Relating Centrifugal Pump Efficiency, Variable Speed Drives

and Hydraulic Improvements to Energy Dollars

Steven Bolles Process Energy Services, LLC

Don CasadaDiagnostic Solutions, LLC

Engineers, Operators, and Maintenance specialists all view their pump systems from different
perspectives. The Engineer has traditionally been focused on pump equipment materials and sizing
based
on the most severe conditions that could be encountered at a facility. The Operator is typically
interested
in equipment flexibility, capacity and reliability. The Maintenance specialist concentrates on servicing
the equipment, ensuring spare parts are readily available, proper installation and refurbishment, and
of
course pump system reliability. Even though many of us are focused on one segment of pump
The continuous pressure to reduce costs with less resources has forced many of us to learn more
systems,
about
we have all picked up ideas and practical knowledge from each of these areas.
pump system mechanical and electrical system efficiencies, and how to evaluate our pump systems
to
ensure we have done everything possible to reduce operational costs. The prime component of these
operational costs is often
This paper discusses the cost of energy.
the performance characteristics of the basic components of pumping systems. It
also relates some factors that affect the components individually as well as how the individual
component
efficiency is interrelated with the other components.
Evaluating a pump system to increase system efficiency and reduce energy costs, can be segmented
into
the following major categories:
Energy Costs and Utility Rate Structures
Hydraulic System Efficiency
Mechanical Efficiency
Electrical Efficiency

Before even beginning the process of evaluating each of the above categories, the most important
step is
often to get the “big picture” on your pumping operation. Some of these questions may include:
Is a plant expansion or major process change expected soon? If this change is three years down the
road
it may still be worth making operational changes that can be implemented in the short term.
Can the process be changed to eliminate the need for pumping or reduce the flow rate or head? (In
some
cases a piping change will allow gravity flow for a process where a pump was previously specified)
Do the system flow rate/head requirements change with time? Are the changes in flow requirements
continuously variable and spread over a relatively broad range, or at a few discrete flow rates?

It may seem like the above questions are obvious, However, it is surprising how often these issues are
overlooked. One approach that is often used in characterizing system requirements is the
development of
a flow rate requirement1 vs. time curve, sometimes called a flow duration curve. An example of
duration
curves for three different systems is shown in Figure 1. The total flow required (i.e., the area under
each
curve) is the same for the three systems, but the distribution is obviously quite different. This type of
curve is a good starting point for overall system analysis. Further consideration, such as the
relationship
between system flow and head requirements is needed to fully understand system requirements and
1
optimal design.
It is important to But
makeeven a casual review
the distinction betweenofsystem
the three duration and
requirements curves shown
actual in Figure
operation, in the1case
reveals that
of pre-
existing
the
systems unless there is confidence that the two quantities are the same.
kind of pumping configuration that is suitable for System A (a single-size pump, selected to operate
efficiently at the constant flow requirement) would be less than optimal for either System B or C.

System A
System B
System C
Flow rate

Hours
Figure 1. Example flow duration curves

Energy Costs and Utility Rate Structures

One of the first areas that deserves investigation before focusing on the efficiency details of a pump
system is to understand energy rate schedules and how the cost of energy is determined at a facility.

Although many industries produce their own energy at very reasonable rates, most facilities purchase
their energy from the local electric utilities. Typical utility energy costs include a consumption charge
(kWh) and a demand charge (kW or kVA), and in some cases, a power factor penalty. Energy
consumption charges range from 3 to 6 cents per kWh in the South and Midwest to 8 to 12 cents per
kWh
in the Northeast. Energy demand charges range from $4.00 up to $25.00 per kW of demand. For a
100-
hp water pumping station that operates continuously, a sample monthly energy cost calculation may
be as
follows: Element Monthly value Per unit cost Cost ($)
Service charge Fixed fee $35/month 35
Energy use 57,600 kWh $0.08/kWh 4,608
Peak demand 90 kW $10/kW 900
Total 5,543

Note that there are other costs that are sometimes involved, such as fuel charges and power factor
penalties that are not included here. As indicated in this example, demand cost can be 20% of the
total
energy bill and should always be considered when calculating savings.
An important consideration in energy cost reduction is the option of applying Time-of-Use Electric
Rates. Although the example in Table 1 shows a constant consumption rate of $0.080/kWh, some
electric
utilities reduce the cost of energy during off peak hours (typically evenings and weekends). Municipal
water system that have been designed with adequate storage and efficient use of “pressure zones”
can
often pump only at night to replenish water storage tanks and rely on water storage capacity and
elevation pressure to provide water service during the day. It is important to distinguish between
overall
energy consumption and energy costs here, since pumping at much higher capacity in off-peak hours
(and
consequently at lower capacity during on-peak hours) may actually require more energy consumption
than if the system flow rate was maintained relatively constant.
Hydraulic System Efficiency
An important aspect of evaluating pump systems is to understand how the system is operated. This
includes reviewing the system surrounding the pump rather than just the pump and driver. Some of
these
considerations include:
What flow rate and pressure are required for the process? Is more of either being provided than is
necessary?
How is the pump being controlled? Can level or pressure setpoints be adjusted to higher or lower
values?
Are there restrictions in the piping system (e.g., throttled valves, pressure reducing valves,
unneeded
or oversized check valves, excessive elbows, corrosion and scale build-up in the piping system)? In
a
similar vein, have the system requirements changed significantly so that the existing pipe velocity
is
well beyond what was originally anticipated?
Answers to the above
Are multiple pumpsquestions
operatingoften provide
in series the most
or parallel cost-effective
being improvements
operated efficiently that cansystems?
with control reduce
pumping energy costs by simple operational adjustments or piping improvements.

It is particularly important to first search for savings in the hydraulic system, since savings achieved
at
the system level are amplified at the pocketbook level. This is because the pump and motor are not
100%
efficient; so for every horsepower reduction in fluid power requirements, there is about a 1.4-hp
reduction in electrical power requirements, even for systems in which the pump and motor are
operating
It is useful to develop system head-capacity curve(s) for the most common system alignment
efficiently. This curve is useful for several reasons:
condition(s).

1) It, in conjunction with the pump head-capacity curve, defines where the system will operate. In
order
to gauge the effects of system or pump changes on overall performance, such a curve is essential.
2) It provides a baseline against which to compare future system performance. As the system ages
and
when components are replaced, the baseline curve can be used as a reference value from which
potential savings can be estimated.
3) It helps identify situations where variable speed drives may be most effective (systems that are
dominated by frictional head) or least effective (systems that are dominated by static head).
To help illustrate the merits of developing and trending system performance curves, an example
system
head-capacity curve (with no static head) is shown in Figure 2. The same curve, along with a second
curve representing the same system after scale buildup in the piping is shown in Figure 3. To maintain
a
constant flow rate of 2500 gpm, an increase of system head of about 31 feet must be overcome. The
increased head of 31 feet corresponds to an increase in hydraulic power (assumed fluid specific
gravity
Assuming = a constant combined pump and motor efficiency of 70%, the frictional losses would
1.0) of
translatealmost 20 hp, which is about 25% more than was required in the new, clean system.
into an electrical power requirement of about 21 kWe. If the system was operated continuously at this
condition, and the per unit energy cost of electricity was $0.08/kWh, the annual cost of the increased
friction would be almost $15,000.
Changes in the system curve affect the pump operating point (which in turn, affects the motor
operating
load), so the pump and motor efficiency would not, in a real-world application, remain fixed. But this
simplified example helps to illustrate the merits of understanding and trending system performance.
 200 200

31 ft at
2500 g pm
150 150
Sy s tem Sy s tem
head , ft head , ft

100 100

50
50

0
0
0 1000 2000 3000 4000
0 1000 2000 3000 4000
Flow rate, g pm Flow rate, g pm
Figure 2. Example system head-capacity curve Figure 3. Change in system curve with
increased friction

Pump Efficiency
After developing an understanding of system operation and the operational cost components of your
pump system, the mechanical/hydraulic efficiency of the pump should be investigated. One of the first
places to start is to obtain a copy of manufacturers pump curve. Possible sources of pump curves are:
1) a test facility certified performance curve for the specific pump in question,
2) an in-situ performance curve (if good quality instrumentation is used, this is a preferred method,
since it captures the actual motor and pump characteristics),
3) generic performance curve from the manufacturer’s catalog, or
4) software packages that include manufacturer curves.

The performance curves will provide a graphical understanding of the relationship of flow rate, head,
efficiency, and shaft input power of the pump. In the case where an in-situ performance curve is
generated, the electrical input power to the motor replaces the shaft power. An example set of
performance curves for a pump that might be used in the system shown in Figures 2 and 3 is shown
in
Figure 4.
200 100
H ead Efficiency
Head , ft
an d Sh aft
p ower, bh
150 75
p Efficien cy,
%

100 50

Brake hp
50 25

0 0

0 1000 2000 3000 4000 5000

Flow rate, g pm
Figure 4. Example pump performance curves

A couple of key questions to be answered when considering pump performance curves

are:
1) Is the pump operating reasonably close to its best efficiency point (BEP), at each required
operational
head and flow condition?
2) Has pump or system performance changed over time (for situations where a baseline set of pump
performance curves are developed and compared against subsequent performance data)?
Once both the pump and system performance curves are available, the system and pump curves can
be
jointly plotted to identify the operating point(s). This exercise is especially important in answering the
second question above.
Figure 5 shows the head-capacity curves for the system(s) and pump from Figures 3 and 4. The shift
in
flow rate and head that occur due to the increased frictional losses in the system are annotated.
Figure 6
demonstrates possible pump performance degradation in a fixed system. The key point to be made is
that
baseline measurements of both pump and system performance characteristics provide the reference
value
against
200 which subsequent measurements can be evaluated200 and thereby help in the identification of
corrective actions.
New s y s tem: New pu mp :
2500 gp m 2500 gp m,
150 125 ft 150
125 ft
A g ed s y s tem:
Head , ft Head , ft W o rn p u mp :
2300 gp m,
2420 gp m,
100 133 ft 100
117 ft

50 50

0 0

0 1000 2000 3000 4000 5000 0 1000 2000 3000 4000 5000
Flo w rate, g p m Flo w rate, g p m
Figure 5. Fixed pump with system Figure 6. Fixed system with pump
degradation degradation

In cases where multiple pumps are operating in parallel or in series, graph paper or a pump-modeling
program is useful for developing composite pump curves.

If the existing pump is operating well away from its BEP flow rate, there is prima facie evidence that
efficiency and operational cost reduction opportunities exist. The suitability of alternative courses of
action depends on a variety of factors, including:
1) the range of flow requirements, and whether the flow varies continuously or in discrete steps;
2) the distribution of the system static and frictional head components;
3) the distribution of time operating at the various flow rates;
4) whether there is evidence that the pump is degraded.

Electrical Efficiency
The electrical efficiency of a pump system is, in essence, the motor output mechanical shaft power
divided by the power measured at the utility meter. It includes the electric motor, switchgear, and
supply
leads, and where applicable, variable speed drive, filters, and transformers. Overall, the efficiency of
these electrical components is usually very high. While there may be opportunities to gain a few
efficiency points in the electrical components themselves, the primary improvements usually
achieved in
the electrical area are to improve the efficiency – or, perhaps more importantly reduce the power
requirements of either the pump or the fluid system.
Typical motor performance efficiency vs. load and power factor vs. load curves for two motor sizes are
shown in Figures 7 and 8. The curves shown are for four-pole, energy efficient ac-induction motors.

96 90

94 M o tor p
M o tor ower80
efficien factor, %
cy , %92
100-hp 70 100-h p
25-hp 25-hp
90

60
88

86 50

20 40 60 80 100 20 40 60 80 100
Load , % o f rated Load , % o f rated
Figure 7. Typical motor efficiencies vs. load Figure 8. Typical motor power factors vs. load

An important point to recognize from Figure 7 is that efficiency for induction motors is essentially
constant over the normal load range. It is only if the motor is extremely lightly loaded that sizing is an
issue. In fact, if a power factor penalty is part of the rate structure (either explicitly or implicitly, based
on historical experience), the low power factor at light load might be a more significant cost factor
than
the slight reduction in efficiency.
When an adjustable speed drive is used, the motor speed can be reduced to accommodate the
system flow
requirements with reduced energy consumption over these alternative means:
1) Not controlling flow at all that is, running the system at a higher flow rate than is necessary to
meet
the system requirements,
2) Controlling flow in a batch mode that is, starting and stopping the pump for limiting conditions,
such
as when filling or depleting a tank or reservoir,
3) Controlling flow by valve throttling, or
Of course, likeflow
4) Controlling any rate
otherbyactive component,
recirculating the drive
a portion of theefficiency
flow. is not 100%. Figure 9 shows motor,
drive, and combined efficiency for a modern pulse width modulated drive operated at rated speed
conditions (on a two-pole, 50-hp motor). The motor efficiency alone, when driven directly from the
power supply (i.e., without the drive) is also shown for comparison. As can be seen, the drive
efficiency
is in the upper 90-percent range. The drive also causes the motor to operate at a slightly lower
efficiency
than
Figure when the motor
10 shows is driven
the same motordirectly across
efficiency the as
curve line.
Figure 9, but for the combined drive and motor
curve, uses data that represents centrifugal loads (i.e., output power is proportional to the speed
cubed).
 100 100

M o t or direct
95 M o t or, 95
Driv e
direct on line
Efficien Efficien
cy , %90 cy , %90

85 85

Com bin ed drive

80 M o t or, Co m bined drive 80 an d m ot o r,
on driv e an d m ot o r cen t rifugal lo ad

75 75
20 40 60 80 100 20 40 60 80 100
Load , % o f rated Load , % o f rated
Figure 9. Motor and adjustable speed drive Figure 10. Motor (rated speed) and adjustable
speed drive (centrifugal-type loadefficiencies at rated speed
speeds) efficiencies

But simply considering efficiency alone doesn’t capture the essence of variable speed drives. The real
saving derives from the drop in power that accompanies speed reduction. For centrifugal loads, such
as
pumps, the shaft power is approximately proportional to the cube of the speed. Figure 11 contrasts
the
difference in shaft power between fixed and variable speed driven when the pump whose
performance
curves are shown in Figure 4 is used in the system whose head-capacity curve is shown in Figure 2.
The 100
shaft power for the variable speed case was calculated using the pump affinity laws (discussed
below).
80
Sh aft p
ower, bh p
60
Sh aft p ower with
fixed s peed mo to r
40

20 Sh aft p ower with

v ariab le s peed d riv e

0 500 1000 1500 2000 2500 3000

Flow rate, g pm
Figure 11. Shaft power requirements for a fixed and variable speed driven pump

Overall Pumping System Calculations and Field Measurement Considerations

Pumping system electric power and energy consumption can be calculated as
follows:
QH QHT
P= and E=
5310 p m s 5310 p m s
Component Abbreviation Units
Power P kW
Energy  kWh
System flow rate Q gpm
System head H ft
Fluid specific gravity Dimensionless
Run time  hours
Pump efficiency T Dimensionless fraction
Motor efficiency p Dimensionless fraction
Electrical supply efficiency m Dimensionless fraction
s

Since it is none of the efficiency values can be practically measured in the field, the efficiency terms
can
be combined to a single overall value, T, resulting in the following relationships:
QH QHT
P= and E=
5310 T 5310 T

Some notes regarding the field measurement of these

parameters:
1. In many cases (for example, where ambient temperature water is the pumped fluid), the specific
gravity can be assumed constant; but it is important to note that other situations, such as where
variable consistency slurries are being pumped, the specific gravity should be measured.
2. By definition, the head and flow rate for the system and the pump are equal, since the intersection
of
the pump and system head-capacity curves define the operating points for both. Measurement of
the
pump head requires that suction and discharge pressure measurements be made (and gages
referenced to a common elevation). True pump head measurement also requires an accounting of
differences in suction and discharge velocity heads, but as a general rule in field measurements, the
differential velocity head can be disregarded because it is such a small component of head.
3. The two parameters that are usually most readily measured are pump head and electrical input
power.
If an accurate pump performance curve is available, the head can be used to graphically determine
flow rate. It is important to emphasize the word accurate here, since it is not unusual for pump
performance to deviate from generic manufacturer curves; furthermore, it is all too commonly found
that the user doesn’t possess even a generic performance curve.
4. It is important to measure rotating speed (e.g., with a portable strobe light). Pump performance
curves (even certified curves) are usually performed at a particular speed. The test speed may be
one
or two percent different from the actual speed. This has a significant impact, and the performance
curve needs to be adjusted for actual speed conditions (see discussion on pump affinity laws below).
5. When the only two parameters that can be measured are head and motor power, and a generic
pump
performance curve is available, there is a great deal of value to using these two measurements to
do a
confirmatory cross check. For example, use the measured pump head and the head vs. capacity
curve
to estimate the flow rate. Next, estimate electrical train efficiency. Unless a variable speed drive is
being used, the motor is the only component that usually needs to be considered. On newer motors,
the motor nameplate will include a NEMA nominal and/or guaranteed motor efficiency. Multiply the
measured motor input power by the estimated motor efficiency to estimate shaft power. Then
compare this value to that derived for the estimated flow rate the power vs. capacity curve. If there
is
a significant difference, further consideration is warranted. Some possible reasons are:
a.Particularly if the measured power suggests that the pump is not as efficient as the curve
indicates (i.e., if the measured power is greater than the curve-based power at the estimated
flow
rate), it may be safe to assume that the pump is simply not operating at the manufacturer curve
 efficiency. This could be due to poor pipe layout (and resultant fluid geometry) in the field; it
could also be the result of pump wear.
b.If the reference curve is simply a generic performance curve for the particular pump model, it
is important to recognize that there may be significant variation in performance from pump to
pump.
c.The head or power measurements (or estimated motor efficiency) were erroneous. If
permanently installed pressure gages were used for the head determination, their accuracy
should
be verified. In the authors’ experience, field gages often go for years without being calibrated or
even checked. If at all possible, temporary test gages known to be in calibration should be used
if at all possible.

It is usually necessary to apply the pump affinity laws to account for differences in actual operating
speed
and the speed at which the pump curves were developed. The pump affinity laws are as follows:
2 3
Q = Q *() H = H *() P = P *()
N2 N2 N2
2 1 2 1 2 1

N1 N1 N1
where Q = flow rate, N = rotational speed, H = head, and P = power. The subscripts 1 and 2
represent
two different speeds.
Another form of the affinity laws relates to impeller diameter, D:
2 3
Q = Q *() H = H *() P = P *()
D2 D2 D2
2 1 2 1 2 1

D1 D1 D1
The subscripts 1 and 2 represent two different impeller diameters. The impeller diameter affinity
scaling
relationships have proven useful in some field-based measurement experiences in that the impeller
diameter used as the basis for the performance curves can be modified iteratively to a point where
the
curve-based flow
Opportunities estimates from
for Improving Pumpmeasured head and power are in agreement.
System Efficiency
After a preliminary review of the pump system as been performed, some quick calculations can be
performed to identify the magnitude of savings that each opportunity may provide prior to more
extensive field testing. Some examples of preliminary calculations may include:

Using the pump equation shown above and the manufacturers pump curve to identify the
impact
of lowering the TDH on a system by raising the level in the pump suction (e.g., raising the
average level of the wetwell in a waste water application), reducing pressure on the discharge
side of the pump, or decreasing the piping system head losses.
Using the affinity laws to determine how a reduction in flow by shaving the pump impeller size
or reducing the speed of a variable speed drive will effect energy use.

Performing some preliminary estimates of savings is useful prior to a detailed evaluation to

understand
what kind of testing should be performed.
As each pump system is reviewed, several potential improvements may provide savings in one area
and
reduce efficiency in other areas. One of the most common improvements that this occurs is when
variable
The primary
speed drives advantages
are applied.of AC variable speed drives (also called VSDs, VFDs or ASDs) is to control the
pump flow rate by adjusting pump motor speed. Figure 12 illustrates the change in pump head-
capacity
curves as motor speed is changed (using the affinity laws).
 200

150

Head , ft
100

50
100%
80% 90%
60% 70%
50%
0

0 1000 2000 3000 4000 5000

Flow rate, g pm
Figure 12. Head-capacity curves at varying speeds

It is important to understand, however, that the change in actual flow rate doesn’t necessarily follow
the
affinity laws in fact it only follows the affinity law scaling for systems with no static head. This can be
illustrated by overlaying two system curves on the series of pump curves, as shown in Figure 13. For
the
system with 100 ft of static head, reducing motor speed to 80% of nominal causes a reduction in flow
rate to about 25% of that for full speed. This is in stark contrast to the all frictional (no static head)
system, for which operation at 80% speed results in the flow rate dropping to 80% of that at full
speed.200
Sy s tem with
n o s tatic h ead

150

Sy s tem with
Head , ft
100 ft s tatic h ead
100

50
100%
80% 90%
60% 70%
50%
0

0 1000 2000 3000 4000 5000

Flow rate, g pm
Figure 13. System and pump head-capacity curves at varying speeds

Several factors that should be weighed when considering application of drives

are:
1) A fundamental requirement for considering variable speed drives should be that flow (or in a more
limited sense, head) requirements change over time. If the system flow and head requirements are
essentially fixed, efforts should concentrate on ensuring a properly sized pump (which should also
be
a consideration, regardless of whether a drive is employed).
2) The greater the proportion of the system head that is static, the less effective will be the variable
speed drive. This is particularly true if the flow requirements vary significantly. Parallel pumps are,
on the other hand, most effective in systems that are dominated by static head. Parallel pumps
need
not all be identical; in fact there are some systems, such as waste treatment plant applications,
where
 a small pump could meet system requirements 80-90% of the time, and supplemented by a
parallel,
3) larger pump for the occasional higher flow condition.
The pumps must operate enough hours during the year to justify applying VSDs (when based on a
simple payback, a pump station with multiple pumps may only have the lead pump equipped with
a
4) VSD).
In situations where there is significant static head, the effective range for a variable speed driven
pump can be increased by selecting the pump such that the pump operating point is to the right
of
5) (i.e., greater than) the BEP when operating at the highest anticipated flow requirement. This
allows
the pump efficiency to initially increase as speed is reduced.
6)
The motor and electrical system should be verified to be suitable for drive applications. This is
particularly important if a VSD is being added as a retrofit.
If a variable speed drive is being retrofitted to systems with switch-based controls, such as a float-
7) type level control system that causes the pump to be started or stopped at distinct levels,
continuously
variable instrumentation, such as a level indicator, will need to be added for control purposes.
In a similar vein, systems that currently employ control valves with automatic valve positioners
can
usually be more easily retrofitted by using the existing valve positioning signal source as the
Evaluating Pump
controlling Systems
signal. It might be noted that valves that are normally significantly throttled are a good
Evaluating pump systems
first indicator begins
of a system with an
for which understanding
a drive of how the pump is operated. Often it is
may be effective.
useful to
assemble a table with several flow intervals that can be used to determine how often the pump
operates at
various flows. Several data collection or estimating options for each aspect of operation are shown in
Table 1.
 Table 1. Measurement considerations
MethodNotes
Flow rate Estimate flow rate from- Depends on accurate pump performance curve;
measured head and/or power- For many pumps, shaft power varies relatively
little with flow rate;
- Motor efficiency must be assumed.
Measure, using permanently- Usually preferred; calibration/accuracy must be
installed flow meterverified
Measure, using portable flow- Ensure that a proper flow profile exists (suitable
meters, such as ultrasoniclength, particularly downstream, of elbows or
other flow disturbances)
Calculate, using- Accuracy Reduced for Low Head Pump Systems
thermodynamic flow meter
Measure, using pump down test - Head changes must be accounted for.

Method Notes
Head Estimate head from of - Depends on accurate pump performance curve;
measured flow and/or power - For many pumps, shaft power varies relatively
little with flow rate;
Motor efficiency must be assumed.
-
Not verified with field data; large uncertainties,
Estimate pressure loss from -
particularly with older systems, unless most of
components, pipe data the head is static
Gage accuracy must be verified. Use of
Measure using suction and - portable test gages is recommended.
discharge taps If significant loss components are between the
- pump and the pressure gage, an estimate of the
associated head loss must be made.

Method Notes
Efficiencies Estimate motor efficiency from - Although there may be some inaccuracy, the
nameplate, manufacturer data, errors will generally be small relative to other
or software sources error sources
Estimate electrical system Accuracy of many power meters is suspect
-
efficiency from power when used between the VSD and motor.
measurements on each side of Monitor on the line side of the drive to
transformer, VSD, etc.. minimize this problem.
Measurements on the high side of transformers
- (above 600V) will require the availability of
already installed instrumentation.
Accuracy depends on individual measured
component accuracies. This is the authors’
Calculate overall pump and -
preferred approach.
electrical train efficiency using
measured input power, head
and flow rate
 Table 1. Measurement considerations (continued)
MethodNotes
kW Calculated from head and flow - Not recommended unless there are no
rate measurements, pumpalternatives.
performance curve and
assumed electrical efficiencies
Estimated from measured- Can be reasonably accurate, provided that
currentrepresentative motor performance curves are
available.
Estimate from measured speed- Not recommended. Errors can be particularly
large on newer, low-slip motors.

After selecting the methodology of data collection, a table can be developed for each flow interval as
shown below in Table 2. Only 5 flow intervals are shown, however each pump system reviewed may
have more or less intervals depending on the application

Table 2. Flow Interval Data Collection Table

Interval Pump Flow Head PumpMotor Calculated Annual kWh

(gpm) (ft) Efficiency Efficiency kW Hours of
Operation

1
2
3
4
5

Totals -- -- -- -- -- 8760

From the collected data, it is often useful to plot head and flow measurements from field collected
data
compared to the original pump curve data. This can be done with pump modeling programs, or with
the
graphing option of spreadsheet programs. This modeling is especially useful to determine how output
and
Examples of Cost Saving Opportunities
efficiency performance is effected in a multiple pump system.
Case Study A: Municipal Wastewater Pump System
Ten 3,500-hp pumps equipped with A.C. variable speed drives were evaluated to identify potential
areas
of savings. A review of pump system operation revealed the following:
The suction tank (wetwell) average operating level could be increased to reduce total system
head
Multiple pump system efficiency could be improved with automatic control systems
A review of the pump system operation provided the following data:
Existing Pump Operation (one flow interval shown over 36 performed for full analysis of this
system)
Vertical single stage centrifugal pumpsPump Type:
85 ft.Total Head (TDH):
110 million gallons per day (76,000 gpm)Flow:
90% (from certified curve)Pump Efficiency:
92% (from electrical calculations)Combined motor and electrical
supply efficiency:
The product of the pump and combined electrical efficiencies is 82.8%. From the pump equation,
electrical power, in kW (P) can be calculated with the following equation:
 QH
P=
5310 T

For one of the existing flow intervals:

76,000 gpm * 85 ft * 1 = 1.47 x 103 kW
P=
5310 * 0.828

Proposed Operation
The potential of raising the level on the suction side of the pump by 2 feet provided an excellent
opportunity to reduce the total head on the pump system. Although the adjustment seems small, the
ability to ramp down the variable speed drive slightly to produce the same flow at a lower drive speed
provided the following savings (pump and electrical efficiencies did not change significantly):

76,000 gpm * 83 ft * 1 = 1.43 x 103 kW

P=
5310 * 0.828

Power, energy, and cost savings were calculated:

Reduction in motor power: 0.04 x 103 kW = 40 kW
Reduction in annual energy usage2: 40 kW * 8760 hours/year = 3.5 x 105 kWh
Annual cost savings: 3.5 x 105 kWh * $0.070/kWh = $24,500

This simple control system adjustment did not require an investment and paid for itself
immediately.
Case Study B: Municipal Water System
An existing 200-hp finish water pump (with the motor driven directly off the line) was reviewed to
determine the cost effectiveness of repairing or replacing the existing pump. Repair estimates to
refurbish the worn pump was approximately $10,000.

A review of the pump system provided the following generic nameplate

information:
Pump Type: Split case horizontal pump Design flow rate, head:1,300 gpm, 415 ft
Pump efficiency from curve: 70% Nameplate motor efficiency: 94% (2-pole motor)

Other electrical system losses were assumed negligible.

The nameplate pump efficiency of 70% was compared with information included in Hydraulics
Institute
(HI) standard ANSI/HI 1.1-1.5-1994, which indicated an achievable efficiency of 80%. To determine the
potential savings of replacing the pump with a higher efficiency unit, the 70% pump efficiency was
confirmed by taking flow, head and kW measurements (electrical efficiency losses were assumed as
noted above). Rearranging the pump curve confirmed the efficiency noted on the pump curve as
shown
below: QH
p =
5310 Pηm

The motor power, measured at the nameplate conditions, was 155 kW. The pump efficiency was then
estimated to be:

p = 1,300 gpm * 415 ft * 1

= 69%
5310 * 157 kW * 0.94

After a review of energy rate schedules and system configuration, the pump efficiency was reviewed
in
more detail to determine if other manufacturers could provide better efficiency at the some flow
2
To simplify this example, the operating condition is assumed to be constant throughout the
year.
conditions. Based on a few phone calls, several pumps with the same configuration were found that
could
achieve the same flow and head at 80% efficiency (which, incidentally, was consistent with the HI
standard noted above). Combining this improved pump design with a premium efficient motor (96%
efficiency) produced the following:
QH
P=
5310 p m

1,300 gpm * 41.5 ft * 1

P= = 132 kW
5310 * 0.80 * 0.96

Reduction in motor power:25 kW

Annual kWh Savings: 25 kW * 8760 hours/year = 219,000 kWh
Annual cost savings: 219,000 kWh * $0.090/kWh = $19,710 annual savings

This project paid for itself within 2 years (not including the avoided $10,000 maintenance expense on
the
old pump).
Summary
It is our hope that this paper has successfully presented useful information to help operators,
engineers
and maintenance staff identify cost saving opportunities for their pump systems. As demonstrated in
the
case studies above, some quick calculations and data collection can help discover improvements that
are
cost effective and help justify investing in new pump equipment, control systems and efficient
electrical
systems.