ASHRAE Journal

The Wal-Mart experience: Part one

The Wal-Mart experience: Part one

Michael MacDonald

In 2005, Wal-Mart opened experimental stores in McKinney, Texas (hot climate), and Aurora, Colo. (cold climate). With these projects Wal-Mart can:

* Learn how to achieve sustainability improvements;

* Gain experience with the design, design process, and operations for some specific advanced technologies;

* Understand energy use patterns in their stores more clearly;

* Lay groundwork for better understanding of how to achieve major carbon footprint reductions; and

* Measure the potential benefits of specific technologies tested.


The stores incorporate several experiments using recycled materials and energy-saving technologies, and include a number of experimental projects. Over the three-year period of 2006-2008, the progress of the experiments and lessons learned will be evaluated. The National Renewable Energy Laboratory (NREL) is monitoring the Aurora location (near Denver), while the Oak Ridge National Laboratory (ORNL) evaluates results in McKinney (near Dallas).

With thousands of stores around the world, Wal-Mart is a leader in the construction and operation of large-scale retail outlet buildings. Wal-Mart is committed to environmental responsibility and leadership in building and operating retail facilities that minimize the use of energy and natural resources.

This article describes the results for HVAC and refrigeration experiments at McKinney and Aurora. In next month’s issue, Part 2 discusses lighting and renewable energy experiments.

Energy Use for “Typical” Supercenters

A supercenter offers an interesting application of the latest energy use index data in the 2007 ASHRAE Handbook–HVAC Applications, Chapter 35 on energy use and management, Table 2. The table provides weighted energy use indices (EUIs), kBtu/[ft.sup.2] per year of site energy (no electric or other losses included), for key distributional percentiles, and the mean for about 50 building types in the U.S. Department of Energy’s Commercial Buildings Energy Consumption Survey (CBECS) microdata for 2003 (

The space uses in a typical supercenter are close to those shown in Table 1, although additional refinement is possible. The space percentage breakouts can be applied to the mean EUIs for each space type corresponding to a building type from the Handbook. A comparison can be calculated for “mean” or “average” building energy use for a supercenter, based on average EUI data for the United States for similar space types. The average comparable type of space mix in the United States calculates to a total EUI of 148 kBtu/[ft.sup.2] (1681 MJ/[m.sup.2]) per year.

Comparison with medians, or the 50th percentile data, is considered important, since the average values often are skewed by very high energy users at the top end of each distribution. A similar calculation with medians leads to a calculated median total EUI of 115 kBtu/[ft.sup.2] (1306 MJ/[m.sup.2]).

Based on data from energy measurement efforts in 2001-2002, combined with current energy measurement data, the energy use for a typical Wal-Mart supercenter of 206,000 [ft.sup.2] (19 000 [m.sup.2]) is estimated. The notion of “typical” becomes complicated when a variety of climates exists, which is the case for supercenters. The typical supercenter was evaluated as having gas-fired dehumidification media regeneration in the 100% outdoor air units, so gas use occurs year-round for heating and dehumidification in the data shown here.

The typical electric use is calculated to be 5.8 GWh/yr, and the typical natural gas use is calculated to be 1.7 GWh/yr, for an EUI of 123 kBtu/[ft.sup.2] (1397 MJ/[m.sup.2]) per year, or 17% less than the average comparable CBECS EUI calculated in Table 1, and 7% more than the median, which is respectable for 24/7 operation.

Figure 1 shows the monthly profile of electricity use, and Figure 2 shows the profile of gas use. Keep in mind that supercenters operate 24/7 and generate a fair amount of heat internally.

A breakout of fuel use by end use also was developed. Gas use is calculated to be 15% for cooking (including a food service tenant), 36% for gas dehumidification and 49% for heating.


The breakout of electric use is shown in Table 2. Refrigeration and lighting are the larger end uses, but plug loads are significant and appear to be growing over time. HVAC fan and cooling energy are controlled reasonably well.

A cost percentage breakout and calculated estimates are shown in Figure 3 for the total energy costs for electricity plus natural gas.

The HVAC, refrigeration, and microturbine experiments performed at the stores are described, which are followed by the measured energy use compared with the typical data, as these experiments affect the comparisons significantly.

HVAC and Microturbine Generation Systems

Standard features in supercenters include separate outdoor air systems (100% OA) to handle tougher outdoor air-conditioning needs such as dehumidification in Texas, and preheating winter air in Colorado. Supercenters typically have carbon dioxide demand-controlled ventilation at 900 ppm absolute. Ventilation paths are provided in the rooftop space conditioning units–although these units are not intended to handle peak loads–and in the 100% outdoor air units.

Many experiments were installed in the stores. Table 3 is a list of HVAC and coupled cogeneration experiments, as well as the refrigeration system experiments for the Aurora store.

The cooling, heating, power (CHP) system at Aurora introduces high complexity, including the introduction of a chiller (absorption), which is not typically seen in a supercenter.


A diagram of the component arrangement for the CHP system in Aurora is shown in Figure 4. Photo 3 shows the microturbine end of the CHP system before being walled in. Photo 4 shows the solar wall air preheater on the south side of the Aurora supercenter.

The main sales floor is served by nine indirect evaporatively cooled units (ECUs). The primary airstream is cooled in an air-to-air heat exchanger by a secondary airstream that is cooled by direct evaporative cooling. The secondary airstream, in this case, is exhaust air from the space.

The air-distribution system uses fabric ducts (Photo 5), which have many small holes that distribute an even airflow along the entire length of the duct, rather than from a single register. The ducts are mounted 11 ft (3.4 m) above the floor and use low velocity fans to distribute the supplied air, with the intent to maintain stratification between cooled air supplied and the upper part of the store.

Rooftop units (RTUs) supplement the ECUs in Aurora, providing additional space conditioning if needed. In the back, 13 rooftop units have chilled water coils; and in the front, where tenants are located, 15 RTUs have refrigerant coils with air-cooled condensers.

Several RTUs along the south-face of the building, along with two ECUs, are connected to the transpired solar collector ventilation air preheater. The wall consists of a perforated metal cladding that creates an air cavity between it and the main outside wall of the building. Solar radiation heats the cladding, and ventilation air drawn in through the perforations is heated by the warm cladding.

Hot water for heating is generated from two waste oil boilers, the microturbine cogeneration package, and a gas boiler, in order of priority. Chilled water is generated by the absorption chiller in the CHP system.

The CHP system in the Aurora store burns natural gas in the microturbine to produce electricity, and the exhaust from the microturbines is directed to the heat exchanger and/or the absorption chiller. The CHP system can achieve an overall efficiency of 75%, as compared to typical utility power plant efficiency, which is in the range of 40%.

Table 4 is a list of HVAC and refrigeration experiments for the McKinney supercenter.



The McKinney store has a complicated hydronic loop, that integrates a gas-fired boiler loop, waste oil boiler loop, refrigeration heat recovery loop, and a heating loop that serves multiple HVAC units and the radiant floor loops.

The supercenter in McKinney has RTUs to provide most heating and cooling, while the outdoor air-handling units (AHUs) temper outdoor air as needed, by cooling and dehumidification or preheating. The heating hydronic loop supplies heat to these units. Cooling systems are electric air-cooled DX. The desiccant dehumidification system is regenerated using electric cooling compressor heat. On startup, the AHUs were found unable to supply tempered air at a cool enough temperature, so additional RTUs had to be added to precool the air supplied to the AHUs (Photo 6).


Waste oil boilers are used as the first source of heating hot water for the water loop when oil is available. However, these systems represent a small part of the total hot water supply. Waste oil comes primarily from the automotive section, but also from the kitchen.

Refrigeration Systems

Supercenters have medium-temperature (refrigerating) and low-temperature (freezing) systems. Heat recovery from refrigeration for heating domestic hot water is standard in supercenters and used in Aurora, although the McKinney store has a more complicated heat recovery loop.

In Aurora a small medium-temperature refrigerant loop cools a secondary glycol loop. The glycol is then used to cool the cases, which can be more efficient, since the internal diameter of the primary heat exchanger coil tubing is fully wetted by the glycol. Traditional coils that use DX technology have been found less efficient because much less of the coil surface area can perform useful work. (Note: Wal-Mart continues to pursue testing of secondary loops at additional stores. For more information, see Reference 1.)

Refrigeration heat rejection in Aurora is by evaporative coolers, and in McKinney it is by cooling towers (Photo 7). The cases and coolers in both experimental stores have enhanced coil surfaces in the fixtures and electrically commutated high efficiency fan/motor packs in the medium-temperature cases. Doors were added for most medium-temperature cases, instead of leaving them open.

HVAC and CHP Results

At Aurora gas is used for HVAC and for electricity generation in the microturbines. The microturbines at Aurora produced 2.6 GWh of electricity, which was 52% of the annual use from June 2006 to May 2007. The microturbines (six 60kW) and double-effect absorption chiller-heater (120 ton [422 kW]) CHP system has performed reasonably well.

Microturbines electricity production has been close to expected, with output ranging from 200 kW in the summer to 355 kW in the winter, and electrical generation efficiencies ranging from 22% to 28%. The annual average efficiency over the last year has been 26%. The average efficiency of utility-supplied electricity in the United States is about 30% overall, including distribution and transmission losses. The overall efficiency of electricity generation for the microturbines depends strongly on the need for heating and cooling and can reach more than 70%. Monthly average values are about 50%.

Comparisons of energy use between buildings with CHP, and without, is complicated by the failure to include electric generation losses for electricity purchased from the utility (only electrical output used by the building is considered). Thus, some adjustments are needed to allow informative comparisons. Comparison fuel totals for Aurora and McKinney for June 2006–May 2007, together with the typical supercenter totals, without gas for cooking, are shown in Table 5. If only purchased fuel is considered, both experimental supercenters use more fuel than the typical store, with higher electric and lower gas use in the warmer climate of McKinney, and much higher gas use in Aurora when microturbine gas is included.

But since input fuel for generation of electricity is ignored in the purchased electricity, one other possible comparison is to look at total electricity use at the store and gas that does not include the input gas to the microturbines. Gas for cooking also is excluded from these values. If all gas input to the microturbines is excluded, the Aurora supercenter would be said to use 0.8 GWh less than the typical supercenter. Readers can use these values to make other comparisons.

McKinney and Aurora produce electricity from photovoltaic (PV) panels and a wind turbine. Total PV and wind production in the “total electricity use” data in Table 5 is 0.16 GWh for Aurora and 0.07 GWh for McKinney.

The CHP results are most interesting for Aurora thus far, and other experiments are showing interactive issues. Results for Aurora HVAC experiments are shown in Table 6.

Fabric ducting must be kept inflated to maintain an attractive appearance, which incurs an energy penalty of about 0.4 GWh/yr occurs in Aurora relative to a typical store. Retrofitting to variable air volume is planned to reduce this penalty.

Results in McKinney, thus far, are not good, with the same fabric duct penalty causing about a 0.8 GWh/yr energy penalty relative to a typical store. The large HVAC units in McKinney that have fabric duct have already been retrofitted with lowered air return intakes and variable air volume drives, but results are not available yet. Experimental results for the McKinney HVAC experiments are summarized in Table 7.

The fabric duct is intended to allow thermal stratification in the experimental stores, and eliminate the need for cooling the store above a height of 12 ft (3.7 m). Potential cooling savings were estimated to be fairly high, possibly 0.5 GWh/yr. Temperature rakes were installed in the McKinney store, where this strategy appeared most important, to measure the temperature profiles in the main store area. All the temperature rakes inside the store indicate stratification does not occur, probably because the air returns to the rooftop units are at roof level and pull air up to that level. Data from the two most interior rakes are shown in Figures 5 and 6 for the first few days of August 2006, at heights from 5 ft (1.2 m) above the floor up to the roof level. The temperatures indicate that little stratification occurs and little difference in load should be expected for these small temperature differences. Outdoor temperatures during these days were 80[degrees]F (27[degrees]C) at night, with 100[degrees]F to 105[degrees]F (38[degrees]C to 41[degrees]C) as daytime highs.

Refrigeration Results

The Aurora store uses evaporative condensers for refrigeration heat rejection, while the McKinney experimental store has cooling towers. The typical store uses air-cooled condensers. The Aurora refrigeration systems have an annual consumption of 1.2 GWh/yr of electricity, 0.6 GWh/yr less than the typical store. Checks on electricity use in Denver area stores, together with detailed measurements at another store in the Denver area, indicate typical supercenter refrigeration energy use in this region to be 1.5 to 1.6 GWh/yr.

The McKinney refrigeration system uses 1.8 GWh/yr, or about the same as the typical store, and also the same as an air-cooled refrigeration system at another supercenter being measured in McKinney.

The refrigeration energy at the experimental store in Aurora measures consistently lower than an air-cooled refrigeration system at another supercenter being measured in the Denver area, while the McKinney system has been shown to have lower use in the summer but higher use in the winter, when compared to a similar store in McKinney. The McKinney cooling tower control has not been adapted to try to benefit from lower outdoor air temperatures.



Refrigeration experiment results are summarized in Table 8. Improper refrigeration control settings have been found to be an issue at Aurora and McKinney, but adjustments are still to be made, so potential benefits have not been evaluated.


Although much has been learned from the HVAC and refrigeration experiments in Aurora and McKinney, major improvements in energy savings are still needed.

Lessons being reinforced by these experiments include:

* The primary issue for HVAC systems in supercenters appears to be creating HVAC systems that can significantly adapt to low occupancy periods and seasonal changes. Too much ventilation is provided at the experimental stores now, due to the fabric duct. Variable speed drives have been installed on the HVAC fabric duct systems in McKinney, but no energy data are available yet. Testing of control changes in a fairly typical supercenter in McKinney that also is being monitored have shown a reduction of about 1.0 GWh over the past year in HVAC energy, primarily due to major reductions in runtime on the outdoor air ventilation units.

* Hydronic-based heating and cooling systems can work well when multiple heat recovery opportunities exist, but the pumping energy should be carefully evaluated, especially in a large building like a supercenter.

* The transpired solar collector has shown some promise for cold climates showing savings potential of 0.1 GWh/ yr, but the system needs to be sized for the ventilation load and controlled properly to make the economics work. Wal-Mart is considering further testing of transpired solar collectors for other locations.

* Several HVAC control issues have been found in Aurora and McKinney, and as often happens, are a challenge to correct due to larger system issues. Potential additional savings from correcting control issues in both locations may be about 0.1 GWh/yr.

* Evaporative cooling for space conditioning reduces energy use and peak demand in dry climates. The water consumption and maintenance issues need to be carefully considered. Further studies on water consumption are needed to quantify this issue.

* Evaporative cooling for heat rejection from the refrigeration system has shown to save energy and reduce summer peak demand by up to 80 kW in a dry climate.

* Some improvements in refrigeration system efficiencies have been found, but more would be useful. Proper control of refrigeration systems remains an important issue. Refrigeration control tuneups are being considered in both locations, and potential savings may also be 0.1 GWh/yr in each location from such tuneups.


(1.) Kazachki, G.S., D.K. Hinde. “Secondary coolant systems for supermarkets.” ASHRAE Journal 48(9)34 – 46.

Part Two of “The Wal-Mart Experience” will appear in the October issue.

Michael MacDonald is a staff member at Oak Ridge National Laboratory in Oak Ridge, Tenn.; and Michael Deru, Ph.D., is a senior engineer at the National Renewable Energy Laboratory in Golden, Colo.

Table 1: Space: breakout and mean CBECS EUI comparison.

Space Type Percent 2007 Handbook EUI Allocation

of Total Mean EUI, kBtu/[ft.sup.2]

Floor Area kBtu/[ft.sup.2] per yr

Grocery 30% 213 64

Restaurant 1% 302 3

Other Retail 50% 120 60

Non-refrigerated 8% 34 3


Other Service 11% 168 18

Total 100% — 148

Table 2: Electricity end-use breakout.

Electric Percent


Lighting 28%

Plug Loads 22%

Refrigeration 31%

HVAC Fans and Cooling 19%

Total 100%

Table 3: Aurora experiment list.

Aurora HVAC Experiment Description

Radiant Floor Heating A radiant floor provides warm floors in

And Cooling colder weather at the checkout area,

entrances, grocery frozen food aisles,

garden center, and auto center. When

cooling, the radiant floors are kept at

58[degrees]F (adjustable).

Cooling, Heating, Power Two 415 kBtu/h waste oil boilers, a micro-

(CHP) System turbine exhaust heat exchanger, and a 5.6

million Btu/h gas boiler are the heat

sources, listed in order of priority of

use. Hot water coils are used in all

rooftop units. A CHIP system, with six 60

kW microturbines, a 120 ton double-effect

absorption chiller, and a cooling tower

are the cooling and power generation


Indirect Evaporative Main sales floor is served by nine indirect

Cooling System evaporatively cooled units (ECUs). ECUs

also have cooling coil supplied by chiller

for hottest days.

Air-Distribution Main sales floor has special fabric ducts

System instead of ceiling plenum-box discharge

grilles to maintain stratification in

store to reduce conditioned volume by


Transpired Solar Several space conditioning units along the

Collector Air Preheater south wall of the store receive preheated

air from the transpired solar collector

in colder weather.

Recovered Fires on used vehicle oil from auto shop

Waste Oil Boilers and recovered cooking oil.

Aurora Refrigeration Description


Evaporatively Evaporative condensers for heat rejection,

Cooled Condenser instead of air-cooled condensers.

Secondary Loop System The medium-temperature refrigerant loop is

small and cools a secondary glycol loop.

Alternative Freezer Cases have high-efficiency, enhanced coil

And Cooler Units surfaces and electrically commutated

high-efficiency fan/motor packs in the

medium-temperature cases. Also used doors

on medium-temperature cases.

Table 4: McKinney experiment list.

McKinney MVAC Experiment Description

All-Electric Desiccant Compressor heat desiccant regeneration

Regeneration instead of gas-fired burners.

Air-Distribution System Use special fabric ducts instead of

ceiling plenum-box discharge grilles to

maintain stratification in store to

reduce conditioned volume by half.

Radiant Floor Heating Use radiant floors to improve customer

comfort in checkout and grocery

refrigeration areas, and worker comfort

in the auto center.

Recovered Waste Oil Boiler Fires on recovered cooking oil and used

vehicle oil from auto shop.

McKinney Refrigeration Description


Water-Cooled Refrigeration Cooling towers for heat rejection,

instead of air-cooled condensers.

Alternative Freezer and Cases have high-efficiency, enhanced

Cooler Units coil surfaces and electrically

commutated high efficiency fan/motor

packs in the medium-temperature cases.

Also used doors on medium-temperature


Captured Refrigeration Complicated hydronic loop is designed

Waste Heat to return refrigeration compressor heat

to boiler room.

Table 5: Fuel totals for comparison.

Aurora McKinney Typical

Fuel GWh/yr GWh/yr GWh/yr

Purchased Electricity 2.1 6.3 5.7

Purchased Gas, Less Cooking 7.8 1.2 1.4

Total Purchased Fuel 9.9 7.5 7.1

Electricity Use and Ignoring Fuel Input for Electricity Generation

Total Electricity Use 4.7 6.4 5.7

HVAC Gas (Less All Micro- 1.6 1.2 1.4

turbine Gas and Cooking)

Total 6.3 7.6 7.1

Table 6: Aurora HVAC experiment results.

Aurora HVAC Experiment Results

Radiant Floor Heating Comfort is improved, but energy

And Cooling savings not expected.

Cooling, Heating, Energy use could be claimed to

Power System be reduced 0.8 GWh/yr, or 11%,

from typical store, although total

purchased fuel is 39% higher. Elec-

tricity generation on site complicates


Indirect Evaporative Savings are not known and are con-

Cooling System founded by higher fan energy for fab-

ric duct and high chilled water pump

use due to cooling zoning issue.

Air-Distribution System No stratification has occurred; ap-

parently since returns and possibly

leaks pull air toward the ceiling. Fans

must run continuously to keep the

fabric duct inflated, and drives are

not variable speed, so electric use

is about 0.4 GWh/yr higher in Aurora

than for the typical store.

Less than expected, providing 0.05

Transpired Solar GWh over past year. Control issues

Collector Air Preheater have kept the system from perform-

ing to the potential of approximately

0.1 GWh/yr.

Recovered Waste Oil Boilers Savings of 0.07 GWh/yr expected,

and savings of 0.03 GWh achieved

over last year. Motor oil supply was

limited and the cooking oil system

was shut down most of the winter for

several reasons. A change to non-

hydrogenated cooking oil should

help performance next winter.

Table 7: McKinney, HVAC experiment results.

McKinney HVAC Experiment Results

All-Electric Desiccant Energy not measured directly, and is

Regeneration affected by air-distribution system

(Figures 5 and 6).

Air-Distribution System No stratification has occurred,

apparently since returns and possibly

leaks pull air toward the ceiling. Fans

must run continuously to keep the

fabric ducting inflated, and drives are

not variable speed, so electric use is

about 0.8 GWh/yr higher in McKinney

than for the typical store.

Radiant Floor Heating Comfort is improved, but energy savings

were not as expected.

Recovered Waste Oil Boiler Waste oil boiler is providing heat

acceptably, at about 0.07 GWh/yr.

Table 8: Refrigeration experiment results.

Aurora Refrigeration Results


Evaporatively Cooled Savings of 0.1 GWh/yr measured on low

Condenser temperature racks. Savings on medium

temperature racks have been difficult to

sort out but are estimated to be 0.1 GWh/yr.

Another important benefit is reduction in

demand up to 80 kW in the summer.

Secondary Loop Energy savings are too small to measure.

System Main benefit is reduced refrigerant charge

and potential to reduce refrigerant leakage.

Leakage differences have not been quanti-

fied yet.

Alternative Freezer Savings of 0.1 GWh/yr achieved relative to

And Cooler Units, typical store.

Including Medium

Temperature Case


McKinney Refrigera- Results

tion Experiment

Water-Cooled No savings achieved, as condensing tem-

Refrigeration perature is kept fairly constant over the

year, although the typical store can float


Alternative Freezer Savings of 0.1 GWh/yr achieved relative to

And Cooler Units, typical store.

Including Medium

Temperature Case


Captured No savings achieved. Condensing

Refrigeration Waste temperatures on affected racks are

Heat maintained artifically high to allow heat

to be captured, so recovered heat is not


Figure 3: Cost breakout for all fuels for a typical supercenter.

Heating 7%

Gas Dehumidfication 5%


And Cooling 16%

Refrigeration 27%

Plug Loads 19%

Lighting 24%

Cooking 2%

Note: Table made from pie chart.

COPYRIGHT 2007 American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc.

COPYRIGHT 2008 Gale, Cengage Learning