Emerging Technologies

Ammonia/CO2 Refrigeration System

Evaluation at a Food Processing Facility
ET13SCE1230
DR13.07

Prepared by:

Emerging Products
Customer Service
Southern California Edison

April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

Acknowledgements

Southern California Edison’s Emerging Products (EP) group is responsible for this project. It was
developed as part of Southern California Edison’s Emerging Technologies Program under internal
project number ET13SCE1230. This technology evaluation was conducted by Scott Mitchell of
Southern California Edison and researchers at the Electric Power Research Institute (EPRI) with
overall guidance and management from Paul Delaney. Contact [email protected] for more
information on this project.

Disclaimer

This report was prepared by Southern California Edison (SCE) and funded by California utility
customers under the auspices of the California Public Utilities Commission. Reproduction or
distribution of the whole or any part of the contents of this document without the express written
permission of SCE is prohibited. This work was performed with reasonable care and in
accordance with professional standards. However, neither SCE nor any entity performing the
work pursuant to SCE’s authority make any warranty or representation, expressed or implied,
with regard to this report, the merchantability or fitness for a particular purpose of the results of
the work, or any analyses, or conclusions contained in this report. The results reflected in the
work are generally representative of operating conditions; however, the results in any other
situation may vary depending upon particular operating conditions.

Southern California Edison

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

EXECUTIVE SUMMARY
The HVAC&R world is undergoing significant changes, with regulatory bodies moving to
reduce the use of refrigerants with high global warming potential (GWP). In commercial and
industrial applications, ammonia refrigerant is a common solution which has zero GWP, but its
applicability is limited because of mild flammability and toxicity concerns. In these applications
R404A and R507A are common refrigerants today, but because of their high GWPs they face
growing regulatory restrictions.

This report describes a field test of an alternative technology: a system using ammonia as the
primary fluid and CO2 as a pumped volatile secondary fluid. This allows the ammonia charge to
be much smaller than an equivalent all-ammonia system, while distributing only CO2 into the
building. The system was installed in a food production facility in Irvine, California and is
monitored along with existing, baseline conventional-refrigerant equipment, to study
performance. The study documents the findings of differences in energy and power consumption
as well as observations and learnings from the process of transitioning from a baseline system to
a new, alternative-refrigerant approach. The system evaluated is shown in Figure 1.

FIGURE 1 PHOTOGRAPH OF NH3/CO2 REFRIGERATION SKID DURING INSTALLATION (PHOTO CREDIT: CIMCO)

The new refrigeration system was installed to provide cooling for an existing a 2,100 square-
foot, -20°F drive-in freezer. The existing refrigeration equipment, part of a R507A system, was
left in place but shut off. Instrumentation equipment was installed to monitor the performance of

Southern California Edison Page i

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

both the new system and the existing equipment, and the new equipment was disabled
periodically during the study to allow collection of baseline data.

The new system used less energy in similar operating conditions than the baseline equipment, as
shown in the stacked-bar chart in Figure 2. This figure shows the total daily energy consumption
of each sub-component. Comparing days in April with similar weather, the baseline system used
220 kWh more per day than the new equipment, a savings of 21%. During hotter weather in
summer, the savings was 16-25%.

FIGURE 2 STACKED-BAR CHART OF DAILY ENERGY CONSUMPTION FOR ALL REFRIGERATION EQUIPMENT

Southern California Edison Page ii

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

FIGURE 3 SEPTEMBER 20 DR EVENT COMPARED WITH SEPTEMBER NON-DR WEEKDAYS

The project also included simulated demand response events, including pre-cooling and load
shedding. This was performed manually on-site with the help of the host facility personnel. The
temperature set-points in the freezer were adjusted by about 5°F in either direction. The control
was simple set-point adjustment. An example of a result comparing a pre-cool and shed event
with typical baseline days is shown in Figure 3. Several different durations were evaluated. The
average power during the events was 14-21 kW lower during the first hour than the same
baseline hours.

The ability to pre-cool was limited by the pressure setting of the CO2 receiver, which could not
be quickly adjusted for these evaluations. Since the CO2 liquid is held at a fixed pressure, the
supply temperature in the freezer is limited by that pressure; subsequent research efforts should
evaluate the effect of adjusting this pressure setting to further pre-cool.

The ammonia and carbon dioxide system evaluated here provided significant energy savings
while using two refrigerants that are environmentally friendly and not subject to increasing
regulatory and environmental pressures. The system remains in place and the host has, since
project completion, added an additional freezer space which is only cooled using the new
ammonia/carbon dioxide system, with no R507A backup.

Southern California Edison Page iii

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

CONTENTS
EXECUTIVE SUMMARY ______________________________________________________ I

INTRODUCTION __________________________________________________________ 1

BACKGROUND __________________________________________________________ 4
Site Overview ............................................................................................. 4
Technology Overview ................................................................................... 5
New Equipment and Installation .................................................................... 6

TESTING OVERVIEW ______________________________________________________ 11

Objectives ................................................................................................ 11
Approach ................................................................................................. 11
Test Overview ...................................................................................... 11
Instrumentation ........................................................................................ 12

RESULTS_______________________________________________________________ 15
Demand Response ..................................................................................... 25

CONCLUSIONS _________________________________________________________ 35

Southern California Edison Page iv

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

FIGURES
Figure 1 Photograph of NH3/CO2 Refrigeration Skid During
Installation (Photo Credit: CIMCO) .................................. i
Figure 2 Stacked-Bar chart of Daily Energy Consumption for All
Refrigeration Equipment................................................ ii
Figure 3 September 20 DR Event Compared with September Non-
DR Weekdays ............................................................. iii
Figure 4 Photo of the Freezer Under Investigation ............................ 4
Figure 5 High-Level Schematic of the NH3/CO2 Refrigeration
System Under Evaluation .............................................. 5
Figure 6 Photo Showing New Exterior Façade Surrounding
Refrigeration System (Photo Credit: CIMCO) ................... 7
Figure 7 Photograph of NH3/CO2 Refrigeration Skid During
Installation (Photo Credit: CIMCO) ................................. 7
Figure 8 Photo of the New Cooling Tower for the NH3/CO2
Refrigeration System (Photo Credit: CIMCO) ................... 8
Figure 9 New Carbon Dioxide Evaporator Coil being Installed in the
Freezer ....................................................................... 9
Figure 10 Existing Compressor in Mezzanine Compressor Room ......... 9
Figure 11 Simplified Schematic of (Evaporator & CO2 Tank Shown
as One)..................................................................... 10
Figure 12 Stacked-Bar chart of Daily Energy Consumption for All
Refrigeration Equipment.............................................. 15
Figure 13 Daily Energy Consumption vs. Average Outdoor
Temperature for Baseline and New Equipment ............... 16
Figure 14 Load Shape of Equipment with Average
OutdoorTemperature Profile (July, Sundays removed) .... 19
Figure 15 Load Shapes by Weekday for Full Time Period, with New
Equipment................................................................. 19
Figure 16 Power of NH3/CO2 Equipment and Outdoor Temperature
for July 6, 2016 ......................................................... 20
Figure 17 Compressor RPM and Outdoor Temperature for July 6,
2016 ........................................................................ 21
Figure 18 Freezer Air Temperature for July 6, 2016 ........................ 22
Figure 19 Baseline R507A Equipment Power and Outdoor
Temperature for July 15, 2016 ..................................... 23
Figure 20 Freezer Air Temperature for July 15, 2016 ...................... 24
Figure 21 Box and Whisker (Median – Quartiles – 1.5X Inter-
Quartile Range ........................................................... 25
Figure 22 Pre-Cool Simulated DR Event, August 17 ........................ 26

Southern California Edison Page v

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

Figure 23 Coil Air Inlet and Outlet Temperatures for CO 2 Coils On

Day of Pre-Cool DR Simulation ..................................... 27
Figure 24 Load Shed Simulated DR Event, August 18...................... 28
Figure 25 Simulated DR Pre-Cool and Shed, September 20 ............. 29
Figure 26 Simulated DR Pre-Cool and Shed, September 21 ............. 30
Figure 27 Simulated DR Pre-Cool and Shed, September 30 ............. 31
Figure 28 September 20 DR Event Compared with September Non-
DR Weekdays ............................................................ 32
Figure 29 September 21 DR Event Compared with September Non-
DR Weekdays ............................................................ 32
Figure 30 September 21 DR Event Compared with September Non-
DR Weekdays ............................................................ 33

TABLES
Table 1 Traits of Some Common Refrigerants .................................. 1
Table 2 Instrumentation Description ............................................. 13
Table 3 Baseline and New Equipment Daily Summary Data for
Spring ...................................................................... 17
Table 4 Baseline and New Equipment Daily Summary Data for
Summer .................................................................... 18

Southern California Edison Page vi

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

INTRODUCTION
Rules around the acceptable use of refrigerants are rapidly changing in the United States and
internationally, as well as locally in California. The U.S. EPA, acting under the Significant New
Alternatives Policy (SNAP) program has recently changed the status of many high global
warming potential (GWP) refrigerants. In the next several years, commonly-used refrigerants
such as R404A, R507A, R134a and others will be prohibited for use in some types of new or
retrofit commercial refrigeration installations1. The October, 2016 signing of the Kigali
Amendment to the Montreal Protocol will further push the heating, ventilation, air conditioning
and refrigeration (HVAC&R) industry to different refrigerant options as the amendment calls for
the phasing out of hydrofluorocarbon (HFC) refrigerants.2

TABLE 1: TRAITS OF SOME COMMON REFRIGERANTS3

Refrigerant Global Ozone Critical Critical Normal ASHRAE Safety

warming depletion Pressure Temperature Boiling Classification
potential potential (psia) (°F) Point
(GWP) (ODP) (°F)
R22 1790 0.040 723.7 205.1 -41.5 A1
R404A 3700 0 540.8 161.7 -51.2 A1
R507A 3800 0 537.4 159.1 -52.1 A1
R410A 2100 0 711.1 160.4 -60.6 A1
R407C 1774 0 671.5 186.9 -46.5 A1
R290 ~20 0 616.6 206.1 -44 A3
(propane)
R744 (carbon 1 0 1070.0 87.8 -109 A1
dioxide)
R717 <1 0 1643.7 270.1 -28 B2L
(ammonia)

The changing regulations have led many in the industry to search for alternatives. Some of the
most common refrigerants today are shown in Table 1. Of the options shown in Table 1, the first
five face regulatory challenges due to their environmental characteristics. The last three are
considered alternative options for future use, but each faces some degree of challenge to growing
adoption.

1
40 CFR Part 82 Protection of Stratospheric Ozone: Change of Listing Status for Certain
Substitutes Under the Significant New Alternatives Policy Program; Final Rule, 80FR42870,
July 20, 2015.
2
United Nations Environment Programme Ozone Secretariat, The Montreal Protocol on
Substances that Deplete the Ozone Layer
http://ozone.unep.org/en/treaties-and-decisions/montreal-protocol-substances-deplete-
ozone-layer
3
2013 ASHRAE Handbook – Fundamentals

Southern California Edison Page 1

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

Changes are also happening at a local level: California’s Air Resources Board has recently
approved a strategy document proposing aggressive changes which could include GWP limits as
low as 150 for stationary refrigeration and 750 for stationary air conditioning4. The document
generally outlines support for the Kigali Amendment, though the details of how that will be
implemented are still to be determined at the time of this writing.

Natural refrigerants such as ammonia (R717), carbon dioxide (R744) and hydrocarbons such as
propane (R290) are being used to meet the demand for very low GWP refrigeration equipment.
Hydrocarbons have gained traction as the long-term solution for small-charge systems such as
stand-alone refrigeration applications, where the charge level of flammable refrigerant is small
and efficiency is very good. For larger industrial applications, ammonia has long been the
refrigerant of choice, but due to toxicity and flammability, its use in large quantities near highly
populated areas brings risks which must be accounted for, and may add cost. Carbon dioxide is
gaining traction for supermarket refrigeration in the U.S., and current research and development
efforts are focused on overcoming efficiency hurdles under transcritical operation, which is a
particular challenge in warm climates. Solutions that can use these refrigerants while addressing
the technical challenges and safety risks associated with their use could open new possibilities in
significantly increasing the energy efficiency of the national refrigeration fleet, and greatly
reducing the greenhouse gas impact of inevitable refrigerant release.

Ammonia is subject to restrictions on the federal, state and local levels, due to toxicity at high
concentrations. In particular, the U.S. EPA has different regulations applying for site inventory
thresholds of 500 pounds and 10,000 pounds, and requires emergency release notification in the
event of leaks exceeding 100 pounds in a 24-hour period. Similarly, OSHA requirements apply
to ammonia facilities, with additional requirements when exceeding the 10,000 pounds threshold.
State level programs are common, too: most notable (and relevant to the field study discussed
here) is California, where the quantity for increased scrutiny is 500 pounds. Inspections and
reporting are required at regular intervals, and compliance audits must also be undertaken at
regular intervals5. Further restrictions may be applied at the local level, particularly considering
ammonia systems in highly populated areas. For these reasons, ammonia charge quantity
reduction is becoming an increasingly hot topic in the industry. Most ammonia regulations were
intended to deal with large charge systems. There are a number of efforts currently underway to
develop regulations specifically for low charge ammonia systems that can take advantage of
ammonia’s high efficiency while minimizing the risk of harm due to leaks.

For some applications, a combined approach using ammonia and carbon dioxide may provide
beneficial performance while limiting risk factors and performance issues. Many readers are
likely familiar with the cascade cycle, a type of two-stage cooling cycle where a high stage fluid
(ammonia in this case) is paired with a low-stage fluid (CO2); heat rejection from the low stage is
absorbed by the evaporating high-stage refrigerant. Cascade systems are useful particularly with

4
California Air Resources Board, Short Lived Climate Pollutant Reduction Strategy Final
Report. March, 2017
https://www.arb.ca.gov/cc/shortlived/meetings/03142017/final_slcp_report.pdf
5
Chapp, T. 2014, Low Ammonia Charge Refrigeration Systems for Cold Storage: White
Paper. International Association of Refrigerated Warehouses.

Southern California Edison Page 2

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

a large temperature difference and can offer good efficiency, but can be expensive and the cost
and complexity is not needed for moderate cold storage and freezing temperature applications.

Another increasingly popular approach is to use CO2 as a volatile secondary fluid, pumped to the
evaporators where it partially evaporates, and condenses in a chiller. This can be achieved using
ammonia as the primary working fluid. Compared with the common approach of using
water/glycol as a secondary fluid, pumping power for volatile CO2 is drastically lower, about 5%
of the power required to pump water or glycol as pointed out in a 2012 ASHRAE Journal
article6.

This report describes a field study of a system using ammonia as the primary fluid and CO2 as a
pumped volatile secondary fluid. The system was installed in a food production facility in Irvine,
California and is monitored along with existing, baseline conventional-refrigerant equipment, to
study performance. The study documents the findings of differences in energy and power
consumption as well as observations and learnings from the process of transitioning from a
baseline system to a new, alternative-refrigerant approach7.

6
Pearson, S. 2012, Using CO2 to Reduce Refrigerant Charge. ASHRAE Journal, October
2012.
7
Bush, J. and Mitchell, S., Reaching Near-Zero GWP with Packaged Ammonia/Carbon Dioxide
Systems. ASHRAE Journal, February 2017.

Southern California Edison Page 3

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

BACKGROUND
For this research effort, Southern California Edison and EPRI investigated an ammonia/carbon
dioxide refrigeration system for use in an industrial food application in Irvine, California. This
section will detail the site and technology.

SITE OVERVIEW
The Irvine, California site was identified by Southern California Edison and Mayekawa, and was
already in the process of considering a transition from their existing R-507A refrigeration system
to an ammonia/carbon dioxide solution. The host is a Japanese-owned company that makes food
products. The facility includes food processing, freezing, and storage.

FIGURE 4: PHOTO OF THE FREEZER UNDER INVESTIGATION

The food products are produced, then stored in a 2,100 square-foot, -20°F drive-in freezer. The
refrigeration load, approximately 12.2 tons of refrigeration (TR), was met with a single R507A
reciprocating compressor with a three-step mechanical unloader, installed in 2010. There are
three additional R507A compressors for various other refrigerated spaces, all sharing an
evaporative condenser. The production line typically operates 12 hour shifts 5 days per week,
with a morning shift on Saturdays depending on workload. The freezer, shown in Figure 4, is used
for short term product storage before shipping. The facility is situated in a busy, mixed-use area,

Southern California Edison Page 4

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

in close proximity to retail and residential areas. This facility was chosen because of the ability
to keep the existing equipment running for baseline comparisons throughout the test period.

TECHNOLOGY OVERVIEW
The system evaluated in this effort is a new packaged ammonia (NH3)/carbon dioxide (CO2)
system manufactured by Mayekawa. The system, given the model name Newton 3000, uses
ammonia as the primary stage and pumped, volatile CO2 as the secondary stage. It will be
referred to as the NH3/CO2 system, or similar, throughout this reporting. A simplified schematic
of the system configuration is shown in Figure 5.

FIGURE 5 HIGH-LEVEL SCHEMATIC OF THE NH3/CO2 REFRIGERATION SYSTEM UNDER EVALUATION

The main ammonia circuit has a two-stage, screw compressor. The discharge of the screw
compressor goes to a water-cooled condenser. From there, the refrigerant goes to an intercooler
which separates liquid from vapor. The vapor goes to the compressor, mixing and cooling inlet
refrigerant to the second stage of compression. The liquid goes to an economizer. The
economizer works similarly, as vapor refrigerant goes to an economizer port in the first stage of
compression, and liquid goes to the evaporator. The evaporator is an ammonia-to-carbon dioxide
heat exchanger, which cools carbon dioxide. The CO2 is in a liquid tank, from which sub-cooled

Southern California Edison Page 5

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

liquid CO2 is pumped to the process evaporators.

This configuration has a few advantages. First, since the system is skid-mounted, on-site install
time can be significantly less than a site-built system. The centralized ammonia loop is relatively
compact and has a low charge of ammonia. Liquid CO2 is pumped to the evaporators and
undergoes sensible heating and partial phase change. It is cooled by the ammonia system, which
essentially acts as a chiller. This contains the charge of hazardous refrigerant to the central
location. There are other advantages to this pre-engineered skit-mount approach. The intercooler
and economizer stages improve compressor performance and send low-vapor quality refrigerant
to the evaporator enhancing system efficiency. The integrated package allows the system
controls to integrate each of these components and coordinate compressor speed and valve
timing effectively and with little or no on-site tuning required.

NEW EQUIPMENT AND INSTALLATION

The Newton package is made in several configurations, having different compressor sizing and
configurations for larger capacity, or specialized configurations for industrial freezing or ice
rinks. The device tested in this effort was the Newton R-3000, for which the manufacturer
provides the following specifications:
• CO2 Supply Temperature: -25.6°F
• Cooling Capacity (with cooling water at 89.6°F): 26.9 TR
• Motor Power: 45 kW
• Ammonia Charge: 55.1 lbs.
• Power Source: AC 400/440V @ 50/60Hz for motor; AC 200/220V @ 50/60 Hz for
controls
• Compressor: Semi-hermetic, compound screw, VFD driven with IPM motor
• Outer Dimensions: L 9ft 2in; W 6ft 5in; H 7ft 11in (excludes cooling tower)

The manufacturer also uses an interior permanent magnet (IPM) synchronous motor for the
compressor; these motors have higher efficiency than conventional induction motors, and
maintain high efficiency even at low compressor speed.

The installation was performed by CIMCO Refrigeration and included:

• Mayekawa Newton R-3000 NH3/CO2 package
• Closed-circuit cooling tower with water treatment
• Two CO2 evaporator fan coils with electric defrost
• Pump for circulating condenser water
• Installation of EPRI data monitoring equipment

The installation process took place over approximately four weeks in February and March, 2016.
The total time on-site was reduced compared with a conventional installation because of the
skid-mounted system. The installation was not without surprises, however. Satisfying the city
permitting and inspection requirements led to several unanticipated additions to the site plan. A
higher surrounding facade (for aesthetic purposes only) and additional ammonia leak

Southern California Edison Page 6

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

containment measures (the addition of an ammonia diffusion tank for the vent discharge) were
required. The surrounding facade is shown in Figure 6.

FIGURE 6 PHOTO SHOWING NEW EXTERIOR FAÇADE SURROUNDING REFRIGERATION SYSTEM (PHOTO CREDIT: CIMCO)

FIGURE 7 PHOTOGRAPH OF NH3/CO2 REFRIGERATION SKID DURING INSTALLATION (PHOTO CREDIT: CIMCO)

Southern California Edison Page 7

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

The Newton package is shown in Figure 7. The compressor (above) and condenser (below) are
visible in this image, with the condenser connected to water/glycol lines to and from the cooling
tower. This photograph was taken from adjacent to the cooling tower, which is shown in Figure
8.

FIGURE 8 PHOTO OF THE NEW COOLING TOWER FOR THE NH3/CO2 REFRIGERATION SYSTEM (PHOTO CREDIT: CIMCO)

Southern California Edison Page 8

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

FIGURE 9 NEW CARBON DIOXIDE EVAPORATOR COIL BEING INSTALLED IN THE FREEZER

Two new CO2 evaporators were installed in the freezer. Figure 9 shows one of the coils during
the installation process. Similar, R507A evaporators remained installed in the freezer during
testing.

The baseline compressor, in the mezzanine compressor room, is shown in Figure 10. The
compressor room features four R507A compressors for the various refrigeration end loads; an
evaporative condenser for the R507A systems is outside the frame of this photo.

FIGURE 10 EXISTING COMPRESSOR IN MEZZANINE COMPRESSOR ROOM

Southern California Edison Page 9

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

An overall schematic of the installation is shown in Figure 11.

FIGURE 11 SIMPLIFIED SCHEMATIC OF (EVAPORATOR & CO2 TANK SHOWN AS ONE)

Southern California Edison Page 10

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

TESTING OVERVIEW
OBJECTIVES
The objectives of this work can be summarized as follows:

• Measure the performance of an NH3/CO2 refrigeration system in a real-world application.

Considering factors such as weather and schedule, analyze system performance for new
learnings and to enhance understanding of the system characteristics.

• Evaluate performance for the Southern California region. The Irvine, CA site provides
only a relatively narrow band of weather compared with other cities in SCE’s territory,
but expected performance in other cities can be estimated based on the data collected in
Irvine.

• Compare performance against existing, baseline equipment. Using the equipment already
installed on-site as baseline, the energy and demand reduction from the new system can
be determined. Since the system has excess capacity for the current loading, estimates
can also be extrapolated to a possible expansion of the refrigerated space.

• Test the equipment in demand response operation. With the support of on-site personnel,
the system was tested for “pre-cool” and “load shed” operation over several days of
testing. The goal of this effort was to assess the current or near-term capability of this
equipment to perform demand response in a freezer or warehouse type application.

APPROACH
The approach for this evaluation is field-monitoring of equipment under normal operation, with
periodic baseline testing which is performed by disabling the system under test, and turning on
the baseline equipment. This was done several times during the test period, in an effort to get
data across a range of weather conditions while respecting the scheduling/availability of on-site
personnel to make and monitor the changes. Demand response tests were performed similarly,
with several days of testing during which on-site personnel executed changes to the equipment
set-points to simulate a demand response event.

TEST OVERVIEW
The testing described above was performed over the course of 2016 starting after the installation
was complete. To generally sort data, the days are filtered into:

• Baseline - the R507A system is the only one cooling the space

• New Equipment - the NH3/CO2 system is the only one cooling the space

Southern California Edison Page 11

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

• Transition - the day is split, usually because of a mid-day switch between baseline and
new equipment, but also possibly including some limited maintenance which was not
identified

• Other - days where on-site maintenance was taking place, demand response testing was
performed, or other known aberrations from the test schedule

The days tagged as “other” are identified by communication with the host site manager. In initial
data processing, each day is flagged as “baseline” if the baseline compressor uses >100 kWh and
the Newton uses <75 kWh. A day is flagged as “New Equipment” if the Newton uses >100 kWh
and the baseline compressor uses <75 kWh. “Transition” and “Other” days are removed from the
bulk data analysis.

Baseline data was collected on the following dates:

• April 1 - 7
• May 11 - 16
• July 12 - 19
• October 25 - 30
• November 22 - 27
• December 12 - 15

The following dates were filtered as “other” based on a known maintenance, interruption,
inspection, demand response test, or other known interruption:
• May 24
• June 21, 23
• July 17, 18, 20
• August 10, 11, 17, 18, 21, 22, 23, 24
• September 13, 20, 21, 30
• November 1, 6, 7, 23, 24, 25, 26

Also, one other change took place during the testing which will affect the results: the suction
pressure of the R507A system (which includes the compressor monitored here, and others) was
reduced to try to improve an ice cream making process elsewhere in the plant. This change,
which took place around July 20 (just after the July baseline data collection) does not affect the
new equipment, but does change the performance with the R507A compressor running. As will
be discussed in the results section, the main difference was a lower average temperature in the
freezer when the R507A system was running after this change.

INSTRUMENTATION
To capture performance of the refrigeration systems, EPRI designed and built an instrumentation
system for installation in the field. The monitoring points and equipment installed are detailed in
Table 3.

Southern California Edison Page 12

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

TABLE 2 INSTRUMENTATION DESCRIPTION

Reading Description Location(s) Accuracy Device

Six circuit
Multiple, see breaker-level Elkor WattsOn
Power below readings +/- 0.2% power Power Meter
Type-T
Refrigerant CO2 circulating Thermocouple w/
Temperature Temperature supply and return +/- 0.9 °F Well
In freezer; cooling
tower air inlet
Temperature Air Temperature and outlet +/- 0.4 °F 10K-3 Thermistor
Water Cooling tower
Temperature Temperature inlet, outlet +/- 0.4 °F 10K-3 Thermistor
Air Air Temperature +/- 0.9F @ 72F;
Temperature/Humidity /Humidity In Freezer +/- 2% RH Dwyer RHT-R016
Air Air Temperature +/- 0.9F @ 72F;
Temperature/Humidity /Humidity Outdoor +/- 2% RH Dwyer RHP-2R11
liquid +/- 0.10% Micro Motion
CO2 circulating reading; gas +/- Coriolis Flow
Refrigerant Flow CO2 flow rate supply line 0.25% reading Meter CMFS050M
Make-up Water Cooling tower +/- 1.5% of Seametrics MJNR-
Make-up Water Flow Flow make-up water reading 075-20P

The power meters were installed at the circuit breaker level for a total of six readings. They are:
• NH3/CO2 system
• NH3/CO2 system cooling tower and pump
• CO2 evaporator coils including defrost heat
• Baseline compressor
• Baseline evaporative condenser
• Baseline evaporator coils, including defrost heat, under-floor heat, and other

There is some overlap in what power must be considered to compare systems, such as the heaters
on the baseline evaporator coil circuit. So, in comparing total power or energy consumption, all
readings are included.

In addition, the following data points were logged from the output of the Newton system’s
control board:
• Suction pressure
• Discharge pressure
• Intermediate Pressure
• Economizer Pressure

Southern California Edison Page 13

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

• Oil Pressure
• CO2 Receiver Pressure
• Suction Temperature
• Discharge Temperature
• Oil Temperature
• Cooling Inlet Temperature
• Cooling Outlet Temperature
• Compressor Motor Current
• Compressor RPM
• CO2 Differential Pressure
• CO2 Pump Motor Current
• CO2 Saturated Temperature

Since these values are from the manufacturer’s equipment, they are considered as secondary
measures and considered to augment the primary data stream.

Southern California Edison Page 14

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

RESULTS
Data acquisition began on March 14, 2016.

Baseline Suction
Pressure Change Date

FIGURE 12 STACKED-BAR CHART OF DAILY ENERGY CONSUMPTION FOR ALL REFRIGERATION EQUIPMENT

From a high-level perspective Figure 12 shows the energy consumption of all equipment, for
each baseline or “new equipment” day, with the “other” and “transition” days removed. The
graph is a stacked bar chart, so the highest number is the total for all equipment. The baseline
compressor is shown in orange, the baseline condenser in light blue and the baseline coils (as
well as underfloor heat) in light green. The NH3/CO2 system is shown in red, the new cooling
tower in dark blue and the CO2 evaporator coils in dark green. The outdoor temperature,
averaged over each day, is also shown in gray. The results show that, while there is some
overlap, for a given period of similar weather the energy consumption is considerably lower
using the new NH3/CO2 system. This graph also shows that, during the new equipment operating
days, some of the baseline equipment (the heaters captured by the baseline coil power meter, and
the baseline condenser) still run. The baseline condenser energy typically is about the same to
somewhat higher with the baseline compressor running. The measured “baseline coil and other

Southern California Edison Page 15

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

heat” energy is much higher with the baseline compressor running, as would be expected since
on “new equipment” days the coils are off and only the auxiliary heaters are running. All of the
new equipment goes to very near zero energy on baseline days; the compressor does run briefly a
few times per day to maintain the pressure level of the CO2 receiver.

The data is visualized differently in Figure 13, which shows the total energy plotted against daily
average outdoor temperature for baseline and new equipment days. Also, on this graph Sundays
are shown with a hollow circle and all other days are filled. This graph shows an upward trend of
energy with outdoor temperature for both cases, as might be expected. The baseline equipment
generally used more energy across the whole temperature range. The energy consumption on
Sundays is shown separately because it is considerably lower than other days: the plant is usually
closed on Sundays.

FIGURE 13 DAILY ENERGY CONSUMPTION VS. AVERAGE OUTDOOR TEMPERATURE FOR BASELINE AND NEW EQUIPMENT

The following tables show examples of baseline and new equipment daily summary data for
spring and summer. Each shows the daily energy for each sub-meter, the maximum power

Southern California Edison Page 16

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

recorded for each sub-meter (in one-minute intervals), and the average, maximum, and minimum
temperature for the day.

First Table 3 shows late March and early April data. The same weekdays were selected for
comparison, and time before and after the baseline period is shown. As shown on the table, there
is some overlap of each system during the operation of the other; for instance, on April 13, the
baseline coils and other heaters had higher energy draw than typical; close inspection shows that
a defrost heater engaged at some point, even though the compressor was not running at that time.
Similarly, for each system there is occasional brief cycling on of the compressor during “off”
times. Comparing April 4-6 and April 11-13, where average temperatures were similar, the new
equipment (on April 11-13) used 21% less energy, or 220 kWh per day less than the baseline.

TABLE 3 BASELINE AND NEW EQUIPMENT DAILY SUMMARY DATA FOR SPRING

3/27 3/28 3/29 3/30 4/3 4/4 4/5 4/6 4/10 4/11 4/12 4/13

Sun Mon Tue Wed Sun Mon Tue Wed Sun Mon Tue Wed
Baseline Compressor Energy kWh 0 0 0 5 606 607 666 724 0 11 7 11
Baseline Condenser Energy kWh 36 39 39 41 47 66 63 67 36 45 47 46
Baseline Coils & Other Heat Energy kWh 81 86 87 89 305 301 321 331 81 90 92 138
NH3/CO2 Rack Energy kWh 289 402 356 427 15 10 13 11 282 442 474 507
NH3/CO2 Cooling Tower Energy kWh 40 55 49 58 3 3 3 3 41 61 65 65
CO2 Coils Energy kWh 71 132 105 116 0 0 0 0 73 139 143 144
Energy Total kWh 518 713 637 736 976 986 1066 1136 512 788 828 911
Baseline Condenser Max. Power kW 2 10 12 12 10 13 13 14 5 13 14 14
Baseline Coils & Other Heat Max
Power kW 4 5 5 10 36 36 36 36 3 11 11 60
Baseline Compressor Max. Power kW 0 0 0 38 38 38 37 39 0 38 30 31
NH3/CO2 Cooling Tower Max.
Power kW 6 5 5 5 4 4 4 4 5 6 6 6
CO2 Coils Max. Power kW 44 44 45 45 0 0 0 0 44 45 44 45
NH3/CO2 Rack Max. Power kW 47 49 50 48 42 41 41 41 48 48 49 50
Average Outdoor Temp. F 62 59 57 57 63 62 63 63 62 63 62 62
Max. Outdoor Temp. F 71 66 67 66 73 76 78 74 73 76 73 74
Minimum Outdoor Temp. F 57 52 50 51 54 53 53 56 58 55 55 57

Southern California Edison Page 17

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

Table 4 shows the same data for days in July. Again, data from before and after the baseline
period is shown. In this case, the weather prior to the baseline period was similar to the baseline
period, and the weather after was considerably hotter. Comparing the days with similar weather,
the energy consumption was 25% lower, or 287 kWh per day, than the baseline days. Comparing
the hotter period of July 21-23 to the baseline period, the energy consumption was still 16%
lower or 191 kWh per day. The energy consumption was on average 96 kWh per day higher for
the hot days than the mild days with the NH3/CO2 system running.

TABLE 4 BASELINE AND NEW EQUIPMENT DAILY SUMMARY DATA FOR SUMMER

7/7 7/8 7/9 7/14 7/15 7/16 7/21 7/22 7/23

Thu Fri Sat Thu Fri Sat Thu Fri Sat
Baseline Compressor Energy kWh 15 6 0 700 742 707 6 12 0
Baseline Condenser Energy kWh 64 66 59 159 149 96 97 105 79
Baseline Coils & Other Heat Energy kWh 89 87 86 305 308 280 88 91 86
NH3/CO2 Rack Energy kWh 508 586 434 30 14 11 549 562 519
NH3/CO2 Cooling Tower Energy kWh 71 82 63 6 3 3 85 84 83
CO2 Coils Energy kWh 164 158 114 0 0 0 178 156 159
Energy Total kWh 910 985 756 1200 1216 1097 1002 1010 928
Baseline Condenser Max. Power kW 13 13 13 13 13 13 13 13 13
Baseline Coils & Other Heat Max Power kW 10 10 5 36 36 36 10 11 5
Baseline Compressor Max. Power kW 38 37 0 39 39 38 35 37 0
NH3/CO2 Cooling Tower Max. Power kW 8 7 7 7 7 7 7 7 7
CO2 Coils Max. Power kW 44 44 44 0 0 0 45 45 45
NH3/CO2 Rack Max. Power kW 49 49 50 47 42 43 49 48 48
Average Outdoor Temp. F 71 72 72 71 71 71 77 79 77
Max. Outdoor Temp. F 81 82 82 79 81 81 92 94 88
Minimum Outdoor Temp. F 65 66 66 67 64 66 67 66 68

A significant difference in the energy consumption is revealed by examining a 24-hour average

load profile for each case, baseline and new equipment. Figure 14 shows the hourly average
power and temperature profile, for data recorded in the month of July (excluding Sundays). The
baseline data is shown in red, the new equipment in blue. The baseline days had slightly lower
temperature on average. The average baseline power ranged from about 37 kW overnight (and
again in the 10 to 11 AM window) to a high of 56 kW from 12 to 1 PM. For the new equipment,
the power ranges from a low of about 24 kW overnight to a high of 48 kW in the 1 to 2 PM
window. The most notable difference is that the baseline equipment has a relatively flat load
profile in comparison. The new equipment power overnight is half of the peak, compared with
about 66% for the new equipment.

Southern California Edison Page 18

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

FIGURE 14 LOAD SHAPE OF EQUIPMENT WITH AVERAGE OUTDOOR TEMPERATURE PROFILE (JULY, SUNDAYS REMOVED)

Figure 15 shows the load shapes for the new equipment for the entire monitoring period. This
shows the difference in energy consumption of each day, which reflects the plant’s production
schedule. The schedule has two shifts each weekday, and a morning shift only on Saturdays as
needed, with no shifts on Sundays. This is reflected clearly in the load profiles, which show
Saturday having a similar morning usage to weekdays, with low afternoon usage, and Sunday
having a low and flat load profile.

FIGURE 15 LOAD SHAPES BY WEEKDAY FOR FULL TIME PERIOD, WITH NEW EQUIPMENT

Southern California Edison Page 19

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

Figure 16 shows a single typical summer weekday (Wednesday, July 6) of operation with the
new equipment. The outdoor temperature ranged from 65°F to 80°F. The main compressor rack
power is shown in blue, the cooling tower in orange, and the CO2 evaporators and defrost heaters
in green. For visual clarity, the power of the other equipment is not shown here. Overnight, the
power can be seen turning on and off, as the system is cycling at minimum compressor speed.
During these cycles, the compressor turns on, and power is 36 kW, before decreasing to about 29
kW. Right before shutting off, the power briefly increases to about 37 kW. The cooling tower
power also turns off when the main compressor rack does. The cooling coil turns off, but briefly
cycles on and off during “off” periods to circulate air. Shortly before 10:00 AM the first defrost
occurs: a total of about 43 kW of power between the fan coil power and resistance heat power
across the two fan coils. In response, the compressor power after the defrost is briefly higher,
reaching 48 kW and gradually ramping down over the course of about 40 minutes. Later, in the
early afternoon, there is a sustained period of operation where the system appears to run above
minimum power for approximately 3 hours. This is followed by another defrost and again a
period of higher power consumption. In the evening, the cyclic operation resumes.

FIGURE 16 POWER OF NH3/CO2 EQUIPMENT AND OUTDOOR TEMPERATURE FOR JULY 6, 2016

Southern California Edison Page 20

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

The compressor RPM for the same time period, as recorded from the Newton system’s PLC is
shown in Figure 17. This provides some insight into the power profile seen above. The power
profile closely follows the compressor RPM; the maximum speed, 4500 RPM, is reached twice
(both times, as a rebound after defrost); and the compressor always increases to 4000 RPM
briefly prior to shut-down.

FIGURE 17 COMPRESSOR RPM AND OUTDOOR TEMPERATURE FOR JULY 6, 2016

Southern California Edison Page 21

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

Figure 18 shows the temperature measured inside the freezer during the period shown above.
During defrost, the air temperature increases to approximately 25°F. For the first defrost, the
duration from when the temperature begins increasing to when it reaches a low point again
(shortly after 10 AM) is one hour, twenty-one minutes. The amount of time it is above 0°F is
about 32 minutes. During normal cycling, the temperature range is -15°F to -5°F. The cycling
occurs because of the oversized cooling capacity; when the load is higher (between about 1 PM
and 4 PM), the temperature is in a tighter range. The time required to pull down the temperature
by 10°F during an overnight cycle is about 20 minutes.

FIGURE 18 FREEZER AIR TEMPERATURE FOR JULY 6, 2016

Southern California Edison Page 22

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

FIGURE 19 BASELINE R507A EQUIPMENT POWER AND OUTDOOR TEMPERATURE FOR JULY 15, 2016

Figure 19 shows a day of baseline operation with a similar outdoor temperature profile. The
compressor power is shown in green, the condenser in purple, and the fan coils, defrost heaters,
and under-floor heat system in brown. There are a few items to note: first, the defrost happens
more frequently and at lower power; a suspected cause is that the defrost for each coil is
individual, and therefore out of sync. The compressor power operates between three stages, with
power levels of approximately 22 kW, 27 kW, and 36 kW. Since the compressor shares suction
and discharge manifolds with other compressors, there is not much variation from those levels
with temperature. The system does stay at a higher stage during the hotter part of the day, and a
lower stage (even cycling off some in the evening) during evening and early morning hours.

Southern California Edison Page 23

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

The temperature in the freezer, shown in Figure 20, behaves differently with the baseline
equipment. Since the compressor size is well-matched to the load, the temperature oscillates
much less than with the new equipment. The more-frequent but lower-power defrosts mean the
temperature increases more often, but less severely. The exception is around 4 PM, where both
defrost heaters ran in close succession. The air temperature reached approximately 29°F and was
over 0°F for approximately 39 minutes.

FIGURE 20 FREEZER AIR TEMPERATURE FOR JULY 15, 2016

Considering freezer temperatures more generally over the course of the year, Figure 21 shows a
box-and-whisker plot of the hourly average temperatures inside the freezer, for each month and
separately for baseline and new equipment if both ran in a given month. The box shows the first
quartile, second quartile (the median temperature) and the third quartile. The whiskers extend to
1.5 times the inter-quartile range. The individual dots show each hourly measurement. The
median temperature was mostly similar between baseline and new equipment, until October
when the median temperature of the baseline equipment was lower by about 2°F. The cause of
this is most likely the reduced R507A suction pressure. As mentioned above, the manufacturer
reduced the suction pressure for their ice cream freezer; this freezer shares a suction manifold
with the freezer under test, and it is suspected that this change led to the lower average
temperature. Also, the extents of the temperature range for the new equipment, in blue, are
considerably wider. For a typical month, about 13-15% of the hourly readings for the new
equipment and about 7-10% extend above the inter-quartile range. A likely reason for this
difference is the different sequencing of defrost heaters, as was addressed above: with the
NH3/CO2 system, the two coils engage defrost at the same time, and the temperature deviates by
a large amount for every defrost. For the baseline system, deviations are more frequent but

Southern California Edison Page 24

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

smaller in magnitude because the two heaters do not usually defrost at the same time. Depending
on the constraints of the facility, in some cases it may be useful to change the control of the
defrost heaters to operate separately, particularly if large temperature variations are problematic.

FIGURE 21 BOX AND WHISKER (MEDIAN – QUARTILES – 1.5X INTER-QUARTILE RANGE

DEMAND RESPONSE
This section details the results of demand response simulations run with the new equipment.
Tests were performed using simple set-point adjustments within the freezer. The nominal freezer
set-point is -15°F. For a pre-cool event, the set-point could be set lower, then allowed to float up
to normal during the event. For a utility-requested shed event, the set point could be set higher.
Early testing revealed that to provide cooling lower than about -20°F, the settings of the CO2
pressure on the system would have to be lowered. While this is possible via remote control from
the manufacturer’s Japan location (and could be programmed into future controllers for local
control), an option for performing such a test was not available in the timeframe of this effort.
Therefore, all adjustments were simple set-point adjustments without further refrigeration cycle
adjustments.

Southern California Edison Page 25

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

The first test was spread over two days, Wednesday, August 17 and Thursday, August 18. On the
first day, a simple “pre-cool” was performed, where the set-point was set to -20°F and held
inside the freezer. The results are shown in Figure 22. This test was planned from 10:45 AM to
1:45 PM, but ran long, until approximately 3:00 PM. The purpose of this initial test was to set
and hold a new, low set-point and observe the response in the room.

FIGURE 22 PRE-COOL SIMULATED DR EVENT, AUGUST 17

Shortly after the test began, a defrost initiated. After the defrost, a long period of compressor
operation can be observed, and the temperature in the freezer (in this case measured near the
middle of the space) stayed on the low end of its normal cycling range. The temperatures of air
entering and leaving the CO2 evaporator coils are shown, along with rack power for reference, in
Figure 23. During the pre-cool period, the temperatures entering and leaving the coils do not
change much, and from some slight increases (more pronounced in the measurements at Coil 1),
it appears there is some changing load in the freezer during this time. After the pre-cool period,

Southern California Edison Page 26

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

the system resumes operation as normal, and there does not appear to be a noticeable difference
in behavior.

FIGURE 23 COIL AIR INLET AND OUTLET TEMPERATURES FOR CO2 COILS ON DAY OF PRE-COOL DR SIMULATION

These results show that the ability to drop temperature below set-point may be limited by the
temperature of the CO2 being supplied. This is supported by the method of operation: the CO2
reservoir is held as a liquid at a fixed pressure, which is maintained by the ammonia system. The
controls of the Newton serve to maintain the condition of the CO2, meaning that the supply of

Southern California Edison Page 27

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

CO2 is at a roughly fixed temperature and pressure. The thermostat of the freezer only turns the
CO2 coils on and off. So, the temperature of air that can be reached with the coils is limited by
the CO2 temperature, and achieving much lower temperature in a reasonable timeframe would
require adjusting the CO2 conditions. This would be possible, and it would be possible to see
more aggressive demand response measures, with integration of advanced controls to the system.
However, the pre-cool as evaluated still serves a purpose: it ensures that the freezer is at the low
end of the temperature range, and by holding the temperature lower, the temperature of the
product in the freezer is likely lowered, which will in effect store some thermal energy. This
effect was not directly measured, as product simulators were not installed.

FIGURE 24 LOAD SHED SIMULATED DR EVENT, AUGUST 18

The following day, August 18, a test load shed event was run in the morning. The schedule for
this event was approximately 8:15 AM to 10:15 AM. At the beginning of the event, the system
was off as part of normal cycling. During the event, the temperature measured in the freezer
increased from -15°F to 0°F between 7:51 AM and 9:41 AM, during which time the system did
not run. As can be seen in the CO2 evaporator coil power measurement, there is a short power
usage approximately every 15 minutes. This is most likely the defrost heater briefly engaging.
The data resolution is one minute, so power measured for under one minute cannot be observed
in detail. The reason this occurred is not clear, and in subsequent shed tests, it did not occur. The
effect of these spikes is small: the average power of all NH3/CO2 equipment during the period
between cycles is 0.42 kW. Once the temperature in the freezer reached approximately 0°F, the

Southern California Edison Page 28

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

system turned on and cooled until the temperature reached about -12°F (the normal shut-off is
about -15°F). The temperature again begins to increase during the “off’ interval. At
approximately 10:30, the system turns on again, and normal operation resumes. The operation
after the shed is not noticeably different than normal.

Several other demand response tests were run, with the pre-cool and shed in succession. A test
on September 20 is shown in Figure 25. The pre-cool period was initiated at approximately 6:50
AM, and the load shed at approximately 10:00 AM. A defrost occurred during the middle of the
pre-cool period. The shed took place from 10:00 AM to 1:00 PM, during which time the system
was off for one hour, thirty-five minutes, and then there were two cooling cycles, approximately
twenty minutes each. The shed was followed by normal operation with longer run-times to
satisfy the -15°F set-point.

FIGURE 25 SIMULATED DR PRE-COOL AND SHED, SEPTEMBER 20

Southern California Edison Page 29

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

A similar event was run on September 21, with a targeted pre-cool period of 10:00 AM to 12:00
PM and shed from 12:00 PM to 2:00 PM. The results are shown in Figure 26. The behavior is
similar to the test above, though the uninterrupted “off” time is shorter: a 38-minute off period is
interrupted by a 20-minute on period. The system turns on again after 32 minutes, shortly after
which the set-point is set back to normal, so the system remains on.

FIGURE 26 SIMULATED DR PRE-COOL AND SHED, SEPTEMBER 21

An additional test was run on September 30, with pre-cool beginning at 1:00 PM and shed from
5:00 PM to approximately 6:30 PM. The results are shown in Figure 27. During the shed period,
there was one, 23-minute on-cycle, and the temperature in the freezer was held between -12°F
and +2°F. The effect of the pre-cool is not pronounced like previous tests; there may have been a
period of higher activity in the freezer during this time.

Southern California Edison Page 30

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

FIGURE 27 SIMULATED DR PRE-COOL AND SHED, SEPTEMBER 30

The change in load due to the DR events can be compared against the typical for similar weather.
In this case, the hourly-average power for each day is compared with the average of all other
(non-DR) weekdays in September in the following figures. The figures show the power of all
NH3/CO2 equipment, as well as the R507A coil circuit because it includes heaters which affect
the NH3/CO2 system. The DR pre-cool and shed time periods are approximated and drawn on
each graph, in blue for pre-cool and red for shed. The data is hourly averages, with each hour
indicated as, “hour beginning”. For example, hour 13 indicates 1:00 PM through 1:59 PM. Since
each simulated DR event represents only one day, some significant deviations from the monthly
average are to be expected. Also, since the start and stop times were manually determined by
personnel on-site and reported back through email, the start and finish do not necessarily align
exactly with each hour. However, the resulting load shapes give a strong indication of the pre-
cool and shed power profiles.

Figure 28 shows the DR test of September 20 compared with the average September weekday
load shape. The hourly average power was 8-23 kW higher during the pre-cool period than
average for the month; the power during the shed period was 12 kW and 20 kW lower than
average during the first two hours of shed, and 8 kW lower than average during the third. After
the shed, the power was a slightly lower than average for the next several hours.

Southern California Edison Page 31

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

FIGURE 28 SEPTEMBER 20 DR EVENT COMPARED WITH SEPTEMBER NON-DR WEEKDAYS

FIGURE 29 SEPTEMBER 21 DR EVENT COMPARED WITH SEPTEMBER NON-DR WEEKDAYS

Southern California Edison Page 32

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

Figure 29 shows the simulated DR event of September 21. The hourly average power was 13-15
kW higher than average during the pre-cool period, and 16-21 kW lower during the shed period.
After the shed, the power was slightly higher for the next several hours.

FIGURE 30: SEPTEMBER 21 DR EVENT COMPARED WITH SEPTEMBER NON-DR WEEKDAYS

Figure 30 shows the September 30 simulated DR event. For this day, the power was considerably
higher than average before the simulated event started, probably due to higher-than-average
traffic in the freezer. During the designated pre-cool period, the power was between
approximately the same and 8 kW higher than average. During the shed the average power was
14 kW lower during the full hour of shed (5:00 to 6:00 PM) and 6 kW lower for the following
hour, during the first 30 minutes of which the system was set to shed. The average power was
5 kW to 6 kW higher for the next few hours.

The simulated demand response testing performed here shows that a pre-cool and shed can be
performed with this equipment. Due to the system’s control functionality, where CO2 is
maintained at a fixed condition and pumped on-demand to the freezer, the ability to over-cool is
limited. This could be extended by allowing adjustment to the CO2 conditions prior to or as part
of the pre-cool period.

The ability to provide simple “on/off” load shedding is as much a function of the load itself as
the equipment. The system under test was mostly able to fully power off during the shed for
extended periods, with limited cycling to keep temperatures in the adjusted set-point range. This
allowed reductions of average hourly power up to 21 kW, without requiring a large rebound or

Southern California Edison Page 33

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

drastically altering the freezer condition. These shedding patterns could have significant financial
impact, depending on the utility rate structure and incentives.

Southern California Edison Page 34

Emerging Products April 2017
Ammonia/CO2 Refrigeration System Evaluation at a Food Processing Facility ET13SCE1230

CONCLUSIONS
In this field study, a new NH3/CO2 refrigeration system was installed at a food processing facility
in Irvine, California. The facility, which manufactures snack foods, has a walk-in freezer which
previously used a R507A refrigeration system. The old system was left in place, enabling direct
“Baseline/Treatment” testing to compare with the new equipment. Instrumentation was installed
to monitor all power circuits relevant to the freezer refrigeration equipment, as well as
temperatures in the freezer and outdoors, and other measurements.

The energy consumption was significantly lower with the new equipment, in the range of
approximately 200-300 kWh per day lower in typical conditions. Each system also maintained a
similar average freezer temperature in the early part of the study; after July baseline testing, a
change was made to the R507A circuit to improve performance elsewhere in the facility, and
baseline temperatures were slightly lower after that change. The freezer temperature deviated
higher during defrost with the new equipment, for which both defrost heaters ran simultaneously,
than for the baseline equipment which defrosted typically one coil at a time, leading to more
frequent temperature changes but of smaller magnitude. The new system was noted to be
considerably oversized for the existing load, so although it has a variable-speed compressor, it
often ran in cyclic operation, particularly during low-load hours. The host expects to double the
freezer capacity by adding an additional, near-identical freezer, and anticipates only CO2 coils
connected to the new system will be used for the new freezer.

Demand response simulations were also performed. Since the NH3/CO2 system’s operation calls
for maintaining CO2 at a fixed condition, and pumping liquid to the freezer, it was determined
that pre-cool capability was limited since the instantaneous capacity at the evaporator coil is
essentially fixed; to modify this, the CO2 set-point conditions could be changed, but this could
not be implemented in time to include in this evaluation. However, testing was performed by
adjusting the freezer temperature set-point. The testing showed an ability to increase load during
the pre-cool period, and to shed load by increasing the temperature set-point. The shed, which
was executed with a simple 5°F increase in set-point temperature, resulted in average hourly
power in the range of 14-21 kW for the first hour of load shed. Generally, the system did cycle
on and run later in the shed event, but run times were shorter than in typical cycling. Also,
defrost operation interrupted the demand response events (particularly the pre-cool) several
times. The defrost heaters were not controlled as part of the demand response study. The overall
performance of demand response would be improved with increasing control capability: the
ability to shift defrost forward or back in time, and the ability to adjust the CO2 conditions to
facilitate a deeper pre-cool, would both potentially offer a greater resource.

Southern California Edison Page 35

Emerging Products April 2017