trouble-free operation.
The lubrication system in a compressor is a critical necessity in keeping
the machine running. It does this by reducing friction, transferring heat to the
cooling system, sealing against air leakage, and hushing away dirt and debris.
The lubrication systems of present-day air compressors must be able
to provide these vital functions in atmospheres that are usually polluted
with dirt particles as well as with corrosive fumes that pass right through
the inlet air filter and enter the compression chamber.
In addition, modern air compressors are much smaller, and run at
speeds undreamed of a few decades ago, making lubrication requirements
all the more critical.
The packaged, integral-gear, centrifugal, plant air compressor has be-
come important over the past 15 years. Here, the lubrication system has to
deal with individual pinion gears that run at speeds of 30,000 to 50,000 rpm.
A more recent compressor is the oil-flooded rotary-screw machine. It
has brought forth still a different set of requirements, because the oil and
air pass through the compression cycle together, with the oil picking up all
the dirt and contaminants that come through the inlet air filter.
Suppliers of lubricants have had to introduce more sophisticated
products in recent years to meet the requirements of this new generation
of air compressors, operating under increasingly adverse environmental
conditions. The oil industry has not only developed new lubricants to meet
these requirements but has actually been able to extend the service life
between oil changes in most cases.

LUBRICATION PROBLEMS

Many factors should be considered when your oil suppliers visit your plant
to advise you on the proper lubricant for a particular compressor. Among
the significant factors are inlet air temperature, air contamination, con-
densation, operating temperature, discharge temperature, and pressure.
Inlet air contamination is a major factor in compressor wear, oil oxi-
dation, and discharge-valve deposits. Recent years have seen an amazing
increase in the rate of buildup of industrial air contaminants, which
are ­often odorless and invisible. Even on the clearest day, considerable
amounts of vapors and fumes come through an inlet air filter in a typ-
ical industrial plant. It surprises most people to learn that the human
body can tolerate far more air contamination than can a modern air
compressor.
 Lubrication and Compressors  •   149

Condensation seldom occurs in a compressor during normal operation.

However, when the relative humidity of inlet air is high, and the cylinder
temperature during the suction stroke is below that of the inlet air, conden-
sation will occur. For instance, to prevent condensation, the cylinder walls
in the first stage of a two-stage unit have to be kept at about 100°F when
the inlet air is at 80°F. The inexperienced operator who runs too much cold
water through the cooling jackets may cause water to condense within the
cylinders and wash away the oil film. This promotes the formation of rust,
which in turn causes abrasion during compressor operation.
High operating temperatures reduce the viscosity of an oil, thereby
­reducing the oil-film thickness. This film reduction creates a marginal con-
dition where metal-to-metal contact can occur. This is why the viscosity in-
dex of the oil purchased is so important. Prolonged operation at temperatures
in the range of 140° to 160°F also increases the oxidation rate of the lube oil.
Oxidation can change a thin oil to a thick, carbonaceous, sticky mass in a rel-
atively short time. Hence, the oil-change period must be strictly maintained.
Air discharge temperature directly affects the quantity of deposits on
the discharge valves and downstream piping. High discharge temperatures
cause rapid oil oxidation; the oxidized residue plates out on the valves. This
buildup eventually interferes with normal valve motion, causing valve leak-
age. Then, preheated or recompressed air increases the discharge tempera-
ture, further speeding oxidation and promoting more deposits. This cycle can
lead to component failure and, at the extreme, to incandescent hot spots.
Another warning: If someone claims he has the ideal oil for all recipro-
cating compressors, don’t believe him. A single-stage reciprocating machine
has a theoretical discharge temperature of 400° to 450°F, while a two-stage
unit would discharge at about 250° to 300°F. Some reciprocating units are
splash lubricated; others are force-feed lubricated. Some are air cooled; others
water cooled. Some use one grade of oil for all moving parts, while others use
different oils for cylinders and running gear. There are many variables.

ROTARY-SCREW COMPRESSORS

Oil-flooded rotary-screw compressors have different requirements from

reciprocating ones. Because air and moisture sucked in at the inlet carry
through with the oil, and are separated after the compression cycle, the oil
must have an excellent water-separation characteristic. Since the oil is con-
tinually cooled in a screw compressor, the internal temperatures can be kept
quite low. However, this is dangerous; as has been said, condensation pres-
ents a severe hazard. Thus, most manufacturers recommend that discharge
air temperature be maintained between 170°F and 200°F. This concept of
deliberately increasing the discharge temperature of a screw machine by
150  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

reducing oil cooling is difficult for some operators to accept, but it must
be insisted upon to prevent condensation and consequent bearing damage.

TYPES OF LUBRICATING OILS

Most of the premium-grade lubricants advertised to the public are the

so-called Pennsylvania-crude-based oils, which are excellent for automo-
tive service. They come from a large natural pool under parts of Pennsyl-
vania, Ohio, and West Virginia, and are unique in the United States. These
are paraffin-based oils, as opposed to the naphthenic-based oils produced
elsewhere in the country. However, paraffin-based lubricants form a hard,
varnish-like residue when used in compression cylinders; naphthenic oils
form a lighter, huffy, carbon residue that is more easily cleaned off.
Discharge-valve deposits can be the principal problem in some recip-
rocating compressor installations. When this is the case, it is essential to
use a naphthenic oil and minimize discharge temperature. If the air com-
pressor must also be operated under high humidity, high ambient tempera-
tures, and intermittent-duty conditions, then condensation and resulting
sticky valves or rings may occur. A compounded or oxidation-inhibited
naphthenic-based oil is best suited for condensation conditions where
valve deposits are a chief lubrication problem.
Paraffinic oils can be used in air compressors that do not have to con-
tend with valve deposits. They are more resistant to thinning out as tem-
perature increases than are straight naphthenic oils. That is, they have a
better viscosity index.

Packaged compressor using a pressurized lube system.

 Lubrication and Compressors  •   151

Bull gear and pinions of centrifugal compressor.

Bearing protection is one of the primary functions of a frame or

crankcase lubricant. The critical period for any bearing lubricant is when
bearing failure begins. The bearing surfaces during critical periods be-
come irregular, and high-localized pressures occur. Naphthenic oils under
high pressure become viscous. This provides an effective medium for car-
rying the bearing metal until its surface is smooth and the load is evenly
distributed.
Synthetic lubricants were developed to overcome the two most fre-
quent complaints by users of petroleum oils—carbonaceous deposits that
necessitate frequent valve cleaning and danger of fire.

DANGER OF FIRE

The so-called fire triangle—oxygen, fuel, and an ignition source—exists

in any air compressor. Air provides the oxygen, and petroleum oils provide
the fuel. Normal discharge temperatures are never high enough to trigger
a fire or cause an explosion. The ignition temperature of most air-cylinder
lubricating oils is above normal compressor operating temperatures.
However, carbonaceous deposits from lube oils do collect on valves,
heads, and discharge piping. As has been stated, this may cause leaking
discharge valves, so that hot air bypasses and continues to build up heat
until autoignition can occur. This is practically always the result of inad-
equate maintenance, such as dirty fins on air-cooled compressors, dirty
152  •   A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

intercoolers, water jackets clogged with scale, broken or leaking valves,

and the like. Therefore, the best oils should be used in a limited amount.
Keep the compressor clean. Inspect valves frequently. Remove scale for-
mations in the cooling-water jackets.
Phosphate-ester synthetic lubricants were developed to minimize
compressor fires. Synthetics have distinct fire-resistant qualities in that
they have an auto-ignition point that is about 50 percent higher than that
of most petroleum oils and have excellent film strength.
There are some disadvantages to synthetics: Their viscosity indexes
are generally not as high as those of petroleum lubes, and they are more
vulnerable to water washing in the presence of condensate. They are much
more expensive than petroleum products, and the cylinder feed rate must
usually be substantially higher. These synthetics are powerful solvents.
They can remove all paint on the inside of air passages, and this carries
into the valves, clogs oil lines, fouls intercoolers, and raises general havoc
inside the compressor.
These synthetics attack normal gasketing material, and fumes must
be directed away from an electric motor, for they can destroy electrical
insulation. Never use a phosphate-ester synthetic lubricant without first
checking with the compressor manufacturer.
Diester-based lubricants, while not offering quite the same explosion
protection as phosphate esters, have many other properties that are advan-
tageous, particularly a long service life between oil changes. Then, there
are the so-called super-oils, which are petroleum oils with additives. They
provide a somewhat longer service life than standard petroleum oils, at a
somewhat lower cost than the synthetics.

WHICH TYPE OF OIL IS BEST?

The next task is selecting the right oil for your needs. Consider first
the centrifugal compressor. This probably presents the least demanding
­lubrication problem. Remember, since the centrifugal air compressor is a
nonlubricated machine, there is no oil in the air compression chambers;
lubrication is confined to the driving gear.
Here the requirement is for good maintenance of film strength under
load, with added inhibitors to prevent rust, sludge formation, and foaming.
Oil-flooded screw compressors, on the other hand, are very sensi-
tive to both the quality and the condition of the lube oil. There are no
high-temperature conditions to contend with, but the oil is repeatedly
exposed to the air being compressed, plus all the contaminants that pass
through the inlet filter. Since a part of the oil that floods the compression
 Lubrication and Compressors  •   153

chamber is bypassed to go through the bearings, moisture collecting in the

unit can be critical to bearing life.
Since operators are not always alert to the dangers of condensation—
contamination or oil degradation in an oil-flooded screw compressor—
one manufacturer offers an oil-sampling program and urges its use. Small
sample bottles of oil from the compressor are mailed to the compressor
manufacturer on a regular basis, and an analysis is immediately returned
to the user. This gives him adequate warning and enables an operator to
schedule a maintenance check before a bearing failure shuts him down.
Because of the very high temperatures incurred in a piston compressor,
sometimes 300° to 400°F, this may present the toughest lubrication require-
ment. Synthetic lubricants and super-oils have been mentioned. Their selec-
tion, compared with a conventional petroleum oil, becomes largely a matter
of economic choice; that is, are the long-life advantages worth the extra cost?

SETTING THE OIL FEED RATE

Cylinder lubricators on piston compressors have adjustable feed rates. The oil
droplets can easily be counted as they pass through a sight glass. The question
most often raised is, how many drops per minute should be used? Attempts
have been made to relate drops per minute to cylinder diameter, piston speed,
etc. However, it must be noted that such an approach may be dangerous; the
correct oil quantity for one machine may be too much or too little for another.
A new compressor should be broken in with a very heavy feed rate
(three or four times normal) for the first 500 hours of operation, using a
cylinder oil two Society of Automotive Engineers (SAE) grades heavier
than normal, until the piston and cylinder have acquired a glaze. This
break-in routine should be followed using only petroleum oil, even if there
is to be a switch to synthetic cylinder oils for regular operation.
Then the lubricator feed rate should be reduced in small steps, shut-
ting down and inspecting the cylinder bore after a few hours of opera-
tion at each step, until the top inside the surface has just a slight oil film.
­Puddles of excess oil lying in the bottom of the cylinder indicate too heavy
a feed rate; dry spots on the top are a signal of too little. An experienced
operator knows that there is much more art than science in arriving at just
the right lubricator setting.
Lubricating oil is the lifeblood of a compressor. Its selection and use
should not be left to chance. Manufacturers’ publications, such as Mainte-
nance of Reciprocating Compressors and the Compressed Air & Gas Data
book, as well as operating manuals for the particular compressor at hand,
should be studied. Then, call in the expert—your oil supplier.
 About the Author

James Watterson has a PhD in mechanical engineering and is ­currently

serving as independent training consultant with JMW, United Arab
­
Emirates, since January 2003, and as part-time senior lecturer/lead
­
trainer–commissioning and process start-up in the oil and gas refineries
with ­ADNOC since February 2011. He trains groups of maintenance,
­operations, instrumentation, and planning engineers in groups of 8 to
20. This includes public courses and in-house courses. Usually spanning
5 days, these courses are highly interactive and use many types of soft-
ware for vibration analysis, compressors–turbines. There are also many
3D ­animated videos for most rotating equipment. Coupled with this, the
author also provides self-assessment questions. The feedback from these
companies has been very positive, with the author receiving many emails of
gratitude from the course participants and from the training ­departments.
Previously he was managing director at JW Management and ­Marketing
Consultancy from September 1985 to October 2000.
 Index

A process control, interaction with,

Adiabatic efficiency, 55–56 130–131
Aerodynamic whirl, 68 ratio setting, 105
Air discharge temperature, 149 rerating feasibility, 133–139
American Gear Manufacturers surging characteristics, 92–93,
Association (AGMA), 20 114–115
American Petroleum Institute (API), temperature changes, effect of,
compressor specifications, 98–99
60–61 testing equipment, setting
Anti-surge control scheme, for and, 132
centrifugal compressors valves, 123–124
automatic temperature Anti-surge valve, and controller,
compensation, 106 123–124
basic control theory, 93–98 Anti-windup function, 102–103
compression temperatures, ways Asymmetric whirl, 68
to get, 143–145 Augers, 7
control line, surge, 121–122 Automatic temperature
controller, 123–124 compensation, 106
discharge-line flow measurement, Axial compressors, 10, 35–37
107–109 capacity, 38
energy savings, example of, 133 defined, 35
improved surge control, 113–114 head of compression, 38–39
inlet-guide vanes, 107 interpretation of curves, 44
instrument ranges, 103–105 operating characteristics of,
instrumentation, 103 15, 17
and valves, 122–123 ratio of compression, 37–38
integrated system, 131–132 review, 37
isolated anti-surge loop, 124–129 revolution per minute
limit line, surge, 115–121 and break horsepower, 41–42
molecular-weight, changes in, and horsepower, 39–40
99–102 surge control, 44–45
operated in types of, 46
parallel, 110–112 surging, 42–44
series, 109–110 troubleshooting problems, 47–50
158  • INDEX

Axial displacement, of rotor, 48 Compressor efficiency, 54

Axial-flow compressors, 62. adiabatic compression, 55–56
See also Axial compressors bearings
Axial motion, 35 for high-speed machines, 68–70
journal, 70–72
B misalignment, 74–77
Backward-curved blades, 66 thrust, 73–77
Base alignment method. See Cold control systems, 84–89
alignment method couplings, 81–83
Basic control theory, of centrifugal design and configuration, 63–66
compressors, 93–98 driver selection, 63
“Batch” function. See Anti-windup environment, 61–62
function equal-work efficiency ratio, 56–57
Batch-packed lubrication high-speed
technique, 82 gears, 84
Bearings rotating machinery, balancing of,
excessive vibration of, 47 79–81
for high-speed machines, 68–70 impellers, and fabrication
journal, 70–72 techniques, 66–67
misalignment, 74–77 isothermal compression, 54–55
running too hot, 48 layout, 60–61
thrust, 73–74 lube-oil systems, 83–84
Bernoulli’s Theorem, 37 multistage compression, with
Break horsepower (BHP), 41–42 intercooling, 58–59
Butt-welding technique, 66–67 centrifugal compressors,
achieving online availability
C of, 59–60
Capacity rotor dynamics, 67–68
of compressor, 38 seals, 77–79
for rerate feasibility, 135 specification criteria, 60
Centrifugal compressors, 10, types and arrangements, 62–63
12–13, 29–35, 62–63 Compressors
achieving online availability of, axial. See Axial compressors
59–60 centrifugal, 29–35
cost, 15, 17–19 combinations of, 16–18
defined, 33 driving mechanisms, 18–20
drives for, types of, 63 factors, overlooked, 22–24
operating characteristics of, installation engineering, 24–26
14–15, 17 system specifications, 21–22
speed, 15–16 vigilance during installation,
surge control. See Anti-surge 26–28
control scheme defined, 2
Chemical process industries dynamic, 28–29
(CPI), 60 marketing methods, 20–21
Cold alignment method, 74, 77 operating parameters, 50–51
 Index  •   159

in parallel, 110–112 F
rotary-screw, working principle Fillet-welding technique, 66–67
of, 2, 6–8 Fire, danger of, 151–152
failures, 6 Flexible shafts, 67–68
features, 8–10 Flow reversal, 43
lubricated/oil-free, 3 Fossil-fuel drives, 20
oil-free, 4–5 Full penetration fillet-welding
oil-injected, 3–4 technique, 66–67
selection considerations for, 10–13
cost, 15, 17–19 G
operating characteristics, Gas turbines, 18–19
13–15, 17 Gear-type couplings, 81–82
speed, 15–16 Gyroscopic-induced whirl, 68
in series, 109–110
types of, 2, 10 H
Condensation, 149 Half-frequency whirl. See
Continuous lubrication Hydrodynamic whirl
technique, 82 High-speed gears, 84
Control systems, 84–89 Horsepower, 39–40
Couplings, 81–83 for rerate feasibility, 136
Cubic feet per minute (CFM), 9 Hot-alignment techniques, 74, 77
Cylindrical bearing, 70 Hydrodynamic whirl, 68

D I
Diaphragm compressors, 10, 11 Ideal-gas equation, 54
Diester-based lubricants, 152 Impeller, 32–33
Diffuser, 33–34 fabrication techniques, 66–67
Digital controllers, 124 Influence-coefficient technique, 81
Discharge-line flow measurement, Inlet air contamination, 148
107–109 Inlet-guide vanes, 107
Disk-type couplings, 81–82 Installation, vigilance during,
Drives 26–28
mechanisms of, 18–20 Instrument ranges, 103–105
selection of, 63 Instrumentation, 103
speed range for, 15–16 and valves, 122–123
Dry-friction whirl, 68 Integrated process control and
Dynamic compressors, 28–29 protective system, 123,
131–132
E Internal-combustion engines, 18, 20
Electron-beam technique, 66–67 Isolated anti-surge loop, 123–129
Energy, forms of, 28 Isothermal efficiency, 54–55
Energy savings, example of, 133
Equal-work efficiency ratio, 56–57 J
Equipment, setting and testing, 132 Journal bearings, 70–72
“Extreme cold climate”, 61 misalignment, 74–77
160  • INDEX

K advantages, 4
Kinetic energy, 28 disadvantages, 4
working principle of, 3–4
L Oil whirl, 68
Labyrinth seals, 77–79 Overcompleteness, 27–28
Liquid-piston compressor, 11
Liquid-piston type compressors, 10 P
Lube-oil systems, 83–84 Paraffin-based oils, 150
Lubrication, 148 Pennsylvania-crude-based oils,
fire, danger of, 151–152 150
oil feed rate, setting of, 153 Phosphate-ester synthetic
oils lubricants, 152
selection of, 152–153 Pneumatic controllers, 124
types of, 150–151 Positive-displacement compressor,
problems, 148–149 2, 13–14
for rotary-screw compressors, Potential energy, 28
149–150 Pressure-dam bearing, 70
Pressure, for rerate feasibility,
M 136–138
Mechanical-contact seals, 79 Pressure rise–volume curve, 13–14
Molecular-weight, changes in, Process control system, interaction
99–102 with, 130–131
Motor drives, 18–19 Purchase specifications, 21–22
Multistage compression, with
intercooling, 58–59 R
Radial-bladed impeller, 66
N Radial-flow compressors. See
Naphthenic-based oils, 150 Centrifugal compressors
Nonequalizing tilting-pad thrust Ratio of compression, 37–38,
bearing 41–42
with ball pivot, 73–74 Ratio station, setting of, 105
with radial pivot, 73–74 Reciprocating compressors, 2,
11–13
O cost, 15, 18–19
Oil feed rate, 153 speed, 15–16
Oil-film thickness, 72 Recycling/pressure control
Oil-flooded screw compressors, method, for surge control, 46
lube oil for, 152–153 Rerate feasibility, of compressor,
Oil-free rotary-screw compressors 133–135
advantages, 5 capacity, 135
disadvantages, 5 horsepower, 136
working principle of, 4–5 pressure, 136–138
Oil-free/lubricated compressors, 3 speed, 139
Oil-injected rotary-screw Reverse-indicator graphical
compressors plotting, 74
 Index  •   161

Revolution per minute (RPM), 32 Society of Automotive Engineers

and break horsepower, 41–42 (SAE), 153
and horsepower, 39–40 Speed, for rerate feasibility, 139
Rivet technique, 66–67 Stator blades, 35–36
Rotary-lobe compressor, 2, 10–11 Steam turbines, 18, 20
Rotary-screw compressors, 2, 4 Stonewall, for compressor, 45
failures of, 6 Surge characteristics, 114–115
features of, 8–10 Surge control, 44–45
lubrication for, 149–150 line, 121–122
versus reciprocating compressor, 2 types of, 46
working principle of, 2, 6–8 Surge limit line, 115–121
oil-free rotary-screw Surging, 42–44, 49–50, 93
compressors, 4–5 characteristics, 92–93
oil-free/lubricated Synthetics, 152
compressors, 3 System specifications, 21–22
oil-injected rotary-screw
compressors, 3–4 T
Rotor blades, 35–36 Tapered-land thrust bearing, 73
Rotor dynamics, 67–68 Temperature changes, effect of,
RPM control method, for surge 98–99
control, 46 Thrust bearings
misalignment, 74–77
S types of, 73–74
Seal lubrication technique. See Tilting-pad bearing, 70–72
Batch-packed lubrication Torque-induced whirl, 68
technique Tropical climates, 62
Self-equalizing thrust bearing, Troubleshooting compressor
73–74 problems, 47–50
Separator, 8
Shaft seals, 79 V
Shrouded impellers, 66 Valves, instrumentation and,
Single-screw compressors, 7 122–123
Sliding-vane compressors, 10 Vigilance, during installation,
Slot-welding technique, 66–67 26–28
 OTHER FORTHCOMING TITLES FROM OUR
ENGINEERING TECHNOLOGY
• Rotating Equipment: Maintenance and Troubleshooting by James M. Watterson
• Tribology: A Simple Guide To The Study of Friction by James M. Watterson
• Digital Systems Design: Numbering Systems and Logical Operations
by Larry Massengale
• Digital Systems Design: Development Methods and Combinational Logic Circuits
by Larry Massengale
• Digital Systems Design: Flip-Flop Circuits and Characteristics of Digital Circuits
by Larry Massengale

Momentum Press is one of the leading book publishers in the field of engineering,
mathematics, health, and applied sciences. Momentum Press offers over 30 collections,
including Aerospace, Biomedical, Civil, Environmental, Nanomaterials, Geotechnical,
and many others.

Momentum Press is actively seeking collection editors as well as authors. For more
information about becoming an MP author or collection editor, please visit
http://www.momentumpress.net/contact

Announcing Digital Content Crafted by Librarians

Concise e-books business students need for classroom and research

Momentum Press offers digital content as authoritative treatments of advanced engineering

topics by leaders in their field. Hosted on ebrary, MP provides practitioners, researchers,
faculty, and students in engineering, science, and industry with innovative electronic content
in sensors and controls engineering, advanced energy engineering, manufacturing, and
materials science.

Momentum Press offers library-friendly terms:

• perpetual access for a one-time fee
• no subscriptions or access fees required
• unlimited concurrent usage permitted
• downloadable PDFs provided
• free MARC records included
• free trials

The Momentum Press digital library is very affordable, with no obligation to buy in future
years.

For more information, please visit www.momentumpress.net/library or to set up a trial in the

US, please contact [email protected].
EBOOKS A Simple Guide to Understanding

WATTERSON
FOR THE Compressors
ENGINEERING ENGINEERING TECHNOLOGY COLLECTION
James M. Watterson
LIBRARY
Create your own This straightforward guide to compressors seeks to
Customized Content unveil a lot of myths surrounding compressors. In this
Bundle—the more book, we will be looking at most types of compres-
books you buy, sors, including the centrifugal compressors, the air
the greater your compressors, and of course the most troublesome of
discount! all compressors, the reciprocating compressors.
Having a compressor with minimal operating A Simple Guide
to Understanding
THE CONTENT problems does not only depend on the selection of
• Manufacturing the right type and size for your job. Detailed speci-

A SIMPLE GUIDE TO UNDERSTANDING COMPRESSORS

Engineering fications of all auxiliary equipment and operating

Compressors
• Mechanical
conditions, as well as keeping constant vigilance
& Chemical
Engineering
over the engineering and installation is imperative.
• Materials Science The Simple Guide will explain in a simple yet defini-
& Engineering tive manner which compressor type is best used for
• Civil & which job and what it can produce.
Environmental
Engineering
James M. Watterson has a PhD in mechanical en-
• Advanced Energy
Technologies
gineering and is currently serving as independent
training consultant with JMW, United Arab Emirates,
THE TERMS since January 2003, and as part-time senior lecturer/
• Perpetual access for lead trainer–commissioning and process start-up in
a one time fee the oil and gas refineries with ADNOC since February
• No subscriptions or 2011. He trains groups of maintenance, operations,
access fees instrumentation, and planning engineers. Previously
• Unlimited
he was managing director at JW Management and
concurrent usage
Marketing Consultancy from September 1985 to
• Downloadable PDFs
• Free MARC records October 2000.

For further information, James M. Watterson

a free trial, or to order,
contact: 
[email protected]

90000

9 781947 083707