Chiller-Plant Energy

Performance
In large, air-conditioned buildings, the
chiller plant is one of the major energy
consumers. Consequently, the energy
performance of the chiller has been studied
closely. This has usually meant evaluating
the efficiency of a constant-speed chiller
running at design conditions; that is, at full
load and at maximum entering condenser
water temperature (ECWT).
Recently, the perspective on chiller-plant
energy performance has broadened to
encompass the energy-cost impact of off-
design conditions. These conditions exist
whenever the building doesnt demand the
chillers full cooling capacity and/or
whenever the ECWT is lower than its design
value. This wider view also takes into
account the energy consumption of auxiliary
chiller-plant components: the cooling tower,
the chilled-water pump, and the condenser-
water pump.
When cost-effective variable-speed drives
(VSD) are added to the picture, it pays to
look at whether applying VSD to the chiller-
plant components can offer reductions in
energy cost. And what is the trade-off in first
cost and maintenance cost?
Since this is a broad subject, well begin
by examining the energy consumption of the
chiller by itself. Then, well add in the tower,
followed by the pumps. Where energy
decisions also have maintenance
implications, we will discuss those. Finally,
well draw some conclusions about the
operation of the chiller-plant as a whole.
Chiller Performance
In order to understand energy
consumption of the chiller, we must first
appreciate the difference between design
and off-design conditions. To do that, we
start with weather data. Well use the
Baltimore/Washington area as an example,
since this is where the most chillers have
been purchased over a recent 25-year
period
1
.
The Baltimore/Washington weather
shown in Figure 1 is based on 25-year
averages
2
. It shows hours spent at outdoor
ECWT
Dry-bulb Temperature Bins (F)
A
n
n
u
a
l

H
o
u
r
s
50
55
60
65
70
75
80
85
T
y
p
i
c
a
l

E
C
W
T

(

F
)
0
200
94 90
6
72
243
428
925
858
755
688
671
631
84 80 74 70 64 60 54 50
400
600
800
1000
Figure 1 - Baltimore/Washington Weather Above 50F
dry-bulb temperatures above 50F, and the
ECWT typically available at those outdoor
temperatures.
A chiller system in this locale could
potentially run for a total of 5,277 hours if
we assume 24-hour operation whenever the
outdoor temperature exceeds 50F (if
operation below 50F is required because
the facility has no airside economizer, there
would be an additional 3,483 hours).
Chiller design conditions typically occur
when the outdoor temperature and ECWT
are at their peak. Of all those hours, note
how little time is spent at design conditions.
Temperatures are in the 95-99F bin (which
can be considered full-load on the chiller)
for only six hours per year. The other 5,271
hours are off-design hours. Consequently,
the chiller only runs about one-tenth of one
percent of potential operating hours at full-
load and design ECWT.
Note also how quickly the ECWT
declines, falling below 75F for almost 4,400
hours (almost 7,900 hours if there is no
economizer), and reaching as low as 55F.
This is a result of the outdoor air getting
drier as it gets colder, which allows the
cooling tower to cool the condenser water
further.
The bottom line in this case (and most
others) is that over 99% of chiller operating
hours are spent at off-design conditions:
either at part-load, or with colder ECWT, or
both.
Importance of Off-Design Performance
In the past, it has been generally
assumed that by picking a chiller with good
performance at design conditions, youd
also get good performance at off-design
conditions. Unfortunately, that isnt always
true.
To illustrate, lets use the Baltimore/
Washington weather data to compare a
chiller with a design efficiency of .55 kW/TR
and excellent off-design performance
against a chiller with a design efficiency of
.50 kW/TR and poor off-design efficiency.
Well assume a single, 500-TR, constant-
speed, centrifugal chiller cooling water from
54 to 44F using 3 GPM/TR of condenser
water at 85F ECWT. Further, well assume
the chiller runs 24 hours per day, whenever
the outdoor temperature is above 50F, and
that electricity demand and usage charges
result in a blended rate of $.08/kWh.
Table 1 shows the energy consumption
and energy cost for each chiller in each
temperature bin. The interesting point is that
both chillers use virtually the same amount
of energy. While the .55 kW/TR unit uses
more energy at higher loads, there are
fewer operating hours spent there.
Conversely, it uses less energy at lower
loads, where the majority of operating hours
are spent.
Just as interesting as the energy-cost
comparison is the first-cost comparison. The
.50 kW/TR chiller would cost as much as
$25,000 more to purchase. Thats a big
premium for the same energy performance.
Off-design Efficiency Factors
But why do some chillers have better
performance at off-design conditions? To
understand that, we must look at the factors
that affect off-design efficiency. While there
are many, the two with the greatest impact
are the chillers ability to operate efficiently
at lower ECWTs and lower loads.
The ability to efficiently handle lower
ECWTs is a function of the chillers basic
design some do it well, some dont. If a
chiller can use colder ECWTs, the
compressor workload can be reduced
substantially, saving energy.
Table 1 - Chiller Performance Comparison
.50 kW/TR, Poor Off-design .55 kW/TR, Excellent Off-design
Temp. Opg. Chlr. Typ. Power Energy Energy Typ. Power Energy Energy
Bin Time Load ECWT Use Use Cost ECWT Use Use Cost
(F) (hrs) (TR) (F) (kW) (kWh) ($) (F) (kW) (kWh) ($)
95-99 6 500 83 246 1,477 118 83 266 1,598 128
90-94 72 464 81 222 15,957 1,277 81 237 17,066 1,365
85-89 243 427 79 197 47,844 3,828 79 208 50,569 4,045
80-84 428 391 77 175 74,768 5,981 77 183 78,427 6,274
75-79 631 355 75 154 97,307 7,785 75 161 101,864 8,149
70-74 925 319 75 138 127,227 10,178 73 142 131,646 10,532
65-69 858 282 75 121 104,121 8,330 68 120 102,763 8,221
60-64 755 246 75 107 80,422 6,434 63 101 75,936 6,075
55-59 688 210 75 93 63,702 5,096 58 84 57,705 4,616
50-55 671 174 75 80 53,388 4,271 55 72 48,031 3,842
Total 666,213 53,298 665,605 53,247
Savings with .55 kW/TR & excellent off-design performance = $51
2
Over 99% of
chiller operat-
ing hours are
spent at
off-design
conditions.
A simple analogy will help to illustrate the
point. Think of the compressor as a man
trying to throw a rock to the top of a hill. The
rock represents the cooling load, and the
height of the hill represents the pressure
difference which the compressor must
overcome. Lowering the ECWT is like
lowering the height of the hill the man will
need less energy to throw the rock to the
top of the hill, regardless of the load.
In order to establish a basis for future
comparison, lets determine the energy
consumption of a different chiller in the
Baltimore/Washington area. Well assume
this chiller has a design efficiency of .55
kW/TR, average off-design performance,
and that its ECWT is artificially limited to
75F minimum, as has often been done in
the past. Otherwise, well use the same
assumptions as in our previous example
Table 2 shows that this chiller uses
724,711 kWh annually, at a cost of $57,975.
If we now take this example and
substitute a chiller that can use ECWT down
to 55F, rather than requiring an artificial
minimum of 75F, the results are significant,
as shown in Table 3. Our chiller now uses
only 686,260 kWh at a cost of $54,900, a
savings of almost $3,100.
Whats especially attractive about this
low-ECWT capability is that it doesnt
require any additional first cost. An efficient
chiller has this capability built-in. And theres
no additional maintenance cost because
there is no additional hardware.
What about our second factor lower
loads? While our examples show lower
ECWTs and lower loads occurring together,
that is not always the
case (as we shall
see later). In reality,
a chiller must handle
varying ECWTs and
varying loads in
order to perform
efficiently.
Constant-speed
chillers handle lower
loads by gradually
closing a set of
guide vanes on the
compressor inlet,
while keeping the
motor turning at full
speed. But closing
the vanes creates
friction losses in the
refrigerant flow, producing a drag on
efficiency. A simple analogy would be trying
to control a car by stepping on the brake
while keeping the accelerator floored.
A better method utilizes variable-speed
technology. A VSD efficiently decreases its
power draw in response to off-design loads
and ECWTs, similar to controlling a car by
easing off the accelerator. The energy
savings can be as much as 30% annually
Table 2 - Energy Use of Chiller with 75F ECWT Minimum
Temp. Opg. Chlr. Typ. Power Energy Energy
Bin Time Load ECWT Use Use Cost
(F) (hrs) (TR) (F) (kW) (kWh) ($)
95-99 6 500 83 270 1,614 129
90-94 72 464 81 240 17,256 1,380
85-89 243 427 79 212 51,517 4,121
80-84 428 391 77 188 80,237 6,419
75-79 631 355 75 165 104,243 8,339
70-74 925 319 75 148 137,148 10,972
65-69 858 282 75 132 113,159 9,053
60-64 755 246 75 117 88,292 7,063
55-59 688 210 75 103 70,852 5,668
50-55 671 174 75 90 60,393 4,831
Total 724,711 57,975
Table 3 - Effect of Lower ECWT on Chiller Energy Use
75F ECWT Minimum 55F ECWT Minimum
Temp. Opg. Chlr. Typ. Power Energy Energy Typ. Power Energy Energy
Bin Time Load ECWT Use Use Cost ECWT Use Use Cost
(F) (hrs) (TR) (F) (kW) (kWh) ($) (F) (kW) (kWh) ($)
95-99 6 500 83 270 1,614 129 83 270 1,614 129
90-94 72 464 81 240 17,256 1,380 81 240 17,256 1,380
85-89 243 427 79 212 51,517 4,121 79 212 51,517 4,121
80-84 428 391 77 188 80,237 6,419 77 188 80,237 6,419
75-79 631 355 75 165 104,243 8,339 75 165 104,243 8,339
70-74 925 319 75 148 137,148 10,972 73 146 134,783 10,783
65-69 858 282 75 132 113,159 9,053 68 124 106,459 8,517
60-64 755 246 75 117 88,292 7,063 63 105 79,050 6,324
55-59 688 210 75 103 70,852 5,668 58 88 60,292 4,823
50-55 671 174 75 90 60,393 4,831 55 76 50,809 4,065
Total 724,711 57,975 686,260 54,900
Savings with 55F ECWT minimum = $3,075
3
Good perfor-
mance at design
conditions
doesnt guaran-
tee good
performance
at off-design
conditions.
when compared to a constant-speed chiller.
In our example, we now substitute a VSD
chiller that can utilize ECWT down to 55F.
As seen in Table 4, our chiller energy
consumption is now only 521,663 kWh at a
cost of $41,732, a savings of $16,243. With
energy savings of this magnitude, the
added cost of the VSD for the chiller can be
paid off very quickly well within the time-
frame required by most building owners.
Multiple Chillers
Up to this point, the focus has been on a
single centrifugal chiller. But most chiller
plants have more than one chiller, which will
affect chiller loading and energy
performance. Do plants with multiple chillers
negate the importance of off-design chiller
performance? Absolutely not!
The impact of multiple chillers on loading
can be seen in an example of a 1,000-TR
building. Assume a two-chiller plant with one
chiller designated as the lead unit and the
other as the lag unit. The lead chiller is the
first unit on and the last unit off. The lag
chiller provides supplemental cooling
capacity. Each chiller is constant-speed, with
a design capacity of 500 TR, using 275 kW
(.55 kW/TR) to chill water from 54 to 44F.
The condenser water-flow is 3 GPM/TR,
ECWT is 85F at design and is limited to
75F minimum.
As seen in Table 5, when the building
load is below 500 TR, the lag chiller cycles
off and the lead chiller carries the building
load alone. One result of this loading pattern
is that the lead chiller sees many more
hours near full-load than it would have seen
if it were a single chiller cooling a 500-TR
building. Specifically, in the 60-64F temp-
erature bin, the lead chiller is spending 755
hours near its design capacity of 500 TR.
Table 4 - Effect of Variable-speed Drive and Lower ECWT on Chiller Energy Use
Constant-speed Drive Variable-speed Drive
75F ECWT Minimum 55F ECWT Minimum
Temp. Opg. Chlr. Typ. Power Energy Energy Typ. Power Energy Energy
Bin Time Load ECWT Use Use Cost ECWT Use Use Cost
(F) (hrs) (TR) (F) (kW) (kWh) ($) (F) (kW) (kWh) ($)
95-99 6 500 83 270 1,614 129 83 278 1,652 132
90-94 72 464 81 240 17,256 1,380 81 239 17,193 1,375
85-89 243 427 79 212 51,517 4,121 79 202 49,143 3,931
80-84 428 391 77 188 80,237 6,419 77 170 72,896 5,832
75-79 631 355 75 165 104,243 8,339 75 142 89,634 7,171
70-74 925 319 75 148 137,148 10,972 73 119 109,667 8,773
65-69 858 282 75 132 113,159 9,053 68 89 76,712 6,137
60-64 755 246 75 117 88,292 7,063 63 65 49,150 3,932
55-59 688 210 75 103 70,852 5,668 58 47 32,016 2,561
50-55 671 174 75 90 60,393 4,831 55 35 23,600 1,888
Total 724,711 57,975 521,663 41,732
Savings with VSD and 55F ECWT minimum = $16,243
Table 5 - Multiple Chillers with Constant-speed Drive and 75F ECWT Minimum
Building Lead Chiller Lag Chiller
Temp. Bldg. Opg. Typ. Chlr. Power Energy Energy Chlr. Power Energy Energy
Bin Load Time ECWT Load Use Use Cost Load Use Use Cost
(F) (TR) (hrs) (F) (TR) (kW) (kWh) ($) (TR) (kW) (kWh) ($)
95-99 1,000 6 83 500 270 1,614 129 500 270 1,614 129
90-94 928 72 81 464 240 17,256 1,380 464 240 17,256 1,380
85-89 854 243 79 427 212 51,517 4,121 427 212 51,517 4,121
80-84 782 428 77 391 188 80,237 6,419 391 188 80,237 6,419
75-79 710 631 75 355 165 104,243 8,339 355 165 104,243 8,339
70-74 638 925 75 319 148 137,148 10,972 319 148 137,148 10,972
65-69 564 858 75 282 132 113,159 9,053 282 132 113,159 9,053
60-64 492 755 75 492 239 180,168 14,413 0 0 0 0
55-59 420 688 75 420 198 136,409 10,913 0 0 0 0
50-55 348 671 75 348 161 108,269 8,662 0 0 0 0
Subtotal 930,020 74,401 505,174 40,413
Totals for both chillers 1,435,194 114,814
4
Variable-speed
chillers have
excellent perfor-
mance at
off-design
conditions.
But note that more hours at full-load dont
mean more hours at design conditions.
Thats because at lower building loads, the
ECWT is also lower in this case 75F,
down from the design of 85F. So, the lead
chiller is seeing more hours at higher loads,
but at lower ECWTs which is another
form of off-design conditions.
So even in multiple-chiller plants, almost
all operating hours are at off-design
conditions with lower loads, lower ECWT,
or both. Consequently, chillers that are
optimized for off-design conditions are
excellent choices for plants with multiple
chillers, as well.
Suppose we take this same multiple-
chiller plant and substitute variable-speed
chillers that can utilize ECWT as low as
55F. As seen in Table 6, the results are
dramatic.
Equipped with chillers optimized for off-
design conditions, this plant shows a
savings of $25,596. Simple payback for the
VSDs would be very attractive.
What about the maintenance-cost impact
of utilizing these type of chillers? If there is
any impact, it will be that maintenance costs
will be lower. Being able to use 55F ECWT
is inherent to the chiller and will have no
effect. Using VSDs may lower maintenance
costs because they soft-start the chillers,
saving wear-and-tear on the driveline and
extending its life for years.
To summarize our chiller discussion,
weve found that off-design performance is
of critical importance in all types of chiller
plants. Weve also found that chillers
designed to operate efficiently with varying
loads and ECWTs can take advantage of
off-design conditions to offer unsurpassed
chiller efficiency, reasonable first cost, lower
maintenance cost, and longer life.
Next, well connect the chiller to the
cooling tower, and look at their combined
performance.
Cooling Tower Performance
The energy consumption of a cooling
tower is greatly influenced by the chiller to
which it is connected. We stated previously
that some chillers require the ECWT be held
at 75F, while others can save energy by
utilizing ECWT as low as 55F. Lets
examine what that means to the tower.
There are five methods by which ECWT
can be controlled:
1. Have some water bypass the tower and
not be cooled (a.k.a. Cooling Tower
Bypass)
2. Have some water bypass the chiller and
not be heated (a.k.a. Condenser Bypass)
3. Cycle the tower fan on and off to limit the
cooling effect (a.k.a. Fan Cycling)
4. Use a two-speed motor (or pony motor)
on the fan to limit the cooling effect
5. Use a VSD on the fan motor to limit the
cooling effect
Lets connect the tower to the chiller and
look at the energy-cost and maintenance-
cost impact of these five control methods.
Bypassing the water around the tower
has nothing to offer in the way of energy-
cost savings because the tower fan
continues to run full-bore, even though only
partial capacity is required. Even worse, this
method increases maintenance costs
because reducing water flow through the
tower increases scaling on the internal
surfaces, and increases the probability of
freeze-up during winter operation. For these
reasons, cooling-tower bypass is not a
Table 6 - Multiple Chillers with Variable-speed Drive and 55F ECWT Minimum
Building Lead Chiller Lag Chiller
Temp. Bldg. Opg. Typ. Chlr. Power Energy Energy Chlr. Power Energy Energy
Bin Load Time ECWT Load Use Use Cost Load Use Use Cost
(F) (TR) (hrs) (F) (TR) (kW) (kWh) ($) (TR) (kW) (kWh) ($)
95-99 1,000 6 83 500 278 1,652 132 500 278 1,652 132
90-94 928 72 81 464 239 17,193 1,375 464 239 17,193 1,375
85-89 854 243 79 427 202 49,143 3,931 427 202 49,143 3,931
80-84 782 428 77 391 170 72,896 5,832 391 170 72,896 5,832
75-79 710 631 75 355 142 89,634 7,171 355 142 89,634 7,171
70-74 638 925 73 319 119 109,667 8,773 319 119 109,667 8,773
65-69 564 858 68 282 89 76,712 6,137 282 89 76,712 6,137
60-64 492 755 63 492 185 139,641 11,171 0 0 0 0
55-59 420 688 58 420 125 86,097 6,888 0 0 0 0
50-55 348 671 55 348 83 55,712 4,457 0 0 0 0
Subtotals 698,347 55,867 416,897 33,351
Totals for both chillers 1,115,244 89,218
Savings with variable-speed drive & 55F ECWT minimum = $25,596
5
Multiple-chiller
plants also
spend over 99%
of their operat-
ing hours at
off-design
conditions.
recommended control method.
Bypassing the water around the chiller
condenser has nothing to offer in the way of
energy-cost savings, either. While the chiller
is seeing lower ECWT, it is also seeing less
water-flow. The result is that refrigerant
pressure in the condenser remains high,
and chiller energy is not reduced. In
addition, the tower fan continues to run full-
bore, even though only partial capacity is
required. This method also increases
maintenance costs because reducing water
flow through the condenser decreases water
velocity through the tubes, increasing
fouling.
Cycling the tower fan does save energy.
In Table 3, we compared the energy cost of
chillers that kept the ECWT at 75F to those
that could use ECWT as low as 55F. The
55F chillers used $3,100 less in energy. But
to do that, they require that the tower fan
runs at all times. The 75F chillers can cycle
the fan, and in our example, that would
generate tower-fan energy savings of about
$2,700. So the 55F chillers would still be
the better energy-cost choice, but only by
about $400.
However, fan cycling has a major
maintenance-cost drawback. All the starting
and stopping of the fan subjects its driveline
to an incredible amount of wear-and-tear.
The cost of maintenance on belts, pulleys,
gearboxes, and motors quickly outweighs
any energy-cost savings. This is the same
reason the HVAC industry abandoned the
technique of cycling air-handler fans to save
on demand charges.
Putting a two-speed motor on the fan is a
bit better than fan cycling. It cuts down on
the number of on-off cycles the driveline
must incur because often the fan can just
cycle between high and low speed.
However, it doesnt completely eliminate the
cycling, so there are still higher maintenance
costs. Energy costs will be somewhat lower
than with fan cycling because running the
motor at low speed will consume less
energy than cycling the motor on and off.
If the chiller requires that ECWT must be
held above 55F, the best alternative is to
equip the tower fan with a VSD.
Maintenance cost will be minimized because
cycling is eliminated, and even start-ups will
be soft, with a gradual ramp-up in speed.
Energy cost for the tower will be lower, even
when compared to the fan-cycling or two-
speed motor options.
However, if the chiller can utilize ECWT
as low as 55F, then it is better to put a
VSD on the chiller and keep the tower fan
at constant speed. Using our earlier
example, Table 7 shows the energy-cost
comparison.
The combination of the variable-speed
chiller and constant-speed tower yields
$11,100 in energy-cost savings. The
maintenance costs will be a wash because
neither system will cycle the tower fan. The
first cost of the tower VSD partially offsets
the first cost of the chiller VSD, so the
energy savings will pay for the balance of
the chiller VSD very quickly.
In summary, the most economical choice
is to use a variable-speed chiller that is
optimized for off-design conditions, rather
than putting additional money into a tower
VSD and controls. There is, however, one
other situation where a variable-speed
tower does make sense. If the building has
no economizer and the chiller is required to
run year-round, some kind of temperature
control will be required for the tower to
maintain 55F ECWT during the winter. In
this case, VSDs on both the chiller and the
tower is the best solution.
Chilled-Water Pump Performance
The chilled-water pump connects the
chiller and the airside equipment. If the
airside is designed to handle variable water-
flow, a VSD on this pump can be a real
energy-saver. However, there are two
cautions on the chiller-side. The flow rate-
of-change must not be too rapid, or the
chiller will begin to hunt. Care must also be
taken with the minimum velocity of the
water.
Most chillers require that the velocity of
the chilled-water flowing through their tubes
not fall below 3.0 to 3.3 fps. This limit can
be a concern when a high-efficiency chiller
is being utilized. High-efficiency chillers are
often selected to operate at full-load with a
water velocity at or below 4.0 fps, which
Table 7 - Effect of Variable-speed Drive on Tower versus Chiller
Constant-speed Chiller Variable-speed Chiller
Variable-speed Tower Constant-speed Tower
75F ECWT minimum 55F ECWT minimum
Chiller energy cost $58,000 $41,700
Tower energy cost $2,600 $7,800
Total of chiller & tower $60,600 $49,500
Savings - $11,100
6
Chiller control
offers greater
savings than
tower control.
would allow the variable-speed pump to
reduce its speed only 10 to 20% limiting
the energy-saving potential.
One solution is to increase the number of
passes the water makes through the chiller.
This will increase the velocity, allowing
greater turndown of the pump speed.
However, it will also increase the water
pressure drop through the chiller. The added
energy cost of pumping the water through
the chiller will have to be weighed against
the savings with VSD control. If this strategy
is applied to our earlier 500-TR example,
Table 8 shows that the savings are almost
$6,000, making for a very attractive
payback.
Condenser-Water Pump Performance
What about putting a VSD on the
condenser-water pump? Does this offer the
same energy benefits as the chilled water
pump? Potentially yes, but maintenance is a
serious question.
When water velocity is slowed, any
materials suspended in the water have a
greater tendency to drop out. In the chilled-
water circuit, this is less of a problem
because the circuit is closed and the water
stays relatively clean. However, the
condenser-water circuit is open to the
atmosphere in the cooling tower and has a
much greater tendency to foul. If the water
velocity is reduced, the chiller, piping,
valves, and tower will see increased fouling.
Either chiller and pump energy costs will
increase because of the fouling, or the
maintenance cost will increase to control the
fouling.
Even if fouling is controlled, variable flow
will have an adverse effect on the
performance of the tower. If flow is reduced
more than 15%, there will not be enough
water pressure to feed water to all the
nozzles, and performance will fluctuate
wildly. For both these reasons, VSD on the
condenser-water pump is not
recommended.
Reduce Condenser-Water Flow
As a different approach to reducing
chiller-plant costs, it has been suggested
that reducing the condenser-water flow from
the standard 3 GPM/TR to 2 GPM/TR will
save both energy cost and first cost.
Although the lower flow will penalize the
energy performance of the chiller, it is
proposed that the energy savings on the
tower and condenser-water pump will out-
weigh the chiller losses. It is also suggested
that the tower, the pump, and the
condenser-water piping can be smaller,
reducing first cost.
There are several fallacies with this
strategy. In the first place, it may not be
possible to reduce pipe size. Thats because
pipe doesnt come in an infinite range of
sizes. More often than not, the same
diameter pipe used with a 3 GPM/TR
system must be used for a 2 GPM/TR
system. Take our 500-TR example. At 3
GPM/TR, the water-flow would be 1500
GPM. Pipe is sized according to the
ASHRAE friction guideline of 1 to 4 feet of
pressure drop per 100 feet of pipe. 1500
GPM would require 8 pipe. If the flow is
reduced to 2 GPM/TR, or 1000 GPM, could
6 pipe (the next smallest size) be used?
No, because the friction loss would exceed
the upper limit of 4 feet with 6 pipe. So,
there is no first-cost savings on the pipe.
Even on the occasions when the pipe
size can be reduced, there is a trade-off
between first cost and energy cost. You
cant reduce the pipe size and still save the
Table 8 - Effect of Variable-speed Drive on Chilled-water Pump Energy Use
Constant-speed Pump Variable-speed Pump
Two-pass Chiller Three-pass Chiller
Temp. Opg. Chlr. Pump Pump Energy Pump Pump Energy
Bin Time Load Power Energy Cost Power Energy Cost
(F) (hrs) (TR) (kW) (kWh) ($) (kW) (kWh) ($)
95-99 6 500 21.2 127 10 29.7 178 14
90-94 72 464 21.2 1,529 122 23.7 1,706 136
85-89 243 427 21.2 5,159 413 18.5 4,488 359
80-84 428 391 21.2 9,087 727 14.2 6,069 486
75-79 631 355 21.2 13,397 1,072 10.6 6,697 536
70-74 925 319 21.2 19,640 1,571 7.7 7,123 570
65-69 858 282 21.2 18,217 1,457 5.3 4,564 365
60-64 755 246 21.2 16,030 1,282 3.7 2,798 224
55-59 688 210 21.2 14,608 1,169 3.7 2,550 204
50-55 671 174 21.2 14,247 1,140 3.7 2,487 199
Total 112,041 8,963 38,660 3,093
Savings with VSD pump and three-pass chiller = $5,870
7
There are
several fallacies
in the 2 GPM/TR
strategy.
same amount of energy. Reducing the pipe
size will increase the frictional loss. You
cant have it both ways.
Similarly, the size of the cooling tower
cannot be reduced without affecting its
efficiency. If energy savings are to be
maximized, then the tower size must be the
same. If the tower is downsized, the energy
savings will be reduced.
You would think that the condenser-water
pump could be reduced in size without
penalty, but thats not necessarily true,
either. Using our example, it turns out that in
order to keep pump efficiency high, the best
selection to handle 1000 GPM is the same
pump which was selected to handle 1500
GPM, but running at 1150 RPM instead of
1750 RPM. This is not a coincidence, but is
a function of pump laws. So, there is no
first-cost savings on the pump, either.
Lets use our 500-TR example again and
compare a 3 GPM/TR system optimized for
off-design conditions with a 2 GPM/TR
system optimized for design conditions.
Well keep the same size components on
the 2 GPM/TR system, so as to maximize
energy savings. All the components will be
constant speed, except that well make a
third comparison with a VSD on the off-
design-optimized chiller only. The results are
summarized in Table 9.
In the 2 GPM/TR system, the energy
savings available from the cooling tower and
the condenser-water pump are not enough
to overcome the energy penalty on the
chiller. This conclusion may vary depending
upon how the chiller-plant is operated, but
generally a 3 GPM/TR system will out-
perform a 2 GPM/TR system.
However, the addition of a VSD on a
chiller optimized for off-design conditions will
always make good economic sense,
because it takes maximum advantage of the
savings potential available with lower loads
and lower ECWTs, while extending chiller
life-expectancy.
Bottom-Line Chiller-Plant Savings
Traditional efficiency comparisons for
large-tonnage chillers have been based on
energy ratings at design conditions. But off-
design conditions prevail in the real world.
Chillers that are optimized for off-design
conditions can utilize VSDs to take
advantage of lower loads and lower ECWTs,
producing significant energy savings when
compared to chillers designed primarily for
design conditions.
A VSD on the cooling tower makes sense
if the chiller plant will run year-round,
otherwise a VSD on the chiller is a better
investment. A VSD on the chilled-water
pump can be an energy-saver, but care
must be taken not to have the water velocity
fall below the chillers lower limit. A VSD on
the condenser-water pump is not
recommended because of the increased
maintenance cost caused by fouling in the
chiller and tower. Finally, the savings claims
being made about 2 GPM/TR systems are
contradictory, and the reality doesnt
measure up to the theory.
System dynamics will always find a way
to make us pay for our shortcuts. But by
focusing on applying variable-speed drives
to chillers, you will see that you can
maximize your chiller plants energy savings
at an attractive first cost and with no hidden
maintenance costs.
1. ARI data from 1967 - 1992.
2. Engineering Weather Data; Department of the Air Force, the
Army, and the Navy; July 1, 1978.
Table 9 - Effect of 2 GPM/TR versus 3 GPM/TR Condenser Water-flow
2 GPM/TR 3 GPM/TR 3 GPM/TR
Constant-speed chiller Constant-speed chiller Variable-speed chiller
75F ECWT minimum 55F ECWT minimum 55F ECWT minimum
Chiller energy cost $63,900 $54,900 $41,700
Tower energy cost $4,400 $7,800 $7,800
CW pump energy cost $4,500 $7,200 $7,200
Total $72,800 $69,900 $56,700
Savings versus 2 GPM/TR $2,900 $16,100
The reality
of 2 GPM/TR
systems doesnt
measure up to
the claims.