[ << Previous | Contents | Next >> ]

Prior to the 1970s, automobile air conditioning was a largely American phenomenon. By 1980, 80% of American new cars had factory-installed air conditioning and a large portion of the output of the Japanese auto industry was air conditioned as well. Over the 20 years between 1980 and now, penetration of air conditioning in car and light truck sales has approached 100% in both the U.S. and Japan and is becoming increasingly popular in Europe. Not only is vehicle air conditioning a preferred comfort feature, but in many car models total engine power (and fuel consumption) at highway speed is less with the windows up and the air conditioning running than it is with the windows down and the air conditioning off. Worldwide, over 30 million air-conditioned cars are produced annually.

Pre-Montreal Protocol, automobile air conditioning systems used CFC-12 refrigerant- based vapor compression cycles. CFC-12 was universally replaced with HFC-134a in the 1993-1996 time period. A significant investment, estimated by the Mobile Air Conditioning Society at U.S. $5 billion ($3.5 billion by OEMs and $1.5 billion by the service industry) by the automobile and fluorochemical industries in the development and evaluation of compatible lubricants, construction materials, and fine tuning of compressor design and heat exchanger capacities was required to make this change so that traditional performance, reliability, durability and safety levels were maintained. This change was made with full knowledge that the global warming potential of the replacement (HFC-134a) was significantly lower than the original fluid being used (CFC-12) and that the replacement did not deplete stratospheric ozone. However, despite the lower GWP of HFC-134a, the use of either of two so-called natural refrigerants has been advocated by some to avoid the global warming impact of HFC-134a emissions. The two alternatives in questions are: a vapor compression cycle using hydrocarbon refrigerants or a transcritical vapor compression cycle using carbon dioxide. These alternatives are evaluated below.

In addition to the usual objectives of minimum weight and cost that drive the design of all automotive components, automobile air conditioning systems are designed, for a given vehicle, to meet cooling performance criteria that each vehicle manufacturer has established to represent competitive performance, meeting the expectations of their customers. In addition, systems are designed for suitability for assembly line installation into the vehicle, and for life (approximately 2,000 operating hours) and reliability consistent with the operating life of a passenger automobile.

A common automobile air conditioning system configuration has evolved to meet these requirements. Shown in Figure 5-1, the main features are:

• Belt driven, clutch actuated compressor hard mounted to the engine.
• Direct expansion evaporator located in the heating/cooling system interior air ductwork. The typical drawn cup - plate fin evaporator is a brazed assembly of thin (typically 0.020 inch thick) plates, which form the refrigerant passages in the heat transfer core and the refrigerant manifolds, and fin stock.
• Most commonly, fin-tube condensers are used, but alternatives have been adopted, including serpentine flat tube and fin and parallel flow flat tube and fin. To obtain the most effective cooling airflow, the condenser is located at the front of the car, usually in front of the radiator.
• An expansion device, to control the flow of liquid refrigerant from the condenser to the evaporator. Expansion devices in use range from orifice tubes to thermostatic expansion valves. For transcritical CO₂ systems, a different expansion device may be needed to accommodate the high pressure, super-critical state of the CO₂ entering the expansion device.
• Use of numerous mechanical fittings (using O-rings or gaskets) to interconnect the major system components and tubing, creating potential leaks, but facilitating initial assembly and future servicing.
• Use of flexible rubber hoses to connect the compressor to the rest of the system, allowing for engine and vehicle road vibration and assembly tolerances.

Figure 5-1: Conventional Automobile Air Conditioning System Configuration

As discussed above, two alternative technologies to the HFC-134a vapor compression cycle will be evaluated here:

• An otherwise conventional vapor compression cycle with hydrocarbon - HC-290 (propane) or HC-600a (isobutane) or a blend of the two - refrigerant and a secondary coolant loop. In this configuration, shown in Figure 5-2, the evaporator would be located in the engine compartment and would chill a secondary coolant, circulated to a heat exchanger in the passenger compartment (normally a DX evaporator) to cool the air. As discussed in Section 5.4, below, this arrangement does not fully address the safety concerns.
• A transcritical vapor compression cycle using carbon dioxide as the refrigerant. In this alternative, the arrangement of the components would be more or less consistent with conventional practice, but the individual system component designs would reflect the extremely high pressure levels of supercritical carbon dioxide (~ 2000 psig) (15 MPa). An intercooler between the suction line and the high pressure gas line is essential to system performance.
• R&D work is also in process on "low pressure" carbon dioxide. This approach is a hybrid compression/sorption system, where the sorbent reduces the CO2 pressures into the range that is more typical of fluorocarbon refrigerant based vapor compression cycles. The refrigerant system is more complex due to the sorbent in the system. This approach has not yet been developed sufficiently to even know if it can be considered to be a viable alternative and could not be considered in this report.

Figure 5-2: Configuration of automobile Air Conditioning System with Secondary Coolant Loop to Passenger Compartment

In comparing the HFC-134a based vapor compression cycle with the alternatives, it is important to remember that the HFC-134a system is characterized on the basis of the fully developed, mass produced system that is installed in approximately 30 million automobiles annually, and meets the diverse range of competitive cost, performance, reliability, durability, and safety requirements outlined previously. The current HFC-134a based system is an evolutionary adaptation of its CFC-12 based predecessor, whose design evolved over a period of 50 years. Despite the evolutionary nature of the adaptation from CFC-12 to HFC-134a, an enormous resource expenditure was required to resolve material compatibility and performance issues and to ensure that traditional consumer expectations for cooling performance and reliability would be met. The two alternatives, CO2 and hydrocarbon/secondary loop, have been prototyped and some performance test results are available to provide a basis for energy and LCCP analysis, but these designs would require substantial additional development to fully address the cost, performance, reliability, and safety requirements of the automobile customer. Therefore, the impact on the cost of an automobile to a consumer cannot be accurately assessed for either of these alternatives, beyond general expectations. First, to bring either of these alternatives to production would require an enormous investment in new manufacturing equipment and facilities, simultaneously making obsolete much of the existing manufacturing base built up over the past decade, let alone the cost to the service industry. Second, it is reasonable to anticipate that the significantly higher operating pressures typical of CO₂ transcritical AC systems will increase hardware cost, while HC based systems will require additional hardware to address safety concerns.

A detailed comparison of energy use of HFC-134a and the two alternative systems is cited below from the TEWI-3 study and the recently published GMR/ORNL studies. Prior to this, a simple, theoretical comparison of vapor compression cycle COPs is presented to help provide some perspective on the detailed results. Assuming the following, typical operating conditions:

• 30&176;F evaporating temperature, saturated vapor exiting the evaporator, heat transfer from the hot engine compartment to the suction hose results in 25°F of superheat at the compressor inlet;

• 80°F or 100°F ambient air (moderate conditions or extreme conditions, respectively);
• Corresponding vapor cycle condensing temperatures of 110°F or 130°F, respectively;
• 15°F liquid subcooling (approach to within half of the temperature difference between condensing and ambient);
• For the hydrocarbon vapor compression system, evaporating temperature of 20°F (10° lower than the others) to account for heat transfer/transport temperature differences (needed to drive heat transfer from the secondary coolant to the evaporating refrigerant). Ignore the parasitic power of the secondary coolant pump, and heat gain in the engine compartment.
• For the transcritical CO₂ vapor compression cycle:
  • CO₂ gas cooler exit temperature approach to within 10°F of ambient; and
  • CO₂ high side pressures of 1500 and 2000 psi, respectively.
  • An interchanger with 80% heat transfer effectiveness between the vapor leaving the evaporator and the supercritical vapor leaving the CO₂ gas cooler; no additional suction gas superheat at compressor inlet
• Equal, 70% isentropic efficiencies for the compressors that would be used with each of these three working fluids was assumed.

Tables 5-1 and 5-2 summarize the cycle COP calculations for mild and severe ambient conditions, respectively. Thermodynamic property data was taken from ASHRAE thermodynamic property tables.

Table 5-1: Summary of Theoretical Cycle COP Calculations at Mild Ambient Temperatures

Table 5-2: Summary of Theoretical Cycle COP Calculations at Severe Ambient Temperatures

A high COP corresponds to high-energy efficiency, as it indicates that more cooling capacity is provided for the mechanical power input expended. At both mild and severe ambient conditions, the HFC-134a vapor compression system has the highest theoretical COP, exceeding that of isobutane (with secondary coolant loop) by 20% and of transcritical CO₂ by more than 70%. The COP decrement of the isobutane cycle compared to HFC-134a is primarily caused by the lower evaporator temperature needed to accommodate the secondary coolant loop. The large difference in thermodynamic cycle COP between the transcritical CO₂ cycle and the R134a cycle is inherent in a transcritical cycle operating with heat rejection above the two-phase dome and with the evaporator operating within, but near the top of the dome, where the latent heat is small in comparison to the compression work. Note that an interchanger would improve the COP of an HFC-134a or a hydrocarbon vapor cycle and that an interchanger adds to the manufacturing cost. In a real system, of course, other factors affect the energy consumption. A major factor is the transport properties of the refrigerant, which have a strong influence on the evaporating and condensing (or super-critical cooling) heat transfer coefficients. In a mobile air conditioning application, engine fuel consumption is attributable to both the compressor power consumption and to the portion of the traction power to move the weight of the air conditioning system, whether it is operating or not. A detailed analysis is beyond the scope of this study, but results of other studies are discussed briefly below.

The TEWI-3 study took the following into account:

• Refrigerant thermodynamic and transport properties
• A range of climates (using binned hourly TMY temperature data)
• A range of driving cycles
• A/C system component weight impact on traction power
• Parasitic (fans, pumps) power consumption
• Variations in system operating modes with variations in driving and climatic conditions

The estimated energy consumption from TEWI-3 is summarized in Table 5-3, based on backing it out from the indirect TEWI contribution.

Table 5-3: Vehicle Lifetime Energy use for Mobile A/C, in Liters of Gasoline, for Technology Alternatives (based on TEWI-3)

The direct GWP of refrigerant emission is a significant part of the LCCP of HFC-134a emissions, so the assumed emissions over the vehicle lifetime is important. Table 5-4 summarizes three emissions scenarios that were used in the TEWI-3 study. The assumed 1 kg refrigerant charge corresponds to current practice for larger passenger cars. Actual charges range from about 0.5 kg for small cars to more than 1 kg for large vans and SUVs.

Table 5-4: Refrigerant Emissions for automobile Air Conditioning (North America) (Table 59 from TEWI-3)

In June, 1998, the world's major automobile manufacturers held a workshop in Phoenix, Arizona [Baker, 1998]. They developed a consensus estimate of lifetime HFC-134a emissions. For current vehicles, estimated lifetime usage is 1.26 system charges; for future vehicles, estimated lifetime usage is 0.71 charges. The former figure is somewhat higher than range of scenarios used in the TEWI-3 study; the latter falls within the range of scenarios.

Table 5-5 is based on Table 17 from the TEWI-3 study, where the Climate Change 95 GWP value for R134a was used. The LCCP values were derived by updating the TEWI values to include embedded energy and the pro-rata GWP of fugitive emissions from the production of HFC-134a (effective GWP of HFC-134a is increased from 1300 to 1313). Table 5-6 adjusts these LCCP values using the GWP values from the WMO (effective HFC-134a GWP of 1613). The numbers for the direct effect of HFC-134a in Table 5-5 and 5-6 represent the range from as manufactured up to two service additions, as contained in Table 5-4, bottom row.

Table 5-5: LCCP Results for Mobile A/C (based on TEWI-3 Study)

Table 5-6: LCCP Results for Mobile A/C (Based on the TEWI-3 study) Updated with WMO 99 GWP Values

In general, the lower energy use of HFC-134a based systems more than offsets the direct effect of projected HFC-134a emissions. Overall, the LCCP results do not provide a compelling basis to favor one working fluid over the others, based on the conditions and assumptions made. However, when other considerations, such as cost and safety are taken into account, HFC-134a systems demonstrate clear advantages. In addition, HFC-134a systems also have potential for further improvement from an LCCP viewpoint.

General Motors Research and the Oak Ridge National Lab collaborated on an in-depth update of the TEWI-3 study analysis for mobile air conditioning. The results were presented at an SAE meeting on June 28^th, 1999. As was the case in the TEWI-3 study, the impact of climate and driving cycles was taken into account. To include the effect of climate and regional differences in driving patterns, six cities were selected - Phoenix, Miami, Boston, Tokyo, Frankfurt, and Sydney. For each of these cities, hourly weather data was used to generate an annual distribution of temperatures and relative humidity levels. Local driving patterns were taken into account through the use of local fuel economy test cycles, e.g. the Federal Urban Driving Schedule in the U.S. Wind tunnel test data were used to relate condenser performance and CO₂ gas cooler performance to the driving cycle. The study attempts to more realistically model real world air conditioners, including the effect of recirculating hot air from the engine compartment through the condenser at idling conditions, based on wind tunnel data.

The primary focus of the GMR/ORNL study is comparing conventional HFC-134a based systems with transcritical CO₂ systems. The warming impact of HFC-134a emissions are a significant contributor to the TEWI. A range of HFC-134a emission scenarios (E1, E2, E3, and E4) were defined, corresponding to lifetime service additions of approximately 150, 300, 450, and 600 grams, respectively. TEWI values were calculated for CO₂ and HFC-134a systems, for small and mid-sized cars, in each of the six aforementioned cities. The impact on the TEWI of different emission rates and different levels of cooling air reentrainment at idling conditions was calculated, along with the impact of transporting the weight of the air conditioning system. Figure 5-3, which was reproduced from the paper, summarizes these results for a mid-size car ( the results for the small car are similar).

The TEWI/LCCP results in Figure 5-3 are generally consistent with the TEWI-3 results, and show that for North America, where relatively more miles are driven annually in a warmer climate, the energy savings of an HFC-134a based system compared to CO₂ and the associated reduced indirect warming more than offset the direct warming of refrigerant emissions, making HFC-134a the best LCCP alternative. In Northern Europe, where the climate is milder and fewer kilometers are driven annually, the LCCP for CO₂ is less than the LCCP for HFC-134a, although this relationship would likely reverse in a warmer climate.

Figure 5-3: TEWI variations for the HFC-134a and CO2 systems for the Mid-size car. (Figure 12 from the GMR/ORNL study [Sumantran, et al., 1999])

Unlike other common uses of fluorocarbons, automobile air conditioning systems are mounted on a platform that is repeatedly exposed to the risk of road collision damage.

Of the three refrigerants being addressed for mobile air conditioning - HFC-134a, hydrocarbons HC-600a/HC-290, and carbon dioxide - none have high toxicity concerns. However, HC-600a and HC-290 (isobutane and propane) are highly flammable and carbon dioxide operates at significantly higher pressures. The question of safely containing CO₂ pressures during operation, repair/maintenance, collision, and disposal at the end of life requires accurate definition of specific criteria. CO₂ is toxic only in high concentrations, a potential concern if the evaporator were to rupture suddenly. It is known that a rupture in the engine compartment can cause sheet metal to crumple.

In view of the greater risk for collisions of motor vehicles (than for stationary equipment) that could result in damage to the areas of the vehicle where air conditioning system components and tubing are installed, the question of fire safety when using a hydrocarbon refrigerant requires careful examination.

Automobiles already contain a significant quantity of hydrocarbons - the 10 to 25 gallons (40 to 100 liters) of gasoline that fuels the engine. In comparison to this amount of flammable liquid, a 1 to 2 lb. hydrocarbon refrigerant charge could be characterized as adding very little incremental risk. However, there are several fundamental differences between the fuel system and the refrigerant containing parts of the A/C system:

• Under most conditions, the vapor pressure of the refrigerant is well above atmospheric pressure, so that any rupture of refrigerant containing lines and components will result in a rapid release of the refrigerant
• The fuel line follows a simple, direct path from the fuel tank to the engine and is located well in from the vehicle exterior (providing protection from most potential collision damage) and entirely outside the passenger compartment. The refrigerant lines, of necessity, have several paths within the engine compartment. To receive adequate cooling airflow, the condenser must be located at the front of the car, leaving it vulnerable to damage in front end collisions.
• Fuel systems have evolved over many years to meet stringent fire safety requirements. While the same could, in principle, be done with A/C systems using a flammable refrigerant, it would be a significant engineering effort, and the resulting impact on vehicle cost and weight or the system COP is not known at this time.

In the cradle-to-grave lifecycle of an automobile having a hydrocarbon refrigerant in the air conditioning system, there are several distinct events where fire safety issues must be considered:

• Refrigerant Manufacture and Transport. Isobutane and propane are both already commercially available and manufactured in production quantities. Safety procedures for production and transportation are well established, the cost of implementation is included in the market prices, and any residual fire risk is at levels that have been accepted by the producers, their employees, and society at large.
• Vehicle Assembly. There are substantive safety issues associated with introducing a flammable working fluid to the automobile assembly line. System charging on the vehicle assembly line, which is primarily a question of investment in the necessary equipment - explosion proof electricals, hydrocarbon vapor monitors, emerging ventilation systems, and fire suppression - to ensure a safe manufacturing operation. The safety issue in this phase of the product lifecycle is manageable, but at a steep cost.
• Leakage During the Life of the Vehicle. Vehicle air conditioning systems will all eventually lose refrigerant through fittings, the compressor shaft seal, O-rings, and permeation through hoses. Despite environmentally driven efforts to reduce refrigerant leakage, finite rates of working fluid leakage from automobile air conditioning systems remains a fact of life. However, the leakage rate is measured in small fractions of a pound per year, far too low of a leakage rate to allow a significant amount of a flammable refrigerant to accumulate at a flammable concentration in sufficient quantity to cause a serious fire. Even the most tightly constructed garage has sufficient air infiltration to dilute normal leakage/permeation to concentrations several orders of magnitude below the lower flammable limit. Some areas of potential concern are the condition of refrigerant containing components of older cars and gradual corrosion in certain geographical areas having higher than average humidity and local atmospheric concentrations of corrosive species.
• Servicing and Repairs. As with vehicle assembly, there are workplace safety issues to consider with flammable working fluid in automobile air conditioners. The equipment and infrastructure needed to reclaim rather than vent the refrigerant has already been developed for CFC-12 and HFC-134a, but with a flammable refrigerant, fire safety equipment and procedures are needed to ensure that flammable refrigerant releases and ignitions do not occur. A significant factor and difficulty is the large number of automobile service establishments, with different levels of training and staff turnover. Ensuring that appropriate safety standards would be maintained at all times throughout the industry would be a major challenge.
• Collision Damage. This is the most significant risk area and is discussed below in greater detail.
• Disposal. When the vehicle is scrapped, any residual refrigerant can be reclaimed before the vehicle is stripped and shredded or crushed. Refrigerant reclamation equipment has been developed for this purpose. The refrigerant reclamation activity with flammable refrigerants will require fire safety procedures and specialized training. As is the case with servicing, the large number of geographically displaced automobile salvage operations leads to some concern about the consistency with which vigorous safety procedures would be followed.

As illustrated in Figure 5-1, in a typical automobile air conditioning system, interior cooling is provided by a direct expansion evaporator located in the climate control ductwork, within the passenger compartment. An alternative arrangement for a hydrocarbon refrigerant, illustrated in Figure 5-2, locates the evaporator in the engine compartment. The evaporator cools a secondary coolant which is circulated to an air cooling heat exchanger located in the climate control ductwork where the evaporator normally would be located. This arrangement keeps the flammable refrigerant outside the passenger compartment. The arrangement of the compressor, condenser, and interconnecting refrigerant hoses and lines is similar for both configurations.

A large scale program to evaluate the fire safety of a conventional configuration air conditioning system with hydrocarbon refrigerant, or to redesign mobile a/c systems to use a hydrocarbon refrigerant with an acceptable level of fire safety has never been carried out, either within the automobile industry or under public sector sponsorship. Such a program would include fault-tree risk analysis, component testing, and collision testing and redesign and development effort as well, if the objective were to design a fire-safe hydrocarbon refrigerant based system. Some preliminary studies and tests have been undertaken.

In 1991, Dieckmann and Bentley did a preliminary fire risk analysis of using a hydrocarbon refrigerant in a conventional a/c system configuration. The study was focussed on the fire risk associated with collision, not the other phases of the product life cycle. The preliminary conclusion of the study was that fire risks could prove to be acceptably low in some cultures, but unacceptably high in others no matter how low the risk is made. Taking the viewpoint of a burned individual, any risk, no matter how small would be unacceptably high in hindsight. Acceptable risk can also be a moving target over time, such that what is acceptable today may not be acceptable tomorrow. Extensive testing and data collection to identify and validate any design changes needed to reduce fire probabilities can help in understanding the issues, but there is no guarantee that the final product will be acceptable over its full useful lifetime, without risk of an expensive recall. The specific points relative to fire risk in the report conclusions noted the major uncertainties involved, as did the detailed treatment in the report:

• The need for a larger sample of collision damaged cars where the observed breach of refrigerant containment can be related to the collision damage extent and location
• Sampling and analysis for the full range of car model sizes and configurations - this study focussed on a typical mid-sized car
• An experimental data based determination of overall ignition probabilities associated with a release of hydrocarbon refrigerant
• Full scale collision testing would be required to verify the fire safety of a resulting design configuration

Considerably more work would be needed to develop reasonably definitive estimates of passenger compartment and engine compartment fire risks and injury risks. This work would involve a combination of field data collection, characterization of potential ignition sources (both in the passenger compartment and in the engine compartment), and collision testing. Again, the report did not address refrigerant retrofit directly; in view of the level of front end collision damage related refrigerant system rupture in the engine compartment that was observed in the limited field survey work that was undertaken in the study, and the absence of hard data on ignition sources and ignition probabilities, the study did not provide a firm basis to conclude that the resulting occurrence of engine compartment fires would be acceptably low for all cultures.

It is important to note that determination of what constitutes an "acceptably low" level of fire risk was beyond the scope of the study; in the report, the estimated fire probabilities are compared with current rates of automobile fires and collisions and resulting injuries, but the fundamental question of what risk level is acceptable was not addressed and is fundamentally a policy decision involving factors beyond the technical factors addressed in the study. For vehicle manufacturers, however, the risk of a massive recall of AC systems (class-action) in some cultures (even if the risk can be made very low) may be too great to even consider hydrocarbon based systems.

Interestingly, the preliminary results indicated that the primary fire issue with a hydrocarbon refrigerant may be collision-related fires originating in the engine compartment, rather than the passenger compartment. In the limited field sample (10 cars, a sample far to small to be the basis of a firm conclusion) of cars that had suffered severe (intrusion in excess of 12 inches) "A" pillar area passenger side impacts, the evaporators and refrigerant lines had not been punctured at all. In several instances, the evaporator had been displaced from its original position by close to 12 inches.

On the other hand, front-end collisions beyond minor "fender-benders", resulted in a high percentage of refrigerant line ruptures. In the collision damaged cars examined in the limited field study, the hood latch often was driven into the upper rows of the condenser tubing, puncturing one or more tubes. Some relatively simple design charges could potentially reduce the susceptibility of AC system components to front-end collision damage, but no development program to this end has been pursued to date. It must be emphasized that confining flammable refrigerant releases and potential ignitions to the engine compartment does not satisfy the fire safety issue (the inherent assumption behind configuring an HC system with a secondary coolant loop to the passenger compartment). Beyond the obvious increase in property loss, a hydrocarbon refrigerant fire in the engine compartment can spread rapidly, igniting plastic materials, the fuel, and in fairly short order can place at grave risk a passenger who is trapped by collision damage or unable to move due to injury.

For a variety of reasons (in the U.S., product liability risk is one significant reason), the automobile industry has not invested the extensive effort and resources that would need to be invested to address the technical and societal issues outlined above.

The potential safety issues associated with high pressure CO₂ systems have not been studied in depth. Two issues have been identified:

• High pressure gas handling, the potential for tubing and hose ruptures to be quite violent
• The potential consequences of an evaporator rupture and rapid release of the entire CO₂ charge into the passenger compartment. A rough estimate is that a 1 kg CO₂ charge released into the vehicle interior would increase the CO₂ concentration to 20% - a level that results in rapid unconsciousness followed by death. More work is needed to determine the likely duration of high CO₂ concentrations compared to human tolerance for these concentrations.

[ << Previous | Contents | Next >> ]