Pumps and Pump Stations

CHAPTER 4
Pumps and Pump Stations
4.1 INTRODUCTION
The main objective of liquid pipeline operations is to transport liquid petroleum prod-
ucts from the producers to the customers. In order to achieve this objective, energy
is added to the products to increase the pressure at the pump stations for offsetting
the pressure loss in the pipeline. In addition, measurements of pressure and flow are
required for facility control and custody transfer. There are other tasks required to oper-
ate pipeline systems. This chapter discusses such key subjects as pump selection and
sizing, pump operating points, pump station design, and station control.
A pump transforms energy to increase pressure of a liquid and is used extensively
to transport liquid through a pipeline system. The pressure of a liquid has to be in-
creased either to overcome frictional losses or to raise the liquid from one elevation to
a higher elevation. As the flow rate increases, more pumps are required to produce the
required pressure along the pipeline (Figure 4-1).
Depending on the method of adding energy to the liquid, pumps are classified
into two types; centrifugal pumps and positive displacement (PD) pumps. Centrifugal
pumps add kinetic energy to the liquid by increasing the liquid flow velocity, while
PD pumps add energy periodically to the liquid by the direct application of a force to
movable volumes of liquid. The two types of pumps can be compared in general terms,
as listed in Table 4-1.
As shown in this table, centrifugal pumps are most suitable for transmission pipe-
lines transporting most petroleum products. Therefore, they are extensively used in
liquid transmission pipelines and thus this book only discusses the design and opera-
tion of centrifugal pumps and pump stations.
Suction Discharge
system system
Pump
- Pressure - Pressure
- Temperature - Flow
- Specific gravity
- Viscosity
Figure 4-1. Pump suction and discharge systems
159
Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

160 n Hydrocarbon Liquid Transmission Pipeline and Storage Systems
Table 4-1. Centrifugal pumps vs. positive displacement pumps

Centrifugal Pump PD Pump
Operating flow range Flexible Limited
Operating pressure range Low–medium High
Control of pressure and flow Flexible Limited
Pumping efficiency Low–medium High
Viscosity range Low–medium High
Capital cost Low–medium High
Maintenance requirement Low High
Physical size Small Large
4.2 CENTRIFUGAL PUMPS
Centrifugal pumps have prevalent application in liquid hydrocarbon pipeline trans-

mission systems as they are capable of handling variable heads and flow rates [1].
These pumps can handle multiple products over a wide range of viscosities and other
properties [2].
Figure 4-2 shows the cross-sectional and axial views of a centrifugal pump with
an end suction impeller. The main components of a centrifugal pump include an impel-
ler, casing, housing or frame, and shaft and stuffing box/mechanical seal. A centrifugal
pump uses the centrifugal action through the impeller within the pump to transfer
energy. As the pump shaft is rotated by a pump driver, the impeller rotates inside the
pump casing. Liquid flows from the suction piping into the impeller through its eye,
and the rotating impeller imparts kinetic energy to the liquid. As the liquid slows down
while passing through the volute, the liquid kinetic energy is converted into potential
energy or pressure to conserve the total energy.
Figure 4-2. Centrifugal pump with impeller

Pumps and Pump Stations n 161
4.3 CENTRIFUGAL PUMP TYPES
Mainline centrifugal pumps are usually designed to ANSI/API 610 or ISO 13709 (Identical
Standards) – Centrifugal pumps for petroleum, petrochemical and natural gas industries.
Pump types typically used in liquid hydrocarbon process, refining and pipeline
transportation include:
4.3.1 End Suction Single Stage Pumps

These pumps have limited application in pipeline systems but are mostly used in proc-
ess industries (Figure 4-3).
Figure 4‑3. A
PI 610 end suction pump — Courtesy of Flowserve Corporation, all rights
reserved.
4.3.2 Vertical In-Line Single Stage Pumps

These pumps are typically used for product transfer within terminals or as booster
pumps for smaller mainline pumps at initiating stations. API 610/ISO 13709 versions
are used extensively in petrochemical and refinery service (Figure 4-4).
4.3.3 Horizontal Axially Split Between-Bearing Single-Stage Pumps

These pumps are typically used as mainline pumps. These high volume pumps are
often piped in series configuration to provide the high pressures required for pipeline
transmission lines. These double suction, double volute pumps provide high efficiency
over a large range of flow. They are inherently balanced with minimal axial thrust issues
at flows well off the best efficiency point due to the double volute design (Figure 4-5).
4.3.4 Horizontal Axially Split Between-Bearing Multi-Stage Pumps

These pumps have application where higher pressures and lower volume capacity than
the single state pumps are required. This pump design, along with the single stage

Figure 4-4. Vertical in-line. Courtesy of Flowserve Corporation, all rights reserved.
version, allows for ease of maintenance as the pump rotating element can be serviced
or removed for impeller modifications without disturbing the station pressure piping.
These pumps are robust in design and provide long service life. API 610/ISO 13709 is
the guiding standard for design, manufacturing and performance testing requirements
of these pumps. API 610/ISO 13709 designates these pumps as BB1 — axially split
single stage between-bearing pumps and BB3 — axially split multistage between-
bearing pumps (Figure 4-6).
4.3.5 Double–Case (Can) Vertically Suspended Volute Pumps

These pumps are used where there is limited suction pressure available such as in
tankage terminals and where higher viscosity product is transported. Tank farm de-
Figure 4-5. H
orizontal axially split BB single stage pump. Courtesy of Flowserve Corpora-
tion, all rights reserved.

Figure 4-6. M
ulti-stage horizontal axially split pump. Courtesy of Flowserve Corporation, all
rights reserved.
signs usually provide manifolds located near the tanks and incorporate booster pumps
with low NPSH requirements. These pumps are typically can-type vertical multi-stage
centrifugal pumps that accommodate the low suction head available from the tankage.
These booster pumps are designed to provide sufficient pressure to overcome frictional
losses of valves, piping, and fittings throughout the station and to meet the NPSH
requirements of the mainline pumps (Figure 4-7).
Impeller hydraulics can be optimized for individual system requirements with
these pumps. Head rise from design point to shutoff can be as low as 15% over design
Figure 4-7. Vertical Can pump. Courtesy of Flowserve

head. This can result in considerable energy savings particularly under partial loads or
off-peak operating conditions.
The can length will determine the NPSH requirements for the pump and is specific
to the particular application. API 610/ISO 13709 designates this type of pump as VS-7 —
Double casing volute vertically suspended pump.
4.4 PUMP SELECTION AND SIZING
The pump performance is the basis of pump selection. To select the pump, it is most
critical to determine the operating range of the pump and the system curve in the pipe-
line system. The actual procedure to be followed is:
·· Determine the required flow range of pump: It is important to define the

maximum, normal and minimum flow requirements. The normal flow is
the flow at which the pump or pumps will usually operate. The rated flow
is likely to be the maximum flow under current or near term conditions,
whereas the minimum flow limit should be established to accommodate
temperature rise and low efficiency or to install recirculation facility or
multiple pumps. It is also important to clarify the number of days per year
of service at which the pump will operate at maximum, normal and mini-
mum flow rates.
·· Determine the system curves for the maximum, normal, and minimum flow
rates so that the operating envelope of one or more pumps is established. It is
preferred to express the system curves in terms of head because the pump curve
is expressed in head.
·· Select pump type, size, and arrangements, which are interrelated. For example,
two small pumps are arranged in parallel for wide range of flows between the
minimum and maximum flows.
·· Prepare pump data sheets showing pump performance requirements and ser
vice conditions.
·· Solicit for bids and make final selection
4.4.1 Pump Performance

Normally, centrifugal pumps are characterized by efficient performance over a wide
range of pressures and flow rates. Their size is relatively small. They cost less than
other types of pumps and operate reliably. In addition, centrifugal pumps are capable of
pumping high throughput and various products with different liquid densities and vis-
cosities. They can handle products with viscosities up to 350 cSt depending on pump
size and speed before efficiency begins to fall off significantly.
Pump characteristics are represented graphically to describe pump performance.
The pump performance curves are normally provided by the manufacturer. The pump
manufacturer tests the performance with water and only guarantees the performance
at the rated point. Therefore, the pump performance curve, including the shut-off pres-
sure, can vary and this variation must be specified or tested.
Pump performance curves show graphically the relationship of flow rate with
head, efficiency, net positive suction head (NPSH) required, and power requirement
for several impeller diameters. If the driver connected to the pump can vary its speed,
these pump curves are produced for different speeds. The information is used for pump
selection.

Figure 4-8. Typical pump performance curve
4.4.1.1 Pump Performance Curves

Pump manufacturers supply performance curves for their pumps with information on
pump performance over a range of flows. The following diagram (Figure 4-8) dem-
onstrates typical pump curves depicting pump head delivered by the impeller diam-
eter chosen over the recommended flow rate for that pump. The Best Efficiency Point
(BEP) is shown and usually forms the basis of the design flow rate. These pump curves
typically provide the following information:
·· Range of impeller diameters available

·· Head vs. flow rate
·· Pump efficiency vs. flow rate
·· Brake power to drive the pump vs. flow rate
·· Net Positive Suction Head Required (NPSHR) for the pump vs. flow rate
These curves are plotted for the rated speed and different impeller diameters. For
variable speed drivers, the curves are shown at various pump speeds.
Some pumps are equipped with double suction impellers with the impeller eyes
located on both sides. Pump shut-off is the head developed at zero flow, while hydrau-
lic runout is the pump capacity above which the pump should not be operated due to
instability and other operational problems. This point is usually defined at 120% of the
best efficiency point (BEP).
4.4.2 Service Conditions

There are many factors that must be considered in the selection and sizing of pumps for
a liquid hydrocarbon pipeline system. In order to select the proper pump, the following
parameters should be known:
·· Liquids to be pumped — clear liquids or liquids containing solids or vapor

·· Liquid specific gravity/density
·· Liquid vapor pressure at the pumping temperature

·· Viscosity— pump performance drops more rapidly above 110 cSt

·· System throughput — desired pump capacity and expected future changes to
the desired capacity are functions of pump speed and size
·· Pump head requirements — pump head depends on speed and impeller
diameter
·· Pressure conditions — suction and discharge pressures, future pressure condi-
tion, and series/parallel operation conditions
·· Suction requirement — NPSHA must be greater than NPSHR of the pump
·· Pump unit efficiency — energy usage is dictated by pumping efficiency
·· Type of service — number of operating hours and criticality of the service
·· Preferred pump and driver type
·· Operability and maintainability
·· Equipment life cycle cost including the initial purchasing and installation costs,
operating cost, and maintenance cost
·· Specific site conditions and space limitations: topography and elevation above
sea level and NPSHA
·· Other considerations such as codes and regulations
A rule of thumb of sizing and selecting centrifugal pumps is to choose the physi-
cally smallest pump that will satisfy the service requirements. Centrifugal pumps are
sized on the following basis:
·· Impeller diameter: The pumping head is proportional to the square of the im-
peller diameter, while the flow rate varies linearly with the diameter. Therefore,
the larger the impeller diameter, the higher the head and the throughput. Nor-
mally, pump vendors provide a range of impeller diameter suitable for a pump.
Impeller diameters are determined based on required head at design point. The
pump manufacturer will then trim the impeller to the required diameter.
·· Impeller speed: The head and flow varies in a similar manner to the impeller as
described above. However, because of dynamic forces on the impeller, speed
limits impeller size. The speed ranges from 1200 RPM up to 5500 RPM. Refer
to Section 4.6.3 for the Affinity Laws.
·· Suction pressure: The NPSHR of a pump is the limiting factor that affects size,
speed, and capacity. This topic is discussed in Section 4.4.3.
·· Suction and discharge nozzle sizes: Suction nozzles are usually larger than dis-
charge nozzles. The larger the nozzle size, the higher the flow capacity of the
pump. Nozzles sizes are determined by the pump manufacturer.
Pipeline system hydraulic requirements will determine the selection of pumps.

If there are batching requirements with products of varying density and viscosity, it
may be necessary to provide multiple pumps operating in series or parallel, or to use
some combination of series/parallel configuration. Variable speed pumps can sim-
plify the selection process. The pump configuration should be selected to maintain the
maximum pump efficiency over the range of flows expected for the pipeline system.
Initial and future flow rates should be evaluated so that the selected pumps have suffi-
cient flexibility to handle any anticipated flow and product properties. For comparison
purposes, pump efficiencies should be evaluated for both series and parallel pump
configurations.
Once the pump configurations with head requirements at various flow rates have
been determined, it will be necessary to select actual pumps from manufacturers’ pub-
lished pump performance data. Pump manufacturers publish pump performance maps
that depict pump performance for a specific family of pumps over a large range of

Figure 4-9. Pump selection map
flows and differential heads. Pump manufacturers produce charts to show the capaci-
ties of different sizes of pumps. Figure 4-9 shows a typical performance map for cen-
trifugal pumps suitable for pipeline service.
4.4.3 Net Positive Suction Head (NPSH)

Net positive suction head (NPSH) is the total absolute suction pressure at the impel-
ler eye less the absolute vapor pressure of the liquid pumped. NPSH must be of a
magnitude to avoid vapor formation in the liquid and hence cavitation. Another way
to describe NPSH is that it represents the amount of head required to push the fluid
into the pump to suppress cavitation [3]. Available and required NPSH are calculated
as follows:
4.4.3.1 Net Positive Suction Head Required (NPSHR)

The NPSHR specified in the pump performance curves is another important perfor
mance parameter. All centrifugal pumps accelerate fluid and possess a corresponding
internal friction loss. Therefore, every pump requires a certain amount of positive suc-
tion pressure in order to avoid cavitation. The pump NPSHR is determined by actual
tests conducted by the pump manufacturer using procedures established by the Hy-
draulic Institute.
Cavitation is a critical problem in controlling pumps. When a liquid enters into the
pump impeller in such a manner that the local pressure drops below the liquid’s vapor
pressure, the liquid transforms into the vapor phase and bubbles are formed in and
around the impeller. If the local pressure recovers due to centrifugal action, the vapor
bubbles collapse into the liquid phase. Since the bubbles occupy much larger volume
than the liquid, collapsing bubbles release a huge amount of energy. This energy will
hit the surrounding metal, causing physical damage to the impeller and casing. The
other effects of cavitation include vibration and noise and dramatic reduction in pump
performance.

4.4.3.2 Net Positive Suction Head Available (NPSHA)

Net Positive Suction Head Available (NPSHA) is calculated by the following
expression:
NPSHA = ha - hVP + hst - hfs
Or
2.31
NPSHA = hst + ( Pa - PVP ) - h fs
G
where
Pa = absolute pressure at the surface of the liquid supply level
PVP = vapor pressure of the liquid at the temperature being pumped
ha = Pa expressed in equivalent head
hVP = PVP expressed in equivalent head
hst = static elevation of the liquid supply above or below the pump inlet
centerline.
hfs = suction line losses including entrance losses and friction of the piping
G = specific gravity
Figure 4-10 shows that NPSHA must be greater than NPSHR for stable operation
and that NPSHR increases and NPSHA decreases as the flow rate increases.
The NPSHR stated by the pump manufacturer is at a point where the pump is
in full cavitation. Therefore, it is important to allow a margin between NPSHR and
NPSHA. As a rule of thumb, the NPSHA should be at least 10% greater than the NPSH
required by the pump [1].
If the requirement for stable operation cannot be satisfied, either NPSHR should be
reduced, NPSHA increased, or both. The NPSHR reduction can be accomplished by using
double suction impellers or by impeller design with a larger impeller eye area. Also, smaller
Figure 4-10. Stable operating point

pumps can be installed in parallel or a larger suction pipe size can be used to reduce frictional
pressure losses in suction piping. NPSHA can be increased by installing a booster pump in
front of the pump or by reducing the friction pressure losses in suction piping.
4.4.4 Specific Speed

Specific speed (NS) is used to predict pump characteristics for the purpose of classifying
pump impellers according to type, proportions and performance. Specific speed is often used
for comparison purposes on selection of pumps from different manufacturers [3].
Specific speed is expressed as:
N Q
Ns = 3
H4
where
Ns = pump specific speed, dimensionless
N = pump speed, RPM
Q = capacity at best efficiency point, USGPM
H = total head per stage at the best efficiency point, feet
For double suction impellers, one half of the flow is used to calculate the specific
speed. The specific speed (Ns) determines the general shape or class of the impeller. As
the specific speed increases, the ratio of the impeller outlet diameter, D2, to the inlet or
eye diameter, D1, decreases. This ratio becomes 1.0 for a true axial flow impeller.
Radial flow impellers develop head mainly through centrifugal force. Pumps of
higher specific speeds develop head partially by centrifugal force and partially by axial
force. A pump with a higher specific speed generates head more by axial forces and less
by centrifugal forces. An axial flow or propeller pump with a specific speed of 10,000
or greater generates its head exclusively through axial forces.
Typical values for specific speed (Ns) for different designs in US units (gpm, ft)
·· radial flow — 500 < Ns < 4000 — typical for centrifugal impeller pumps with
radial vanes — double and single suction. Francis vane impellers operate in the
upper range
·· mixed flow — 2000 < Ns < 8000 — more typical for mixed impeller single
suction pumps
·· axial flow — 7000 < Ns < 20,000 — typical for propellers and axial fans
To convert between US units (USgpm) and Metric units (m3/h)
Ns(US gpm, ft) = 0.861 Ns(m3/h, m)
The specific speed of an impeller can provide a wide variety of information about
its performance:
Impellers with low specific speed are long and thin and are used for low-flow,
high-head applications. Impellers with high specific speed are short and stubby and are
used for high-flow, low-head applications (Figure 4-11).
Efficiency is determined by considering the losses through pump impeller fric-
tion, ring leakage, and mechanical losses, as well as losses incurred by movement of
the liquid within the pump, referred to as hydrodynamic losses. Specific speed affects
pump efficiency. The lower the specific speed, the lower the efficiency. The reason is

Figure 4‑11. Impeller design vs. specific speed –— Hydraulic Institute
that a higher percentage of energy is lost to overcome the impeller disk friction that is
necessary to generate high heads (Figure 4-12).
Once an impeller is designed for a certain specific speed, it will produce a typical
head capacity curve and efficiency curve shape. A low specific speed impeller has a
flat curve with a wide efficiency range. A high specific speed impeller produces a steep
curve with a narrow efficiency range.
The major use of the specific speed number is to help specify pumps to be as efficient
as possible for the service intended. Maximum pump efficiency is obtained in the specific
speed range of 2000 to 3000. Pumps for high head low capacity occupy the range 500 to
1000 while low head high capacity pumps may have a specific speed of 15,000 or larger.
4.4.5 Suction Specific Speed

Suction specific speed is an index that describes the characteristics of the suction side
of the impeller. It is calculated at the pump’s best efficiency point and maximum impel-
ler diameter. The equation for suction specific speed (designated Nss or S) is:
Figure 4-12. Pump efficiency vs. specific speed

N(Q )
0.5
Nss =
( NPSHR)0.75
where
N = rotating speed (rpm)
Q = flow per impeller eye (m3/second)
NPSHR = Net Positive Suction Head Required (see Section 4.4.3)
For double suction impellers, Q is one half of total flow. Nss derived using SI
units can be converted to US Customary Units by multiplying by a factor of 51.64
(Ref. API 610)
From the equation, we can see that the lower the NPSHR for a pump the higher
the Nss. Nss values for many standard impellers typically range from 7000 to 9000, but
some designs may have an Nss as high as 18,000 to 20,000.
It is important to consider that increasing the Nss of a pump has been shown to
shift the onset of suction or discharge recirculation closer to the best efficiency point
(BEP) flow of the pump. This effectively decreases the window of stable operation for
the pump. Suction recirculation is a reversal of flow in the impeller eye that can lead to
increased noise, surges, and cavitation-like damage to the impeller vane. Discharge re-
circulation is a similar reversal of flow occurring at the discharge of the impeller vane.
It is recommended that pumps should have an Nss of no more than 9000 for water and
11,000 for hydrocarbons [3].
4.4.6 Pump Performance Curve Characteristics

There are a number of pump head-capacity (H-Q) curve shapes that are shown in
Figure 4-13 below. These characteristic curves are listed below and described thereafter:
·· Rising
·· Drooping
·· Steep
·· Flat
Figure 4-13. Pump performance curve characteristics

·· Stable
·· Unstable
Rising characteristic — the head rises continuously as the capacity is decreased to

shut-off.
Drooping characteristic — the head developed at shut-off is less than at some of the
other capacities.
Steep characteristic — the head developed at shut-off is significantly larger than that
developed at the design capacity.
Flat characteristic — the head developed at shut-off is approximately that developed
at the design capacity. The curve can be a slightly rising or drooping.
Stable curve — is a rising curve where only one capacity can be obtained at any one
head. A curve with a rising characteristic would be an example of a stable curve.
Unstable curve — is a drooping curve where more than one capacity can be
obtained at one head. A curve with a drooping characteristic would be an example of
an unstable curve.
Unstable curves where the maximum developed head is at some flow greater than zero
are undesirable in applications where multiple pumps operate in parallel. In such applica-
tions, zero flow head may be less than system head, making it impossible to bring a second
pump on line. It is also possible for pumps in this configuration to deliver unequal flow with
the discharge pressure from one pump determining the flow rate from another [1].
4.4.7 Centrifugal Pump Power and Efficiency

The ideal hydraulic power to drive a pump depends on the flow rate, the liquid density,
and the differential head generated by the pump. This can be calculated as:
qrgh
Ph =
(3.6 ´ 106 )
where
Ph = hydraulic power (kW)
q = flow capacity (m3/h)
r = density of fluid (kg/m3)
g = acceleration due to gravity (9.81 m/s2)
h = differential head (m)
Power is more commonly expressed as kilowatts (kW) or horsepower

(hp = kW ´ 0.746).
The shaft power is the power required transferred from the motor to the shaft of the
pump, and it depends on the efficiency of the pump. Shaft power can be calculated as:
Ps = Ph
h
where
Ps = shaft power (kW)
h = pump efficiency
4.4.8 Performance Modifications For Varying Pipeline Applications

The performance of pipeline pumps often needs to be altered to accommodate varying
liquid transmission conditions. These generally require that the pump be physically mod-
ified to meet the new conditions. The following are considered by the pipeline industry:

Figure 4-14. Changing performance by impeller vane number
Impeller change — In order to change the specific speed, impeller size may be changed
to meet the new demand on performance. Pump manufacturers can usually offer sev-
eral different impeller diameters and vanes that will fit the pump casing without any
further internal modifications. The effect of changing the impeller characteristics is
illustrated in Figures 4-14 and 4-15.
Restaging — Pumps with multi-staging capabilities can be restaged (up or down)
to meet the change in pressure, head or flow requirements. For example, if an entire
pressure range is not needed for a particular period of time, a number of impellers can
be removed to meet the required conditions. Manufacturers provide de-staging kits to
block off the unused pump impeller areas to maintain efficiency (Figure 4-16).
Figure 4-15. Changing pump performance by impeller vane angle

Figure 4-16. Changing performance by re-staging
Figure 4-17. Changing performance with underfiling

Figure 4-18. Changing performance with volute chipping
Figure 4-19. Changing performance with volute inserts

Impeller underfiling and overfiling — This is undertaken to alter the performance of

a pump. It involves modifying the flow area of the impeller by grinding metal off the
impeller outlet vanes (Figure 4-17).
Impeller volute chipping — This is a technique that is used to alter the outlet flow
area of the pump casing in order to modify performance (Figure 4-18).
Impeller volute inserts — This technique (Figure 4-19) involves inserting special
removable volutes into the pump to allow for a wider performance range. It also allows
for more accurate and close control over the performance.
4.4.9 Cavitation [6–10]

Cavitation is the rapid formation and collapse of vapor bubbles that form in a pump
inlet whenever the local absolute pressure of the liquid falls below its vapor pres-
sure. These bubbles collapse rapidly and violently when the local absolute pressure
increases due to kinetic forces being imparted by the impeller (see Figure 4-20 below).
Cavitation is the rapid formation and collapse of these vapor bubbles.
Figure 4-20. Cavitation bubble implosion, arrows indicate fluid pressure [16 modified]
Collapsing vapor bubbles cause noise, vibration, and erosion of material from the
damaged impeller as shown below in Figure 4-21 [4].
Cavitation control is a very important consideration in any liquid system and thus
any cavitation induced conditions must be avoided when operating centrifugal pumps. If
a liquid is accelerated in such a manner that the local pressure falls below the liquid vapor
pressure, the liquid will transform into the vapor phase, which results in the formation
of bubbles. If the local pressure recovers, the vapor bubbles will transform themselves
back into a liquid. There is a tremendous volume change during transformation, because
collapsing bubbles release a large amount of energy. Because the bubbles are very small,
the resulting impact loads on the surrounding metal can be significant. This can result in
the creation of high noise levels and physical damage to the metal [6].
Some liquids (such as water) are more difficult to handle from a cavitation point of
view. When the vapor pressure of a homogenous fluid such as water is reached, the entire
fluid begins to change phase, resulting in the formation of a large number of damage-
causing bubbles. For a non-homogeneous fluid such as a hydrocarbon, only the light ends
(such as condensates) are affected and the impacts of cavitation are reduced (Figure 4-22).
In a centrifugal pump, the fluid is accelerated by the impeller. The area of lowest
pressure in the pump suction system, as shown in Figure 4-23, is the eye of the impeller
at cross section A-A. If the pressure falls below the vapor pressure of the liquid, vapor
bubbles form. As the mixture of liquid and bubbles continue through the pump, the
pressure increases and the bubbles return to the liquid state. Damage to the impeller

Figure 4-21. Pump impellers destroyed by cavitation [10,8]
Figure 4-22. Vapor pressure of some fluids

Figure 4-23. Cavitation formation
Figure 4-24. Result of cavitation on pump performance

occurs where the bubbles collapse as shown at cross section B-B. This location varies
for different impellers and different suction conditions:
The effects of cavitation include:
·· Noise and vibration

·· Pump damage (e.g., pitting of the impeller)
·· Fall off of pump performance and efficiency
Cavitation in centrifugal pumps can be recognized by a characteristic noise, which

sounds just like it is trying to pump gravel. A typical break-off in the performance
curve of a pump due to cavitation is shown in Figure 4-24.
Figure 4-25. P
erformance correction chart for viscous liquids — with permission of Hydrau
lics Institute

4.4.10 Viscous Hydrocarbon Behavior in Pumps [11]

The performance of centrifugal pumps is affected by fluids with higher viscosities in-
cluding fluids that behave in a non-Newtonian manner. An increase in power required
a reduction in head generated, a loss in efficiency and in some cases a loss of capacity
can be expected with the transmission of high viscosity fluids.
Performance correction factors, as depicted in Figure 4-25, can be applied to esti-
mate the actual performance of a pump handling viscous fluids.
Figure 4-25 is to be used only within the scales shown and only for pumps of
conventional design. The chart is applicable only for conventional centrifugal pumps
operating close to their best efficiency point (BEP), having sufficient net positive suc-
tion head available (NPSHA) and transferring Newtonian fluids (such as crude oil). For
description on non-Newtonian fluids refer to Chapter 1.
The following equations are used for determining the viscous performance when
the water performance of the pump is known:
Qvis = CQ ´ QW
Hvis = CH ´ HW
Evis = CE ´ EW
Qvis ´ H vis ´ sp gr
BHPvis =
3960 ´ Evis
where
Qvis = viscous capacity (USgpm)
Hvis = viscous head (ft)
Evis = viscous efficiency (%)
BHPvis = viscous brake horsepower
Figure 4-26. Viscosity increase vs. pump performance

QW = water capacity (USgpm)

HW = water head (ft)
CQ = Capacity correction factor
CH = Head correction factor
CE = Efficiency correction factor
QW = Water Capacity at which maximum efficiency is obtained
Figure 4-26 illustrates graphically the consequences of viscosity increases on the

head, capacity, and brake power requirements. As the liquid viscosity increases, the
head generated by a centrifugal pump is reduced, the pumping efficiency drops, and
as a result the brake power required for pumping increases. Note that the viscosity is
greater than 500 SSU or about 100 cSt, the pump performance drops rapidly.
4.4.11 Temperature Rise

Fluid temperature rises across the pump for two reasons. First, there is an adiabatic
temperature rise due to compression of the hydrocarbon fluid across the pump. As well,
heat is generated due to frictional forces generated within the pump and by recircula-
tion caused by leakage through clearances within the pump. This major component of
the temperature rise is directly related to the efficiency of the pump at the operating
flow rate. The amount of heat generated is the difference between the input energy to
the pump and the delivery energy.
Temperature rise in a pump with a closed discharge valve can occur quickly if
the pump generates high head. Power losses are equal to the shutoff input power, and
all this power goes into heating the small quantity of fluid contained within the pump
casing. Care must be taken to prevent extended shut-in head conditions during pump
startup.
Since the heat is generated as the liquid passes through the pump, the pump dis-
charge temperature is increased and is calculated using the fundamental concept that
the mechanical energy lost in the pump due to mechanical efficiency is converted to
heat energy.
DP æ 1 - h ö
Td = Ts +
rCp çè h ÷ø
where
Td = discharge temperature (oC)
Ts = suction temperature (oC)
Cp = liquid heat capacity (kJ/kg oC)
h = pump mechanical efficiency
r = liquid density (kg/m3)
Δ P = discharge and suction pressure difference (kPa)
When the pump is running normally, the temperature increase is small, in the
o rder of a few degrees Celsius. If the pump discharge is shut off or the flow is too slow,
energy is converted to heat and the heat cannot be carried away quickly. The liquid
in the pump will heat and eventually vaporize. This can result in dramatic failures,
particularly for large multi-stage pumps. Such a situation can be avoided by auto-
matically shutting down the pump as the flow rate drops below the pump’s Minimum
Continuous Stable Flow (see Section 4.4.12) or by providing a recirculation system
(see Section 4.9.4).

Figure 4-27. Performance curve showing MCSF
4.4.12 Minimum Flow

Minimum Continuous Stable Flow (MCSF) is usually provided by the pump manu-
facturer and is the lowest flow rate at which a pump can operate continuously without
exceeding the vibration and/or noise limits specified. These limits are usually refer-
enced to industry standards such as ANSI, API, ISO, and ASME or, in some cases, to
the customer’s own pump specifications (Figure 4-27).
There are many ways by which a manufacturer determines its recommended
MCSF for a specific pump. This can be based on actual test results, historical experi-
ence or the specifics of the pump design. MCSF can be influenced by the properties of
the pumped liquid as well [5].
4.5 PUMP SPECIFICATION AND PURCHASE
A pump purchase requisition must be prepared and should consist of at least two
parts:
·· Completed API 610/ISO 13709 Pump Data Sheets

·· Clarifications/Supplements to API 610/ISO 13709 Standards, if required
4.5.1 Pump Data Sheets

Appendix B of API 610/ISO 13709 contains sample data sheets that should be used
when purchasing an API 610/ISO 13709 pump.
It is important to complete all sections of the pump data sheets. There are many
options in the API 610/ISO 13709 standards and the data sheet will clarify which

options are required and ensure that all vendors are quoting to the same requirements.
Important aspects that the data sheets clarify are:
·· Service conditions (see Section 4.4.2)

·· Liquid properties (see Chapter 2)
·· Site conditions
·· Preferred coupling type
·· Materials of construction
·· Mechanical seal type
·· Mechanical seal flush piping
·· Bearing lubrication system requirements
·· Instrumentation requirements
·· Painting requirements
·· Electrical power available and hazardous area classification
·· Pump testing requirements
·· Material inspection requirements
·· Spare parts requirements
4.6 RETROFITTING CENTRIFUGAL PUMPS FOR CHANGING

SERVICE CONDITIONS
4.6.1 Reduced Pipeline Throughput

When pipeline throughput is reduced, the pumps in effect become technically over-
sized and are therefore operating inefficiently. These pumps then become subject to
problems that are associated with low flows, such as vibration or seal and bearing
failures. There are a number of retrofit solutions:
Change the pump speed to adapt to the desired flow conditions in the pump,
·· reduce the pump impeller diameter to better suit the lower flow rate,
·· install volute inserts; or,
·· combination of the above solutions.
4.6.2 Increased Pipeline Throughput

When pipeline throughput above current system design is required, it will be neces-
sary to add intermediate pump stations to the pipeline system. By adding stations, the
system hydraulic performance curves are modified to allow increased flow rates while
staying within system Maximum Allowable Operating Pressures.
These increased flow rates through existing pump stations may change the pump
operating conditions to the point that pumps operate inefficiently or overload the pump
drivers. Depending on the actual flow rate increases required, there are a number of
modifications that can be adopted to have the existing pumps perform well under the
new operating conditions. Retrofit solutions include:
·· Trimming the pump impeller to allow a constant speed pump driver to remain
within its power limits.
·· Reducing the speed of a variable speed driver to limit the power requirements
of increased flow through the pump.

·· Adding a parallel pump unit capable of delivering the same pump head as the
existing pump/pumps that operate in series. This approach then reduces the
flow to the existing pump/pumps to allow pump performance to be maintained
as originally designed.
The following section on affinity laws outlines the method of calculating impact
on capacities, pressure rise and pump speeds by changing certain parameters.
4.6.3 Affinity Laws [4]

The flow rate and head generated by a centrifugal pump may be changed by varying
either the pump speed or changing the impeller diameter. This results in a change to
the impeller tip speed or velocity of its vanes, which causes a change in the velocity at
which the liquid leaves the impeller. Usually, impellers can be cut down to 80% of their
original diameter without lowering their efficiency significantly.
For centrifugal pumps with radial impellers, the relationships are approximated
as follows:
For diameter change only:
2 3
æD ö æD ö æD ö
Q2 = Q1 ç 2 ÷ , H 2 = H1 ç 2 ÷ , BHP2 = BHP1 ç 2 ÷
è D1 ø è D1 ø è D1 ø
For speed change only:
2 3
N  N  N 
Q2 = Q1  2  , H 2 = H1  2  , BHP2 = BHP1  2  ,
 N1   N1   N1 
Figure 4-28. Affinity laws

For diameter and speed change:

2 3
æD N ö æ D2 N 2 ö æ D2 N 2 ö
Q2 = Q1 ç 2 ´ 2 ÷ø , H 2 = H1 çè D ´ N ÷ø , BHP2 = BHP1 çè D ´ N ÷ø ,
è D1 N1 1 1 1 1
where
D = impeller diameter (in.)
H = head (ft)
Q = capacity (USgpm)
N = speed in RPM
BHP = brake horsepower
1 = original conditions subscript
2 = new design conditions subscript
The affinity laws can be presented graphically as shown in Figure 4-28. The pump-
ing capacity and efficiency increases as the impeller diameter and/or speed increases.
Since the head and flow capacity are higher, the power needed for higher speed and/
or larger diameter is greater. Figure 4-30 illustrates the capacity changes with speed
change on the left and impeller diameter change on the right hand side.
4.7 PIPELINE HYDRAULIC REQUIREMENTS
4.7.1 System Head Curves and Pump Operating Points

A system curve or system head curve for a pipeline demonstrates the head required at
that location for the flow rate range. As demonstrated in Chapter 3, the pressure or head
required to overcome frictional losses increases with the flow rate. The system curve
depends on the following variables:
·· Flow rate
·· Liquid density and
·· Liquid viscosity
·· Pipe diameter, wall thickness, and
·· Pipe roughness
·· Elevation difference
Pressure
or head
Friction
Flow Rate
Figure 4-29. System curve

s olution for base load applications where electrical power is available and reliable. They
have the advantage of low maintenance costs and are simple to operate. Variable speed drive
(VSD) motors are becoming the standard for pump stations that have varying flow or prod-
uct density requirements such as on batched product pipelines. Despite their control systems
being more complex than for a constant speed motor, variable speed motors are much more
energy efficient. This is because pump capacity can be controlled without the disadvantage
of pressure loss incurred by the throttling through a discharge control valve.
Variable speed pumps control the flow and pressures by varying the speed of the driv-
ers with maximum power override. For a pump station that contains both fixed-speed and
variable speed motors, the control strategy is to run the fixed speed units at a base load with
minimal throttling and use the variable speed units to adjust for the required station set point.
Figure 4-58 exhibits the performance curves of a variable speed pump.
In applications that require flow or pressure control, the most energy efficient op-
tion is an electronic VSD, referred to as a Variable Frequency Drive (VFD). The most
common form of VFD is the voltage-source, pulse-width modulated frequency con-
verter. The converter develops a voltage directly proportional to the frequency which
produces a constant magnetic flux in the motor. This type of speed control can be
related to set points of discharge pressure or flow.
4.10.6.1 Throttling for Fixed Speed Pumps
As discussed in Section 4.7.1, there is only one operating point for a fixed speed pump.
As shown in the figure below, a throttling action is required to match the system head
curve to the pumping head curve of a fixed speed pump at a particular flow rate other
than the design flow. Fundamentally, throttling changes the system curve. A control
valve is used to throttle the fluid flow, and is installed downstream of the pump.
Figure 4-51 shows that the pump operates at H1 for the design flow rate Q0. If a
throttle valve is partially closed in the pump discharge line, the throughput drops from
Q0 to QT, and additional friction pressure drop occurs through the partially closed
valve. As a result, the pump will operate at a new operating point, H2.
Figure 4-51. Capacity change with throttling

Pressure
or Head
System
Curve
Friction
Head
Elevation Static
Difference Head
0 Flow Rate
Figure 4-31. System curve with elevation difference
These changes could move the system curve and consequently pump operating
point. Therefore, the range of these changes in operating point and their consequences
in the power requirement, capacity and NPSH has to be considered in designing and
selecting pumps.
A system curve is a graphic representation plotted on an x–y graph, where x-axis
represents the flow rate and y-axis the pressure or head caused by the frictional pressure
drop along the pipeline. Figure 4-31 shows a typical system curve for a flat pipeline
system in which a single product is transported.
Figure 4-30 below shows system curves for a level pipeline system in which two
products are transported, assuming that the two products have different densities and/
or viscosities. The higher specific gravity and viscosity product requires greater pres-
sures compared to the liquid with lower gravity and viscosity, generating different sys-
tem curves. Figure 4-30 shows two examples of system curves; one for heavier product
Pressure
or Head
System
Pump Curve
Curve
PO
Operating
Point
Flow Rate Q0
Figure 4.32. Pump operating point

causing higher frictional pressure drop and the other for lighter product causing lower
frictional pressure drop. It is assumed that the pipe size remains constant and the eleva-
tion profile is flat. If the pipe size changes, so does the slope of the system curve where
the pipe size changes. If the elevation changes at two different locations, the system curve
simply moves up or down as shown in Figure 4-31.
If the next pump station or delivery point is higher in elevation, the system curve
is shifted upward as shown in Figure 4-33.
The pump operating point is the point where the pump curve meets the system
curve, as shown in Figure 4-32 below.
4.7.2 Hydraulic Performance in Batched Pipeline Systems with

Constant Speed Pumps
If more than one product is batched in a pipeline, the operating points will change
depending on their positions with respect to pipeline system and pump curves. As
an example, refer to Figure 4-33, showing the system curves for the heavy and
light oil. The operating points of these two liquids are located at H for the heavy
oil and L for the light oil for the pump. Even though these two products have
different densities, the pump curve in head is the same. With all heavy oil in the
pipeline, the operating point is located at H, the flow rate being QH and the head
being HH. When batching products, the positions of the batched products shift
with respect to a pump and pipeline, filling the pipeline with the heavy oil and the
rest with the light oil. Therefore, the flow rate will be at some point between QH
and QL, and the operating point is determined with a new system curve. In this
example, as the light oil moves into the pump and the pipeline, the operating point
moves to L, eventually reaching equilibrium at an operating point where the flow
rate is located at QL and the head at HL.
Figure 4-34 below shows the pump and system curves plotted in pressure instead
of head. Since the pressure changes with density, two pump pressure curves are dis-
played with four operating points. If all heavy oil is in the pipeline and passes through
Head
Pump Curve
H
HH
Heavy Oil
L
HL
Light Oil
Flow Rate QH QL
Figure 4-33. Operating points in head for multiple products

Pressure
Pump curve Heavy oil
for heavy oil
Pump curve
HH
for light oil Light oil
PHH
HL
PHL LH
PLH
LL
PLL
Flow Rate QHL QHH QLL QLH
Figure 4-34. Operating points in pressure for multiple products — constant speed pumps
the pump, the operating point is located at HH in the figure. As the light oil enters the
pump, while the heavy oil is still flowing in the pipeline, the operating point slowly
moves from HH to HL, but with reduced flow rate. In reality, this movement does
not occur abruptly because the batch interface enters first and the density gradually
changes before the light oil completely replaces the interface. The flow reduction is
caused by the low density of the light oil passing through the pump and thus reducing
the differential pressure. As the light oil moves into the pipeline, the operating point
gradually moves to the new operating point, LL, because the light oil requires lower
system curve. As the light oil fills the pipeline, the flow rate increases but the pressure
decreases. The new point represents the operating point with the light oil in both the
pipeline and the pump.
In the same process as described in the previous paragraph, as the heavy oil moves
into the pump and the pipeline is filled with light oil, the operating point moves to
LH, and eventually the operating point moves back to HH as the heavy oil fills both
the pump and the pipeline. In batch operations, the flow rates vary between QHL and
QLH, so does the pressure assuming the pump is not throttled. Note that the positions
of these four operating points may be altered depending on the viscosity of the heavy
oil because pump performance curves are changed with higher viscosity liquids (refer
to Section 4.4.10).
4.7.3 Hydraulic Performance in Batched Pipeline Systems with

Variable Speed Pumps
With the introduction of economically available variable speed drivers for centrifugal
pumps (see later descriptions in this chapter); it is possible to provide more capacity to
batched product pipelines very efficiently. By increasing pump speed in the example
shown above (Figure 4-34) as light oil enters the pump with the pipeline full of heavy
oil, it is possible to maintain pressures and thus flow rates at QHH with more rapidly
increasing flow rates to QLH than in the previous example. In this case, the design point
for the pumps would be at LH with heavy oil density and viscosity determining the
power requirements of the variable speed pump.

4.7.4 Pump Configurations

4.7.4.1 Parallel Operation
If a station has more than one pump unit, it is operated in one of two modes: parallel
operation and series operation. Multiple pumps in a pump station can be arranged in
parallel, series, or combination of parallel and series arrangement. Some combination
of these arrangements is needed to increase both flow and pressure.
The primary purpose of operating units in parallel is to allow a wider range
of flow than would be possible with a single unit for systems with widely vary-
ing flows. Generally, pumps are arranged to operate in parallel to increase sys-
tem flow rates by maintaining discharge head through a much greater pipeline
throughput. The following alternatives should be considered in selecting pump
configurations:
·· One operational pump with one spare unit in parallel;

·· Two operational pumps in parallel with one parallel pump as spare;
·· Two pumps in parallel without a 100% spare, if less capacity can be tolerated
at times.
When pump units are arranged in parallel, loads are shared by more than one unit. The
following load sharing strategies are available:
·· Base loading: one or more units may be operated at a constant load while other
units are operated to handle additional flow rate requirements;
·· Optimum load sharing: set points for each unit are determined based on know-
ing the individual operating curves and allocating load to optimize pump
efficiency and minimize overall energy consumption;
·· Equal load sharing: each unit shares load equally. This strategy is used for
identical units which are arranged in parallel.
There are several areas of concern when operating centrifugal pumps in parallel. If
adequate flow control is not provided, hydraulic imbalance could occur resulting in
high flow rates in one of the parallel pump units and low flow rates in the other unit.
Possible issues include:
Station Station
Station Discharge
Suction valve Block valve valve
Check
Filter Bypass valve
check valve Isolation
valve
V-23 V-24
Control
valve
Isolation
Pump 1 Bypass valve
check valve
V-23 V-24
Pump 2
Figure 4-35. Parallel operation

Figure 4-36. Operating points for parallel operation
·· High flow rate induced cavitation;

·· Pump driver overloading;
·· Prolonged operation of one of the pumps at a flow below its minimum accept-
able continuous flow rate.
Selection of appropriate control systems is essential with pumps operating in parallel

configuration.
Figure 4-35 shows a parallel arrangement of two pumps, where more than one unit
can be operated at the same time. Figure 4-36 demonstrates the operating points of one
pump and two identical pumps when they are arranged in parallel. When two or more
units operate in parallel, all units have common suction and discharge pressures.
Station Station
Station discharge
Suction valve valve
Block valve
Check
Filter valve
Isolation
valve
Control
valve
Isolation
Bypass Bypass valve
check valve check valve
Suction Discharge Suction Discharge

valve valve valve valve
Pump 1 Pump 2
Figure 4-37. Two pumps in series

Figure 4-38. Operating points for series operation
4.7.4.2 Series Operation

The main reason for operating units in series is to increase the pumping head from
what would be possible with a single unit. Pumps operating in series add head and
capacity to the system output. In areas of large pipeline elevation rise, if two pumps
are arranged in series and one shuts down, the remaining pump alone may not be able
to provide the necessary head for the static lift necessary to maintain pipeline flow. In
this case, appropriate station design would require that a spare pump unit be installed
to be activated in the event of a loss of a pump unit operation.
Figure 4-37 shows a series arrangement of two pumps, and Figure 4-38 demonstrates
the operating points of one pump and two identical pumps when they are arranged in
series. In series operation, the flow through all of the units is equal and the discharge
of one pump feeds the suction of the next unit.
4.8 PUMP DRIVERS
A mainline horizontal centrifugal pump can be driven by an electric motor, gas turbine
or a diesel engine. Liquid hydrocarbon transmission pipelines are typically driven by
electric motors where electrical power is available. This is primarily due to their lower
initial capital cost and inherent reliability. As well, electric motor drivers have benefits
over gas turbine and diesel drivers due to ever more stringent emission level limits.
This section will consider pump station operation using both constant speed elec-
trically driven pumps and variable speed electrically driven pumps. Fixed-speed elec-
tric motors provide a cost-effective solution for base load applications where electrical
power is available and reliable. They have the advantage of low-maintenance costs and
are simple to operate. Variable speed drive (VSD) motors are becoming the standard

Figure 4-39. Variable speed pump performance curve
for pump stations that have varying flow or product density requirements such as on
batched product pipelines. Despite their control systems being more complex than for
a constant speed motor, variable speed motors are much more energy efficient. This
is because pump capacity can be controlled without the disadvantage of pressure loss
incurred by the throttling through a discharge control valve.
Variable speed pumps control the flow and pressures by varying the speed of the
drivers with maximum power override. For a pump station that contains both fixed-speed
and variable speed motors, the control strategy is to run the fixed speed units at a base
load with minimal throttling and use the variable speed units to adjust for the required
station set point. Figure 4-39 exhibits the performance curves of a variable speed pump.
In applications that require flow or pressure control, the most energy efficient
option is an electronic VSD, referred to as a Variable Frequency Drive (VFD). The
most common form of VFD is the voltage-source, pulse-width modulated frequency
converter [12]. The converter develops a voltage directly proportional to the frequency
which produces a constant magnetic flux in the motor. This type of speed control can
be driven by set points of discharge pressure or flow rate.
As energy cost increases, VFDs and thus variable speed pumps are becoming
more cost-effective. This is because they allow pressure and flow control without wast-
ing energy incurred by throttling a control valve to control the discharge pressure of
the pipeline. Even though variable frequency drives are more expensive, they have the
advantage of reduced energy consumption and are efficient over a wide range of flow.
Energy savings of between 30% and 50% have been achieved in many installa-
tions by installing VFDs [13].
The following simplified graph, Figure 4-40 shows the significant reduction of
power required by variable speed pumps as compared to fixed speed pumps through a
range of flows.

Figure 4-40. Power requirements — constant speed drive vs. variable speed drive
Figure 4-41. Constant vs. variable speed power requirements

Figure 4-41 below shows a pump operating at the operating point 1X, where the
flow rate and head are Q1X and H1X, respectively. Here, for illustrative purposes, it can
be assumed that the operating point 1X is the best efficiency point (BEP), where the
efficiency of the pump is highest. Then, the power required by the pump is PW 1X =
Q1X ´ P1X/(dens ´ ηhBEP), where ηhBEP is the pump efficiency at the BEP.
To achieve a lower flow rate QY, the control valve is partially closed for the fixed
speed pump or the pump speed is reduced for the variable speed pump. These are il-
lustrated in Figure 4-41. 1Y is the operating point of the fixed speed pump and 2Y is
that of the variable speed pump for the lower flow rate. At 1Y, the flow rate is reduced
to Q1Y but the pump head increases to H1Y for the fixed speed pump where the pump
efficiency, h1Y, is considerably lower than the pump efficiency at BEP. In the case of
the variable speed pump, the operating point 2Y for the same flow rate generates only
the head required for the system curve, H2Y and the pump efficiency, h2Y, is only mar-
ginally lower than the best efficiency point.
In summary, energy requirements are directly proportional to head generated and
are factored by changes in pump efficiency. As demonstrated by Figure 4-41 above, there
is considerable energy savings by use of variable speed drives. The above example also
does not take into account the continuous pressure loss that occurs in a fully open con-
trol valve and therefore is conservative in its energy savings by use of variable speed
motors and elimination of the control valve.
See Section 4.10.6.2 for further discussion on the energy savings of variable speed
pumps.
4.9 PUMP STATION DESIGN
As described in Section 3.4, intermediate pump station locations on a pipeline system

are determined by the hydraulic design of the transmission pipeline and the frictional
pressure losses that occur with the product proposed for transportation. Adjustments
to station location from its ideal hydraulic location are often required because of such
Figure 4-42. Typical pump station diagram

factors as land access issues or topography constraints. Final locations for intermedi-
ate pump stations may result in some slight reduction in pipeline capacity because of
some hydraulic imbalance in the final station sites. This section addresses several key
equipment and features required for designing a pump station.
4.9.1 Pump Station Diagram

A simple pump station diagram is shown in Figure 4-42. In this diagram, the main com-
ponents include two pump units, station piping connecting various components, station
isolation valves, bypass valves, pressure control valve, and check valves. In addition,
many intermediate pump stations are installed with a pig receiver and launcher set.
Pump stations pumping heavy crudes may be equipped with heaters.
Figure 4-42 shows a typical pump station, which is composed of the following
equipment and instruments:
·· Station isolation valves are MOV-103 and MOV-104. These valves are used to
isolate the pump station for safety, maintenance, or other operating purposes.
·· Check valves are shown as CKV-101, CKV-102, and CKV-103.
·· Bypass valves include the station bypass valve along the main line, MOV-103
on the suction side, and MOV-104 on the discharge end. The station bypass
valve is open and the other two valves are closed when the pump station is shut
down, so that the fluid flows through the station bypass valve.
·· Instrumentation for pressure measurement (Ps, Pc, and Pd) is essential for sta-
tion control. Certain pump stations are equipped with a flow meter for control
purposes, not for custody transfer.
·· A pressure control valve (PCV-101) is installed on the station discharge piping
to control the discharge pressure and flow rate.
Pumps are driven by pump drivers which can be either fixed speed type or variable
speed type. In some cases, combined fixed and variable speed drivers have been used
to take advantage of low cost fixed speed drivers for base load and of low energy cost
variable speed drivers for extra load.
4.9.2 Pump Station Piping

Station suction piping design is important to ensure that NPSHA be maintained above
the NPSHR for the pump units. Piping must be designed with sufficient pressure to
withstand potential surge pressure changes during pump start-up and shut-down opera-
tions. Piping sizing should be determined based on minimizing pressure losses to pump
suction. As a rule of thumb, flow velocity on the suction side should be in the 1.5 m/sec
to 2.5 m/sec range. Higher velocities increase the frictional pressure loss and potential
surge pressure. Due to potential surge when a pump shuts down, the suction pressure
can suddenly increase significantly. Therefore, the suction line pipe must be able to
withstand the increased pressure.
Since centrifugal pumps generate a performance curve that rises as the flow de-
creases, the discharge pressure should be determined at shut-off.
If the liquid is fed from a tank, the amount of entrained vapor or air must be kept to
an absolute minimum. Entrained vapor causes not only vibration and possibly cavita-
tion but also reduced capacity and efficiency. Normally, a booster pump is installed to
avoid this condition.
Station piping in pipeline systems that are intended for batched products should
be designed to minimize any areas of potential product contamination. All fittings

should be close-coupled and piping should be designed to eliminate any connections

that could include isolated pockets of product that would contaminate a batch of dif-
ferent products.
For pipelines transporting batches of different product, pump units that are not
in operation should have their suction and discharge valves left in the open position
so that all products flowing through the station piping are continuously being flushed.
When there is a need to put the pump unit into operation, the control logic for the unit
start should first close the discharge valve before the pump unit is to be started. Once
the pump is on-line, the discharge valve is opened and the flow through the pump is
re-established with its added head.
In addition, a control valve is required for fixed speed pumps and some variable
speed pumps. Some initiating, intermediate pump and pressure reducing stations are
equipped with a pig launcher and receiver. In certain situations, heaters are installed
to heat heavy crude oil. In cases where the flow rate through a pump can drop below
the minimum continuous flow limit, recirculation piping and valves may need to be
installed (See Section 4.9.4 Station Flow Recirculation).
4.9.3 Control Valve and Sizing

The proper selection of a valve for petroleum liquid pipelines depends on factors such
as liquid properties, the system curve, the pressure drops to be controlled, and cost. The
liquid properties to be specified are its phase (whether or not it contains solids or vapor)
and its corrosion/erosion property. Temperature may not be a critical factor for most
liquid pipeline applications. The most suitable type of valve required for throttling the
flow or controlling the pressure includes globe, control ball, butterfly, and rotary plug
valves.
Valve size is obtained from a basic liquid sizing equation, which can be written
as follows:
DP
Q = Cv
g
where
Q = flow rate
Cv = v alve sizing coefficient determined experimentally for each type and size of
valve
DP = pressure differential
g = specific gravity of liquid
To calculate the expected Cv for a valve controlling liquid flows, the above equation is
re-arranged for Cv in terms of the flow rate and pressure differential.
g
Cv = Q
DP
We can obtain the valve’s Cv requirements by inputting into this equation the minimum
and maximum flow rates together with valve upstream and downstream pressures. This
value is used to select from the manufacturer’s data the size of the valve required. The
valve size should be compared to the pipe size in which the valve is to be installed. As
a rule of thumb, the control valve should not be smaller than two nominal pipe sizes
below the nominal pipe size.

4.9.4 Station Flow Recirculation

Pumps can be operated at or near zero flow for a short time at start-up without adverse con-
sequences to the pumping system. To avoid recirculation problems, at least 20% of BEP
flow is required for small pumps, and 50% or even 60% of BEP flow for large pumps.
However, if centrifugal pumps operate at low or no flow conditions frequently and
the duration of such operation is long, most pumps will need to be provided with a flow
recirculation system to protect them from potential damage or unstable flow conditions
in a low flow operation. The following conditions can develop at low flow rates:
·· Temperature in the pump rises significantly due to low pump efficiency;

·· Unstable flow conditions occur, resulting in surging pulsations and pipe vibration.
Figure 4-43. Pump station with a recirculation system

If the low flow conditions persist for a prolonged period or occur frequently, pumps
and other equipments can be damaged. If a pump station has to pump low flows fre-
quently, there are several options to address this problem:
·· Install two or more small pumps in parallel;

·· Install variable speed pumps;
·· Install a recirculation system with flow measurement, recirculation valve and
piping.
The overall economics will dictate the choice of these options.

Several alternative recirculation systems are available. Figure 4-43 shows one
such alternative recirculation system, which includes a flow meter, recirculation con-
trol valve and piping. When the flow rate gets closer to a set minimum flow, a bypass
control valve is activated and the flow downstream of the pump is recycled back to the
pump suction through the recirculation piping. As shown in this figure, the recircula-
tion system consists of a flow sensing device, a minimum flow control valve, a check
valve, and a bypass valve. Note that a pressure safety valve (PSV) is installed down-
stream of each pump to prevent the discharge piping from over-pressuring.
Recycling flow rate is controlled by a control valve with the flow meter located in
the suction line. As the flow rate decreases, the flow meter sends a signal to open the
recirculation control valve to keep the combined flow rate at or above the minimum
required flow rate. In general, this system works well in maintaining the minimum flow
rate required by the pump. Note that the take-off for the recirculation line is upstream
of the check valve.
4.9.5 Pig Launcher and Receiver

Pipeline pigs are extensively used in petroleum pipelines. They are intended to clean
and/or inspect the inside of pipelines; or, in some cases, to separate batch interfaces
in a multiproduct pipeline to reduce interface mixing. So-called smart pigs can detect
pipe corrosion by measuring pipe wall thickness and cracks in the pipe wall by means
of ultrasonic responses.
The pig is inserted into the pipeline through a pig launcher. After a pig is
loaded into the pig launcher, the MOV-107 valve connected to the pig launcher
(refer to Figure 4-42) is opened. When MOV-108 is closed, the flow pushes the pig
out of the launcher and the pig travels along the pipeline until it reaches the next
pig receiving station or pig trap where the pig is retrieved. The valves along the
Figure 4-44. Pig movement

pipeline from the launcher to the trap must be full bore valves to allow the pig to
pass through them.
Pigging operations are not usually automated because they do not take place
frequently. They seldom interrupt normal transportation, but some liquids can be
spilled into the sump tank when the pig is retrieved from the pig trap. When a pig is
launched, a launch signal can be generated to track its movement until it is received
at the designated pig trap.
Pigs usually do not travel at the same velocity as the liquid flow in the pipeline,
slipping gradually as the liquid passes by the pig. Some SCADA systems estimate the
location and the estimated time arrival to the receiving trap based on the pig slippage
factor and current pipeline conditions such as flow velocity. The pig tracking function
is useful for the pig retrieving schedule. Figure 4-44 shows a pig movement with pig
slippage.
When a pig arrives at the upstream location of the pig receiver, MOV-106 is
opened to equalize pressure across MOV-101. MOV-101 is then opened and MOV-
105 is closed and the pig moves into the trap. Once the pig arrival has been con-
firmed, MOV-105 is re-opened and MOV-101 and MOV-106 are closed. The trap is
then drained into sump and the pig is removed.
Where batching pigs are utilized, there is a trend toward fully automating
the pig operations. Full automation can be justified because the batching opera-
tions are required frequently and manpower involvement and operating costs are
reduced.
4.9.6 Pump Station at a Tank Farm

A tank farm includes not only multiple tanks and manifolds but also booster pumps
and a meter station with a prover. If it is the initiating lifting point, mainline pumps are
installed there. Figure 4-45 shows the tanks, booster pumps, meter station and mainline
pump station. Normally, a pig launcher is also installed.
The operation of a tank farm and booster pump start-up are relatively simple, un-
less the tank farm is very large and complicated. Assuming that the pipeline is initially
shut-in, the tank farm control system opens the valve connected to the tank containing
the desired batch. The density of the batch is measured to confirm that it is the cor-
rect batch. After the confirmation, the suction valve and then the discharge valve are
opened to prime the pump. When the pump is primed, it starts rotating and the batch
Figure 4-45. Tank farm booster and mainline pumps

Figure 4-46. Side stream injection
starts flowing. The discharge pressure of the booster pump should be higher than the
suction pressure of the mainline pump(s), Ps, which should be greater than or equal to
its NPSHR.
A side stream injection takes place on the suction side of an intermediate pump
station or in the middle of the pipeline section. Usually, a booster pump and flow
measurement system including density, temperature, and pressure measurement are
installed in the side stream injection location, but it may not be equipped with a pig
launcher. The figure below illustrates a simple side stream injection facility, injecting
into the suction piping of an intermediate pump station (Figure 4-46).
4.9.7 Pump Station Heater

As discussed in Chapter 3, heavy oil is often heated to reduce its viscosity and lower fric-
tional pressure drop. If the liquid needs to be heated, heaters are installed at the lifting tank
farm and at certain intermediate pump stations along the pipeline. They are installed down-
stream of the tanks and booster pumps, but on the suction side of the mainline pumps.
Figure 4-47. Heater, pump and pig launcher/receiver

Figure 4-47 shows a SCADA display including the heating units, pumping units
and pig launcher/receiver arrangement at an intermediate pump station. These units
can be controlled from this screen. Under certain operating conditions, the heater op-
erations may not be required, particularly at intermediate stations. The operating condi-
tions include high ambient temperature and/or low flow rate:
·· If the ambient temperature is high, the fluid temperature may remain high
enough to keep the viscosity low.
·· The liquid may be able to flow within the pipeline pressure limits without fur-
ther reducing the viscosity at low flow rates, because the frictional pressure
drop is low at the low flow rate.
Normally, heater operations take place before pump operations. In other words, crude
oil is heated to increase the pumping efficiency before pumping into the pipeline starts,
while heaters are shut down before the pumps are shut down to save the heating cost.
4.10 PIPELINE SYSTEM CONTROL
Pipeline system control requires the selection of a control strategy. The strategy de-
pends on the type of liquid, type of prime mover (fixed vs. variable speed), type of
controlling station (meter station, pump station, backpressure controller, etc.), and its
location and the pipeline system (delivery junction, steep terrain, permafrost zone,
etc.). The variables for controlling pump stations are pressures, flow, and possibly tem-
perature. Temperature control is not required for most liquid pipelines except when the
fluid is heavy crude with high viscosity or when the pipeline runs along a permafrost
zone. For batch operation, product density and the interface is a control variable.
For a pump with a constant speed prime mover, the pump head is fixed for a given
flow and thus the discharge pressure can be reduced by throttling the flowing liquid.
The throttling action is performed by a pressure control valve, installed downstream
of the pumps. The pressures discharged from pumps with variable speed drivers are
controlled by the speed of the drivers with maximum power override. A pump does not
require flow control as long as the flow is within the pump’s capacity.
Side stream delivery may disrupt the main line pressure. To avoid potential pres-
sure disruptions, the main line pressure is controlled by holding the delivery pressure.
If a liquid pipeline runs in terrain with a steep elevation drop, the pressure around
the peak elevation point will drop below the vaporization point, creating a slack flow
condition, unless a high backpressure or high pressure downstream of the peak point is
allowed. A backpressure controller, restricting flow, is needed at a location downstream
from the peak point to avoid this condition.
Pipeline system control is accomplished by means of a set point mechanism.
In other words, the dispatcher sets pressure, flow or temperature at the desired level
and the control system responds to reach the set point. Since pressure is the primary
control variable, several pressure set points are discussed below. The controlling
pressures, that can be monitored and changed by the dispatchers through the SCADA
system, are:
·· Suction set point: the desired suction pressure at the station. During normal
operation, the suction pressure is equal to or higher than the suction set point.
The control system will not function properly if the suction pressure is less than
the set point, unless the pressure measurement is erroneous. For liquid pipe-
lines, suction pressure control with discharge pressure override is commonly

used to maintain the pressure above the vapor pressure and at the same time
keep the pressure below the maximum allowable operating pressure (MAOP).
Normally, the minimum suction set point is higher than the station trip pressure
below which the station automatically shuts down.
·· Discharge set point: the desired discharge pressure at the station. The dis-
charge set point is the pressure that the station control system tries to maintain
as a maximum value. No control action takes place if the discharge pressure
is below the discharge set point. For a pump with a constant speed driver, a
control valve is used to control the discharge pressure. The discharge pressure
is equal to or lower than the pump casing pressure, and the difference between
the casing and discharge pressures is called the throttle pressure. Normally,
the maximum discharge set point at a station is lower, say by 200 kPa,
than the maximum operating pressure, in order to avoid an accidental station
shutdown.
·· Holding pressure: the holding pressure is set to maintain a desired main line
pressure at the junction where a side-stream delivery may take place or where
there is a delivery point but no pump station. The holding pressure for the lat-
ter case is called the delivery pressure, while for the former case the pressure
measurement device is installed on the suction side of a station where delivery
takes place.
4.10.1 Pump Station Operation

This section presents an overview of the key aspects of the pump station control.
The following basic control operations are involved in controlling a pump station
properly:
·· Control of pumping units including their driver,

·· Control of station valves and possibly line valves,
·· Alarm annunciation to correct an operation for limit violation,
·· Emergency shutdown to prevent possible damage to facilities.
The first two control operations are usually interlocked for starting and stopping opera-
tions of stations and/or pipeline system.
Nowadays, pipeline system control is automated to provide the capabilities of operat-
ing pipeline systems reliably, efficiently and thus economically. Automated system control
enables monitoring and control of pipelines, pump stations, metering stations, and other
facilities through a SCADA system. In addition, it can be extended to other applications such
as storage management, energy optimization, volume accounting, leak detection, etc.
It is now generally an accepted practice that stations are automated and operated
under remote control from a central SCADA control center. This is possible because a
centralized system provides the capability to monitor the complete pipeline system and
control the entire pipeline system in a safe and efficient manner. Only under abnormal
conditions or during some maintenance tasks will the station be under local control.
Some stations may be completely unmanned whereas others will have maintenance
staff on site but who will not normally be in control of the station equipment. Figure
4-48 shows a typical implementation of a pipeline system and station control.
The key criteria of deciding if the station should be unmanned are that the sta-
tion can be operated reliably and robustly and the control system has a high level of
availability. In addition, transfer of control from “remote” to “local” must be easily
supported in the event of an abnormal situation.

The following factors are usually taken into account in deciding if the station is
to be unmanned:
·· Safety can be enhanced by minimizing potential human errors;

·· Reliable and quick control response can be achieved;
·· Stations can be more efficiently operated remotely to reduce the operation cost.
Overall, the costs of installing the automation system and of maintaining the unmanned
station must be compared against the costs of manning the station.
A typical station control system consists of several components; station control,
unit control, driver control, and other auxiliary unit control. A station can be operated
locally as well as interfaced to a SCADA system to enable remote control from a cen-
tral control center. References [14, 15] discuss the control in more detail.
A properly designed remote control system will provide the ability to:
·· Monitor all equipment associated with the station including station auxiliary
systems;
·· Provide two-way communication between the station and the host;
·· Monitor the starting and stopping sequence of the drivers and pump units;
·· Control and monitor sequencing of station valves;
·· Initiate an emergency shutdown of the station or unit.
These extra system control capabilities can meet the following objectives for station
operation:
·· Operate the station safely and reliably, while maintaining cost efficiency;
·· Allow constant monitoring of critical components of the station;
·· Shorten response time to potential problems;
·· Eliminate mundane tasks for the station operators.
Figure 4-48. Implementation of station control system

Unplanned outages can cost a pipeline company tens of thousands of dollars. Therefore,
the station control system must be reliable, robust and have a high level of availability in
order to minimize business interruptions and maintain a safe environment for personnel
and equipment. It must also be able to transfer control from “remote” to local in the event
of an emergency or an abnormal situation. From the perspective of the overall operation of
a pipeline, a pump station can be viewed as a “black box” that maintains product flow by
offsetting pressure losses in the pipeline. The pipeline operator may only be interested in set-
ting pressures at the various stations and not be concerned with the control of the individual
units. In this situation, the station control system would receive station set points rather than
individual unit set points from the SCADA system. It would then determine how many units
should be operating and the set points for each unit.
An alternate control scheme is to include the station control system within the
SCADA system. The system operator would then be initiating start/stop commands
and relaying them to individual units as well as sending them the required set points.
Through the SCADA system, the following data required for monitoring and control-
ling pump units are displayed:
·· Pump unit status

·· Flow rate
·· Product name and density or API gravity
·· Suction pressure and its set point
·· Discharge pressure and its set point
·· Throttle pressure or difference between the pump casing pressure and the sta-
tion discharge pressure
·· Holding pressure
·· Station electrical load
·· Communication status
·· Alarms
In practice, the local station control system is implemented with a programmable logic con-
troller (PLC). PLCs are now the heart of station control for station equipment (pumps, driv-
ers, lube oil systems). PLCs consist of a programmable microprocessor unit, communication
modules, and input/output modules for connection to field devices. The PLC has overall
control capability for the station. This includes all equipment not under the direct control of
a unit control system. It ensures that the station operates within the parameters for the station
and mainline piping (above minimum inlet pressure, below maximum allowable operating
pressure, etc.). In addition, the PLC determines the required set points for the operating units
based on the required station set points received from the pipeline operator via the SCADA
system. The individual set points sent to each unit will be determined based on a load shar-
ing strategy. This will vary depending on the type of units installed and the overall pipeline
operating strategy. They may include such strategies as:
·· Base loading: one or more units may be operated at a constant load while other
more efficient units are used to compensate for small changes.
·· Optimum load sharing: set points for each unit are determined based on know-
ing the individual unit operating curves and allocating load to minimize overall
energy consumption.
With the increase in computing capability, it is now more common for pipeline com-
panies wanting to optimize their pipeline operations to consider having a system opti-
mizer that would optimize pumping usage on the entire pipeline. This is discussed in
more detail in Reference [14].

4.10.2 Pump Control Strategy

An optimum design of a station control system includes factors such as pump costs,
spare parts requirement, load, operation and maintenance, and flexibility of operation.
The pump control strategy must incorporate the following criteria:
·· Pump suction pressure must be above the minimum Net Positive Suction Head
(NPSH) for the pump in order to prevent cavitation of the pump.
·· Pump discharge pressure must be below the maximum allowable operating
pressure (MAOP) of the station discharge piping to avoid pipe and associated
equipment damage.
·· Station discharge pressure must be below the MAOP for the pipeline to avoid
damage and to ensure the pipeline is operating within the acceptable limits ap-
proved by the regulatory agency.
·· Station suction pressure must be above the minimum allowable operating
pressure to meet contractual requirements and in the case of liquid lines
to avoid “slack line flow” or column separation upstream of the pump
station.
·· Driver power must be kept within acceptable limits to avoid tripping of the
driver.
A pump driven by a constant speed electric motor driver requires a discharge control
valve to control pump throughput; the system controlling this valve must have a station
suction pressure (or station discharge pressure) control loop. Set points for maximum
station discharge pressure, minimum station suction pressure, and maximum motor
power are set on the controller. The controller will satisfy the set point for station
discharge until the suction pressure or driver power limits are reached then these will
override the discharge pressure set point.
Pressure switches are set to provide a trip signal in the event of controller failure.
The final backup is a pressure relief valve in the event of a complete control system
failure.
For a pump station, that contains both fixed-speed and variable speed motors, the
control strategy is to run the fixed-speed units at a base load with minimal throttling
and utilize the variable speed unit(s) to adjust for the required station set points.
Generally, there are three major levels of monitoring and station control in the
hierarchy of automated pipeline stations, namely:
Local: In this mode, command control of all local devices and

skids is passed to the station control system. This allows a
local operator to control the complete station and all auxil-
iary equipment from a single location at the station. In this
mode, command control is limited to the local device or the
skid control panel.
Remote: If the station control system is in “remote” mode, then in
effect all control is from the SCADA control center.
Control center/SCADA: In this mode, command control of the station is passed to
the central control center via the SCADA system. No local
control is possible. This is essentially the “remote” mode
for the station control system. Process values and status
may still be sent to SCADA for monitoring and logging
purposes.

It is important to realize that the control levels described here affect the state of op-
erator control. In all modes, the local device is always being controlled by its control
equipment. The change of mode describes from where, for example, set points or com-
mands to the local controller will originate.
4.10.3 Station Control

An automatic station controller has the ability to start, stop, and adjust units based on
the corresponding command sent by the pipeline system operator usually at the control
center or an operator occasionally at a station if communication is lost. In addition,
it can control pressure or load sharing as flow rate changes, based on the set points
entered by the operator. Proper control logic including operation sequence and time is
essential to control stations and to reduce pipeline transients. In general, a care must be
exercised when a valve is closed. If the valve is closed too rapidly, large pressure surge
can be generated and the surge can damage the pump and other pipeline components.
Therefore, the surge effect should be minimized by closing the valve slowly.
4.10.3.1 Pump Station Valve Control
One of the key aspects of the pump station control system is the station valve control.
Each valve in the station is controlled remotely, but if necessary can be controlled
locally by a field operator. The valve control logic incorporates interlocks with motor-
operated valves to ensure proper sequencing and to avoid damage to equipment. Some
sequencing scenarios that the control system contains include:
·· Station start up and shutdown;

·· Scraper launching and receiving;
·· Station by-passing;
·· Batch receiving and batch launching.
In addition, there may be some control logic required to help minimize or reduce pipe-
line surges (transients) depending on the results of the pipeline hydraulic studies.
The station controller directs the sequences of station valves by changing their
position for the following station operations:
·· Open and/or close valves in pre-arranged sequences for starting a pump station
or bringing additional units on line;
·· Close and/or open valves in pre-arranged sequences for shutting down a pump
station, by-passing a station, or bringing additional units off line;
·· Close and/or open valves in pre-arranged sequences for launching and receiv-
ing a pig or scraper;
·· Partially close or open a control valve to adjusting the station discharge pres-
sure for fixed speed pumps. Control valve operation may not be needed for a
variable speed pump, unless the station is installed with both types of pumps
(Figure 4-49).
Described below are the station start-up and shut-down sequences (refer to Figure 4-51):
·· Start the station from an initially shut-in state.

·· Assuming the station isolation valves, MOV-101 and MOV-102, are closed,
they are opened and Pump 1 is primed.
·· The pump starts rotating; the suction and discharge valves are opened as the
pump casing pressure increases, and the liquid starts flowing.

Figure 4-49. Pump station operation
If the flow rate needs to be increased, the same sequence is followed for Pump 2.
·· MOV-101 starts closing.

·· MOV-103 and MOV-104 start closing, as the pump slows down or shuts down.
·· MOV-102 is closed and the station is shut-down.
4.10.4 Injection/Delivery Station Control

Injection of products from and delivery to tank farms or other pipelines requires valve
operation to properly control pressure level and flow direction. The pressure control
is required to avoid damage to equipment, while flow control is needed to deliver
products to their correct destinations. Valve sequencing operation is also necessary for
launching and receiving batches.
4.10.5 Pump Unit Control

Pump units are controlled in one of two ways: discharge pressure control by throttling
a control valve and pump speed control. Throttling is a common way of controlling the
discharge going through a centrifugal pump and the control valve is installed on the dis-
charge side of the pump. Another way to control the discharge flow is to control the speed
of the pump, which in turn is controlled by the speed of the pump driver. Unlike valve
closing, reducing the flow by reducing the pump speed does not waste energy. This con-
trolling action results in changes in flow through the pump and pressure in the pipeline.
In operating a pump unit, several problems occur: the violation of the required min-
imum flow and the presence of vapour in the pump. The minimum flow requirements of
the pump must be carefully taken into account during the design and operation of pumping
units. If the flow is slow, energy is converted to heat due to low pumping efficiency and the
heat cannot be carried away quickly. The liquid in the pump will heat and eventually vapor-
ize. Typically, the pump manufacturer will place a minimum flow requirement of about 40%
of design flow for pumps associated with the pipeline industry. For most of the time, this
does not limit operations but care must be taken during the start-up of the line.

Typically, at the inlet station, some method of recirculation is provided so that the inlet
pumps can be brought on-line safely. When the pipeline is running at a large base flow, this
recirculation valve may be manual, but for lines where the product is stopped and started, it
may be controlled automatically by the unit control logic so that it is open until the minimum
flow requirement down the line has been established. The goal is to have the pump operate at
or near the most efficient point-labeled Best Efficiency Point (BEP) in Figure 4-50.
The pump may be allowed to continue to run for a short time (at most a few minutes)
after the valve is closed and thus the discharge flow is zero. If the valve is closed or even
slightly opened, but the pump keeps running, energy is wasted, resulting in an overheated
and highly pressurized pump and subsequently shortening the life of the pump.
Another problem is the presence of vapor in a pump. If the pressure drops below the
vapor pressure of a liquid at the pump suction, the liquid vaporizes and the vaporized
liquid forms bubbles. These bubbles move with the flow into the pump impeller and
volute where the pressure of the liquid increases sharply. Then, the bubbles collapse in
the high pressure area and the collapsing bubbles can generate localized high-pressure,
causing damage in the form of surface pitting. The problem is normally avoided by
increasing the pressure of the flow on the suction side of the pump — making the avail-
able NPSH higher than the required NPSH.
However, when a pump is shut down, vapor can fill in the pump unit and station piping.
If the pump is started under such a condition, the pump impeller will be spinning without
liquid flowing through the pump, and thus the liquid cannot be drawn into the pump fully
and the flow is slow. As a result, the pump can be overheated if such an operation lasts a
long time. To prevent this from happening, the pump must be primed with liquid before
starting. After the priming is done, the flow is allowed to increase until it reaches the desired
flow level. If a control valve is installed for a fixed speed pump, the control valve should be
opened gradually after the flow starts flowing through the pump.
4.10.6 Throttling vs. Speed Controls

This section will consider pump station operation using both constant speed and variable
speed electrically-driven pumps. Fixed-speed electric motors provide a cost-effective
Figure 4-50. Best efficiency point

s olution for base load applications where electrical power is available and reliable. They
have the advantage of low maintenance costs and are simple to operate. Variable speed drive
(VSD) motors are becoming the standard for pump stations that have varying flow or prod-
uct density requirements such as on batched product pipelines. Despite their control systems
being more complex than for a constant speed motor, variable speed motors are much more
energy efficient. This is because pump capacity can be controlled without the disadvantage
of pressure loss incurred by the throttling through a discharge control valve.
ers with maximum power override. For a pump station that contains both fixed-speed and
variable speed motors, the control strategy is to run the fixed speed units at a base load with
minimal throttling and use the variable speed units to adjust for the required station set point.
Figure 4-58 exhibits the performance curves of a variable speed pump.
In applications that require flow or pressure control, the most energy efficient op-
tion is an electronic VSD, referred to as a Variable Frequency Drive (VFD). The most
common form of VFD is the voltage-source, pulse-width modulated frequency con-
verter. The converter develops a voltage directly proportional to the frequency which
produces a constant magnetic flux in the motor. This type of speed control can be
related to set points of discharge pressure or flow.
4.10.6.1 Throttling for Fixed Speed Pumps
As discussed in Section 4.7.1, there is only one operating point for a fixed speed pump.
As shown in the figure below, a throttling action is required to match the system head
curve to the pumping head curve of a fixed speed pump at a particular flow rate other
than the design flow. Fundamentally, throttling changes the system curve. A control
valve is used to throttle the fluid flow, and is installed downstream of the pump.
Figure 4-51 shows that the pump operates at H1 for the design flow rate Q0. If a
throttle valve is partially closed in the pump discharge line, the throughput drops from
Q0 to QT, and additional friction pressure drop occurs through the partially closed
valve. As a result, the pump will operate at a new operating point, H2.
Figure 4-51. Capacity change with throttling

The throttle pressure is the difference between the case pressure and the dis-
charge pressure. The casing pressure of a pump is the available pressure generated by
a pump, and the discharge pressure is the pipeline pressure on the discharge side of
the pump station. The discharge is the pressure required to transport the liquid to the
next pump station or terminal. The throttle pressure is unused pressure developed by
the pump and thus results in wasted power. Figure 4-52 illustrates the energy losses
caused by throttling; the greater the throttle pressure, the greater the energy loss.
4.10.6.2 Speed Control for Variable Speed Pumps
Pipeline systems operate at flow rates different from the design conditions, because
supply or demand changes, liquid properties also can vary as in the case of batch opera-
tion, or other operating conditions. Such varying conditions demand flow control. De-
pending on the varying conditions, there are several ways of controlling flow rates:
·· Install a control valve at each pump station to throttle the flow rate;
·· Install multiple pumping units to provide sufficient discharge head that can be
matched to the flow requirement;
·· Install variable speed pumps.
ers with maximum power override. For a pump station that contains both fixed-speed
and variable speed motors, the control strategy is to run the fixed speed units at a base
load with minimal throttling and use the variable speed units to adjust for the required
station set point.
Even though variable speed pumps are more expensive, it is advantageous to in-
stall and operate variable speed pumps because energy cost can be saved and it they
easily applicable to a wide range of flow changes.
Compared to fixed speed pumps, variable speed pumps can produce significant
energy or power savings as illustrated in Figure 4-52 [16]:
Figure 4.53 shows a pump operating at the operating point B, where the flow rate
and pressure are QH and PB, respectively. Here, for discussion purposes, it can be as-
sumed that the operating point is the best efficiency point (BEP), where the efficiency
Figure 4-52. Energy losses due to throttling

of the pump is highest. The power required by the pump for the high flow condition
is PWH = QH ´ PB/hBEP, where hBEP = 84%, the pump efficiency at the BEP, and is
represented by the area of rectangle, PB0QHB, in the figure.
To achieve the lower flow rate QL, the control valve is partially closed for the fixed
speed pump or the pump speed is lowered for the variable speed pump, as illustrated in
the figure; C is the operating point of the fixed speed pump and A that of the variable
speed pump for the lower flow rate. At C, the flow rate is reduced to QL but the pump
pressure is increased to PC for the fixed speed pump, and thus the power required by
the fixed speed pump is PWC = QL ´ PC/hC, where the pump efficiency, hC is lower
than the pump efficiency at BEP. On the other hand, the operating point of the vari-
able speed pump is A for the same flow rate. There, the power is PWA = QL * PA/hA,
where the pump efficiency, hA may be lower than the pump efficiency at BEP but will
be higher than hC. As shown by rectangles in the figure, PA0QLA, the power required
by the variable speed pump is lower than the power required by the fixed speed pump,
because the pressure requirement is lower and the pump efficiency is higher.
In summary, the fixed speed pumps waste energy by throttling the flow to achieve
a lower flow rate, because:
·· The pump operates at a reduced efficiency,

·· The pump is required to produce an increased pressure.
Energy savings can result from using variable speed pumps, and thus it is advantageous
to use them from the viewpoint of reducing the energy cost. In addition, variable speed
pumps offer the following advantages:
·· Pressure surge can be small, particularly during pump start-up and shut-down
operations, because changes in flow and pressure occur gradually.
·· They provide flexibility of controlling flow over a wide range.
Figure 4-53. Power required at operating points A, B, C

4.11 STATION ELECTRICAL CONTROL
A pump station using electric motor drivers requires a reliable source of electricity.
This may be supplied from a commercial source or generated at the station. Economic
and reliability considerations usually determine the choice of power source.
The electrical supply usually will have high voltage feeders, voltage reduction
equipment, and be a multi-bus operation with its associated transfer equipment. All the
bus and equipment protection required to support such a system is normally provided
with the electrical equipment. Controls for this equipment may be incorporated into
stand-alone control equipment or they may be part of the station control system.
The electrical protection is always contained in stand-alone, specialized equip-
ment package that will protect against:
·· Over and under voltage

·· Over and under frequency
·· Over current and short circuits
·· Ground fault
·· Voltage imbalance
·· Phase reversal
·· Transformer gas and high temperature
The electrical supply control system monitors the electrical system and sends the fol-
lowing information back to the station control system:
·· Voltage and current values

·· Real power, power factor
·· Electrical energy consumption
·· Circuit breaker and disconnect position
·· Frequency
4.11.1 Station Auxiliary Systems

The station control system controls and monitors the functioning of all station auxiliary
systems common to the operation of all units. These systems include some or all of the
following equipment, depending on the specific station requirements:
·· Auxiliary (emergency) electrical generator

·· DC Battery charger(s)
·· Inverter
·· Security system
·· Boiler (if required)
·· Air-conditioning
·· Commercial AC power monitor
·· Ground fault detection
·· Starting Air System(s) for a gas turbine driver
·· Fuel system (for non-electric drivers)
·· Vent fans and louvers
·· Inlet air filter system
·· Central lube oil conditioning (filter/cooling)
·· Fire and gas detection system
·· Station Emergency Shutdown Device (ESD)

Generally, the design of the station control system allows for the complete control of
the station to be from a local control room, with the option of passing control to the
pipeline controller via the SCADA system. This would allow the station to be operated
remotely and be unattended.
4.11.2 Shutdown Modes

A typical arrangement for station controls is to have different levels or severity of
shutdowns such as:
Normal Shutdown: This will shut down the equipment through a normal
shutdown sequence. The unit can be restarted normally.
This would be initiated by an operator command or may
be required if process conditions exceed limits. Once
process conditions have been restored, the unit can be
restarted.
Shutdown Lockout: This is activated to stop a unit due to a serious problem
such as loss of lube oil, etc. “Lockout” means the unit
cannot be restarted until manually reset locally. This
ensures that the site is visited by a technician/operator,
who must evaluate the situation before the unit can be re-
started. This can apply to individual units or the complete
station. Wherever possible, the shutdown will follow
normal shutdown procedures to minimize hydraulic
disturbances.
Emergency Shutdown This condition requires immediate shutdown of all units
(ESD): and will initiate a hydraulic isolation of the station. In a
natural gas pipeline, this will also result in the activation
of associated blow-down valves. Following an emergency
shutdown, all controls will be in a lockout state and require
local resetting.
4.11.2.1 Emergency Shutdown System

The purpose of an Emergency Shutdown System (ESD) is to provide a fail-safe in-
dependent control system that can shut down a station and isolate it in the event of a
pipeline rupture, station piping rupture or a fire at the station.
From a design perspective, ESD systems should be hardened against the explosive
forces and fire associated with this type of system failure. Indeed, to be fail-safe, the
ESD feature should be capable of automatically isolating the flow of product to an ac-
cident site until it has been verified that it is safe to reactivate normal operations. The
ESD system overrides any operating signals from the station or local controls and its
design therefore, needs to meet the requirements of both the regulatory regime and the
owner’s own design philosophies and criteria.
The ESD is the last line of defence to shut down a station and must be able to per-
form its function even if the station has lost normal power supply, has lost the ability
to communicate with SCADA or in the case of local control, system failure. Normal
designs of an Emergency Shutdown Controller (ESD) provide for them to be independ-
ent of the station controller itself. It should also be possible to test the ESD system on
a regular basis without interrupting normal operations. They will typically include re-
dundancy control capability to ensure that no single point of failure in the ESD system
will disable the capability to properly detect and execute an ESD action.

A station ESD system has associated shutdown valves to isolate the station. If ESD
valves close too quickly a pressure transient can be generated that could damage facili-
ties. Hydraulic studies are usually undertaken to determine ESD valve closure times in
order to limit pressure transients along the pipeline from the station.
4.12 APPLICABLE CODES AND STANDARDS
·· Hydraulic Institute Standards

·· American Petroleum Institute (API) 610 11th Edition, September 2010 — Cen-
trifugal pumps for petroleum, petrochemical and natural gas industries
·· ISO 13709:2009 — Centrifugal pumps for petroleum, petrochemical and nat-
ural gas industries (Identical to API 610)
·· API 614, Fourth Edition: General Purpose Lube Oil System components For
Rotating Process Equipment.
·· API RP 1110, 2007: Pressure Testing of Steel Pipelines for the Transportation
of Gas, Petroleum Gas, Hazardous Liquids, Highly Volatile Liquids or Carbon
Dioxide. This RP applies to all parts of a pipeline or pipeline facility including
line pipe, pump station piping, terminal piping, etc.
·· BS 5136: Acceptance Tests for Pumps for Centrifugal Mixed Flow and Axial
Pumps
·· ISO 13709:2003: Centrifugal pumps for petroleum, petrochemical and natural
gas industries, (Metric)
·· ISO∕TR 17766:2005: Centrifugal pumps handling viscous liquids — Perfor
mance corrections
·· ISO 21049:2004: Pumps — Shaft sealing systems for centrifugal and rotary
pumps
REFERENCES
[1] Lobanoff, V. R. R., 1992, Centrifugal Pumps: Design and Application, 2nd Edition, Butterworth,
Heinemann, 1992.
[2] Karassik, I., 1976, Pump Handbook, McGraw Hill, New York, NY.
[3] Mohitpour, M., Golshan, H., and Murray, M., 2007, Pipeline Design & Construction, A Practical
Approach, 3rd Ed., ASME Press, New York, NY.
[4] Chaurette, J., 2004, “Centrifugal Pump Specific Speed Primer and the Affinity Laws,” Pump-Flo,
pp. http://pump-flo.com/pump-library/pump-library-archive/jacques-chaurette/centrifugal-pump-
specific-speed-primer.aspx, November.
[5] Karassik, I., 1994, “Setting the Minimum Flows for Centrifugal Pumps,” Pumps and Systems
Magazine, March.
[6] Andrews, D., 2004, “Cavitation - Intelligent Maintenance of Pumps,” Run Times - Lawrence
Pumps Inc., October.
[7] A. D. B., 2007, “Discharge Recirculation and Cavitation,” Run Times - Lawrence Pumps,
April.
[8] Andrews, D., 2007, “Intelligent Maintenance Management of Pumps,” Run Times - Lawrence
Pumps Inc., April.
[9] Andrews, D. B., 2004, Cavitation, Run Times - Lawrence Pumps Inc., http://www.lawrencepumps.
com/Newsletter/news_v01_i5_oct.html, October.
[10] Lohrberg, H., Stoffel B., 2000, “Avoiding Cavitation Erosion,” Pump Users International Forum,
Karlsruhe, Germany.
[11] Andrews, D., 2007, Viscocity and Pump Performance, Run Times Vol 4 Lawrence Pumps Inc.,
February.

[12] Hydraulics Institute, 2004, Variable Speed Pumping - A Guide to Successful Applications - Executive
Summary, ISBN 1-85617-449-2.
[13] Hydraulics Institute, 2000, Pump Life Cycle Costs: A Guide to LCC Analysis for Pumping Systems
ISBN 1-880952-58-0.
[14] Yoon, M., Warren, C., Adam, S., 2007, Pipeline System Automation and Control, ASME Press,
New York, NY.
[15] Mohitpour, M., Van Hardeveld, T., Peterson, W., and Szabo, J., 2007, Pipeline Operation and
Maintenance — A Practical Approach, ASME Press, New York, NY.
[16] Lohrberg, H., Stoffel B., 2000 Intelligent Maintenance Management of Pumps; Pump Users
International Forum.

Pumps and Pump Stations

Uploaded by

Copyright:

Available Formats

Pumps and Pump Stations

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Pumps and Pump Stations

Uploaded by

Copyright:

Available Formats

CHAPTER 4

Pumps and Pump Stations

Figure 4-1. Pump suction and discharge systems

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

Table 4-1. Centrifugal pumps vs. positive displacement pumps

4.2 CENTRIFUGAL PUMPS

Centrifugal pumps have prevalent application in liquid hydrocarbon pipeline trans-

Figure 4-2. Centrifugal pump with impeller

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

4.3 CENTRIFUGAL PUMP TYPES

4.3.1 End Suction Single Stage Pumps

4.3.2 Vertical In-Line Single Stage Pumps

4.3.3 Horizontal Axially Split Between-Bearing Single-Stage Pumps

4.3.4 Horizontal Axially Split Between-Bearing Multi-Stage Pumps

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

4.3.5 Double–Case (Can) Vertically Suspended Volute Pumps

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

Figure 4-7. Vertical Can pump. Courtesy of Flowserve

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

4.4 PUMP SELECTION AND SIZING

·· Determine the required flow range of pump: It is important to define the

4.4.1 Pump Performance

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

Figure 4-8. Typical pump performance curve

4.4.1.1 Pump Performance Curves

·· Range of impeller diameters available

4.4.2 Service Conditions

·· Liquids to be pumped — clear liquids or liquids containing solids or vapor

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

·· Viscosity— pump performance drops more rapidly above 110 cSt

Pipeline system hydraulic requirements will determine the selection of pumps.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

Figure 4-9. Pump selection map

4.4.3 Net Positive Suction Head (NPSH)

4.4.3.1 Net Positive Suction Head Required (NPSHR)

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

4.4.3.2 Net Positive Suction Head Available (NPSHA)

NPSHA = ha - hVP + hst - hfs

Figure 4-10. Stable operating point

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

4.4.4 Specific Speed

Specific speed is expressed as:

To convert between US units (USgpm) and Metric units (m3/h)

Ns(US gpm, ft) = 0.861 Ns(m3/h, m)

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

Figure 4‑11. Impeller design vs. specific speed –— Hydraulic Institute

4.4.5 Suction Specific Speed

Figure 4-12. Pump efficiency vs. specific speed

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

4.4.6 Pump Performance Curve Characteristics

Figure 4-13. Pump performance curve characteristics

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

Rising characteristic — the head rises continuously as the capacity is decreased to

4.4.7 Centrifugal Pump Power and Efficiency

Power is more commonly expressed as kilowatts (kW) or horsepower

4.4.8 Performance Modifications For Varying Pipeline Applications

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

Figure 4-14. Changing performance by impeller vane number

Figure 4-15. Changing pump performance by impeller vane angle

4.3.3 Horizontal Axially Split Between-Bearing Single-Stage Pumps

Figure 4-21. Pump impellers destroyed by cavitation [10,8]

4.6 RETROFITTING CENTRIFUGAL PUMPS FOR CHANGING

4.7.2 Hydraulic Performance in Batched Pipeline Systems with

4.7.3 Hydraulic Performance in Batched Pipeline Systems with