IIR Short Course Presentation PDF

Energy Efficiency Opportunities in
Industrial Refrigeration Systems
Douglas T. Reindl, Ph.D., P.E.

Director, Industrial Refrigeration Consortium
University of Wisconsin-Madison
dreindl@wisc.edu
© 2003 Reindl University of Wisconsin-Madison
Learning Objective:
The learning objective of this module is to review the common configurations of single and multi-stage
vapor compression refrigeration systems that we first studied in the Introduction course. At the end of
this segment you should recall the following:
• theory of operation for single stage and two-stage vapor compression refrigeration systems
• direct-expansion, flooded, and liquid overfed evaporator configurations
• two-stage systems with direct & indirect liquid expansion
---
This training module has been developed for presentation at the IIR meeting in Washington, DC, USA
by Douglas T. Reindl, Director of the Industrial Refrigeration Consortium at the University of
Wisconsin-Madison (dreindl@wisc.edu). It is intended solely for use by short course attendees and
may not be reproduced, wholly or partially, without written permission of the author. Please contact
the author for information on use (dreindl@wisc.edu).
Page 1
Our Tasks During This Workshop
Review of systems
single stage compression
systems
multi-stage compression
systems
Performance Analysis
(Benchmarking)
Energy Efficiency
Opportunities
© 2003 Reindl
The refrigerant of choice in the industrial sector is anhydrous ammonia; consequently, the materials
prepared and data shown in this series of presentations assumes that ammonia is the refrigerant.
Page 2
Direct-Expansion System
Evaporative
Condenser
2
Suction trap or
Reboiler
Expansion Valve
1
T
Equalizer line
Evaporator Compressor 3
4
King Valve
External Equalize Line

Hi Pressure Receiver
© 2003 Reindl
Theory of Operation:
High pressure and relatively high temperature liquid is drawn from the high pressure receiver and throttled into the evaporator to meet
refrigeration loads. The expansion valve (typically a thermostatic expansion device) functions to modulate the flow of refrigerant
supplied to evaporator in response to changes in the demand for refrigeration. The demand for refrigeration is determined indirectly by
“monitoring” the state of refrigerant superheat at the outlet of the evaporator. Refrigerant superheat is sensed by a temperature-
sensitive bulb attached to the evaporator outlet. When the evaporator load increases, the bulb will sense an increase in refrigerant
superheat and call for the expansion valve to open; thereby, increasing the supply of refrigerant to the evaporator. As the evaporator
load decreases, the refrigerant superheat decreases and the sensing bulb notifies the expansion valve to reduce the flow of refrigerant
supplied to the evaporator. If an evaporator is equipped with a refrigerant distributor or if the pressure drop across the evaporator is
high (in excess of 2 F saturation temperature equivalent pressure loss which corresponds to about 2 psi at an evaporator temperature
of 10 F), then an external equalizing line is needed to properly compensate the temperature sensing bulb. Always connect the
equalizer line at the top of the line leaving the evaporator to avoid any oil or other contaminants from clogging the equalizer line.
Since expansion valves in ammonia systems are especially prone to “wire-drawing” (accelerated wear of valve’s needle and seat), they
require frequent service. In situations where valves are worn, they will not be able to adequately modulate flow at low load conditions;
thereby, admitting more refrigerant to the evaporator than is needed for the present load. The net result is an increased likelihood for
liquid refrigerant to carry-over to the compressor. To prevent this, a “suction trap” (also known as a “knock-out drum” or “reboiler” is a
vessel to separate liquid from vapor prior to compressor suction) is positioned between the outlet of the evaporator(s) and the
compressor suction. The trap functions to catch any liquid refrigerant that is carried-over; thereby, protecting the compressor from
ingesting liquid refrigerant. A subcooling circuit is typically included in this vessel to boil off any liquid trapped. Alternatively, a number
of other liquid transfer methods/strategies can also be applied to the trap. All methods have to limit the accumulation of liquid from the
trap to prevent harming the compressor. In addition, many systems have high level alarms and controls to shutdown compressors
when the liquid level in the suction trap could endanger the compressors by drawing liquid into the suction line.
Page 3
Gravity Flooded Recirculation
Evaporative
Hand-expansion 2 Condenser
valve Suction Trap
Solenoid
valve
4 1
Evaporator
Equalizer line
via transfer station (not shown) 3
Fill float
4’
King Valve
Surge Drum
Hi Pressure Receiver
© 2003 Reindl
Low pressure low temperature vapor (state 1) enters the compressor. The compressor raises the pressure of the refrigerant and the
superheated vapor travels to the condenser (state 2). Heat is rejected from the refrigerant and, in the process, liquefies (state 3).
Saturated liquid refrigerant at high pressure drains into the high pressure receiver. On-demand, the high pressure liquid is “metered” or
throttled to the low-side of the system to meet refrigeration requirements. The throttling process is accomplished by a float sensing
liquid level. As the liquid level drops, the float will energize a single pole double throw contact that powers open a liquid feed solenoid
valve to feed high pressure liquid across a downstream hand-expansion valve. When the liquid level rises to the upper control of the
float (usually a 2 inch change in level), the float opens the single pole double throw switch; thereby, closing the solenoid valve by
denergizing it.
In a gravity recirculation or flooded system, each evaporator is fitted with a surge drum that supplies low temperature saturated liquid
refrigerant to the bottom of an evaporator coil (state 4’). As the coil absorbs heat, the refrigerant on the tube-side of the evaporator
begins to boil. Since refrigerant, in a vapor state, is significantly lighter than the liquid state, the refrigerant vapor will tend to rise and
induce a natural circulation of refrigerant flow upward through the circuits of the evaporator to the outlet line that connects to the vapor
space in the surge drum.
At sufficiently high loads, the circulation rate through the evaporator could be high enough to entrain and return un-boiled liquid from the
evaporator back to the surge drum. Under this operating circumstance, the surge drum serves to separate out liquid and vapor which
allows the liquid to recirculate back to the evaporator and the vapor (State 1) to the compressor suction (usually through a transfer
station).
When the liquid feed solenoid valve is open, the portion of high pressure liquid throttled into the surge drum that flashes to vapor will
short-circuit the evaporator and return directly to the compressor suction via the suction trap. The remaining cold liquid falls to the
bottom of the surge drum and is made available for circulating through the coil.
Typically, a suction trap (also known as a dump trap, trap, knock-out drum) is provided to protect compressors from liquid carry-over.
There are a number of possible configurations to return accumulated liquid to the system including gas pumping, mechanically pumped,
gravity drain, and others. A dashed line is shown in the figure indicating that any accumulated liquid would be transferred back to the
high pressure receiver using the appropriate equipment.
Page 4
Overfeed System layout
Evaporator
Wet suction
2
Evaporative
Compressor Condenser
T
1
Evaporator 4
Dry suction
Level probe
Equalizer line
T 3
Pulse-width 4’
valve
via transfer station
(not shown)
4’’
Trap
Low Pressure
Pump
Accumulator Hi Pressure Receiver
© 2003 Reindl
In an overfeed system, a refrigerant pump removes cold saturated liquid refrigerant (state 4’) from a low pressure accumulator and
pumps it out to a multiplicity of individual evaporators (state 4’’). As evaporators call for refrigeration (usually by a triggered temperature
sensor in the refrigerated space or the process being cooled), a solenoid valve will open admitting liquid refrigerant to flow through the
evaporator. By design, more liquid refrigerant is pumped through the evaporator than can be evaporated in a single pass - hence the
name “overfeed system”. As a result, a mixture of liquid and vapor refrigerant will flow back to the low pressure accumulator through
the “wet suction return” lines.
The low pressure accumulator functions to separate the refrigerant into its liquid and vapor components. The liquid falls to the bottom
of the vessel to be pumped back out to the evaporators while the saturated vapor (state 1) is sent back to the compressor suction. As
refrigerant is “consumed” by the evaporators, additional refrigerant is supplied by throttling high pressure liquid into the accumulator.
There are a number of alternative strategies that can be used to throttle the high pressure liquid. One strategy is to use a simple fill
float. As the liquid level drops, the fill float closes a single pole double throw switch and energizes a liquid feed solenoid valve to
throttle high pressure liquid across a hand-expansion valve. A second alternative is to monitor the liquid level using a capacitance
probe (sometimes referred to as a “level master”). The capacitance probe signal can be used to drive a “pulse-width modulating” valve
that provides continuous liquid make-up to the vessel. As the level in the accumulator drops, the pulse width valve will remain open for
longer periods of time increase the make-up rate. As the level rises, the pulse width valve will have a shorter dwell period to decrease
the rate of refrigerant make-up to the accumulator. Finally, the capacitance probe can be used to drive a motorized valve. As the level
rises, the motorized valve is driven to move in a direction of closing. As the level falls, the motorized valve is driven open.
Page 5
The need for 2-stage Refrigeration Systems
Lower evaporator temperatures

requires lower evaporator pressures
leading to increased compressor compression ratios
limitations of specific compression technologies
increased refrigerant discharge superheat
© 2003 Reindl
Page 6
Two Stage
(single temperature, two-stage compression, single stage liquid expansion)
2 Evaporative
Evaporator Condenser
Intercooler High Stage
T
Compressor
7
6 1
Fill float
Equalizer line
3
LPA 5
4’ 4 King
5’5’ Fill float
5’’
Low Stage Hi Pressure Receiver
Compressor
© 2003 Reindl
The above figure illustrates a refrigeration system consisting of a single temperature two stage compression with a single stage liquid
expansion. The single temperature-level evaporator is configured in a mechanically-pumped overfeed arrangement. Other evaporator
configurations are possible as are multiplicity of evaporators.
Operation:
The high-side of the system is virtually identical to the single stage systems already discussed. Starting from the compressor suction
(state 1), intermediate pressure vapor refrigerant is drawn into the compressor and raised in pressure. High pressure superheated
gaseous refrigerant leaves the compressor (state 2) and travels to the evaporative condenser. The evaporative condenser rejects heat
and liquefies the refrigerant (state 3) storing it in the high pressure receiver. Liquid refrigerant leaving the high pressure receiver
follows one of two paths. One path delivers high pressure liquid refrigerant (make-up) to the low pressure accumulator (LPA). The
flash gas generated (saturated vapor at state 6) in the throttling process returns to the low-stage or booster compressor working
between the lowest pressure in the system and an intermediate pressure in the system. The low pressure saturated liquid (state 5’) is
available to be pumped-out to all low temperature evaporators (state 5’’). All vapor generated on the low-pressure side of this system is
handled by the booster compressor (at state 6). The booster compressor is responsible for raising the pressure of the refrigerant from
the low-side of the system (state 6) to an intermediate pressure (state 7). Since the discharge state of the refrigerant from the booster
compressor is superheated (even though it is at an intermediate pressure), it must be cooled (or desuperheated) prior to raising its
pressure from the intermediate system pressure to the system condensing pressure.
The job of desuperheating the discharge gas from the booster compressor(s) is accomplished by the intercooler. In other words, the
total heat of rejection on the low-temperature side of the system is imposed or becomes a load on the intercooler. Cool liquid
refrigerant (state 4’) in the intercooler absorbs that total heat of rejection and boils to meet the heat load placed on it by the boosters.
The level of liquid refrigerant in the intercooler is maintained by throttling additional liquid in from the high-pressure receiver. The high-
stage compressor draws intermediate pressure saturated vapor (state 1) from the intercooler and raises its pressure to the system
condensing pressure. In the system arrangement shown above, there are two sources of flash gas generation in the intercooler. The
first is a consequence of dumping the total heat of rejection from the low-side of the system into the intercooler and the second is a
result of throttling high-pressure liquid into the intermediate pressure of the intercooler.
Comments:
In the above system arrangement, we pay a penalty for throttling the high-pressure liquid directly to the low-side of the system. An
alternative would be to first throttle liquid into the intercooler then throttle liquid from the intercooler to the LPR. This can significantly
improve system efficiency for low pressure (temperature) systems since the liquid being throttled to the LPR is precooled as a result of
originating from the intercooler rather than the high-pressure receiver. The next system diagram shows this change.
Page 7
Two Stage
(single temperature, two-stage compression, two stage liquid expansion)
2 Evaporative
Evaporator Condenser
Flash
Intercooler
T
7 1
6
High Stage
Equalizer line
Compressor 3
5
Fill float
LPA
King
5’ Fill float 4’
4
5’’
Low Stage Hi Pressure Receiver
Compressor
© 2003 Reindl
The above figure illustrates a system with single temperature two stage compression and two stage direct liquid expansion. The
system operation is similar to that discussed for the single liquid expansion with one important exception. All the liquid make-up to the
low-pressure accumulator (LPA) now comes from the intercooler. This precooled liquid reduces the fraction of liquid that flashes to
vapor in the LPA and improves system efficiency.
Page 8
Two Stage System
(two temperature level, two-stage compression, two stage liquid expansion)
Hi Temp 2 Evaporative
Lo Temp Evaporator Condenser
Evaporator
T
7 Fill 1
6 float
Equalizer line
5 High Stage 3
Compressor
LPA
King
5’ Fill float 4’
4
5’’
Low Stage Direct Hi Pressure Receiver
Compressor Flash
Intercooler
© 2003 Reindl
The above figure illustrates a two-temperature level two stage compression with a two stage direct liquid expansion. The system
operation is similar to that discussed for the single temperature two-stage liquid expansion system but in this case, we have
refrigeration requirements at the intermediate pressure (temperature) of the system. The evaporators could be configured as DX,
flooded, or overfeed in this arrangement.
Page 9
Two Stage System
(two temperature level, two-stage compression, single stage indirect liquid expansion)
Hi Temp 2 Evaporative
Low Temp Evaporator Condenser
TXV
Evaporator
T
7 1
6
Fill float
Equalizer line
5 High Stage 3
Compressor
LPA
4’
King
Fill float
5’ 4
5’’
Low Stage Indirect Hi Pressure Receiver
Compressor Flash
Intercooler
© 2003 Reindl
The above figure illustrates a two-temperature level, two stage compression with indirect liquid expansion. The system operation is
similar to that discussed for the single temperature two stage direct liquid expansion system but in this case, we have a heat exchanger
immersed below the liquid level in the intercooler. The heat exchanger serves to subcool the high-pressure liquid prior to throttling it to
the low-side of the system. Although this is an improvement over the single stage direct expansion arrangement, it suffers slightly
worse performance when compared to the two-stage direct case shown previously. The principle reason for slightly worse performance
lies in the fact that the liquid being expanded to the low temperature evaporator will not be as cool as if liquid was taken directly from
the intercooler (due to limitations of the heat exchanger). As a result, the indirect liquid expansion system will have a greater
percentage of flash gas relative to the direct liquid expansion case. The principle advantage of the indirect liquid expansion case lies in
the greater degree of subcooling obtained in the high pressure liquid. This subcooling will counter pressure drops in long liquid line
runs. Again, the evaporators could be configured as DX, flooded, or overfeed in this arrangement.
A significant disadvantage of an indirect-intercooler is that repairing a leak in the heat exchanger is virtually impossible (not to mention
the havoc it wreaks on your system when that heat exchanger springs a leak). Although heat exchanger leaks for indirect intercoolers
are not a frequent occurrence, they have happened. Another possible method to accomplish subcooling the high press liquid is to
configure an external subcooling heat exchanger (consider it an evaporator). The high pressure liquid flows through the tube side and
the cold liquid from the intercooler flows on the shell-side. In this arrangement, repairing or replacing the heat exchanger can be
accomplished without scrapping the intercooler vessel. It is a nice method for generating high pressure subcooled liquid.
Page 10
Let’s Look at Performance Analysis
© 2003 Reindl
Page 11
Is your system “efficient”?
Compared to what?
© 2003 Reindl
A challenge in evaluating energy efficiency for industrial refrigeration systems is the fact that each one is a
custom-engineered field-erected system. This makes cataloging system performance impossible. As a result,
evaluating their performance is difficult and challenging. We will look at some of the factors that influence the
performance of industrial refrigeration systems as a preface to investigating opportunities to improve their
efficiency.
Page 12
Performance Analysis Helps You:
Compare a system's efficiency to others
Determine the cost of refrigeration
Assess a system's ability to meet added loads
Quantify benefits of system modifications
Verify predicted performance was achieved
© 2003 Reindl
Page 13
Measures of Performance
Efficiency
COP, BHP/ton
Capacity or productivity
tons of product manufactured per hr or per day
Annual energy cost
$ per year
Normalized energy cost
$ per ton of product manufactured
© 2003 Reindl
There are a number of possible methods and quantities that can be organized to quantify the performance of a
refrigeration system. In some cases, it may make sense to quantify system performance based on an efficiency
expressed as the total brake horsepower (BHP) divided by the system refrigeration capacity (expressed in units of
tons). In other cases, it may be appropriate to determine the total productivity of a refrigeration system in terms of
the tons of cold or frozen product manufactured per hour or per day. Financial managers might be inclined to look
only at the annual energy cost for refrigeration. As we will see later in this section, this measure alone can be very
deceiving. In a manufacturing environment, it is often preferred to express the annual cost by dividing the energy
cost by the total product manufactured.
Page 14
Efficiency
Engineering definition:
"the ratio of the useful output to the energy
input to a process or machine“
General usage:
the relationship between useful output and
purchased input
Refrigeration:
typically COP or hp per ton
© 2003 Reindl
Whenever you talk about the term “efficiency”, it usually represents what you get out over what you put in. Take a
boiler for example. An efficiency rating of 75% for a boiler represents the useful output (steam flowing at the
required conditions) divided by the amount of energy input (in the form of a fuel such as natural gas).
In general, an efficiency is expressed in terms of a useful output (e.g. number of pizzas produced) divided by the
maximum theoretical output if the line ran at its maximum speed without interruption.
For refrigeration systems, the efficiency is often quantified as the total horsepower input (compressors, condenser
fans, condenser water pumps, refrigerant pumps, and evaporator fans) divided by the actual tons of refrigeration
produced. In general, measuring the refrigeration-related electrical energy in a plant is simple. The difficult
quantity to measure is the tons of refrigeration. Often, other quantities besides “tons of refrigeration” may be
tracked such as “tons of hot dogs”, “millions of pizzas”, etc. This process of selecting a quantity to divide into the
total work input to the system is often called “normalization”. Selecting an appropriate normalizing variable is
often the difference between having a meaningful measure of system performance and having a measure that is
useless.
Page 15
Component vs. System Efficiency
Efficient components does
not necessarily mean an
efficient system
Energy use is primarily

determined by system
configuration and
operation
© 2003 Reindl
You should be interested in SYSTEM performance since the operating cost of refrigeration system is dependent
on the efficiency of all components – not just one. It is possible to select and install the most efficient system
components but that is no guarantee that the operation of those components when put together in a system will be
efficient. As we will see, the interaction of components greatly influences how the system as a whole operates.
Page 16
Instantaneous vs. Long-term
Efficiency
Annual efficiency is not the same as performance
at design conditions
Variations in loads
Compressor part-load performance
Compressor sequencing
Head pressure control
© 2003 Reindl
Many of the components in a refrigeration system have efficiency characteristics that are variable. We have
already looked at how changing head pressure can significantly influence the HP/ton of a compressor. We will
look at the performance of compressors are part-load conditions in a future chapter.
Page 17
Capacity
Possibly the most important performance
measure
measure of useful refrigeration
more capacity means more production capability
Efficiency improvements often increase
capacity
© 2003 Reindl
Page 18
Annual Energy Cost
Bottom-line measure of refrigeration system
efficiency
However,
Comparisons should consider
Weather
Production
Utility rates
© 2003 Reindl
Page 19
Annual Energy Cost Example
800000 1800
1600
Monthly Peak Demand (kW)

700000
Monthly Energy Use (kWh)
1400
600000
1200
500000 1000
800
400000
600
300000
400
200000 200
N-93 M-94 N-94 M-95 N-95 M-96 N-96 M-97 N-97
Month
© 2003 Reindl
We conducted an investigation aimed at improving the energy efficiency of a doughnut factory. Our visit was
prompted by the large increase in energy usage by the plant during the prior four years. When we arrived, we
reviewed energy cost and energy use data during the previous five years. The energy use and demand of
electricity for the facility had in fact increased during the prior four years. In fact, the demand increased from six
hundred kW to well over 1200 kW. The monthly electrical usage increased from approximately 300,000 kWh per
month to over 700,000 kWh/month.
Page 20
Normalized Energy Cost
3000000 $0.0400
Monthly Unit Electric Cost ($/doz)

$0.0350
Monthly Production (doz)
2500000
$0.0300
2000000
$0.0250
1500000 $0.0200
$0.0150
1000000
$0.0100
500000
$0.0050
0 $0.0000
N-93 M-94 N-94 M-95 N-95 M-96 N-96 M-97 N-97
Month
© 2003 Reindl
After pressing the vice president of manufacturing for more details on production, we developed the above graph.
First, the monthly production of doughnuts over the prior four years rose substantially. Over the same time period,
the unit cost ($ per dozen doughnuts) dropped by more than 50%.
Page 21
Factors Affecting Performance
A refrigeration system’s performance is affected by:
Loads and the system’s response to meet the loads
© 2003 Reindl
Page 22
Factors Influencing Refrigeration Loads
Weather
Mix of products
Production rates
Production schedules
Process constraints
Operating procedures
© 2003 Reindl
Page 23
Factors Influencing System Response
Weather
System design
Operating procedures
Equipment performance
© 2003 Reindl
Page 24
Weather Influences on Loads
Weather affects:
cold storage warehouses
refrigerated docks
blast freezing systems in unconditioned production
areas
Weather has minimal effect on:
many food processing applications
industrial process cooling
© 2003 Reindl
Page 25
Weather Influences on the System
Evaporative condensing
sensitive to outside air wet-bulb temperature
lower system head pressure
Air-cooled condensing
sensitive to outside air dry-bulb temperature
higher systems head pressure
© 2003 Reindl
Page 26
Using Results
Determine cost of refrigeration per pound of
product
Track energy use by month
Justify system improvements
Verify savings of installed projects
© 2003 Reindl
Page 27
Benchmark Comparisons -
Refrigerated Warehouses
12,000,000
10,000,000
8,000,000
kWh/yr
6,000,000
4,000,000
2,000,000
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
volume, million cu ft
© 2003 Reindl
Since our refrigeration systems do not come with a “MPG” meter, we are stuck trying to identify appropriate
measures for quantifying system performance. Once we have those measures quantified, how do we know
whether or not the energy use is out of line with what can be expected.
We have been looking hard to start a database of energy use statistics. As this database matures, owners can
quickly ascertain whether or not their facility is performing at a reasonable level. The above plot shows the annual
energy use of refrigerated warehouses according to their size. This plot, by it self, does not provide a completely
clear picture.
Page 28
Benchmark Comparisons -
Refrigerated Warehouses
5.0
4.5
4.0
3.5
kWh/cu ft-yr
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
volume, million cu ft
© 2003 Reindl
Taking the previous data and normalizing the energy use by including the warehouse size leads to the above plot.
In this case, it appears that the annual energy use on a per cubic foot basis is relatively constant at approximately
1.3 kWh/ft3-yr. If you owned the warehouse that consumed 4.5 kWh/ft3-yr, would you be convinced that
opportunities to improve the efficiency of the facilities’ refrigeration system abound?
Page 29
Billing Analysis
Review 1-3 years’ monthly utility bills
Look at demand and energy use profiles
Seasonal, month-to-month, year-to-year
variations
© 2003 Reindl
Page 30
Billing Analysis Example - Warehouse
800,000 2,500
700,000
2,000
600,000
Energy Use [kWh]
500,000
Demand [kW]
1,500
400,000
1,000
300,000
200,000
Energy (kWh) 500
100,000
Demand (kW)
0 -
Nov-99 Feb-00 May-00 Aug-00 Nov-00 Feb-01 May-01 Aug-01
© 2003 Reindl
Page 31
Billing Analysis – Dairy Plant
700,000 3000
600,000
2500
500,000
Consumption [kWh]
2000
Demand [kW]
400,000
1500
300,000
1000
200,000
Consumption
(kWh) 500
100,000
Demand (kW)
- 0
Aug-99 Dec-99 Mar-00 Jun-00 Oct-00 Jan-01 Apr-01 Jul-01 Nov-01
© 2003 Reindl
Page 32
Isolating Refrigerating System Energy
Estimate use for other loads:

lighting
Battery chargers
Air compressors
Production equipment
Underfloor heat
Offices
Consider peak load, partial load and
schedule of operation
© 2003 Reindl
Page 33
End-Use Breakdown
Lights, misc
Air compressor 1%
14% Compressors
21%
Pumps
12%
Heaters
52%
© 2003 Reindl
Page 34
Normalizing Results
Account for variations in weather, production
rates, etc.
For example:
$/ft2·degree day
Energy cost per unit produced
kWh per square foot
© 2003 Reindl
Page 35
Accounting for Weather Influences
Does system performance depend on
weather?
Do outside air conditions influence system
loads?
If “yes” to either question, normalize to
account for weather
Average monthly temperature, degree-days,
other
© 2003 Reindl
Page 36
Using Results
Determine cost of refrigeration per pound of
product
Track energy use by month
Justify system improvements
Verify savings of installed projects
© 2003 Reindl
Page 37
Uncovering Energy Efficiency Opportunities
© 2003 Reindl
Learning Objective:
The objective of this program segment is to provide you with a ideas for energy efficiency
improvements to your systems. At the end of this program segment you should understand the
following:
• the need for baselining the performance of your systems
• Energy efficiency improvement strategies and their impact potential
• importance of following-up after implementation to insure energy improvement targets are met
Upon completion of this program segment you should be able to:
• determine appropriate benchmarks for industrial refrigeration systems
• consider, evaluate, implement, and verify energy conservation measures based on a range of potential
candidate approaches
Page 38
Barriers to Realizing Efficient Systems
“Selecting efficient components

leads to efficient systems – right?”
“I can’t afford to design an efficient system”
“The owner does not want an efficient system”
© 2003 Reindl
As I continue working in the area of energy efficiency and industrial refrigeration systems, I find it interesting to hear all of the
reasons why efficiency improvements are not possible with “my system”. There are a lot of fallacies out there. One owner
that I talked to had a brand new refrigeration system. After looking over the drawings for a short time and talking with the
system operators, I was convinced that the system, in its current state, could stand for some changes to improve its
efficiency. I went back and asked the owner – “In your opinion, what are the strengths of your current system?” One of the
first items stated by the owner was that he had an efficient system. I asked the owner how he knew he had an efficient
system. After thinking for a moment, he responded – “I don’t know but I think the system is pretty efficient”.
This owner shares a lot in common with other owners. In order to assess efficiency of a system, it is absolutely essential to
have performance benchmarks or reference points. Without a reference point, you cannot possibly make accurate
statements about the efficiency of a system. In addition, benchmarking your system is critical to determine whether or not
changes that are made to improve system efficiency have had the desired effect.
Page 39
Barriers to Realizing Efficient Systems
The owner can’t afford an efficient system
Energy is too cheap – not cost-effective
“I don’t know how”
© 2003 Reindl
Page 40
Establishing Efficiency Improvement Goals
What are the steps?
1. baseline the performance of the current system
2. compare baseline w/benchmarks (competitors, published, …)
3. identify potential opportunities
4. estimate impact/benefit to system
5. estimate capital cost
6. assess operational risks/constraints
7. prioritize by greatest benefit to cost
8. follow-up, quantify benefit, & continuously improve
© 2003 Reindl
With a solid understanding of the basic system types and their operation as well as utility rates and benchmarking, your are
now ready to take the next step. The next step is to pursue, identify, and implement energy conservation measures (ECMs).
Upon implementing ECMs, you should follow-up and verify that the savings you have projected are actually being realized.
This is a fundamental element of any quality improvement process.
Page 41
Where are the opportunities?
Floating head pressure control Companion
Strategies
Oversize evaporative condenser
Raise suction pressure
Compressor selection & operational sequencing
Intercooler pressure reset (two-stage systems)
Convert gas-pumped systems to mechanically-pumped
© 2003 Reindl
Let’s look at a long list of potential energy conservation measures. Many of the results and analyses presented in
this section are based on past projects that we have tackled. Some of the energy conservation measures are very
cost-effective to implement while others are not so financially attractive. We will not have sufficient time to cover
all of these opportunities in depth. We will cover as many as time allows.
The first two strategies above: floating head pressure and oversized evaporative condensers can be considered
“companion strategies”. For example, it would be foolish to install an oversized evaporative condenser(s) and not
lower the system’s head pressure.
Page 42
Eliminate or reduce parasitic losses/gains
Defrosting (improved methods & optimized strategies)
Reduce envelope & infiltration loads
Break-out suction levels (eliminate EPRs)
Secondary fluid system improvements
High efficiency motors
Liquid/Flash gas management
© 2003 Reindl
Page 43
Single stage vs. multi-stage systems
Thermal energy storage (passive & active)
Better system integration
Pipe sizing & valve selection (esp. suction side)
Maintenance-related issues
Suction gas desuperheating
Liquid subcooling
Oil cooling strategies
© 2003 Reindl
The remainder of this section will examine each of these energy efficiency improvement strategies in more detail. Review
them all carefully and then take one and proceed with evaluating and implementing the ECM. Finally, follow-up with verifying
that the projected savings are actually being realized.
Page 44
Head Pressure Control
© 2003 Reindl 45
Page 45
How do we control head pressure in industrial

refrigeration systems?
© 2003 Reindl
Recall, system head pressure is a controlled variable. We control head pressure by changing the system’s heat rejection
rate. Increasing the heat rejection rate causes the head pressure to drop while decreasing the heat rejection rate will cause
the head pressure to rise. Systems will have controls and control sequences to maintain system head pressure at a target
level.
Typically, systems have a head pressure control set-point such as 150 psig (a saturation temperature of 84 F). Controls will
modulate the capacity of the heat rejection system in attempts to maintain the head pressure at or near the desired setpoint.
Depending on the available heat rejection capacity, the evaporative condenser may or may not have sufficient capacity to
reject heat from the system and achieve the head pressure setpoint. For example, let’s assume that our evaporative
condenser has an approach temperature (difference in temperature between the saturated condensing refrigerant and the
outside air wet bulb temperature) of 18 F. This means that if the outside air wet bulb is 68 F, the saturated condensing
temperature would be no lower than 86 F (or ~155 psig). In this case, the evaporative condenser would be running at its
maximum capacity and the setpoint condensing pressure would not be satisfied until the outside air wet bulb temperature
drops further.
At low outside air wetbulb temperatures, the evaporative condenser would have excess capacity. In this case, the effective
capacity of the evaporative condenser(s) needs to be modified to maintain the head pressure in a desired range. The
following are methods used for modulating condenser capacity: on/off single speed fan control. two-speed fan control, and
variable speed fans.
For the first two options, a deadband around the desired system head pressure set-point is established e.g. 3 psi. For the
single speed fan option, as the head pressure rises to the upper range of the deadband (in this case 153 psig), the condenser
fan(s) are energized; thereby, increasing the heat rejection capacity and driving the head pressure down until it reaches the
lower deadband limit (147 psig) at which time the condenser fans are cycled off.
With the two-speed fan option, the control strategy is basically the same. From an energy perspective, the two-speed fan
offers benefits as it attempts to maintain head pressure below the upper deadband by first cycling fans on low-speed and if
head pressure continues to rise cycling on fans to high-speed. Since the fan horsepower varies with the cube of the fan
speed, any reduction in fan speed can result in considerable fan horsepower savings.
The third option is a variable speed fans. In this case, the fan speed is modulated (from its minimum to its maximum) to
maintain the desired head pressure. This option offers the benefits of stable head pressures and most efficient condensing
performance; however, it comes at a cost premium over the discrete fan speed options.
Page 46
Our heat rejection system controls head pressure

Evaporative condenser fan controls
on/off (single speed fans)
two-speed fans
variable speed fans
© 2003 Reindl
Page 47
Design Condensing Pressure
Most common design pressure
196 psia [1,351 kPa] (95 F [39 C] saturation temperature)
Alternatives to consider
good for cold climates and situations that allow head
pressure to float during most months of the year
good for moderate to cold climates and system designs that
allow head pressure to float during most months of the
year
© 2003 Reindl
Page 48
Control Options
Single speed fan with on/off control

most common method of head pressure control
need to set cut-in (e.g. 150 psig [1,138 kPa]) & cut-out
pressures (e.g. 145 psig [1,103 kPa])
simple control method but results in
higher energy consumption compared to two-speed or
variable speed
higher maintenance (fan motors & belts)
liquid management problems in multiple condenser systems
© 2003 Reindl
Page 49
Control Options – Cont.
2-Speed fan control
need to set high speed cut-in (e.g. 160 psig [1,207 kPa]),
low-speed cut-in pressure (e.g. 150 psig [1,138 kPa]), and
low-speed cut-out pressure (e.g. 145 psig [1,103 kPa])
relatively simple control method but results in
higher capital cost compared to single speed fan option
lower energy consumption compared to single-speed but slightly higher
energy consumption compared to variable speed
liquid refrigerant management problems with multiple condensers
cycling results in less system transients compared to single speed
sequencing speed controls requires attention
© 2003 Reindl
Page 50
Control Options – Cont.
Variable frequency drive
need to set a target head pressure and fan speed is
modulated to maintain head pressure
a very simple principle and method to implement
highest capital cost alternative
lowest energy consumption control alternative
should modulate all condensers the same in systems with
multiple evaporative condensers
results in smoother system operation with minimal
transients
© 2003 Reindl
Page 51
Condenser Fan Control Map
Strategy Mode 1 Mode 2 Mode 3 Mode 4 Mode 5
Small Motor off on off on
1
Large Motor off off on on
Small Motor off off on
2
Large Motor off on on
Small Motor off on on on
3
Large Motor off off half-speed on
Small Motor off half-speed half-speed on on
4
Large Motor off off half-speed half-speed on
Small Motor off variable speed
5
Large Motor off variable speed
© 2003 Reindl
The above map provides five different strategies that could be used for an evaporative condenser that is equipped with twin
motors, two-speed fans, or variable speed fans. The “modes” are indicative of changes in head pressure (either increasing
as one moves from left to right or decreasing as one moves from right to left).
For example, strategy 3 would work as follows. In mode 1 all fans are off. As the head pressure rises, the system responds
by energizing a small fan motor in attempts to maintain system head pressure. If the head pressure continues to rise and the
setpoint is not satisfied, mode three is initiated by the start of the larger fan motor to half-speed. As the head pressure rises
further, mode 4 dictates that the larger fan motor is tripped to run at high speed. The exact opposite sequence occurs as the
head pressure falls.
Page 52
Condenser Fan Control Options
© 2003 Reindl
The above figure illustrates the required fan energy (expressed as a percentage of full-load fan power) as a function of the
evaporative condenser capacity for the five strategies listed previously. The least efficient option is the on/off control
(strategy 1) while the most efficient option is the variable speed drive option. The two-speed fan option yields nearly all of the
part-load power and capacity benefits of the variable speed option but with much less costly equipment.
Notice that at zero fan power for all options, the capacity of the evaporative condenser is not zero. This is due to the fact that
natural convection will occur drawing air through the condenser coils and rejecting heat yielding about 10% of the
condenser’s heat rejection capacity while the fans are idle. This assumes that the condenser coils are running wet i.e. water
continues to flow over the condenser coils.
Page 53
Floating Head Pressure Control
This strategy:
allows head pressure to drop with decreasing outside
air wet bulb temperature
takes advantage of excess evaporative condenser
capacity during cool outside air conditions
head pressure allowed to drop to a pre-determined
minimum (for example Pcond,min = 110 psig [862 kPa])
© 2003 Reindl
Floating head pressure control is an ECM that often has the greatest cost/benefit when compared with other ECMs for a wide
range of system types. All too often, systems in the field operate with excessively high head pressure leading to higher than
nessary operating costs.
The above items describe the “whats” of floating head pressure control. The thing to realize is that all systems have “floating
head pressure” to some extent. Many systems are controlled to artificially maintain head pressures high (150 psig or higher)
year-’round regardless of outside air temperatures.
An extreme example of this is the ice arena case study that we will discuss. This R22 system had its controls to maintain
head pressure between 220-235 psig year-’round (this corresponds to a condensing temperature of between 108-113 F!).
Not only was this system using an excessive amount of energy, the compressor’s life will be shortened considerably as a
result of being forced to work overtime year-’round. The belts on the evaporative condenser fans as well as the motors had
their lives shortened as well due to the extreme short-cycling of the evaporative condenser fans during cold outside air
temperatures. This system’s head pressure was reset to 150 psig without any adverse problems yielding significant energy
and maintenance cost savings.
Page 54
Consequences of lowering head pressure

increased evaporative condenser energy usage
decreased compressor energy usage
reduced high stage compression (on average)
© 2003 Reindl
Page 55
Benefits
improved system efficiency
~1.3% for each °F reduction in saturated condensing
temperature
increased system capacity

compressor life is prolonged
oil cooling loads decrease
© 2003 Reindl
The above items outline the benefits or the “whys” of floating head pressure control.
As a rule-of-thumb, you can expect that a compressor will realize about a ~1.3% improvement in efficiency
(BHP/ton) for each degree lower in condensing temperature. For example, a compressor operating at 0 F (15.7
psig) saturated suction temperature and 95 F (180 psig) saturated condensing temperature would operate at
about 1.68 BHP/ton. If the condensing temperature were allowed to decrease to 85 F (152 psig), the compressor
would now operate at 1.46 BHP/ton - nearly a 13% improvement in compressor efficiency. The actual compressor
performance enhancement with lower condensing temperatures will depend on the compressor technology
(reciprocating, screw, etc.) and its individual performance character.
The “hows” of implementing floating head pressure control will be discussed next including limitations. Basically,
floating head pressure control is as simple as resetting the condensing pressure control setpoint (every system
has one).
Page 56
Optimum Head Pressure
160
Axial Fan
Toa,wb=78°F
Compressor+Condenser
140
120
100
Power (kW)
Compressor
80
Tcond,opt = 87.1 F
60
40
Condenser
20
0
82 84 86 88 90 92 94 96
© 2003 Reindl Saturated Condensing Temperature (F)
Page 57
300
Centrifugal Fan
Toa,wb=78°F
250
200
Power (kW)
Compressor+Condenser
150
100
Compressor
Tcond,opt = 89.9 F
50 Condenser
0
82 84 86 88 90 92 94 96
© 2003 Reindl Saturated Condensing Temperature (F)
Page 58
Head pressure limits dictated by:
hot gas defrost requirements
setting of defrost relief valves
sizing of hot gas main
condensate management in hot gas main
DX evaporators
most thermostatic expansion valves need at least 75 psig
[517 kPa] differential pressure to function properly
liquid injection oil cooling
check manufacturer’s requirements for TXV pressure
differential
© 2003 Reindl
As with most things, there are limits to lowering system head pressure. We do not want to create problems by trying to
improve the efficiency of our systems. The above items are some of the more common factors constraining or limiting our
ability to lower system head pressure. Keep in mind that these items may not necessarily be unmovable barriers; however,
changes in components or system arrangements may be required to overcome their limiting effects on the system.
Hot Gas Defrost:
Many industrial refrigeration systems utilize hot gaseous refrigerant to defrost evaporators. In cases where defrost relief
valves are installed, a sufficient pressure differential (e.g. 75 psig) across the valve must be created to open the valve. Sizing
of the hot gas main may also impose constraints. If a hot gas main is undersized, hot gas (at a sufficient rate) will not be
delivered to the evaporator without a high differential pressure. For larger size hot gas mains, a much lower differential
pressure will allow adequate flow of hot gas to defrosting evaporator(s). Finally, if condensate is not properly managed in hot
gas mains, hydraulic shock can cause catastrophic failures of hot gas piping on a call for defrost. Also, the condensate
effectively decreases the pipe size causing similar symptoms as an undersized line with regard to head pressure
requirements. All of these deficiencies can be overcome in the long run; however, they do create real barriers to lowering
head pressure in the short run.
DX Evaporators:
In systems that utilize direct-expansion evaporators, a minimum differential pressure is required across the thermostatic
expansion valves (TXV). The minimum pressure differential is dependent on the specific valve selection but is routinely on
the order of 75 psig. When we drop the pressure differential across the TXV below the minimum, we lose controllability of the
valve (control engineers call this “control authority”). What results is an inability to properly modulate refrigerant to the
evaporator. Since our evaporator pressure i.e. downstream of the TXV the pressure is fixed (to satisfy our temperature
requirements for meeting load), the head pressure i.e. upstream of the TXV is limited in its ability to float. If we have DX
evaporators on a dock being controlled to 48 psig (32 F saturation temperature), our minimum head pressure would be in the
range of 123 psig (48 psi+75 psi).
LIOC:
Liquid injection oil cooling (also commonly referred to as SOC or screw oil cooling) utilizes high pressure liquid and expands it
directly (using TXVs) into the compressor to accomplish oil cooling in screw compressors. The TXVs for liquid injection oil
cooling tend to be small and require larger pressure differentials for controllability – usually on the order of 100 psi. This
further limits many screw compressor’s from floating head pressure. One approach for circumventing this limitation is to
utilize thermosiphon oil coolers (TSOC).
Page 59
evaporative condenser selection
oversized evaporative condensers usually result in an
optimum head pressure that depends on outdoor air
temperature (more on this momentarily)
evaporative condenser fan controls
VFD fans are preferred but 2-speed fans yield
considerable benefits
© 2003 Reindl
Evaporative Condenser Selection:

If an evaporative condenser is too small, the system head pressure will rise until its heat rejection capacity is sufficient to
reject the needed heat from the system.
Fan Controls:
Although fan controls themselves do not necessarily limit head pressure, there are methods of fan controls that lead to more
stable and efficient system operation. Two speed condenser fans or variable frequency drive (VFD) fans have better capacity
modulating capability and result in more stable head pressures – leading to more stable system operation. In addition to their
stability, two-speed and VFD controlled fans will result in improved energy due to their better part-load performance as
compared to single speed fans.
Thermosiphon Oil Cooling (TSOC):

TSOC improves compressor efficiency by using a thermosiphon effect coupled with the system’s evaporative condenser to
reject heat from the compressor’s oil.
Page 60
hand expansion valve settings
significantly lowering head pressure will likely require seasonal HEV
adjustments
this constraint can be overcome by the use of motorized valves or
pulse width valves
oil separator sizing
gas driven systems (transfer systems and gas pumpers)
controlled-pressure receiver setpoints
heat recovery
engineering and operations (knowledge and willingness)
© 2003 Reindl
Page 61
Case Study: Madison Ice Arena
New system (1996)
City-owned and operated
Rink is operated year-’round
Capacity of 103 tons
Six compressors: max power = 240 kW
Refrigerants R22 and ethylene glycol
Evaporative condenser
Annual electrical operating cost: $45,600
© 2003 Reindl
Just another case study to illustrate the impacts of head pressure and head pressure control on system performance. This
particular system consists of a built-up refrigeration system that serves a local municipal ice arena. The system consists of
two separate refrigeration circuits using R-22 as the refrigerant. Ethylene glycol moves through a common shell-and-tube
evaporator with two refrigerant direct-expansion refrigerant circuits. Each circuit is served by three Carrier Carlyle
reciprocating compressors (i.e. a rack) operating in parallel.
When we started on the project, the system was approximately one year old (a renovation) and the owner (City of Madison)
was disappointed that operating cost savings had not reached expected levels. We were invited to evaluate the system and
make recommendations to the owner for improving the system performance, efficiency, and operating costs. As we will see
shortly, the single largest opportunity to reduce energy costs in this system involved a head pressure reset.
(NOTE: This case study was conducted by Brownell, K. A. in partial fulfillment of the requirements for a MS degree
in Mechanical Engineering under the direction of Professor’s Reindl, D. T., and Klein, S.A. during 1997-1999.
Portions of the thesis prepared by Brownell titled “Investigation of the Field Performance for Industrial Refrigeration
Systems”, 1998 have been excerpted in this case study. A complete copy of the Brownell thesis is available for
download at: http://www.irc.wisc.edu/publications
Page 62
Case Study: Madison Ice Arena
As-installed, head pressure controlled 220-235 psig
[1,620 – 1,724 kPa]
Proposed: allow condenser pressure to ‘float’ with
varying outdoor temperature
Low pressure limit reset to 150 psig [1,138 kPa]
required change: fan controller setpoint
Advantages
quieter; lower maintenance
21% operating cost savings $9,600/yr
© 2003 Reindl
Page 63
Oversized Evaporative Condensers
© 2003 Reindl 64
Page 64
Oversizing Evaporative Condensers
Design point (historical)
95 F [35 C] saturated condensing temperature (196
psia, 1,351 kPa) at design outdoor air wet bulb
temperature
Selecting a larger condenser allows head
pressure to be lowered during design day
decreases compressor energy consumption
increases condenser fan energy consumption
85 F [29 C] design saturated condensing
temperature is possible, 90 F [32 C] is practical
© 2003 Reindl
Page 65
Design Outside Air Design Conditions
Based on 1997 ASHRAE HOF

Use the design wet bulb/mean coincident dry
bulb data
Three choices of design conditions: 0.4%, 1%, or 2%
Alternatively – could use ASHRAE Extremes
computer program
© 2003 Reindl
In 1997, ASHRAE updated its climatic data that designers commonly use for determining outside air conditions during extreme
weather conditions as a basis for selecting system components. Prior to 1997, designers had three choices of peak weather
conditions that corresponded to 1%, 2.5% and 5% conditions. The % conditions list a temperature that is exceeded only X%
(i.e. either 1%, 2.5% or 5%) of the time during the four summer months (June-September) for the location in question. The
lower the percentage, the higher the outside air condition.
For example consider the design summer weather conditions for Atlanta, GA (ASHRAE 1993 Handbook of Fundamentals).
The first column is the ASHRAE climatic design condition, the second column gives the corresponding design dry bulb
temperature and mean coincident wet bulb (MCWB) temperature, and the last column gives the number of hours in the year
(on average) that the design condition will be exceeded.
%Condition Tdb/MCWB (F) hours exceeded (annually)
1% 94/74 30
2.5% 92/74 75
5% 90/73 150
The new ASHRAE design conditions are based on annual percentiles (rather than just percentiles from the summer months).
The 1, 2.5, and 5% conditions have been replaced by the 0.4, 1, and 2% conditions. ASHRAE made the change in an effort to
better provide design conditions that represent the same probability of occurrence at any location regardless of the distribution
of temperature and humidity conditions throughout the year. In addition to the design dry bulb with mean coincident wet bulb,
ASHRAE has also developed data that gives the design wet bulb with mean coincident dry bulb. The following data is for
Atlanta (ASHRAE 1997 Handbook of Fundamentals).
%Condition Tdb/MCWB (F) WB/MDB (F) hours/yr exceeded
0.4% 93/75 77/88 35
1% 91/74 76/87 88
2% 88/73 75/85 175
Page 66
Example Design Conditions
WB/MDB (F)
0.4% 1% 2%
Location
WB MDB WB MDB WB MDB
Madison, WI 76 86 74 84 72 82
Miami, FL 80 87 79 87 78 86
Phoenix, AZ 76 97 75 96 74 95
San Fran, CA 64 79 63 75 62 72
Portland, OR 69 87 67 84 65 80
© 2003 Reindl
Page 67
Model Selection
Most manufacturers discuss two methods

Evaporator ton method
Heat rejection method
Recommend the heat rejection method
determine the total system heat rejection and
match to appropriate models
© 2003 Reindl
Page 68
Evaporative Condenser Ratings
Manufacturers catalog nominal condenser

ratings
independent of refrigerant type
independent of outside air conditions (wet bulb)
Nominal rating has to be adjusted for
desired design condensing temperature
design outside air wet bulb temperature
© 2003 Reindl
Page 69
Evaporative Condenser Ratings
Nominal Capacity
Capacity =
HRF (Twb, Tcond )
Where:
Capacity is total heat rejection capacity needed
Nominal Capacity is the catalog rated value
HRF is the heat rejection factor
© 2003 Reindl
Page 70
Example
A single stage NH3 refrigeration system has total
refrigeration load of 390 tons in Madison and uses an
FES 23L which generates 398 tons and requires 587 BHP
at its design condition of 0 F suction and 95 F
condensing.
1. Select an Evapco condenser to do the job.

2. How would the condenser size change if the design
condensing temperature were 85 F?
© 2003 Reindl
Page 71
Example
Capacitycondenser,actual = Refrigeration Load + Compressor Work
Btu Btu
Capacity condenser , actual = 390 tons ⋅ 12 ,000 + 587 BHP ⋅ 2545
ton − hr hp − hr
Btu mBh
Capacitycondenser,actual = 6,173,915 ⋅ = 6,174 mBh
hr 1000Btu / hr
© 2003 Reindl
Page 72
Example
HRF –
Tcond = 95 F
Twb = 76 F (0.4% conditions)
HRF = 1.34
Nominal Capacity = HRF ⋅ Capacity = 1.34 ⋅ 6,174 mBh

= 8,273 mBh
© 2003 Reindl
Page 73
Example
From the Evapco catalog:
Model 560 – 8,232 mBh
Model 580 – 8,526 mBh
© 2003 Reindl
Page 74
Example
If the design condensing temperature is lowered

to 85 F, the new compressor selection gives:
At 85 F condensing, the 23L would deliver 410 tons

and require 524 BHP
At 95% of full-load, the compressor would deliver
the required 390 tons and demand 506 BHP
© 2003 Reindl
Page 75
Example
Capacitycondenser ,actual = Refrigeration Load + Compressor Work
Btu Btu
Capacity condenser , actual = 390 tons ⋅ 12 ,000 + 506 BHP ⋅ 2545
ton − hr hp − hr
Btu mBh
Capacitycondenser ,actual = 5,967,770 ⋅ = 5,968 mBh
hr 1000 Btu / hr
© 2003 Reindl
Page 76
Example
HRF –
Tcond = 85 F
Twb = 76 F (0.4% conditions)
HRF = 2.94
Nominal Capacity = HRF ⋅ Capacity = 2.94 ⋅ 5,968 mBh

= 17,546 mBh
© 2003 Reindl
Page 77
Condenser Selection Summary
Design SCT Load Avail. Cap. Nom. Heat

BHP/Tonfull BHP390 ton Rej. Cap.
(F) (tons) (tons) load
(mBh)
95 390 398 1.51 587 8,273
85 390 410 1.28 506 17,546
© 2003 Reindl
Page 78
Case Study: Cold Storage Warehouse
Size 4 Compressors Available
34°F 39,000 (ft²)
Instrumentation
0°F 52,000 (ft²)
600,000 (lbs/day, food) Temp, Pressure,
Type Mass Flow!

ammonia, single-stage Defrost Strategies
compression, liquid
overfeed evaporators Head Pressure Control
Operating Costs
9,000 ($/month)
© 2003 Reindl
The refrigeration system examined as part of this case study is a cold storage warehouse facility located near Milwaukee, WI.
The facility contains four types of refrigerated spaces – low temperature freezer (0 F), cooler (34 F), docks (45 F), and
ripening rooms (45-64 F). From a thermal mass perspective, the warehouse construction type can be considered
“lightweight” for all spaces. There is mostly insulation and very little mass in the walls and roofs.
The freezer and cooler with its loading dock are separate buildings located adjacent to each other. The banana and tomato
ripening rooms are located in a heated space adjacent to the cooler. The refrigerant used throughout this system is ammonia
(R-717). Evaporators in the freezer are top fed, pumped liquid overfeed. Cooler, and cooler dock evaporators are all bottom
feed pumped liquid overfeed where as the evaporators in the banana and tomato ripening rooms are direct expansion
controlled by thermal expansion valves and back pressure regulators.
(NOTE: This case study was conducted by Manske, K. A. in partial fulfillment of the requirements for a MS degree in
Mechanical Engineering under the direction of Professor’s Reindl, D. T., and Klein, S.A. during 1998-1999. Portions
of the thesis prepared by Mankse titled “Performance Optimization of Industrial Refrigeration Systems”, 1999 have
been excerpted for this section. A complete copy of the Manske thesis is available for download at:
http://www.irc.wisc.edu/publications
Page 79
Qreject
HPR
Condenser
PLO
DX Evap 23°F
Evap BPR
Qspace
45-55°F Qspace PLO

Evap
-10°F
Design Loads Yearly Average Loads
Qspace
Fruit Ripening = 90 tons Fruit Ripening = 43 tons
Cooler = 107 tons Cooler = 58 tons
Freezer = 106 tons Freezer = 71 tons
© 2003 Reindl
There are three main vessels in the system as shown above. The first is the high pressure receiver where liquid refrigerant
draining from the condenser is stored. Liquid refrigerant from the high pressure receiver is then throttled to either the
intermediate pressure receiver or to the direct expansion evaporators in the banana and tomato ripening rooms. The
temperature of the refrigerant in the banana/tomato room evaporators is regulated at a desired level by use of a back-
pressure regulator. The back-pressure regulator then throttles the refrigerant gas to the intermediate pressure receiver which
is at a lower temperature/pressure. Liquid in the intermediate pressure receiver is then either pumped to the cooler and
cooler dock evaporators or throttled again to the low pressure receiver. Liquid refrigerant from the low pressure receiver is
pumped to freezer evaporators with a mechanical liquid recirculating pump. Liquid levels in the intermediate and low pressure
receivers are maintained at a near constant level by a pilot operated, modulating expansion valve controlled by a float switch
located on the receiver tank.
A single-screw (Vilter model# VSS 451 connected to the low temperature vessel) and reciprocating compressor (Vilter
model# VMC 4412 connected to the high temperature vessel) operate in parallel, each compressing to a common discharge
header and a single evaporative condenser. The suction line from the low pressure receiver leads to the screw compressor.
The suction line from the intermediate pressure receiver leads to the reciprocating compressor. Additional compressors, in
parallel piping arrangements to the primary compressors, can be brought on-line if the load exceeds the capacity of the
primary compressors.
Page 80
Baseline System Performance
Performance High Low
Combined
Measures Temperature Temperature
ton-hr / ft²-yr 19.4 11.6 15.2
kWh / ton-hr 1.2 1.8 1.4
COP 3.0 1.9 2.4
hp/ton 1.6 2.5 1.9
hp(comp.)/ton 0.9 1.6 1.2
hp(cond.)/ton 0.1 0.2 0.2
hp(evap.)/ton 0.2 0.3 0.2
OnPeak kWh / yr 379107 426176 805271
OffPeak kWh / yr 655582 736493 1392118
Peak kW 194.4 218.4 412.8
$ / ft² -yr $0.92 $0.88 $0.90
$ / cu.ft -yr $0.05 $0.04 $0.04
$ / ton-hr -yr $0.05 $0.08 $0.06
© 2003 Reindl $ / yr $42,158 $47,402 $89,667
The cold storage warehouse considered in this case study is already very efficient. Prior to considering any alternative
operating strategies or system modifications to improve performance, it is important to establish a baseline of the system.
The above table illustrates the performance of the cold storage warehouse system using a number of different metrics
including system COP, hp/ton (for system, compressor, condenser, evaporators), energy costs (total, and on a unit basis) for
the high temperature system, low temperature system, and total warehouse.
The electrical costs are calculated on a monthly schedule from four parameters dealing with electrical usage. The first two
parameters are the on-peak and off-peak energy charges. This is the total amount of electricity [kWh] consumed during on or
off-peak hours. The charge is $0.0327 and $0.0203 per kWh for on and off-peak respectively. The third charge is referred to
as the billing demand charge. This is the peak demand electrical load [kW] in the month that occurs during the on-peak hours
Monday through Friday. The cost for billing demand is $8.24 per kW. The final electrical charge is referred to as the customer
demand charge. This is a monthly charge based on the highest electrical demand reached in the last 12 months of operation.
The customer demand charge is $0.65 per kW. These electric rates are assumed constant for the whole year.
Page 81
Condenser Fan Control Options
© 2003 Reindl
The above figure illustrates the required fan energy (expressed as a percentage of full-load fan power) as a function of the
evaporative condenser capacity for the five strategies listed previously. The least efficient option is the on/off control
(strategy 1) while the most efficient option is the variable speed drive option. The two-speed fan option yields nearly all of the
part-load power and capacity benefits of the variable speed option but with much less costly equipment.
Notice that at zero fan power for all options, the capacity of the evaporative condenser is not zero. This is due to the fact that
natural convection will occur drawing air through across the condenser coils and rejecting heat yielding about 10% of the
condenser’s heat rejection capacity while the fans are idle. This assumes that the condenser coils are running wet i.e. water
continues to flow over the condenser coils.
Page 82
Comparison of Condenser Fan Controls
© 2003 Reindl Source: Manske, K., 2000
Of course we do not want to just minimize the power of the evaporative condenser at the expense of the system;
consequently, we must look at the impacts or tradeoffs associated with spending more energy on evaporative condenser fans
vs. the reduction in compressor power that accrues due to the lower head pressure.
The case study system had an oversized evaporative condenser. As a result, it was possible to drive head pressures
extremely low in the system. So low in fact that the incremental expenditure of fan energy was not compensated for by an
incremental reduction in compressor energy demand.
The above plot shows the comparison between heat rejection system control strategies. The point furthest to the left on the
curves in the figure represents the system balance point head pressure at which the condenser is operating at 100 percent
capacity (for a given outdoor air wet bulb and system load during a peak hour on a average day in May). Any further
decrease in condensing pressure would prevent the condenser from rejecting the required amount of energy from the
system. The figure shows that VFD fan control could save the system nearly 8% in combined compressor and condenser
energy requirements if the head pressure were raised to 125 psia. VFD fan control looses its advantages at low head
pressures because the fans must run at near full speed most of the time anyway. At high head pressures the fans in on/off
control don’t stay on long because of the high rate of heat transfer that occurs. However, at high head pressure an on/off
control strategy would cycle the fans on and off frequently which would cause excessive wear on the motors and fan belts.
The figure also shows that there is a different optimum head pressure for each type of condenser fan control. It is also
interesting to note that half-speed fan motors have energy requirements that are only approximately one percent above the
VFD motors at elevated head pressures. Since this system has a minimum allowed head pressure of 130 psia, VFD and half-
speed motors may have very similar energy requirements for most of the year.
Page 83
Optimum Head Pressure Control
This plot illustrates the preferred control head pressure control strategy for two different evaporative condenser sizes. With
an evaporative condenser sized for 95 F saturated condensing temperature on a design day, the optimum head pressure is
the lowest head pressure achievable by running the evap condenser fans “full out”. If the condenser is oversized (i.e. an
oversized evap condenser is defined as one that yields a saturated condensing temperature of 85 F on the design day),
there is an optimum head pressure (i.e. a head pressure greater than the minimum achievable that will minimize the
combined power of the compressor and condenser). In this case, the optimum head pressure is likely a function of the
outside air wetbulb temperature.
The dark set of lines is for the condenser that is currently installed in the system. The current condenser is large enough to
allow the system to balance out with a saturated condensing temperature of 85°F on the design day. The
compressor/condenser power with a smaller condenser is given by the lighter colored line. The point furthest to the left on
each line represents the pressure at which the evaporative condenser has reached 100 percent capacity. Given that the load
is constant, it would be physically impossible to achieve a lower head pressure without adding additional condensing
capacity. Note, the above case assumes that the refrigeration load is progressively decreasing during the winter months;
however, refrigeration load has little influence on the optimum head pressure.
Because of the presence of high temperature direct-expansion coils in the case study system, the head pressure is not
allowed to go below 130 psia. Therefore, the system cannot possibly be operated at its ideal head pressure except for the
months of June through September. It must be operated above its optimum head pressure resulting in a slight excess of
compressor power.
Page 84
6
230 3.4x 10
Curve Fit (Variable Evaporator Load)
6
Total System Heat Rejection [Btu/hr]

220 Calculated Condenser Heat Rejection (Variable Evaporator Load) 3.2x 10
Optimum Head Pressure [psia]
210 Calculated Condenser Heat Rejection (Constant Evaporator Load) 6

3.0x 10
Calculated Ideal Head Pressure (Variable Evaporator Load)
200 6
Calculated Ideal Head Pressure (Constant Evaporator Load) 2.8x 10
190 6
2.6x 10
180
6
2.5x 10
170
6
2.3x 10
160
6
2.1x 10
150
6
1.9x 10
140
minimum head pressure 6
130 1.7x 10
as required by dx txv
6
120 1.5x 10
50 55 60 65 70 75 80
© 2003 Reindl Outside Air Wet Bulb Temperature [°F] Source: Manske, K., 2000
When performing the calculations to identify the optimum condensing pressure for the year, we discovered that the optimum
condensing pressure had a near linear relationship with the outside air wet bulb temperature. The above curve illustrates the
relationship between optimum head pressure and outside air wetbulb temperature (lower curve) over a range of evaporator
load conditions (corresponding variability in heat rejection is shown by the points above). In the case of this system, a very
simple linear relationship was developed that allows a supervisory reset on the system head pressure given the prevailing
outside air wet bulb temperature according to the following:
Phead,opt = -27.6 + 2.55 * Twb
where Phead,opt is the head pressure corresponding to minimum system power in psia and Twb is the outside air wet bulb
temperature in F. This relationship assumes that the condensers have variable speed drives. Keep in mind that the above
relationship needs to have a lower bound as dictated by the characteristics of each given system.
Page 85
Optimizing Head Pressure
1. Measure the outdoor air wet bulb temperature
2. Note the current condensing pressure and system electrical demand
3. Reset the condensing pressure down 5 psig (35 kPa) & allow system to equilibrate
4. Note the new system electrical demand
5. Continue steps 3 and 4 until the lower condensing pressure limit set point is reached
6. Plot the system electrical demand vs. the condensing pressure and note the
condensing pressure corresponding to point of minimum system electrical demand
7. Plot that single “optimum” condensing pressure point on a optimum condensing
pressure vs. outdoor air wet bulb temperature curve
8. Repeat the procedure from 1-7 to more fully develop a curve analogous to the figure
given on the previous page.
© 2003 Reindl
Procedure for Determining Optimum Relation Between Condensing Pressure and Outdoor Wetbulb
The trajectory of optimum condensing pressures for corresponding outside air wet bulb temperatures as shown on the
previous page is specific to the existing ammonia system. Each system will have its own unique trajectory. However, the
following procedure can be used to empirically develop the trajectory of optimum condensing pressures. Note, this procedure
needs to be executed during off-design periods of the year (during relatively lower outside air wet bulb conditions). The
procedure also requires the ability to continuously monitor the outdoor air wet bulb temperature, condensing pressure, and
the engine room total electrical demand. We also recommend that other system state variables (such as suction pressures,
superheat – if applicable, etc.) be monitored to ensure reliable system operation during the procedure.
1. Measure the outdoor air wet bulb temperature
2. Note the current condensing pressure and system electrical demand
3. Reset the condensing pressure down 5 psig (35 kPa) and allow the system to equilibrate
4. Note the new system electrical demand
5. Continue steps 3 and 4 until the lower limit in condensing pressure setpoint is reached
6. Plot the system electrical demand vs. the condensing pressure and note the condensing pressure that corresponds to the
point of minimum system electrical demand
7. Plot that single “optimum” condensing pressure point on a optimum condensing pressure vs. outdoor air wet bulb
temperature curve
8. Repeat the procedure from 1-7 to more fully develop a curve analogous to the figure given on the previous page.
Once the optimum condensing pressure trajectory curve is developed, it can be programmed into a system PLC or
supervisory controller to yield optimum system performance throughout the year. Bear in mind that the procedures 1-6 above
need to be executed in a relatively short period of time (1-2 hrs) as the outside air wet bulb will change throughout the day. In
general, the outside air wet bulb temperature has a daily range of between 7-10°F (4 – 5.5°C). Step 3 above is important.
The period to achieve equilibrium operation will be longer for larger systems (on the order of tens of minutes). Finally,
constrain the condensing pressure from dropping below a lower limit that will degrade the operation of a system (due to
expansion valves, hot gas defrost, etc.).
Page 86
Current System - On/Off Condenser Fan
Current System - VFD Condenser Fan
Control, Minimum Floating Head
Performance Control, Ideal Floating Head Pressure
Pressure
Measures
High Low High Low
Combined Combined
Temperature Temperature Temperature Temperature
ton-hr / ft²-yr 19.4 11.6 15.2 19.4 11.6 15.2
kWh / ton-hr 1.2 1.8 1.4 1.1 1.8 1.4
COP 3.0 1.9 2.4 3.2 2.0 2.6
hp/ton 1.6 2.5 1.9 1.5 2.4 1.8
hp(comp.)/ton 0.9 1.6 1.2 0.9 1.5 1.1
hp(cond.)/ton 0.1 0.2 0.2 0.1 0.1 0.1
hp(evap.)/ton 0.2 0.3 0.2 0.2 0.3 0.2
OnPeak kWh / yr 379107 426176 805271 361974 407223 769193
OffPeak kWh / yr 655582 736493 1392118 626423 704629 1331056
Peak kW 194.4 218.4 412.8 187.7 211.6 399.3
$ / ft² -yr $0.92 $0.88 $0.90 $0.88 $0.84 $0.86
$ / cu.ft -yr $0.05 $0.04 $0.04 $0.05 $0.03 $0.04
$ / ton-hr -yr $0.05 $0.08 $0.06 $0.05 $0.07 $0.06
$ / yr $42,158 $47,402 $89,667 $40,319 $45,412 $85,811
A performance comparison between the way the system is currently operated and if optimized head pressure and VFD
condenser fan control were used is shown in the above table. Optimized head pressure control along with VFD condenser
fan control would save 97,140 kWh or $3,856 per year in electrical operating costs.
Page 87
Floating Head Pressure Summary
Fixed head pressure is simple but expensive
Optimized head pressure is usually a function of
outdoor wet bulb temperature
Controlling to optimum head pressure with “over-sized”
evap. condensers
Condensers w/VFD’s or 2-speed fans offer energy
savings and more stable systems
© 2003 Reindl
Page 88
Compressor Selection & Sequencing
© 2003 Reindl 89
Page 89
Compressor Selection & Operation
Recognize advantages, disadvantages, and limitations
of compressor selections
Make wise choices for fixed Vi screw compressors in
high-stage or single stage systems
Recips vs. screws?
Lead screw and lag recip. or lead recip. and lag screw?
Recognize part-load characteristics of compressors
© 2003 Reindl
For industrial refrigeration applications, the two most commonly used compression technologies are reciprocating and screw
compressors. Today, the market has taken-up and gravitated toward selecting, installing, and operating screw compressors
in industrial refrigeration in lieu of reciprocating compressors. Although there are a number of advantages that screw
compressors afford, many have overlooked their drawbacks (at the expense of increased system operating costs). It is
important to consider the operation and sequencing of compressors in a SYSTEMS context to maximize the performance and
energy efficiency of the compressors in the system.
Page 90
Screw Compressor Volume Ratios
6.5
6.0
o
ati
nR
Volume or Compression Ratio
5.5 s sio
re
omp
5.0 C
4.5
4.0
atio
me R
3.5 Volu
3.0
k = 1.37
2.5
Saturated Suction Temperature = 0 F
2.0
120 140 160 180 200
© 2003 Reindl
Discharge Pressure (psia)
Page 91
Required Volume Ratio
© 2003 Reindl
Page 92
Fixed Volume Ratio Efficiency
© 2003 Reindl
Page 93
© 2003 Reindl
Page 94
© 2003 Reindl
Page 95
Fixed vs. Variable Volume Ratio
© 2003 Reindl
Page 96
Selection Considerations
1. Expected range of operating suction & discharge pressures
a) single stage or two stage operation (booster or high-stage)
b) swing duty (boosters operating as a single stage)
c) load variability over time (large pull-down loads vs. relatively
constant loads)
2. Climate type and system minimum head pressure
constraints
3. Oil separator sizing/selection
4. Oil cooling methods
5. System and package losses for check valves, service
valves, strainers installed around the compressor
6. Expected maintenance costs over machine’s life
© 2003 Reindl
Page 97
Fixed Vi Selection Ranges
Head Pressure Range1
Sat. Suction
Temperature High Medium Low
[°F] 180 – 100 psig 180 – 115 psig 180 – 135 psig
(95 - 65°F SCT) (95 – 70°F SCT) (95 - 80°F SCT)
-40 5.0 or higher 5.0 or higher 5.0 or higher
-20 3.5 – 5.0 3.5 – 5.0 4.0 – 5.0
0 2.5 – 3.5 2.7 – 3.5 3.0 – 4.0
20 1.5 – 2.7 1.7 – 3.0 2.0 – 3.5
40 1.4 – 2.5 1.5 – 2.7 1.5 – 3.0
© 2003 Reindl
Page 98
Frequency Analysis for Madison
4000
Madison, WI
3500
3000
2500
Hours per Year
Dry Operation
Wet Operation
2000
1500
1000
500
0
63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95
Saturated Condensing Temperature [F]
© 2003 Reindl
Page 99
Typical Part-Load Compressor Performance
©Source: Manske, K. et al., 2000

2003 Reindl
Large screw and reciprocating compressors typically have the capability for reducing their capacity to match the required
refrigeration demand by the system. Screw compressors typically accomplish this task by the use of a slide valve that,
effectively, changes the point where the compression process begins along the axis of the screw. Most screw compressors
have the ability to continuously modulate capacity between 10 to 100% of its available full load capacity. Some screw
compressors are also equipped with variable speed drives that gives further flexibility to their part-load operation – these are
not considered in the following analysis.
Reciprocating compressors can be equipped with cylinder unloaders. Unloaders consist of hydraulically or electrically-
actuated push rods that hold open suction valves on individual or groups of cylinders. By holding open the suction valves,
the number of cylinders that are providing active gas compression is reduced; thereby, reducing the compressor’s capacity.
As screw compressors are unloaded, their power and oil cooling requirements decrease, but not necessarily in direct
proportion to capacity. Reciprocating compressors tend to unload more linearly. Unloading curves for both the screw and
multi-cylinder reciprocating compressors are shown in the figure above. These curves give the fraction of full load power
(%FLP) the compressor will require when operated at a specific percent of its full load capacity (%FLC) or part-load ratio.
The part-load characteristic for the reciprocating compressor does not pass through the origin of the above figure because of
the additional compressor power requirement (approx. 3%) to overcome parasitic losses associated with frictional and
windage effects on the unloaded cylinders. The part-load characteristics for both compressors were obtained from the
compressor manufacturer, Vilter. These characteristics are expected to be representative, but the characteristics of
compressors from different manufacturers may differ from those above due to different compressor designs and
manufacturing processes.
Note, the above unloading curve and subsequent analysis does not consider economized operation.
Page 100
Compressor Part-Load Efficiency Comparison
Source: Manske, K. et al., 2000

© 2003 Reindl
If multiple compressors are used to meet the refrigeration load, it is desirable to operate the compressors at the lowest
combined power while still meeting the system loads. In refrigeration systems with variable loads, the delivered capacity of
the compressors must be modulated by unloading the compressors in order to balance the compressor(s) capacity with the
refrigeration demands of the system. Each compressor, depending upon type and manufacturer, may have a different
unloading characteristic. The following results are strictly valid only for the particular screw and reciprocating compressors
investigated in this study; however, the general concepts can be applied to all refrigeration systems with multiple
compressors.
A theoretical refrigeration system utilizing two compressors operating in parallel was used to explore optimum compressor
operation. Compressor performance can be characterized in terms of specific power, i.e., the dimensionless ratio of the
compressor power to refrigeration capacity at a particular set of operating conditions (saturated suction temperature,
saturated discharge temperature, and part-load ratio). The specific capacity is the inverse of the coefficient of performance
(COP). The figure above shows a performance comparison, in terms of the compressor specific power, between the screw
and reciprocating compressors for several different saturated suction temperatures over a range of part-load conditions
assuming a fixed saturated discharge temperature of 29.4°C (85°F). Reciprocating compressors unload nearly linearly and
their performance curve should, theoretically, be flat for a fixed suction temperature. The slight increasing trend
(corresponding to a decrease in compressor performance) from left to right in the figure for the reciprocating compressor is a
result of increasing pressure drop in the dry suction line due to increasing refrigerant mass flow rate. The additional (3%)
increase in total compressor power discussed above also contributes to the increasing trend in specific power. Several
observations can be made about compressor operation from the figure.
• A single screw compressor unloaded to 25 percent of its full load capacity has nearly a 50 percent increase in specific
power when compared to a reciprocating compressor.
• Screw compressors perform better than reciprocating compressors when operated near full load. The performance
advantage increases as the suction pressure drops.
• Reciprocating compressors are better suited in refrigeration systems where significant unloading, i.e. load following, is
required.
• From an energy standpoint, is more important to size screw compressors correctly as compared to multi-cylinder
reciprocating compressors.
Page 101
Compressor Part-Load Efficiency
A screw compressor unloaded to 25% of its full load
capacity has nearly 50% higher BHP/ton vs. recip.
Multi-cylinder reciprocating compressors unload nearly
linearly
Screw compressors perform better than reciprocating
compressors when operated near full load
Unloading penalty for screw compressors is greater
with increasing pressure lifts
© 2003 Reindl
Page 102
Refrigeration Load Allocation
Assuming we have all screw compressors, how should

we distribute or allocate the refrigeration load?
© 2003 Reindl
Page 103
Efficiency of Two Compressor Operation
Source: Manske, K., et al., 2000
© 2003 Reindl
The above plot assumes that two 110 ton compressors are available to meet varying proportions of a 220 ton peak
refrigeration load. The saturated suction temperature is 0 F and the saturated condensing temperature is 85 F. The part-
load characteristics are based on the Vilter single screw.
The following can be observed from the above plot. If the total system load is greater than 60% of the full load (0.6*220= 132
tons), it is preferable to fully load one of the compressors and use the other compressor to “load-follow”. If the load is less
than 60% of the design full load, it is preferable to equally unload the two compressors. This behavior is dictated by the
unloading characteristic of the screw compressor and, in this case, independent of the condensing temperature.
Page 104
Efficiency of Two Compressor Operation
©Source:
2003 Manske,
Reindl K., et al., 2000
Screw compressors unload non-linearly and their parallel operation must be treated quite differently compared to
reciprocating compressors. The figure above shows a plot of the aggregate specific power for a system with two equally
sized screw compressors operating in parallel. Each separate line on the plot represents a different total system load. The
system load is expressed in terms of the system part load ratio (PLR) defined as the ratio of the actual delivered system
capacity to the total available capacity of both compressors at full load. Starting at the far right of the plot for a given load,
one compressor is fully loaded and the other compressor is operating at part load such that the combined capacity of the two
compressors matches the total system load. The abscissa is the ratio of the capacity of the lead (more heavily loaded)
compressor to the capacity of lag (less heavily loaded) compressor. By progressing from right to left along a constant system
load line, the capacity of the lead compressor is reduced while that for the lag compressor is increased.
The figure shows that, for part load ratios below approximately 0.65, optimal operation results when the load is split equally
between the two compressors (a compressor capacity ratio of 1). When the system has a part load ratio above 0.65, the
system performance is optimized when one compressor is fully loaded and the remaining compressor is loaded to make up
the difference. This behavior can be explained by the previously-shown part-load characteristics, which illustrates that the
specific power of an individual screw compressor begins to increase very rapidly if it is operated with a compressor part load
ratio below 0.5. The values of the aggregate specific power shown above depend on several factors, including the
compressor performance characteristics and pressure drop in the compressor suction lines. However, the conclusions
relating on how to optimize the combined operation of the two compressors are independent of these factors.
Page 105
Efficiency of Unequally Sized Compressors
Source: Manske, K., et al., 2000

© 2003 Reindl
The above example illustrates how unequally-sized compressors function to meet system loads. At system loads equal to or
greater than about 72% of design (209 tons or greater), it is preferable to fully load the larger screw compressor and use the
smaller compressor to load-follow. As the system load continues to drop below 72%, it is advantageous to fully load the
smaller compressor and begin unloading the larger; however, if the system load drops below about 58% (197 tons) both
screw compressors should operate at equal part-load ratios.
Page 106
Variable Speed Screw Compressor
55
FES 315S Booster Compressor
CF Industries
50
Albany Terminal
Compressor Power [kWE]
Compressor C-2
45
Fixed Vi=2.6
Fixed Speed
July 17, 2003
40
35
30 Variable Speed
25
20
15 kw=9.55645 + 15.576·PLRCalculated + 26.2308·PLRCalculated2
kw=22.5113 + 11.5401·PLRCalculated + 16.0278·PLRCalculated2
10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
© 2003 Reindl Part-Load Ratio
Page 107
Compressor Operation Conclusions
When screws & recips are used, unload recip
first and screw last
Always try to operate screw compressors at
part-load ratios greater than 50%
Remember operating systems with unequal
sized compressors differs from systems with
equally sized compressors
© 2003 Reindl
Compressors in industrial refrigeration systems regularly operate at part-load conditions. Part-load compressor operation is
required in order to balance the refrigeration capacity of the compressors with the refrigeration demand from the system.
Reciprocating compressors have near-linear unloading curves and therefore introduce relatively small performance penalties
when operated at low part-load ratios. This is not the case however for screw compressors. As screw compressors are
unloaded they require more power per unit of cooling capacity. It is recommended that screw compressors be sized
appropriately so they can be operated at or near full capacity as much as possible. Screw compressors should be used for
base loading and reciprocating compressors should be used to meet the transient portion of a varying load.
If two screw compressors are sharing a load, there exists a point where it is better to fully load one compressor rather than
split the load equally. In the case of two equally sized screw compressors, the optimal situation occurs when the compressors
share the load up to an identifiable crossover point which occurs when the load on the system is about 66 percent of the
combined available capacity of the compressors. At loads above that point it is best to fully load one of the screws and make
up the difference with the other. The crossover point is a characteristic of type and size of compressors used. When load
sharing between two unequal sized screw compressors is required, it is best to first fully load the smaller of the two, then at a
certain identifiable crossover point, fully load the larger of the two compressors and make up the difference with the smaller of
the two. Calculations with unequal sized screw compressors (the larger compressor has 64% greater capacity compared to
the smaller compressor), indicated that this cross-over point occurred when the load was 76 percent of the available capacity
of both compressors. Crossover points are independent of compressor suction and discharge conditions.
Page 108
Intercooler Pressure Reset
© 2003 Reindl 109
Page 109
Intercooler Pressure Reset
How do we determine intercooler pressure?
requirements of “high” temperature loads
equal compression ratio on boosters and high
stage machines
designer’s selection of intermediate temperature
other strategies
© 2003 Reindl
Page 110
Determining Intermediate Temperature
A commonly applied rule-of-thumb for

approximating an optimum intermediate
pressure level is given by:
Pint = Pcond ⋅ Pevap
where all pressures are expressed as absolute pressures

© 2003 Reindl
Pint = Intercooler pressure (psia)

Pcond = Condensing pressure (psia)
Pevap = Low temperature evaporator pressure (psia)
Keep in mind that the pressures in the above relationship need to be in absolute terms. Recall the conversion between
absolute and gage pressures:
Pabs = Pgage + Patm
Where Patm is the local atmospheric pressure (14.7 psi at sea level).
Page 111
Determining Intermediate Pressure
As head pressure floats, the optimum intermediate

pressure changes
Many systems establish their interstage pressure and
leave it fixed
this strategy can lead to significant energy penalties on an
annual basis
alternative strategies aim to reset the intermediate pressure
The previous relationship is an approximation, how
good is it?
© 2003 Reindl
Page 112
Case Study
Anhydrous ammonia
storage terminal
system charge – 30,000 tons
low stage pressure ~0.5 psig – critical control pt
intermediate stage pressure: 55 psig
a total of 560 HP of compressors
G
© 2003 Reindl
Page 113
Evaporative
Condenser
Evaporative
Condenser
KO Drum
HS-1 HS-2 HS-3

BC-1 BC-2 BC-3
Low Stage High Stage
Compressor Compressors
Intercooler
To Purge
Hi Press. Rec
Two Stage Compression with

82 kW heater Single Temperature Level
© 2003 Reindl DTR, 8/27/00
Page 114
Intercooler Pressure
50
Pcond = 175 psig
48
System Power [hp]
Pcond = 150 psig
46
Pcond=125 psig
44
42
40
35 40 45 50 55 60 65 70 75 80
© 2003 Reindl Intercooler Pressure [psia]
The above plot illustrates the optimum interstage pressure for the two-stage compression single temperature level
system. The curves illustrate the total system horsepower requirements as the intercooler pressure is reset in a
range from 40 – 80 psia based on a relatively simple mathematical model of the system. The model of this
particular system neglects piping pressure drops (piping pressure drop effects will influence the optimum
interstage pressure as will be illustrated in the section on single stage vs. two stage).
Each curve in the above plot represents a fixed condensing pressure range from 125 psig to 175 psig. The solid
dots show the points of optimum intercooler pressure as estimated by the square root relationship previously
presented. As is apparent, the square root relationship provides good estimates of the true optimum. Note that
the best operating strategy involves resetting the intercooler pressure as head pressure changes.
Page 115
The Need for Multi-Stage Systems
High compression ratios
recip compressor limits: ~8:1
screw compressor limits: ~18:1
High fixed head pressures

floating head pressure decreases the attractiveness
of multi-stage systems
Loads at multiple temperature levels

© 2003 Reindl
As we operate systems at lower and lower suction pressures (to achieve lower temperature operation), we eventually reach
limits of the compression devices in our systems. Reciprocating compressors have a compression ratio limit of approximately
8:1 which means that if our design condensing temperature was 196 psia (95 F), our lowest possible evaporator
pressure/temperature for a single stage system would be 24.5 psia (about –8.5 F). In the case that our application required
colder temperatures, staging would be required.
It is widely stated that the two stage system is more efficient than a single stage system. Although this is true in a number of
circumstances, its applicability needs to be carefully evaluated for the application and control setpoints in a proposed or
actual system. The question is: when does 2-stage make sense (from an energy efficiency standpoint)? The corollary to this
question is: When should I operate a two stage system as a single stage? This is what we would call a “swing” system.
Page 116
Case Study: -10 F System
Percent Savings of Compressor Horsepower
10.0
8.0 Actual Simulation Results

over Single-Stage Compression
High-Stage - Recip.
6.0
Low-Stage - Screw
4.0
2.0
0.0
-2.0
-4.0 Two-staging will save on compressor

power above a temp. of 80°F.
-6.0
-8.0
-10.0
65 70 75 80 85 90 95 100
Minimum Saturated Condensing Temperature Allowed [°F]
We considered the operation of a two stage system with a low stage suction temperature of –10 F (pretty much on the edge if
we were to use recips for a single stage application). The above plot represents a system that uses screw compressors
as the boosters and recip compressors for the high stage. The low stage suction pressure is fixed at -10 F with the
intermediate stage temperature fixed at 23 F. Interestingly, at minimum saturated condensing temperature below 80 F, a
two stage system will consume more energy than a single stage system. This is due to a few contributing factors.
1. The second stage includes additional piping that imposes suction line pressure drops (approximately twice the aggregate
pressure drop as compared to a single stage system).
2. The above system has a fixed intermediate temperature that will not be optimum as the condensing temperature
changes over the course of the year.
Page 117
Case Study: -10 F System
20
Percent Savings of Compressor Horsepower
2 stage reciprocating compressors full load
15 economizer screw s ideal int. temperature
over Single Stage Compression
23°F int. temperature

full load
10 part load
0
50% part load
2 stage screw s w ith ideal
-5 intermediate temperature
-10 2 stage screw s w ith 23°F

intermediate temperature
-10°F Suction Temp.
-15
65 70 75 80 85 90 95
Saturated Condensing Temperature [°F]
© 2003 Reindl
Source: Manske, K., 2000
The above plot shows the percent savings of various two-stage refrigeration system options over a single stage system at full
load (upper set of curves) and ½ load (lower set of curves). The two cases for each load represent each a fixed intermediate
temperature level of 23 F and a floating intermediate temperature level which minimizes the system power required.
The lower set of lines in the above plot shows that two stage systems using screw compressors have a much higher
saturated condensing temperature at which a two stage system becomes competitive as compared to a system using recip
compressors (the upper set of lines). Also included in each of the cases are a fixed intermediate temperature (23 F) and an
optimized intermediate temperature (pressure).
Page 118
Intermediate Temperature
Source: Manske, K., 2000

© 2003 Reindl
Selecting intermediate pressures other than the optimum will reduce the effectiveness of a two-stage compression system.
The penalty of using a non-ideal intermediate pressure is most significant with screw compressors. This effect can best be
explained with the help of the above figure. The figure shows the saturation temperature that the intermediate pressure
receiver should operate at to minimize the compressor horsepower per ton for the composite two-stage system. The optimum
intermediate pressure was found using optimization routines in a simulation environment. Curves are shown for both single
screw and reciprocating compressors. It is interesting to note that the optimized temperatures for the screw and reciprocating
compressor systems straddle the “rule of thumb" ideal intermediate suggested by Mitchell and Braun, 1998.
Page 119
Multi-Stage System Conclusions
Only consider multi-stage for system temps
below 0 F [-17.8 C] (practically speaking, below -28 F, -33 C)
Use care in selecting intermediate temperature
level (reset if possible)
Systems with floating head pressure erode
advantages of multi-stage systems
© 2003 Reindl
Page 120
Next Steps
© 2003 Reindl
Page 121
Establish Goals for Efficiency Improvement
What are the steps?

1. baseline system performance
2. compare baseline with benchmarks (competitors, published, …)
3. identify potential opportunities
4. estimate impact/benefit to system
5. estimate capital cost
6. assess operational risks/constraints
7. prioritize by greatest benefit to cost
8. follow-up and quantify benefit & continuously improve
© 2003 Reindl
Page 122
Baseline system perform ance
Does perform ance yes

exceed benchm arks?
no
Identify potential energy

efficiency im provements
Estimate im pact/benefit
to system
Econom ic criteria no
Satisfied?
yes
Assess operational risks
Operational risks no
criteria satisfied?
yes
Im plement and follow-up with

© 2003 Reindl next opportunity
Page 123
Energy Efficiency Measures
Approaches for efficiency improvements
floating head pressure control
demand-shifting strategies
defrosting
compressor operation/sizing
multi-stage systems
maintenance issues
© 2003 Reindl
Page 124
Achieving Energy Conservation
Opportune times to evaluate system
in conjunction with proposed expansions
bi-annually or on a set schedule
upgrades (as part of management-of-change)
Choose ECM projects that offer good

cost/benefit and low risk
© 2003 Reindl
Page 125
Validation/Verification
Compare post-ECM system performance with
pre-ECM system performance
use normalization to account for differences in
weather
production
compare with estimated impact from feasibility
assessment(s)
is the trend “as-expected”
is the magnitude of the impact “as-anticipated”
© 2003 Reindl
Page 126
How & When
Don’t try to do everything at once
make changes slowly and incrementally to minimize risk to
system operation
be sure to involve safety, operations, mechanics, and
production personnel in proposed system changes
include changes in your management-of-change (MOC)
procedures as required by your plant’s process safety
management (PSM) program
training, training, and training
© 2003 Reindl
Page 127
It takes effort!
Baseline, implement, & verify

strive for continuous efficiency
improvements
© 2003 Reindl
Page 128
Good Luck!
Innovation – Knowledge – Progress
© 2003 Reindl www.irc.wisc.edu
Page 129

IIR Short Course Presentation PDF

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

IIR Short Course Presentation PDF

Uploaded by

Copyright:

Available Formats

Energy Efficiency Opportunities in

Industrial Refrigeration Systems

Douglas T. Reindl, Ph.D., P.E.

© 2003 Reindl University of Wisconsin-Madison

External Equalize Line

Lower evaporator temperatures

Energy use is primarily

Monthly Peak Demand (kW)

Monthly Unit Electric Cost ($/doz)

A refrigeration system’s performance is affected by:

Loads and the system’s response to meet the loads

Estimate use for other loads:

“Selecting efficient components

“I can’t afford to design an efficient system”

“The owner does not want an efficient system”

The owner can’t afford an efficient system

Energy is too cheap – not cost-effective

“I don’t know how”

3. identify potential opportunities

4. estimate impact/benefit to system

5. estimate capital cost

6. assess operational risks/constraints

7. prioritize by greatest benefit to cost

8. follow-up, quantify benefit, & continuously improve

How do we control head pressure in industrial

Our heat rejection system controls head pressure

Single speed fan with on/off control

Consequences of lowering head pressure

 increased system capacity

© 2003 Reindl Saturated Condensing Temperature (F)

Evaporative Condenser Selection:

Thermosiphon Oil Cooling (TSOC):

 21% operating cost savings $9,600/yr

Based on 1997 ASHRAE HOF

Most manufacturers discuss two methods

Manufacturers catalog nominal condenser

1. Select an Evapco condenser to do the job.

Capacitycondenser,actual = Refrigeration Load + Compressor Work

Nominal Capacity = HRF ⋅ Capacity = 1.34 ⋅ 6,174 mBh

If the design condensing temperature is lowered

 At 85 F condensing, the 23L would deliver 410 tons

Capacitycondenser ,actual = Refrigeration Load + Compressor Work

Nominal Capacity = HRF ⋅ Capacity = 2.94 ⋅ 5,968 mBh

Design SCT Load Avail. Cap. Nom. Heat

95 390 398 1.51 587 8,273

85 390 410 1.28 506 17,546

Type Mass Flow!

45-55°F Qspace PLO

© 2003 Reindl Source: Manske, K., 2000

© 2003 Reindl Source: Manske, K., 2000

Total System Heat Rejection [Btu/hr]

210 Calculated Condenser Heat Rejection (Constant Evaporator Load) 6

Phead,opt = -27.6 + 2.55 * Twb

© 2003 Reindl Source: Manske, K., 2000

©Source: Manske, K. et al., 2000

Source: Manske, K. et al., 2000

Assuming we have all screw compressors, how should

Source: Manske, K., et al., 2000

Source: Manske, K., et al., 2000

15 kw=9.55645 + 15.576·PLRCalculated + 26.2308·PLRCalculated2

kw=22.5113 + 11.5401·PLRCalculated + 16.0278·PLRCalculated2

© 2003 Reindl Part-Load Ratio

© 2003 Reindl 109

increased system capacity

21% operating cost savings $9,600/yr

At 85 F condensing, the 23L would deliver 410 tons

Type Mass Flow!