[ << Previous | Contents | Next >> ]

In much of the developed world, air conditioning has become a near necessity. In the U.S., more than 90% of newly constructed housing units are centrally air-conditioned [ACHRN, 1999]. Two thirds of all dwelling units in the U.S. have central air conditioning and another one-third have one or more room air conditioners. [Appliance, 1997] [U.S. Census Bureau, 2000] Virtually all-commercial building space in the U.S. is air-conditioned. The rapid population growth of the Sun Belt in the U.S. was facilitated, if not enabled, by the universal use of air conditioning. Acceptance of air conditioning varies throughout the developed world, but in general is increasing rapidly in all but the coolest climates. Use of air conditioning for commercial buildings is growing rapidly, even in cooler climates.

The majority of both the existing installed capacity and new production of air conditioning equipment is unitary equipment. Unitary air conditioning equipment is a broad category of air-to-air air conditioning systems and heat pumps, including:

• Residential central air conditioning systems and heat pumps, both single package and split systems, generally between 5 kW (1 1/2 tons) and 18 kW (5 tons) cooling capacity
• Packaged air-to-air systems and split systems for commercial air conditioning, ranging in cooling capacity from 10 kW (3 tons) too more than 350 kW (100 tons). The ubiquitous commercial rooftop air conditioner falls into this category
• Ductless split systems, both mini-splits for one room and larger systems having multiple indoor evaporator/fan units connected to a single outdoor unit
• Water-source air conditioners and heat pumps
• Room air conditioners (a.k.a. window air conditioners)
• Packaged Terminal Air Conditioners (PTAC), functionally similar to room air conditioners

Note that while ductless split systems, room air conditioners, and PTACs are air to air systems, they are not always classified as "unitary" in published market data. Water source air conditioners and heat pumps are used in approximately 2% of new air conditioner applications [U.S. Census Bureau, 1995]. In the TEWI-3 study, TEWI comparisons were developed for the first three of these types of unitary equipment, in a variety of climatic locations and at several different efficiency levels.

In the present study, the focus is limited to the first two of the above categories of unitary equipment, and energy and LCCP comparisons are presented for a typical residence and a typical small office building in Atlanta, a location whose climate is somewhat warmer than the U.S. average. The prototypical systems that have been analyzed are two of the example systems used in the TEWI-3 study:

• 10.5 kW (3-ton) residential central air conditioning system (cooling only) or heat pump (heating and cooling). The baseline system (HCFC-22, 10 SEER, 6.8HSPF) is representative of a significant portion of the current (2002) market, meeting the current NAECA minimum SEER of 10 Btu/Watt-hr. This baseline system is compared to current/future systems using HCFC-22 or alternate refrigerants and with higher efficiencies (12 to 14 SEER), within the likely range of future efficiency requirements and of premium efficiency levels to qualify for electric utility rebate programs. Central air conditioners and heat pumps with efficiencies in this range are available commercially now.
• 26.4 kW (7.5 ton) single package commercial rooftop air conditioner (cooling only). The baseline system (HCFC-22, 10.3 EER) is representative of a significant portion of the current (2002) market. This baseline system is compared to current/future systems using HCFC-22 or alternate refrigerants and with higher efficiencies (11 EER).

For several decades, virtually all of the unitary air conditioning types described above have operated on a reverse Rankine cycle, commonly referred to as the standard vapor compression cycle, using HCFC-22 as the refrigerant. Some unitary air conditioning products are now being produced with one of two HFC blends, R407C or R410A, and with HFC-134a. While the majority of unitary air conditioning in production today still uses HCFC-22 as the refrigerant, in the post-ODS phase-out context of this study the two HFC blends R407C and R410A are the baseline refrigerants. In the Alternative Refrigerant Evaluation Program (AREP), HFC-134a was also evaluated as a replacement for HCFC-22. HFC-134a requires a larger volume flow rate of refrigerant for a given cooling capacity, requiring a larger compressor displacement and larger diameter tubing throughout, increasing the cost. Consequently, limited unitary equipment is likely to be produced in the future with HFC-134a refrigerant, more in large commercial unitary equipment than in smaller capacities.

Alternatives to the vapor compression cycle include:

• Trans-critical vapor compression cycle with carbon dioxide refrigerant. There has been considerable research activity with CO2 cycles in the past few years. Pressure levels are considerably higher with CO2 than with conventional (e.g., HCFC-22, HFC-407C, HFC-410A) refrigerants. Consequently, experimental system designs have used microchannel heat exchangers for both the condenser and evaporator. The higher pressures require heavy walled compressor shells. Basic refrigerant flow control between the condenser (gas cooler) and evaporator is more complex than with conventional vapor compression, because optimum control of the high side pressure needs to be accomplished, simultaneously with metering the optimum flow of refrigerant to the low side. An additional heat exchanger --- an intercooler --- is beneficial in improving the performance of CO2 systems. Some component manufacturers have begun to develop prototype compressors, flow controls, and heat exchangers for this application, but large-scale commercialization would take a long time to happen and is not assured. Due to the higher operating pressures, it is expected that CO2 systems will cost about 20% to 40% more than conventional systems.
• Vapor compression cycle with ammonia. Aside from toxicity and flammability considerations, ammonia is not well suited for use in the typical, cost-effective, unitary configuration that has evolved. Ammonia in the presence of water attacks copper, which is used extensively for refrigerant tubing, and many of the materials used in hermetic motors. Ammonia has not been evaluated further for unitary equipment in this study, however, it is expected that the higher discharge temperature of this fluid would further complicate its use in this application. Large commercial unitary equipment conceivably could be replaced with ammonia chillers, which are covered in the chiller section
• Vapor compression cycle with propane, or blends of hydrocarbons. Propane has a negligibly low GWP and performs well as a refrigerant, but due to its flammability, the cooling and heating capacity of a propane based air conditioner would need to be coupled to the interior space with a secondary coolant loop. A propane/secondary loop system is included in the comparison for residential sized equipment.
• Other cycles such as reverse-Brayton and reverse-Stirling. Reverse-Brayton tends to be considerably lower in efficiency than the standard vapor-compression cycle, while reverse-Stirling equipment would likely be considerably more expensive than conventional vapor compression cycle equipment, without offering any efficiency advantage. Both cycles have been well known for a very long time and neither has made significant inroads in the market for stationary air conditioning equipment. Neither technology is likely to become commercially viable in the foreseeable future.

For the purposes of comparison of the energy consumption of the refrigerant alternatives, the prototypical 3 ton residential systems and the comparable capacity technical alternatives described above have been analyzed for a representative, 1800 sq. ft. residential application in Atlanta. The prototypical 26.4 kW (7.5-ton) commercial rooftop unit has been analyzed for a representative light commercial application in Atlanta. The results have been calculated using the heating and cooling loads calculated in the TEWI-3 study analysis. In the TEWI-3 study, cooling and heating loads and performance were analyzed for several other locations as well, with comparable results.

Tables 6-1 and 6-2 summarize the electric energy consumption of 10.5 kW (3 ton) central air conditioners in Atlanta and 10.5 kW (3 ton) heat pumps in Atlanta, respectively. Note that in the NAECA-efficiency-standards-driven market for this equipment, energy consumption does not vary with refrigerant choice, per se. For example, if the future NAECA minimum efficiency were to be 12 SEER (note that the rulemaking to determine this level is ongoing and the new minimum has not been determined yet), a large portion of the air conditioners and heat pumps manufactured would meet this level without exceeding it by much, absent significant market demand for higher efficiency systems. The inherent efficiency characteristics of each refrigerant alternative will impact the design (and cost) required to provide the required minimum level of performance.

The differences in the inherent efficiency characteristics of the fluorocarbon alternatives are due to differences in thermodynamic and heat transport properties, and system operating pressure. Among R-22, R-407C, and R410A, these differences, and differences in refrigerant prices, can lead to "comparatively small" differences in manufacturing cost. An analysis of these differences is beyond the scope of this project, but it is recognized that even "comparatively small" differences in manufacturing cost can have a meaningful impact on gross and net profit given the highly competitive nature of a market having more than a half dozen strong manufacturers and scores of smaller manufacturers seeking greater market share.

To use propane as the refrigerant, it is assumed that a secondary loop must be used to deliver the cooling and heating capacity to the building interior. The secondary loop consumes parasitic pumping power and adds a heat transfer temperature difference to the overall thermodynamic lift. To overcome the resulting efficiency loss and meet a minimum efficiency level requires significant offsetting design modifications (e.g., larger heat exchangers, more efficient, but higher cost fan motors) to increase the efficiency. The associated costs, along with other costs associated with the safe use of propane will result in a significant (hundreds of U.S. dollars) increase in manufacturing costs and a larger increase in end-user prices.

Results reported by [Hrnjak, 1999] at the International Congress of Refrigeration indicate that it is technically feasible to make a CO2 transcritical cycle air conditioner with a single speed compressor having an SEER of approximately 10 to 11 (based on measuring the steady state EER at the DOE "B" --- 82oF outdoor temperature test point and accounting for the compressor drive motor efficiency and air moving power). However, the EER of a CO2 cycle decreases with increasing outdoor temperature more rapidly than occurs with conventional vapor cycle air conditioners. As a result, at more stringent (e.g., 95oF) outdoor ambient temperatures, a CO2 based air conditioner will be approximately 10% less efficient than a conventional HCFC-22 vapor-compression cycle having a comparable SEER, an issue where peak electric utility loads are a concern. Development of CO2 based air conditioners is still in the early research stage, providing little basis for projecting manufacturing cost vs. efficiency for comparison with conventional alternatives. Qualitatively, higher costs would be expected due to the higher pressure compressor shell requirement, larger and more costly microchannel heat exchangers, and the need for additional components such as an inter cooler. Additionally, there would be costs for developing and implementing an infrastructure capable of servicing carbon dioxide based equipment, to include new tools, service equipment, and training for technicians.

Table 6-1: Energy Consumption for a Representative Residential Air Conditioning Application in Atlanta (Cooling Only, Annual Cooling Load 33.8 million Btu, per TEWI-3)

Table 6-2: Energy Consumption for a Representative Residential Heat Pump (Heating and Cooling) Application in Atlanta (Annual heating and cooling loads of 34.8 million Btu and 33.8 million Btu, respectively, per TEWI-3)

The annual electric energy consumption of a 7.5 ton commercial rooftop air conditioner in a typical application in Atlanta is summarized in Table 6-3. The general comments, above, on inherent efficiency differences among refrigerants in residential air conditioning equipment are applicable here as well.

Table 6-3: Energy for Rooftop Air Conditioner in Atlanta

The Life Cycle Climate Performance (LCCP, see Section 1.3) for unitary cooling or unitary cooling and heating is made up of the indirect warming associated with the energy consumption summarized above plus the direct warming associated with refrigerant emissions (the warming effect of the refrigerant plus the embodied energy and fugitive emissions associated with manufacturing). The refrigerant charge size and charge loss rates that were assumed in the TEWI-3 study are summarized in Table 6-4. The GWP values are summarized in Table 6-5, which includes both the GWP of the refrigerant and the equivalent GWP of the energy and fugitive emissions associated with manufacturing and transporting the refrigerant.

Table 6-4: Refrigerant Charge Size and Charge Losses for Baseline Unitary Equipment

Table 6-5: GWP of Refrigerants and Warming Impact of Energy and Fugitive Emissions During Refrigerant Manufacturing

The LCCP for residential cooling only units in Atlanta is summarized in Table 6-6, for R22, R407C, R410A, propane, and CO2 at several SEER levels. The LCCP for residential heating and cooling is summarized in Table 6-7.

Table 6-6: LCCP for Residential Air Conditioning in Atlanta (2005 Technology)

Table 6-7: LCCP for Residential Heating and Cooling in Atlanta (2005 Technology)

Tables 6-6 and 6-7 are based on an efficiency standards-driven scenario, where the energy efficiency and energy consumption are the same, regardless of refrigerant. As a result the two "natural" refrigerant alternatives, CO2 and propane, with their low GWPs, have the lowest LCCP. However, at an SEER equal to a conventional fluorocarbon refrigerant based system, both propane and CO2 based systems are more costly to manufacture, propane due to the fire safety measures that must be included and CO2 due to the higher pressure, more complex system and larger heat exchangers required to achieve a given efficiency level. If this added cost was spent to produce a higher efficiency fluorocarbon refrigerant based unit, lower TEWI might be achieved through the reduced lifetime energy consumption and the associated reduction of energy related CO2 emissions. Work reported by Keller, et. al. [Keller, 1996] and [Keller, 1997] show that this is the case. With a conventional 12 SEER, HCFC-22 system as the baseline, a hypothetical fire-safe propane based system reduced the warming impact by 10%. With the same incremental cost of the propane system applied to a higher efficiency R-410A system, the TEWI was reduced by at least 12%, providing a lower lifetime warming impact for the same investment (without any fire safety uncertainties).

The LCCP for a single package rooftop in Atlanta is summarized in Table 6-8.

Table 6-8: LCCP for Commercial Rooftop in Atlanta (7.5 Ton Rated Capacity)

The basic observations that can be drawn from the LCCP values in Tables 6-6 through 6-8 are:

• The direct warming effect due to refrigerant emissions is < 5% of the total LCCP, for fluorochemical refrigerants
• Differences in efficiency have a much greater effect on LCCP than the direct effect of refrigerant emissions
• While the direct warming effect with propane is insignificant, the added energy due to the secondary loop must be offset by larger heat exchangers and other efficiency enhancements, so the cost would be increased significantly, by on the order of $1,000 to a residential end-user.
• While the direct warming impact of carbon dioxide refrigerant emissions is negligible, much higher manufacturing cost and price to the consumer would be expected. Substantial investments would be required to convert to CO2 which would essentially obsolete the current capital equipment base for manufacturing compressors and finned-tube coils.

With the exception of its high level of flammability, propane appears to be a suitable replacement for R-22 in all respects. Because of the flammability, and the several kg charge size of typical residential unitary equipment, to use propane for residential air conditioning, it is necessary to restrict the propane charge to outdoor equipment, and couple the cooling and heating to the interior with a secondary loop. While this might adequately eliminate the possibility of an explosion occurring within the confined interior space, further fire-safety measures would be needed in the outdoor unit:

• Propane vapor detection and alarm
• Placement of all electrical contacts (and any other potential ignition sources) clear of any potential propane leaks and/or in an enclosure that could contain an ignition --- either explosion proof or sealed with a flame arresting vent.
• To reach a given efficiency level, e.g., a NAECA minimum SEER, larger heat exchanger coils and other efficiency enhancements are needed to offset the losses of efficiency due to the secondary loop.
• In addition, fire-safety risks must be addressed in the manufacturing environment and service personnel must be properly trained to service propane-based systems safely. The latter is particularly challenging, given the high turnover rate of HVAC service technicians.

All of the measures described above are technically feasible and involve well-known technology. However, they would add significantly to the cost. When Lennox evaluated this option in the early 1990's, they concluded that the necessary fire safety measures would add 30% to the cost of a residential central air conditioning system.