- Email: [email protected]
Performance analysis of a feasible air-cycle refrigeration system for road transport
International Journal of Refrigeration 28 (2005) 381–388 www.elsevier.com/locate/ijrefrig
Performance analysis of a feasible air-cycle refrigeration system for road transport Stephen W.T. Spencea,*, W. John Doranb, David W. Artta, G. McCulloughc a
School of Mechanical and Manufacturing Engineering, Queen’s University of Belfast, Ashby Building, Stranmillis Road, Belfast BT9 5AH, Northern Ireland b School of Engineering, Letterkenny Institute of Technology Port Road, Letterkenny, Co. Donegal, Ireland c School of Mechanical Engineering, Queen’s University of Belfast, Belfast, Northern Ireland Received 5 February 2004; received in revised form 26 July 2004; accepted 12 August 2004 Available online 30 November 2004
Abstract The performance of an air-cycle refrigeration unit for road transport, which had been previously reported, was analysed in detail and compared with the original design model and an equivalent Thermo King SL200 vapour-cycle refrigeration unit. Poor heat exchanger performance was found to be the major contributor to low coefficient of performance values. Using stateof-the-art, but achievable performance levels for turbomachinery and heat exchangers, the performance of an optimised aircycle refrigeration unit for the same application was predicted. The power requirement of the optimised air-cycle unit was 7% greater than the equivalent vapour-cycle unit at full-load operation. However, at part-load operation the air-cycle unit was estimated to absorb 35% less power than the vapour-cycle unit. The analysis demonstrated that the air-cycle system could potentially match the overall fuel consumption of the vapour-cycle transport refrigeration unit, while delivering the benefit of a completely refrigerant free system. q 2004 Elsevier Ltd and IIR. All rights reserved. Keywords: Refrigerated transport; Air-cycle system; Research; Thermodynamic cycle; Air; Performance; COP
Analyse de la performance d’un syste`me frigorifique a` cycle a` air utilise´ dans le transport routier Mots cle´s: Transport frigorifique ; Refroidisseur d’air ; Recherche ; Cycle thermodynamique ; Air ; Performance ; COP
1. Introduction The concept of air-cycle refrigeration was identified in the early 1800s and the first commercial air-cycle machine appears to have been in service in 1844. A succinct
* Corresponding author. E-mail address: [email protected] (S.W.T. Spence). 0140-7007/$35.00 q 2004 Elsevier Ltd and IIR. All rights reserved. doi:10.1016/j.ijrefrig.2004.08.005
historical account of developments in the field of air-cycle refrigeration is provided by Bhatti [1]. The reciprocating compression and expansion machinery used for early air-cycle machines rendered the systems inefficient and they were replaced by CO2 vapour compression systems prior to the development of chlorofluorocarbon refrigerants. However, awareness of the environmental risks associated with using HCFC and HFC refrigerant fluids has spurred interest in alterative, natural refrigerant fluids that can deliver safe and
382
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
Nomenclature/Abbreviations COP CAU
coefficient of performance cold air unit
sustainable refrigeration in the future. Today’s highly efficient turbomachinery, which was not available to early air-cycle systems, has enhanced the performance of the air-cycle. A previous paper by Spence et al. [2] has reported the design, construction and testing of an air-cycle refrigeration unit for road transport. The programme was supported by Enterprise Ireland and Thermo King (Ireland). The unit constructed was a demonstrator plant that was directly comparable to Thermo King’s SL200 trailer refrigeration unit. The demonstrator incorporated commercially available components that were not optimised for the air-cycle system and consequently the system would not be capable of achieving the optimum performance. Testing of the demonstrator unit on Thermo King’s calorimeter test facility confirmed that the original objective had been met, which was to demonstrate that an air-cycle system could fit within the existing restrictive physical envelope of the SL200 unit and develop an equivalent level of cooling power to the existing vapour-cycle unit. The measured performance of the air-cycle demonstrator is summarised in Table 1. For comparison; the standard SL200 vapour-cycle unit delivered 7.2 kW of cooling duty at K20 8C and 12 kW at 0 8C. As previously reported, the fuel consumption of the air-cycle demonstrator was much greater than that of the SL200 vapour-cycle unit. At full load operation, the air-cycle fuel consumption was over three times greater than the vapour-cycle unit, although at part load operation the fuel consumption penalty reduced from over 200% to around 80%. Since the air-cycle demonstrator had been constructed using modified commercially available turbomachinery and compromised heat exchanger configurations, achieving good energy efficiency was never an expectation. This paper reports detailed measurements taken throughout the air-cycle demonstrator system and identifies the potential performance improvements necessary for the air-cycle system to compete on energy efficiency terms with the standard vapour-cycle unit.
QUB LYIT
Queen’s University Belfast Letterkenny Institute of Technology
2. Air-cycle demonstrator plant Instrumentation was attached to the air-cycle demonstrator plant to measure temperature and pressure at each step around the cycle. Fig. 1 shows the two-stage compression open air-cycle system used for the demonstrator plant, which is referred to as the ‘boot-strap’ configuration. The diagram indicates the location of the measurement stations, each of which is numbered. Table 2 reports the average values of temperature and pressure measured at each of the three operating conditions of interest. Unfortunately an instrumentation problem during testing meant that the pressure at station 6, the aftercooler outlet, was not measured correctly. The air-cycle demonstrator plant did not represent an ideal configuration for measuring turbomachinery efficiency, mainly because of heat transfer effects and temperature gradients in ducts. Prior to construction of the demonstrator plant, the turbomachinery components had been tested in isolation to determine their performance characteristics. Consequently, while efficiencies for turbomachinery components could be calculated directly from the measurements in Table 2, the measured pressure ratio and speed were used to determine the turbine and compressor efficiencies from the previously obtained performance maps. The mass flow rate of air through the system was also determined in this way, since it had not been measured directly on the demonstrator plant. A thermodynamic model was developed based on the measured conditions in the demonstrator plant and the known performance characteristics of the various components. The model was used to check parameters such as the work balance between the CAU turbine and compressor, and to calculate the power input to the primary compressor and power dissipated through bearing friction in the CAU. Table 3 reports the component performances at each operating condition, which were determined through the use of both the model and the experimental measurements. Due to the problem with the
Table 1 Measured performance for the air-cycle demonstrator plant
Cooling capacity (W) Ambient temperature (8C) Trailer temperature (8C) Discharge air temperature (8C) Engine speed (rpm)
Full-load, K20 8C
Part-load, K20 8C
Full-load, 0 8C
7800 29.3 K20.0 K46.4 2210
3400 29.9 K20.0 K35.8 1760
9500 30.6 0.0 K29.4 2210
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
383
Fig. 1. Schematic layout of air-cycle demonstrator unit with measurement stations numbered.
pressure measurement at the aftercooler outlet, the pressure drop recorded for the aftercooler in Table 3 was determined from isolated heat exchanger flow tests rather than from the measurements taken during the tests of the air-cycle demonstrator plant. Table 3 also includes the assumed component performances used in the design phase for the plant, which was described by Spence et al. [2]. Using the model, the power input at the primary compressor of the demonstrator plant was estimated at 22,140 W, which was 3272 W larger than the 18,868 W of input power calculated for the original design analysis. Fig. 2 was produced using
the model to estimate the performance impact of each stage of the air-cycle demonstrator unit at the K20 8C full-load condition. The plot shows the percentage of the 3272 W power input discrepancy that can be attributed to over/under-performance of each component. A negative value in Fig. 2 indicates aspects of the cycle that were underperforming and corresponded to additional input power, while positive values show power savings relative to the design model. From comparison of the values in the ‘Design prediction’ column of Table 3 with those determined from the
Table 2 Measurements at each step around the air-cycle system Full load, K20 8C, 7800 W
Part load, K20 8C, 3400 W
Full load, 0 8C, 9500 W
Measurement location
Station no.
Pressure and temperature
Ambient
1
Recuperator inlet from cold space Recuperator outlet/ first compressor inlet First compressor outlet/second compressor inlet Second compressor outlet/aftercooler inlet Aftercooler outlet/ recuperator inlet Recuperator outlet/ turbine inlet Turbine outlet
2
P (kPa abs) T (8C) P (kPa abs) T (8C) P (kPa abs) T (8C)
99.2 28.7 99.1 K20.3 96.9 43.2
98.8 29.1 98.8 K20.3 97.6 34.8
99.2 29.3 99.7 0.0 97.4 48.1
4
P (kPa abs) T (8C)
178.6 119.1
147.6 84.9
177.2 123.5
5
P (kPa abs) T (8C)
236.5 162.7
185.4 116.6
238.8 168.7
6
8
Cold space inlet
9
P (kPa T (8C) P (kPa T (8C) P (kPa T (8C) P (kPa T (8C) (rpm)
– 61.0 230.5 K2.5 99.1 K48.3 99.1 K47.2 59,515
– 48.6 181.0 K5.0 98.8 K37.4 98.8 K36.7 47,396
– 62.7 232.3 12.8 99.7 K32.1 99.7 K30.8 59,515
58,678 0.29
49,401 0.21
60,381 0.30
First compressor speed Cold air unit speed Air mass flow rate
3
7
(rpm) (kg/s)
abs) abs) abs) abs)
384
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
Table 3 Individual component performance at each operating condition Component
Performance parameter
First compressor Second compressor Turbine Aftercooler
Efficiency (%) Efficiency (%) Efficiency (%) Effectiveness (%) DP (tubes) (Pa) Effectiveness (%) DP (tubes) (Pa) DP (fins) (Pa) Power loss (W)
Recuperator
CAU bearings a
Design prediction, K20 8C, 7500 W
Full-load, 0 8C, 9500 W
Full-load, K20 8C, 7800 W
Part-load, K20 8C, 3400 W
Optimised air-cycle system
78.0 78.0
79.0 77.0
78.0 75.0
79.0 77.0
81.0 81.0
80.0 80.0
80.5 76
79.0 77
80.5 76
85.0 93
1300 80.0
724a 78
515a 78
775a 78
150 93
1500 500 2000
5352 2270 1073
3892 1189 462
5504 2270 1073
600 250 100
The aftercooler DP values were determined from isolated flow tests rather than tests of the complete air-cycle demonstrator.
performance test at full load at K20 8C, and from observation of Fig. 2, it is evident that while the turbomachinery met and exceeded the original performance objectives, the anticipated heat exchanger performance was not realised. The efficiency deficit of the demonstrator unit by comparison with the original design analysis was due to three main areas; low heat exchanger effectiveness, high recuperator pressure loss and a significant amount of heat soakage into the cold air.
Fig. 2. Distribution of the power discrepancy existing between the tested air-cycle demonstrator unit and the design model. Negative values indicate additional power absorbed by the demonstrator unit compared with the original design model.
The impact of the under-performing heat exchangers on the cycle efficiency was substantial. Although the pressure loss incurred by the aftercooler was lower than had been anticipated at the design stage, this benefit was far outweighed by the excessive pressure losses in the recuperator. The low aftercooler effectiveness was responsible for one third of the additional 3272 W of input power necessary to produce the duty, the recuperator effectiveness cost a further 17% and the combined effect of excessive heat exchanger pressure loss contributed 52% of the additional power input. No allowance was made during the original design analysis for performance deterioration due to heat soak into the cycle. Places where heat flow into the air would have had a detrimental effect included the entire recuperator and the entire turbine stage including any downstream ducting. On the demonstrator plant, the recuperator had been mounted just a few millimetres above the engine, and in close proximity to the exhaust manifold, with only a thin layer of insulation between the two. Considering that the recuperator inlet temperature was K20 8C while the diesel engine water temperature was 90 8C, heat flow into the aluminium recuperator was inevitable. Heat flow into the recuperator would have caused a temperature rise at turbine inlet (station 7 in Fig. 1) increasing the turbine outlet temperature, and would have had the effect of reducing the measured value of recuperator effectiveness. The turbine inlet was connected directly to the outlet header from the recuperator to minimise heat flow into the cold air stream. The compressor of the CAU had an outlet temperature of 163 8C, so heat would flow from the compressor through the bearing housing to the turbine. Precautions were taken during the design of the CAU to reduce the potential for heat conduction from the compressor. The turbine efficiency measured during the turbine performance testing was 80.5%, while the value determined from the temperature measurements during the
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
demonstrator tests was lower at 79.0%, suggesting that heat was being conducted into the turbine stage and reducing the temperature drop across the turbine. The rate of heat conduction corresponding to this 1.5% discrepancy in efficiency was calculated at 260 W. Heat transfer was also possible between the turbine inlet air, through the shroud and exhaust diffuser, into the turbine outlet air. Any heat transfer from these means would have been reflected in the original 80.5% efficiency value measured during turbine tests. Use of alternative materials, such as plastics, may have reduced heat conduction and increased the efficiency of the turbine stage above 80.5%. An exhaust silencer was fitted at outlet from the turbine exhaust diffuser. The sound absorption material also served as heat insulation for the exit duct between turbine outlet and the point where the air discharged from the demonstrator unit. Temperature was measured at the diffuser outlet/silencer inlet (station 8 in Fig. 1) and also at exit from the silencer (station 9 in Fig. 1). The temperature measurements in Table 2 show a temperature rise through the exhaust silencer of 0.9 8C at the design operating condition of K20 8C. This temperature rise, which was also apparent at the other operating conditions, corresponded to a heat flow of 260 W into the cold air. Heat flow into the cold air stream through the turbine stage and the exhaust silencer appeared to represent a significant performance penalty to the demonstrator unit since around 520 W of cooling duty was being lost. The cost of compensating for this 520 W of lost duty was around 1300 W of additional input power. Although it was not possible to quantify from the measurements, it was likely that an even greater amount of duty was lost through heat flow into the large recuperator from the proximity of the hot engine. The values obtained for power dissipation in the CAU bearings were determined by using the model to balance turbine and compressor work, and are not to be regarded as accurate measurements. The levels of bearing power loss were lower than those used in the design analysis but are similar to typical values for a commercial turbocharger of the same shaft size and speed. The design analysis had anticipated higher bearing losses since it was assumed that the oil would be colder in the CAU than in a turbocharger application and consequently lead to higher viscous losses. The lower than anticipated power losses in the CAU bearings offset some of the performance penalty resulting from the heat exchangers. Thus far, only the power consumed within the refrigeration cycle has been considered. The complete unit incorporated two fan systems. In the original SL200 refrigeration unit a pair of fans was used to force ambient air through the condenser while a further fan was used to circulate air over the evaporator and around the trailer. In the air-cycle demonstrator unit, the condenser fans were used to force ambient air through the aftercooler heat exchangers, but the evaporator fan was excluded from the system. Since the evaporator fan was absent, the
385
demonstrator unit was not powering the fan or producing the refrigeration duty necessary to overcome the power dissipated by the evaporator fan into the cold air. In a final application, while not having an evaporator, the air-cycle unit would need a circulation fan (or some other development of the turbine exhaust system) to achieve the necessary air velocity and volumetric flow rate in a refrigerated trailer. However, less fan power would be needed by comparison with conventional units due to the absence of the pressure drop across the evaporator. In order to account for the power consumed by the air-cycle demonstrator unit by comparison with the standard SL200 vapour-cycle refrigeration unit, Table 4 breaks down the power consumption for the various different systems being discussed. In order to protect the confidentiality of the manufacturer’s data, fan and compressor power values for the SL200 have not been included in Table 4. However, a value calculated by the authors for the overall coefficient of performance (COP) of the SL200 has been included to allow comparison with the air-cycle unit. The table includes the power consumption of the aircycle system that was predicted using the thermodynamic model during the design phase, although the power values now correspond to a duty of 7800 W in order to allow effective comparison. As highlighted, there was no evaporator fan included in the design model. The power input determined for the demonstrator unit following testing is also listed. The condenser fans on the air-cycle demonstrator absorbed considerably more power than on the standard SL200 unit because the diesel engine was operated at a higher speed. The power absorbed by the condenser fans on the demonstrator plant was calculated at 4389 W. Table 4 shows that the COP of the air-cycle demonstrator is almost half the COP for the SL200 unit. However, the engine power figures listed take no account of losses in the step-up gearbox used to drive the primary compressor. It was not possible to determine the gearbox loss from the test measurements. However, judging from the amount of heat rejected by the gearbox, the losses were significant and probably amounted to several kilowatts. Therefore, the actual COP of the air-cycle demonstrator unit was less than half that of the SL200 unit. The measured fuel consumption of the air-cycle demonstrator was approximately three times greater than the SL200 unit. However, the diesel engine was very heavily loaded, leading to poorer thermal efficiency and a disproportionate increase in fuel consumption.
3. Optimised air-cycle unit for road transport Section 2 analysed and discussed the performance of an air-cycle demonstrator which, while it matched the cooling duty of the existing vapour-cycle system, was clearly unacceptable on the basis of fuel efficiency. Having developed a satisfactory model of the system and used it to analyse the deficiencies of the air-cycle demonstrator unit, the study was carried further to assess the potential of
386
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
Table 4 Power breakdown for different configurations
First comp power (W) Effective cooling duty Total cooling duty Ref. cycle COP Evap. fan power (W) Cond. fan power (W) Total engine power Overall COP
Thermo King SL200
Design prediction scaled to 7800 W
Air-cycle demo. unit
–
18,868
22,140
9640
10,876
–
7800
7800
7800
7800
– – –
7800 0.413 0
7800 0.352 0
7800 0.809 0
8800 0.809 1000
–
2950
4389
2950
2950
– 0.563
21,818 0.358
an optimised air-cycle unit. The demonstrator unit was not optimised because it used turbomachinery that had been modified from other applications and heat exchangers that were severely compromised by the existing layout of fans and belts within the SL200 chassis. An optimised unit would use specifically designed turbomachinery to achieve stateof-the-art efficiencies, a more efficient bearing arrangement, higher-performance heat exchangers and a chassis that was designed to accommodate the new refrigeration cycle efficiently. The thermodynamic model was used to calculate the performance of an optimised system using the component performance figures listed in the last column of Table 3. With the optimised component efficiencies listed, the power consumption of an optimised air-cycle unit was calculated and listed under two headings in Table 4. Firstly, the optimised system is listed with no evaporator fan power to allow direct comparison with the power consumption of the existing air-cycle demonstrator unit. Secondly, a circulation fan with an absorbed power of 1 kW is included into the optimised air-cycle system, which should comfortably meet the published air circulation capabilities of the existing SL200 unit. While the proposed air-cycle system is an open system and does not have a heat exchanger on the low temperature side, a circulation fan is still necessary to achieve the required level of air circulation in a refrigerated trailer packed with produce. The power of the proposed circulation fan is substantially lower than the power of the existing evaporator fan in the SL200 unit since the air-cycle unit does not have the air pressure loss across the evaporator. In both cases, the condenser fan power, used for the aftercooler heat exchangers, has been calculated at 2950 W. With the inclusion of the circulation fan, the COP of the optimised air-cycle unit is now just 7% less than the SL200 vapour-cycle unit, although this figure does not include any allowance for power transmission losses to drive the primary compressor. Fig. 3 shows the different areas of the system where the efficiency improvements are made by comparing the
26,529 0.294
Optimised model
12,590 0.620
Optimised model with circulation fan
14,826 0.526
optimised air-cycle with the original design analysis. The graph shows how much each area of improvement contributes to the 8 kW primary compressor power saving of the optimised system compared with the original prototype design model. Heat exchanger effectiveness was clearly the most important parameter, with further improvements in pressure losses bringing only minor benefits. Improvements in compressor performance were also overshadowed by the benefits of additional turbine efficiency and reduced bearing losses. Improvements in component performance mean that the air-cycle COP has increased to 0.809, which is achieved at an overall system pressure ratio of 1.834, corresponding to a primary compressor pressure ratio of 1.4. Lower pressure ratios would yield further improvements in COP, but the volumetric flow of air in the cycle would increase as a
Fig. 3. Distribution of the power savings achieved in the optimised air-cycle system by comparison with the original design model.
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
consequence making the heat exchanger design more challenging. The critical components in the optimised air-cycle unit are clearly the recuperator and the aftercooler. The challenge is to achieve very high levels of heat exchanger effectiveness with modest pressure losses, within the available physical envelope for a trailer refrigeration unit. Plate fin heat exchangers are the technology most likely to meet this demanding application. Adoption of air bearings in the CAU would ensure low losses, eliminate the possibility of any oil contaminating the air stream and enhance turbomachinery efficiency by permitting smaller turbine and compressor blade clearances. However, forces from road vibration may pose a problem for an air bearing system, particularly if the system is not operating and the shaft is static. A high speed electric drive would seem like an attractive option for powering the primary compressor. Using a diesel engine coupled to a generator would remove the mechanical link between the engine and the primary compressor. The permanent magnet motors used in high speed electric drives typically have very high efficiencies (O95%). Although the efficiency of the electric generator also needs to be considered in the overall power transmission system, the efficiency of the electric drive would probably exceed what could be achieved by a conventional gearbox; mechanical reliability would also surpass that of a gearbox. The use of electrically powered condenser and circulation fans could complement the electrically driven compressor; providing the potential for more flexible energy saving control of fan speeds and releasing valuable space within the chassis through the elimination of mechanical belts and pulleys. The potential benefits of the electric drive option would need to be analysed in greater detail. The performance tests of the air-cycle demonstrator unit revealed that the large fuel consumption penalty by comparison with the standard vapour-cycle unit at fullload operation was substantially reduced at part-load operation. With correct design, the efficiency of the turbomachinery in the optimised air-cycle plant should be maintained at part-load conditions, while the heat exchanger performance would increase with the lower air flow rates at part-load. Therefore, the full-load COP of 0.526 for the optimised air-cycle system would be maintained at part-load operation. By contrast, a 56% reduction in the duty of the standard vapour-cycle system corresponded to a fuel consumption reduction of just 26%. Consequently, the aircycle system could undercut the fuel consumption of the vapour-cycle unit at the part-load condition by 35%. For long haul refrigerated transport, the trailer refrigeration system spends much of its operating time at part-load. Clearly the balance of part-load and full-load operation varies widely between different applications. However, the preceding consideration of an optimised air-cycle unit shows that such a system has the potential to match the overall fuel-consumption of conventional vapour cycle
387
systems, but with a complete absence of harmful refrigerant fluids. There is a need for a comprehensive life-cycle study of such a system to determine for a typical operating pattern the life-cycle costs of the air-cycle system compared with a vapour cycle system. The analysis should also consider the total exhaust emissions and the absence of any refrigerant fluid and the associated recharging. The management of moisture and ice deposits has not been discussed in detail here. During testing of the air-cycle demonstrator plant, ice deposits were only apparent downstream of the turbine rotor. No ice deposits were evident in the recuperator heat exchanger. The formation of water or ice particles upstream of the turbine could lead to blade damage; however, under normal operating conditions the air temperature at turbine inlet would always be higher than the air temperature in the cold space, so there should not be any condensation at turbine inlet. Assessing the impact of moisture on an air-cycle system would require further tests over longer periods at a range of ambient and trailer temperatures and humidity levels. Bringing the potential of the optimised air-cycle system to fruition will require several new technologies to be introduced to the transport refrigeration industry: 1. 2. 3. 4.
High efficiency air–air heat exchangers High speed electric drives Air bearings Turbomachinery
Each of the technologies identified exists, although the first three are still relatively low volume and high cost technologies. The turbocharger industry successfully manufactures small-scale turbomachinery at very competitive costs. Ideally, the development of an optimised air-cycle unit for transport refrigeration would be through a collaborative group comprising an established company in the transport refrigeration market and specialised companies bringing together the necessary new technologies. While the technology is available, the amount of development work and the associated investment to bring a range of air-cycle transport refrigeration units to market would be considerable. As with automotive emission control systems, the compulsion for developing such a refrigerant free unit is likely to be legislation.
4. Conclusions The performance of an air-cycle refrigeration unit for road transport, which had been previously reported, was analysed in detail and compared with the original design model. Overall, the turbomachinery was found to satisfy the original design requirements, but the heat exchangers proved to be a major performance handicap. The heat exchanger effectiveness values were several percentage points lower than anticipated and the heat exchanger
388
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
pressure losses were excessive. As a consequence, the aircycle demonstrator unit absorbed around 25% more power than the design value and more than twice as much power as the equivalent vapour-cycle unit. Using state-of-the-art, but achievable performance levels for turbomachinery and heat exchangers, the performance of an optimised air-cycle refrigeration unit was predicted. The power requirement of the optimised air-cycle unit was 7% greater than the equivalent vapour-cycle unit at full-load operation. However, at part-load operation the air-cycle unit was estimated to absorb 35% less power than the vapour-cycle unit. The analysis demonstrated that the air-cycle system could potentially match the overall fuel consumption of the vapour-cycle transport refrigeration unit, while delivering the benefit of a completely refrigerant free system. Bringing the optimised air-cycle to fruition would require the introduction of several new technologies into the transport refrigeration sector including air bearings, plate-fin heat exchangers, turbomachinery and high speed electric drives. The amount of development
work and the associated investment to bring a range of air-cycle transport refrigeration units to market would be considerable, and would be unlikely in the absence of legislation to encourage such refrigerant free systems.
Acknowledgements The authors would like to thank Enterprise Ireland and Thermo King for funding this project, and particularly Mr John Gough for his enduring support.
References [1] M.S. Bhatti, Open air cycle air conditioning system for motor vehicles, Society of Automotive Engineers, 1998. SAE 980289. [2] S.W.T. Spence, W.J. Doran, D.W. Artt, Design, construction and testing of an air-cycle refrigeration system for road transport, Int J Refrig 27/5 (2004) 503–510.
Performance analysis of a feasible air-cycle refrigeration system for road transport Stephen W.T. Spencea,*, W. John Doranb, David W. Artta, G. McCulloughc a
School of Mechanical and Manufacturing Engineering, Queen’s University of Belfast, Ashby Building, Stranmillis Road, Belfast BT9 5AH, Northern Ireland b School of Engineering, Letterkenny Institute of Technology Port Road, Letterkenny, Co. Donegal, Ireland c School of Mechanical Engineering, Queen’s University of Belfast, Belfast, Northern Ireland Received 5 February 2004; received in revised form 26 July 2004; accepted 12 August 2004 Available online 30 November 2004
Abstract The performance of an air-cycle refrigeration unit for road transport, which had been previously reported, was analysed in detail and compared with the original design model and an equivalent Thermo King SL200 vapour-cycle refrigeration unit. Poor heat exchanger performance was found to be the major contributor to low coefficient of performance values. Using stateof-the-art, but achievable performance levels for turbomachinery and heat exchangers, the performance of an optimised aircycle refrigeration unit for the same application was predicted. The power requirement of the optimised air-cycle unit was 7% greater than the equivalent vapour-cycle unit at full-load operation. However, at part-load operation the air-cycle unit was estimated to absorb 35% less power than the vapour-cycle unit. The analysis demonstrated that the air-cycle system could potentially match the overall fuel consumption of the vapour-cycle transport refrigeration unit, while delivering the benefit of a completely refrigerant free system. q 2004 Elsevier Ltd and IIR. All rights reserved. Keywords: Refrigerated transport; Air-cycle system; Research; Thermodynamic cycle; Air; Performance; COP
Analyse de la performance d’un syste`me frigorifique a` cycle a` air utilise´ dans le transport routier Mots cle´s: Transport frigorifique ; Refroidisseur d’air ; Recherche ; Cycle thermodynamique ; Air ; Performance ; COP
1. Introduction The concept of air-cycle refrigeration was identified in the early 1800s and the first commercial air-cycle machine appears to have been in service in 1844. A succinct
* Corresponding author. E-mail address: [email protected] (S.W.T. Spence). 0140-7007/$35.00 q 2004 Elsevier Ltd and IIR. All rights reserved. doi:10.1016/j.ijrefrig.2004.08.005
historical account of developments in the field of air-cycle refrigeration is provided by Bhatti [1]. The reciprocating compression and expansion machinery used for early air-cycle machines rendered the systems inefficient and they were replaced by CO2 vapour compression systems prior to the development of chlorofluorocarbon refrigerants. However, awareness of the environmental risks associated with using HCFC and HFC refrigerant fluids has spurred interest in alterative, natural refrigerant fluids that can deliver safe and
382
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
Nomenclature/Abbreviations COP CAU
coefficient of performance cold air unit
sustainable refrigeration in the future. Today’s highly efficient turbomachinery, which was not available to early air-cycle systems, has enhanced the performance of the air-cycle. A previous paper by Spence et al. [2] has reported the design, construction and testing of an air-cycle refrigeration unit for road transport. The programme was supported by Enterprise Ireland and Thermo King (Ireland). The unit constructed was a demonstrator plant that was directly comparable to Thermo King’s SL200 trailer refrigeration unit. The demonstrator incorporated commercially available components that were not optimised for the air-cycle system and consequently the system would not be capable of achieving the optimum performance. Testing of the demonstrator unit on Thermo King’s calorimeter test facility confirmed that the original objective had been met, which was to demonstrate that an air-cycle system could fit within the existing restrictive physical envelope of the SL200 unit and develop an equivalent level of cooling power to the existing vapour-cycle unit. The measured performance of the air-cycle demonstrator is summarised in Table 1. For comparison; the standard SL200 vapour-cycle unit delivered 7.2 kW of cooling duty at K20 8C and 12 kW at 0 8C. As previously reported, the fuel consumption of the air-cycle demonstrator was much greater than that of the SL200 vapour-cycle unit. At full load operation, the air-cycle fuel consumption was over three times greater than the vapour-cycle unit, although at part load operation the fuel consumption penalty reduced from over 200% to around 80%. Since the air-cycle demonstrator had been constructed using modified commercially available turbomachinery and compromised heat exchanger configurations, achieving good energy efficiency was never an expectation. This paper reports detailed measurements taken throughout the air-cycle demonstrator system and identifies the potential performance improvements necessary for the air-cycle system to compete on energy efficiency terms with the standard vapour-cycle unit.
QUB LYIT
Queen’s University Belfast Letterkenny Institute of Technology
2. Air-cycle demonstrator plant Instrumentation was attached to the air-cycle demonstrator plant to measure temperature and pressure at each step around the cycle. Fig. 1 shows the two-stage compression open air-cycle system used for the demonstrator plant, which is referred to as the ‘boot-strap’ configuration. The diagram indicates the location of the measurement stations, each of which is numbered. Table 2 reports the average values of temperature and pressure measured at each of the three operating conditions of interest. Unfortunately an instrumentation problem during testing meant that the pressure at station 6, the aftercooler outlet, was not measured correctly. The air-cycle demonstrator plant did not represent an ideal configuration for measuring turbomachinery efficiency, mainly because of heat transfer effects and temperature gradients in ducts. Prior to construction of the demonstrator plant, the turbomachinery components had been tested in isolation to determine their performance characteristics. Consequently, while efficiencies for turbomachinery components could be calculated directly from the measurements in Table 2, the measured pressure ratio and speed were used to determine the turbine and compressor efficiencies from the previously obtained performance maps. The mass flow rate of air through the system was also determined in this way, since it had not been measured directly on the demonstrator plant. A thermodynamic model was developed based on the measured conditions in the demonstrator plant and the known performance characteristics of the various components. The model was used to check parameters such as the work balance between the CAU turbine and compressor, and to calculate the power input to the primary compressor and power dissipated through bearing friction in the CAU. Table 3 reports the component performances at each operating condition, which were determined through the use of both the model and the experimental measurements. Due to the problem with the
Table 1 Measured performance for the air-cycle demonstrator plant
Cooling capacity (W) Ambient temperature (8C) Trailer temperature (8C) Discharge air temperature (8C) Engine speed (rpm)
Full-load, K20 8C
Part-load, K20 8C
Full-load, 0 8C
7800 29.3 K20.0 K46.4 2210
3400 29.9 K20.0 K35.8 1760
9500 30.6 0.0 K29.4 2210
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
383
Fig. 1. Schematic layout of air-cycle demonstrator unit with measurement stations numbered.
pressure measurement at the aftercooler outlet, the pressure drop recorded for the aftercooler in Table 3 was determined from isolated heat exchanger flow tests rather than from the measurements taken during the tests of the air-cycle demonstrator plant. Table 3 also includes the assumed component performances used in the design phase for the plant, which was described by Spence et al. [2]. Using the model, the power input at the primary compressor of the demonstrator plant was estimated at 22,140 W, which was 3272 W larger than the 18,868 W of input power calculated for the original design analysis. Fig. 2 was produced using
the model to estimate the performance impact of each stage of the air-cycle demonstrator unit at the K20 8C full-load condition. The plot shows the percentage of the 3272 W power input discrepancy that can be attributed to over/under-performance of each component. A negative value in Fig. 2 indicates aspects of the cycle that were underperforming and corresponded to additional input power, while positive values show power savings relative to the design model. From comparison of the values in the ‘Design prediction’ column of Table 3 with those determined from the
Table 2 Measurements at each step around the air-cycle system Full load, K20 8C, 7800 W
Part load, K20 8C, 3400 W
Full load, 0 8C, 9500 W
Measurement location
Station no.
Pressure and temperature
Ambient
1
Recuperator inlet from cold space Recuperator outlet/ first compressor inlet First compressor outlet/second compressor inlet Second compressor outlet/aftercooler inlet Aftercooler outlet/ recuperator inlet Recuperator outlet/ turbine inlet Turbine outlet
2
P (kPa abs) T (8C) P (kPa abs) T (8C) P (kPa abs) T (8C)
99.2 28.7 99.1 K20.3 96.9 43.2
98.8 29.1 98.8 K20.3 97.6 34.8
99.2 29.3 99.7 0.0 97.4 48.1
4
P (kPa abs) T (8C)
178.6 119.1
147.6 84.9
177.2 123.5
5
P (kPa abs) T (8C)
236.5 162.7
185.4 116.6
238.8 168.7
6
8
Cold space inlet
9
P (kPa T (8C) P (kPa T (8C) P (kPa T (8C) P (kPa T (8C) (rpm)
– 61.0 230.5 K2.5 99.1 K48.3 99.1 K47.2 59,515
– 48.6 181.0 K5.0 98.8 K37.4 98.8 K36.7 47,396
– 62.7 232.3 12.8 99.7 K32.1 99.7 K30.8 59,515
58,678 0.29
49,401 0.21
60,381 0.30
First compressor speed Cold air unit speed Air mass flow rate
3
7
(rpm) (kg/s)
abs) abs) abs) abs)
384
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
Table 3 Individual component performance at each operating condition Component
Performance parameter
First compressor Second compressor Turbine Aftercooler
Efficiency (%) Efficiency (%) Efficiency (%) Effectiveness (%) DP (tubes) (Pa) Effectiveness (%) DP (tubes) (Pa) DP (fins) (Pa) Power loss (W)
Recuperator
CAU bearings a
Design prediction, K20 8C, 7500 W
Full-load, 0 8C, 9500 W
Full-load, K20 8C, 7800 W
Part-load, K20 8C, 3400 W
Optimised air-cycle system
78.0 78.0
79.0 77.0
78.0 75.0
79.0 77.0
81.0 81.0
80.0 80.0
80.5 76
79.0 77
80.5 76
85.0 93
1300 80.0
724a 78
515a 78
775a 78
150 93
1500 500 2000
5352 2270 1073
3892 1189 462
5504 2270 1073
600 250 100
The aftercooler DP values were determined from isolated flow tests rather than tests of the complete air-cycle demonstrator.
performance test at full load at K20 8C, and from observation of Fig. 2, it is evident that while the turbomachinery met and exceeded the original performance objectives, the anticipated heat exchanger performance was not realised. The efficiency deficit of the demonstrator unit by comparison with the original design analysis was due to three main areas; low heat exchanger effectiveness, high recuperator pressure loss and a significant amount of heat soakage into the cold air.
Fig. 2. Distribution of the power discrepancy existing between the tested air-cycle demonstrator unit and the design model. Negative values indicate additional power absorbed by the demonstrator unit compared with the original design model.
The impact of the under-performing heat exchangers on the cycle efficiency was substantial. Although the pressure loss incurred by the aftercooler was lower than had been anticipated at the design stage, this benefit was far outweighed by the excessive pressure losses in the recuperator. The low aftercooler effectiveness was responsible for one third of the additional 3272 W of input power necessary to produce the duty, the recuperator effectiveness cost a further 17% and the combined effect of excessive heat exchanger pressure loss contributed 52% of the additional power input. No allowance was made during the original design analysis for performance deterioration due to heat soak into the cycle. Places where heat flow into the air would have had a detrimental effect included the entire recuperator and the entire turbine stage including any downstream ducting. On the demonstrator plant, the recuperator had been mounted just a few millimetres above the engine, and in close proximity to the exhaust manifold, with only a thin layer of insulation between the two. Considering that the recuperator inlet temperature was K20 8C while the diesel engine water temperature was 90 8C, heat flow into the aluminium recuperator was inevitable. Heat flow into the recuperator would have caused a temperature rise at turbine inlet (station 7 in Fig. 1) increasing the turbine outlet temperature, and would have had the effect of reducing the measured value of recuperator effectiveness. The turbine inlet was connected directly to the outlet header from the recuperator to minimise heat flow into the cold air stream. The compressor of the CAU had an outlet temperature of 163 8C, so heat would flow from the compressor through the bearing housing to the turbine. Precautions were taken during the design of the CAU to reduce the potential for heat conduction from the compressor. The turbine efficiency measured during the turbine performance testing was 80.5%, while the value determined from the temperature measurements during the
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
demonstrator tests was lower at 79.0%, suggesting that heat was being conducted into the turbine stage and reducing the temperature drop across the turbine. The rate of heat conduction corresponding to this 1.5% discrepancy in efficiency was calculated at 260 W. Heat transfer was also possible between the turbine inlet air, through the shroud and exhaust diffuser, into the turbine outlet air. Any heat transfer from these means would have been reflected in the original 80.5% efficiency value measured during turbine tests. Use of alternative materials, such as plastics, may have reduced heat conduction and increased the efficiency of the turbine stage above 80.5%. An exhaust silencer was fitted at outlet from the turbine exhaust diffuser. The sound absorption material also served as heat insulation for the exit duct between turbine outlet and the point where the air discharged from the demonstrator unit. Temperature was measured at the diffuser outlet/silencer inlet (station 8 in Fig. 1) and also at exit from the silencer (station 9 in Fig. 1). The temperature measurements in Table 2 show a temperature rise through the exhaust silencer of 0.9 8C at the design operating condition of K20 8C. This temperature rise, which was also apparent at the other operating conditions, corresponded to a heat flow of 260 W into the cold air. Heat flow into the cold air stream through the turbine stage and the exhaust silencer appeared to represent a significant performance penalty to the demonstrator unit since around 520 W of cooling duty was being lost. The cost of compensating for this 520 W of lost duty was around 1300 W of additional input power. Although it was not possible to quantify from the measurements, it was likely that an even greater amount of duty was lost through heat flow into the large recuperator from the proximity of the hot engine. The values obtained for power dissipation in the CAU bearings were determined by using the model to balance turbine and compressor work, and are not to be regarded as accurate measurements. The levels of bearing power loss were lower than those used in the design analysis but are similar to typical values for a commercial turbocharger of the same shaft size and speed. The design analysis had anticipated higher bearing losses since it was assumed that the oil would be colder in the CAU than in a turbocharger application and consequently lead to higher viscous losses. The lower than anticipated power losses in the CAU bearings offset some of the performance penalty resulting from the heat exchangers. Thus far, only the power consumed within the refrigeration cycle has been considered. The complete unit incorporated two fan systems. In the original SL200 refrigeration unit a pair of fans was used to force ambient air through the condenser while a further fan was used to circulate air over the evaporator and around the trailer. In the air-cycle demonstrator unit, the condenser fans were used to force ambient air through the aftercooler heat exchangers, but the evaporator fan was excluded from the system. Since the evaporator fan was absent, the
385
demonstrator unit was not powering the fan or producing the refrigeration duty necessary to overcome the power dissipated by the evaporator fan into the cold air. In a final application, while not having an evaporator, the air-cycle unit would need a circulation fan (or some other development of the turbine exhaust system) to achieve the necessary air velocity and volumetric flow rate in a refrigerated trailer. However, less fan power would be needed by comparison with conventional units due to the absence of the pressure drop across the evaporator. In order to account for the power consumed by the air-cycle demonstrator unit by comparison with the standard SL200 vapour-cycle refrigeration unit, Table 4 breaks down the power consumption for the various different systems being discussed. In order to protect the confidentiality of the manufacturer’s data, fan and compressor power values for the SL200 have not been included in Table 4. However, a value calculated by the authors for the overall coefficient of performance (COP) of the SL200 has been included to allow comparison with the air-cycle unit. The table includes the power consumption of the aircycle system that was predicted using the thermodynamic model during the design phase, although the power values now correspond to a duty of 7800 W in order to allow effective comparison. As highlighted, there was no evaporator fan included in the design model. The power input determined for the demonstrator unit following testing is also listed. The condenser fans on the air-cycle demonstrator absorbed considerably more power than on the standard SL200 unit because the diesel engine was operated at a higher speed. The power absorbed by the condenser fans on the demonstrator plant was calculated at 4389 W. Table 4 shows that the COP of the air-cycle demonstrator is almost half the COP for the SL200 unit. However, the engine power figures listed take no account of losses in the step-up gearbox used to drive the primary compressor. It was not possible to determine the gearbox loss from the test measurements. However, judging from the amount of heat rejected by the gearbox, the losses were significant and probably amounted to several kilowatts. Therefore, the actual COP of the air-cycle demonstrator unit was less than half that of the SL200 unit. The measured fuel consumption of the air-cycle demonstrator was approximately three times greater than the SL200 unit. However, the diesel engine was very heavily loaded, leading to poorer thermal efficiency and a disproportionate increase in fuel consumption.
3. Optimised air-cycle unit for road transport Section 2 analysed and discussed the performance of an air-cycle demonstrator which, while it matched the cooling duty of the existing vapour-cycle system, was clearly unacceptable on the basis of fuel efficiency. Having developed a satisfactory model of the system and used it to analyse the deficiencies of the air-cycle demonstrator unit, the study was carried further to assess the potential of
386
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
Table 4 Power breakdown for different configurations
First comp power (W) Effective cooling duty Total cooling duty Ref. cycle COP Evap. fan power (W) Cond. fan power (W) Total engine power Overall COP
Thermo King SL200
Design prediction scaled to 7800 W
Air-cycle demo. unit
–
18,868
22,140
9640
10,876
–
7800
7800
7800
7800
– – –
7800 0.413 0
7800 0.352 0
7800 0.809 0
8800 0.809 1000
–
2950
4389
2950
2950
– 0.563
21,818 0.358
an optimised air-cycle unit. The demonstrator unit was not optimised because it used turbomachinery that had been modified from other applications and heat exchangers that were severely compromised by the existing layout of fans and belts within the SL200 chassis. An optimised unit would use specifically designed turbomachinery to achieve stateof-the-art efficiencies, a more efficient bearing arrangement, higher-performance heat exchangers and a chassis that was designed to accommodate the new refrigeration cycle efficiently. The thermodynamic model was used to calculate the performance of an optimised system using the component performance figures listed in the last column of Table 3. With the optimised component efficiencies listed, the power consumption of an optimised air-cycle unit was calculated and listed under two headings in Table 4. Firstly, the optimised system is listed with no evaporator fan power to allow direct comparison with the power consumption of the existing air-cycle demonstrator unit. Secondly, a circulation fan with an absorbed power of 1 kW is included into the optimised air-cycle system, which should comfortably meet the published air circulation capabilities of the existing SL200 unit. While the proposed air-cycle system is an open system and does not have a heat exchanger on the low temperature side, a circulation fan is still necessary to achieve the required level of air circulation in a refrigerated trailer packed with produce. The power of the proposed circulation fan is substantially lower than the power of the existing evaporator fan in the SL200 unit since the air-cycle unit does not have the air pressure loss across the evaporator. In both cases, the condenser fan power, used for the aftercooler heat exchangers, has been calculated at 2950 W. With the inclusion of the circulation fan, the COP of the optimised air-cycle unit is now just 7% less than the SL200 vapour-cycle unit, although this figure does not include any allowance for power transmission losses to drive the primary compressor. Fig. 3 shows the different areas of the system where the efficiency improvements are made by comparing the
26,529 0.294
Optimised model
12,590 0.620
Optimised model with circulation fan
14,826 0.526
optimised air-cycle with the original design analysis. The graph shows how much each area of improvement contributes to the 8 kW primary compressor power saving of the optimised system compared with the original prototype design model. Heat exchanger effectiveness was clearly the most important parameter, with further improvements in pressure losses bringing only minor benefits. Improvements in compressor performance were also overshadowed by the benefits of additional turbine efficiency and reduced bearing losses. Improvements in component performance mean that the air-cycle COP has increased to 0.809, which is achieved at an overall system pressure ratio of 1.834, corresponding to a primary compressor pressure ratio of 1.4. Lower pressure ratios would yield further improvements in COP, but the volumetric flow of air in the cycle would increase as a
Fig. 3. Distribution of the power savings achieved in the optimised air-cycle system by comparison with the original design model.
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
consequence making the heat exchanger design more challenging. The critical components in the optimised air-cycle unit are clearly the recuperator and the aftercooler. The challenge is to achieve very high levels of heat exchanger effectiveness with modest pressure losses, within the available physical envelope for a trailer refrigeration unit. Plate fin heat exchangers are the technology most likely to meet this demanding application. Adoption of air bearings in the CAU would ensure low losses, eliminate the possibility of any oil contaminating the air stream and enhance turbomachinery efficiency by permitting smaller turbine and compressor blade clearances. However, forces from road vibration may pose a problem for an air bearing system, particularly if the system is not operating and the shaft is static. A high speed electric drive would seem like an attractive option for powering the primary compressor. Using a diesel engine coupled to a generator would remove the mechanical link between the engine and the primary compressor. The permanent magnet motors used in high speed electric drives typically have very high efficiencies (O95%). Although the efficiency of the electric generator also needs to be considered in the overall power transmission system, the efficiency of the electric drive would probably exceed what could be achieved by a conventional gearbox; mechanical reliability would also surpass that of a gearbox. The use of electrically powered condenser and circulation fans could complement the electrically driven compressor; providing the potential for more flexible energy saving control of fan speeds and releasing valuable space within the chassis through the elimination of mechanical belts and pulleys. The potential benefits of the electric drive option would need to be analysed in greater detail. The performance tests of the air-cycle demonstrator unit revealed that the large fuel consumption penalty by comparison with the standard vapour-cycle unit at fullload operation was substantially reduced at part-load operation. With correct design, the efficiency of the turbomachinery in the optimised air-cycle plant should be maintained at part-load conditions, while the heat exchanger performance would increase with the lower air flow rates at part-load. Therefore, the full-load COP of 0.526 for the optimised air-cycle system would be maintained at part-load operation. By contrast, a 56% reduction in the duty of the standard vapour-cycle system corresponded to a fuel consumption reduction of just 26%. Consequently, the aircycle system could undercut the fuel consumption of the vapour-cycle unit at the part-load condition by 35%. For long haul refrigerated transport, the trailer refrigeration system spends much of its operating time at part-load. Clearly the balance of part-load and full-load operation varies widely between different applications. However, the preceding consideration of an optimised air-cycle unit shows that such a system has the potential to match the overall fuel-consumption of conventional vapour cycle
387
systems, but with a complete absence of harmful refrigerant fluids. There is a need for a comprehensive life-cycle study of such a system to determine for a typical operating pattern the life-cycle costs of the air-cycle system compared with a vapour cycle system. The analysis should also consider the total exhaust emissions and the absence of any refrigerant fluid and the associated recharging. The management of moisture and ice deposits has not been discussed in detail here. During testing of the air-cycle demonstrator plant, ice deposits were only apparent downstream of the turbine rotor. No ice deposits were evident in the recuperator heat exchanger. The formation of water or ice particles upstream of the turbine could lead to blade damage; however, under normal operating conditions the air temperature at turbine inlet would always be higher than the air temperature in the cold space, so there should not be any condensation at turbine inlet. Assessing the impact of moisture on an air-cycle system would require further tests over longer periods at a range of ambient and trailer temperatures and humidity levels. Bringing the potential of the optimised air-cycle system to fruition will require several new technologies to be introduced to the transport refrigeration industry: 1. 2. 3. 4.
High efficiency air–air heat exchangers High speed electric drives Air bearings Turbomachinery
Each of the technologies identified exists, although the first three are still relatively low volume and high cost technologies. The turbocharger industry successfully manufactures small-scale turbomachinery at very competitive costs. Ideally, the development of an optimised air-cycle unit for transport refrigeration would be through a collaborative group comprising an established company in the transport refrigeration market and specialised companies bringing together the necessary new technologies. While the technology is available, the amount of development work and the associated investment to bring a range of air-cycle transport refrigeration units to market would be considerable. As with automotive emission control systems, the compulsion for developing such a refrigerant free unit is likely to be legislation.
4. Conclusions The performance of an air-cycle refrigeration unit for road transport, which had been previously reported, was analysed in detail and compared with the original design model. Overall, the turbomachinery was found to satisfy the original design requirements, but the heat exchangers proved to be a major performance handicap. The heat exchanger effectiveness values were several percentage points lower than anticipated and the heat exchanger
388
S.W.T. Spence et al. / International Journal of Refrigeration 28 (2005) 381–388
pressure losses were excessive. As a consequence, the aircycle demonstrator unit absorbed around 25% more power than the design value and more than twice as much power as the equivalent vapour-cycle unit. Using state-of-the-art, but achievable performance levels for turbomachinery and heat exchangers, the performance of an optimised air-cycle refrigeration unit was predicted. The power requirement of the optimised air-cycle unit was 7% greater than the equivalent vapour-cycle unit at full-load operation. However, at part-load operation the air-cycle unit was estimated to absorb 35% less power than the vapour-cycle unit. The analysis demonstrated that the air-cycle system could potentially match the overall fuel consumption of the vapour-cycle transport refrigeration unit, while delivering the benefit of a completely refrigerant free system. Bringing the optimised air-cycle to fruition would require the introduction of several new technologies into the transport refrigeration sector including air bearings, plate-fin heat exchangers, turbomachinery and high speed electric drives. The amount of development
work and the associated investment to bring a range of air-cycle transport refrigeration units to market would be considerable, and would be unlikely in the absence of legislation to encourage such refrigerant free systems.
Acknowledgements The authors would like to thank Enterprise Ireland and Thermo King for funding this project, and particularly Mr John Gough for his enduring support.
References [1] M.S. Bhatti, Open air cycle air conditioning system for motor vehicles, Society of Automotive Engineers, 1998. SAE 980289. [2] S.W.T. Spence, W.J. Doran, D.W. Artt, Design, construction and testing of an air-cycle refrigeration system for road transport, Int J Refrig 27/5 (2004) 503–510.