International Journal of Refrigeration 27 (2004) 503–510

www.elsevier.com/locate/ijrefrig

Design, construction and testing of an air-cycle refrigeration

system for road transport
Stephen W.T. Spencea,*, W. John Doranb, David W. Artta
a
School of Mechanical and Manufacturing Engineering, Queen’s University Belfast, Ashby Building, Stranmillis Road,
Belfast BT9 5AH, Northern Ireland
b
School of Engineering, Letterkenny Institute of Technology, Port Road, Letterkenny, Co., Donegal, Ireland
Received 9 September 2003; received in revised form 5 February 2004; accepted 18 February 2004

Abstract
The environmental attractions of air-cycle refrigeration are considerable. Following a thermodynamic design analysis, an air-
cycle demonstrator plant was constructed within the restricted physical envelope of an existing Thermo King SL200 trailer
refrigeration unit. This unique plant operated satisfactorily, delivering sustainable cooling for refrigerated trailers using a
completely natural and safe working fluid. The full load capacity of the air-cycle unit at 2 20 8C was 7,8 kW, 8% greater than
the equivalent vapour-cycle unit, but the fuel consumption of the air-cycle plant was excessively high. However, at part load
operation the disparity in fuel consumption dropped from approximately 200% to around 80%. The components used in the air-
cycle demonstrator were not optimised and considerable potential exists for efficiency improvements, possibly to the point
where the air-cycle system could rival the efficiency of the standard vapour-cycle system at part-load operation, which
represents the biggest proportion of operating time for most units.
q 2004 Elsevier Ltd and IIR. All rights reserved.
Keywords: Design; Construction; Air conditioner; Refrigerated transport; Thermodynamic cycle; Air; Centrifuge compressor; Turbine
expander

Système frigorifique à cycle à air pour le transport routier:

conception, fabrication et essais
Mots-clés: Conception; Fabrication; Conditionneur d’air; Transport frigorifique; Cycle thermodynamique; Air; Compresseur centrifuge;
Turbine à détente

1. Introduction refrigeration is preferred over air-cycle refrigeration is

simply that in the great majority of cases vapour-cycle is the
The current legislative pressure on conventional refriger- most energy efficient option. Consequently, as soon as
ants is well known. The reason why vapour-cycle alternative systems, such as non-HFC refrigerants or air-
cycle systems are considered, the issue of increased energy
consumption arises immediately.
Abbreviations: COP, coefficient of performance; CAU, cold air
unit; QUB, Queen’s University of Belfast; LYIT, Letterkenny Concerns over legislation affecting HFC refrigerants and
Institute of Technology. the desire to improve long-term system reliability led to the
* Corresponding author. examination of the feasibility of an air-cycle system for
E-mail address: s.w.spence@qub.ac.uk (S.W.T. Spence). refrigerated transport. With the support of Enterprise Ireland
0140-7007/$35.00 q 2004 Elsevier Ltd and IIR. All rights reserved.
doi:10.1016/j.ijrefrig.2004.02.003
504 S.W.T. Spence et al. / International Journal of Refrigeration 27 (2004) 503–510

would not be optimal. In practical terms, this meant using

Nomenclature the chassis and panels for an existing SL200 unit along with
the standard diesel engine and circulation fans. The
PR Compressor or turbine pressure ratio
turbomachinery used for compression and expansion was
TA Heat exchanger side A temperature (K)
adapted from commercial turbochargers.
TB Heat exchanger side B temperature (K)
Tinlet Inlet temperature (K)
Toutlet Outlet temperature (K)
hcomp Compressor isentropic efficiency 2. Thermodynamic modelling and design of the
hturb Turbine isentropic efficiency demonstrator plant
hheat exchanger Heat exchanger effectiveness
The thermodynamics of the air-cycle (or the reverse
‘Joule cycle’) are adequately presented in most thermodyn-
and Thermo King (Ireland), the authors undertook the amic textbooks and will not be repeated here. For anything
design and construction of an air-cycle refrigeration other than the smallest flow rates, the most efficient
demonstrator plant at LYIT and QUB. This was not the machines available for the necessary compression and
first time in recent years that air-cycle systems had been expansion processes are turbomachines. Considerations for
employed in transport. NormalAir Garrett developed and the selection of turbomachinery for air-cycle refrigeration
commercialised an air-cycle air conditioning pack that was systems have been presented and discussed by Spence et al.
fitted to high speed trains in Germany in the 90s [1]. As part [3].
of an European funded programme [2], a range of Fig. 1 shows a typical configuration of an air-cycle
applications for air-cycle refrigeration were investigated system, which is sometimes called the ‘boot-strap’ con-
and several demonstrator plants were constructed. However, figuration. For mechanical convenience the compression
the authors are unaware of any other case where a self- process is divided into two stages, meaning that the turbine
contained air-cycle unit has been developed for the is not constrained to operate at the same speed as the
challenging application of trailer refrigeration. primary compressor. Instead, the work recovered by the
Thermo King decided that the demonstrator should be a turbine during expansion is utilised in the secondary
trailer refrigeration unit, since those were the units with the compressor. The two-stage compression also permits
largest refrigeration capacity but presented the greatest intercooling, which enhances the overall efficiency of the
challenges with regard to physical packaging. Conse- compression process. Fig. 1 shows an ‘open system’ where
quently, the main objective was to demonstrate that an air- the cold air is ejected directly into the cold space, removing
cycle system could fit within the existing physical envelop the need for a heat exchanger in the cold space. In the
and develop an equivalent level of cooling power to the interests of efficiency, the return air from the cold space is
existing vapour-cycle unit, but using only air as the working used to pre-cool the compressed air entering the turbine by
fluid. The salient performance specifications for the existing means of a heat exchanger known as the ‘regenerator’ or the
Thermo King SL200 vapour-cycle trailer refrigeration unit ‘recuperator’.
are listed in Table 1. To support the design of the air-cycle demonstrator
It was not the objective of the exercise to complete the plant, and the selection of suitable components, a simple
design and development of a new refrigeration product that thermodynamic model of the air-cycle configuration shown
would be ready for manufacture. To limit the level of in Fig. 1 was developed. The compression and expansion
resources necessary, existing hardware was to be used where processes were modelled using appropriate values of
possible with the recognition that the efficiencies achieved isentropic efficiency, as defined in Eqs. (1) and (2). The

Table 1
Specifications of existing trailer refrigeration unit. (Specifications obtained from Thermo King sales brochure TK 50246-3-PL E, published
October 1998)

Thermo King SL-200 trailer temperature control system

Cooling capacity at 0 8C trailer temperature (30 8C ambient) 12 kW

Cooling capacity at 220 8C trailer temperature (30 8C ambient) 7,2 kW
Volumetric airflow 4800 m3/h
Engine type 4 cyl., liquid cooled, diesel 21,3 kW rated output
Engine speed 1450 rev/min and 1900 rev/min
Refrigerant charge 6,35 kg of R-404A
Physical dimensions (width £ height £ depth) 2056 £ 2235 £ 430 mm
Weight 739 kg
 S.W.T. Spence et al. / International Journal of Refrigeration 27 (2004) 503–510 505

Fig. 1. Schematic layout of a typical air cycle system with two stage compression (boot-strap configuration).

heat exchange processes were modelled using values of heat TA inlet 2 TA outlet
hheat ex ¼ ð3Þ
exchanger effectiveness as defined in Eq. (3). The model TA inlet 2 TB inlet
also made allowance for heat exchanger pressure drop. The
Cooling power produced
system COP was determined from the ratio of the cooling COP ¼ ð4Þ
power delivered to the power input to the primary Compressor input power
compressor, as defined in Eq. (4). Figs. 2 –6 illustrate air- Fig. 2 illustrates the variation in air-cycle COP and
cycle performance characteristics as determined from the expander outlet temperature over a range of cycle pressure
thermodynamic model: ratios for a plant operating between 220 8C and þ30 8C.

g21  The cycle pressure ratio is defined as the ratio of the
Tinlet PR g 2 1 maximum cycle pressure at secondary compressor outlet to
hcomp ¼ ð1Þ
Toutlet 2 Tinlet the pressure at turbine outlet. For the ideal air-cycle, with no
losses, the cycle COP increases with decreasing cycle
Tinlet 2 Toutlet
hturbine ¼   g21 ! ð2Þ pressure ratio and tends to infinity as the pressure ratio
1 g
approaches unity. However, the introduction of real
Tinlet 1 2
PR component efficiencies means that there is a definite peak
value of COP that occurs at a certain pressure ratio for a
particular cycle. However, as Fig. 2 illustrates, there is a

Fig. 2. Variation of COP and air discharge temperature with overall

air-cycle pressure ratio for a two stage compression air-cycle system
operating between 220 and þ 30 8C. (Assumed efficiencies were Fig. 3. Impact of intercooler and aftercooler effectiveness on the
81% for compression and 85% for expansion.) COP of the air-cycle plant.
506 S.W.T. Spence et al. / International Journal of Refrigeration 27 (2004) 503–510

Fig. 4. Impact of intercooler and aftercooler pressure loss on the Fig. 6. Relationship between air mass flow rate and system pressure
COP of the air-cycle plant. ratio to deliver a fixed cooling duty of 7,5 kW.

broad range of pressure ratio and duty over which the system efficiency of the plant so that the best compromise could be
can be operated with only moderate variation of COP. developed. Figs. 3 and 4 show the impact of intercooler and
The class of turbomachinery suitable for the demonstra- aftercooler effectiveness and pressure loss on the COP of the
tor plant required speeds of around 50 000 rev/min. To proposed plant.
simplify the mechanical arrangement and avoid the need for The two-stage system in Fig. 1 incorporated an
a high-speed electric motor, the two-stage compression intercooler between the two compression stages. By
system shown in Fig. 1 was adopted. The existing Thermo dispensing with the intercooler and its associated duct
King SL200 chassis incorporated a substantial system of work a larger aftercooler could be accommodated with
belts and pulleys to power circulation fans, which severely improved effectiveness and reduced pressure loss. Analysis
restricted the useful space available for mounting heat suggested that the improved performance from a larger
exchangers. A simple thermodynamic model was used to aftercooler could compensate for the loss of the intercooler.
assess the influence of heat exchanger performance on the Fig. 5 shows the impact of the recuperator effectiveness
on the COP of the plant, which is clearly more significant
than that of the other heat exchangers. As well as boosting
cycle efficiency, increased recuperator effectiveness also
moves the peak COP to a lower overall system pressure
ratio. The impact of pressure loss in the recuperator is the
same as for the intercooler and aftercooler shown in Fig. 4.
The model did not distinguish between pressure losses in
different locations; it was only the sum of the pressure losses
that was significant. Any pressure loss in connecting duct
work and headers was also lumped together with the heat
exchanger pressure loss and analysed as a block pressure
loss.
The specific cooling capacity of the air-cycle increases
with system pressure ratio. Consequently, if a higher system
pressure ratio was used the required cooling duty could be
achieved with a smaller flow rate of air. Fig. 6 shows the
mass flow rate of air required to deliver 7,5 kW of cooling
power for varying system pressure ratios.
Since the demonstrator system was to be based on
commercially available turbomachinery, it became import-
Fig. 5. Impact of recuperator effectiveness on the COP of the air- ant to choose a pressure ratio and flow rate that could be
cycle plant. accommodated efficiently by some existing compressor and
 S.W.T. Spence et al. / International Journal of Refrigeration 27 (2004) 503–510 507

turbine rotors. Figs. 2 – 6 were based on efficiencies of 81 Roots blowers and rotary piston blowers were all excluded
and 85% for compression and expansion, respectively. on the basis of poor efficiency.
While such efficiencies are attainable with optimised A one-off gearbox was designed and manufactured as
designs, they would not be realised using compromised part of the project to step-up the engine shaft speed to
turbocharger components. For the design of the demonstra- around 55 000 rev/min. The gearbox was a two stage, three
tor plant efficiencies of 78 and 80% were assumed to be shaft unit which mounted directly on the end of the diesel
realistically attainable for compression and expansion. engine and was driven through the existing centrifugal
Lower turbomachinery efficiencies corresponded to clutch.
higher cycle pressure ratios and flow rates in order to
achieve the target cooling duty. The cycle design point
was also compromised to help heat exchanger perform- 4. Cold air unit
ance. The pressure losses in duct work and heat
exchangers increased in proportion with the square of
The secondary compressor and the expansion turbine
flow velocity. Selecting a higher cycle pressure ratio
were mounted on the same shaft in a free rotating unit. The
corresponded to a lower mass flow rate and also
combination of the secondary compressor and the turbine
increased density at inlet to the aftercooler heat
was designated as the ‘Cold Air Unit’ (CAU). While the
exchanger. The combined effect was a decrease in the
CAU was mechanically equivalent to a turbocharger, a
mean velocity in the heat exchanger, a decrease in the
standard turbocharger would not satisfy the aerodynamic
expected pressure losses in the heat exchanger and duct
requirements efficiently since the pressure ratios and inlet
work, and an increase in the effectiveness of the heat
densities for both the compressor and the turbine were
exchanger. Consequently, a system pressure ratio higher
significantly different from any turbocharger installation.
than the value corresponding to peak COP was chosen in Consequently, both the secondary compressor and the
order to achieve acceptable heat exchanger performance turbine stage were specially chosen and developed to
within the available physical space. The below optimum deliver suitable performance.
performance of turbomachinery and heat exchanger Most turbochargers use plain oil fed journal bearings,
components, coupled with excessive bearing losses, which are low-cost, reliable and provide effective damping
meant that the predicted COP of the overall system of shaft vibrations. However, plain bearings dissipate a
dropped to around 0,41. The system pressure ratio at the substantial amount of shaft power through viscous losses in
design point was 2,14 and the corresponding mass flow the oil films. A plain bearing arrangement for the CAU was
rate of air was 0,278 kg/s. expected to absorb 2 – 3 kW of mechanical power, which
By moving the design point beyond the pressure ratio for represented around 25% of the anticipated turbine power.
peak COP, it was anticipated that the demonstrator plant Also, the clearances in plain bearings require larger blade tip
would yield good part-load performance since the COP clearances for both the compressor and the turbine with a
would not fall as the pressure ratio was reduced. Also, consequential efficiency penalty. Given the pressurised inlet
operating at part-load corresponded to lower flow velocities to the secondary compressor, the limited thrust capacity of
and anticipated improvements in heat exchanger perform- the plain bearing arrangement was also a concern. A CAU
ance. Part-load operation was achieved by reducing the utilising high-speed ball bearings, or air bearings, was
speed of the primary compressor, resulting in a decrease in identified as a preferable arrangement to plain bearings.
both pressure and mass flow rate throughout the cycle. Benefits would include greatly reduced bearing power
losses, reduced turbomachinery tip clearance losses and
increased thrust load capacity. However, adequate resources
were not available to design a special one-off high speed ball
3. Prime mover and primary compressor bearing system. Consequently, a standard turbocharger plain
bearing system was used.
The existing diesel engine was judged adequate to power The secondary compressor stage was a standard
the demonstrator plant. The standard engine was a four turbocharger compressor selected for a pressure ratio of
cylinder, water cooled diesel engine fitted with a centrifugal 1,264. Secondary compressor and turbine selection were
clutch and all necessary ancillaries and was controlled by a linked because of the requirement to balance power and
microprocessor controller. match the speed. Since most commercial turbines are sized
From the thermodynamic model, the pressure ratio for for high temperature (and consequently low density) air at
the primary compressor was 1,70. The centrifugal com- inlet, a special turbine stage was developed for the
pressor required a shaft speed of around 55 000 rev/min. application. Cost considerations precluded the manufacture
Other alternatives were evaluated for primary compression of a custom turbine rotor, so a commercially available rotor
with the aim of obtaining a suitable device that operated at a was used. The standard turbine rotor blade profile was
lower speed. Other commercially available devices such as substantially modified and vaned nozzles for turbine inlet
508 S.W.T. Spence et al. / International Journal of Refrigeration 27 (2004) 503–510

were designed to match the modified rotor, in line with with a magnetic speed counter. No air flow measurement
previous turbine investigations at QUB (Spence and Artt, was included on the demonstrator plant. Instead, the air flow
[4]). An exhaust diffuser was also incorporated into the rate was deduced from the previously obtained turbine
turbine stage in order to improve turbine efficiency and to performance map using the measurements of turbine
moderate the exhaust noise levels through reduced air pressure ratio and rotational speed.
velocity. The exhaust diffuser exited into a specially
designed exhaust silencer.
The performance of the turbine stage was measured 7. System testing
before the unit was incorporated into the complete
demonstrator plant. The peak efficiency of the turbine was
During some preliminary tests a heat load was applied
established at 81%.
and the functionality of the demonstrator plant was
established. Having assessed that it was capable of
delivering approximately the required performance, the
5. Heat exchangers plant was transported to a Thermo King calorimeter test
facility specifically for measuring the performance of
Due to packaging constraints, the heat exchangers had to transport refrigeration units. The calorimeter was ideally
be specially designed with careful consideration being given suited for accurately measuring the refrigeration capacity of
to heat exchanger position and header geometry in an the air-cycle demonstrator plant. The calorimeter was
attempt to achieve the best performance from the heat operated according to standard ARI 1100-2001; the absolute
exchangers. Tube and fin aluminium heat exchangers, accuracy was better than 200W and all auxiliary instru-
similar to those used in automotive intercooler applications, mentation was calibrated against appropriate standards.
were chosen primarily because they could be produced on a The performance capacity of transport refrigeration units
‘one-off’ basis at a reasonable cost. There were other heat is generally rated at two operating conditions; 0 and
exchanger technologies available that would have yielded 220 8C, and both at an ambient temperature of þ30 8C.
better performance from the available volume, but high one- Along with the specified operating conditions of 0 and
off production costs precluded their use in the demonstrator 220 8C, a further part-load condition at 2 20 8C was
plant. assessed. Considering that the air-cycle plant was only
Several different tube and fin heat exchangers were intended to demonstrate a concept and that there were
tested and used to validate a computational model. Once concerns about the reliability of the gearbox and the cold air
validated, the model was used to assess a wide range of unit thrust bearing, it was decided to operate the plant only
possible heat exchanger configurations that could fit within as long as was necessary to obtain stabilised measurements
the Thermo King SL200 chassis. Fitting the proposed heat at each operating point. The demonstrator plant operated
exchangers within the existing chassis and around the satisfactorily, allowing sufficient measurements to be
mechanical drive system for the circulation fans, but while obtained at each of the three operating conditions. The
still achieving the necessary heat exchanger performance recorded performance is summarised in Table 2.
was very challenging. It was clear that potential heat In total, the unit operated for approximately 3 h during
exchanger performance was being sacrificed through the the course of the various tests. While the demonstrator plant
choice of tube and fin construction and by the constraints of operated adequately to allow measurements, some smoke
the layout of the existing SL200 chassis. The final selection from the oil system breather suggested that the thrust
comprised two separate aftercooler units, while the single bearing of the CAU was heavily overloaded and would fail,
recuperator was a large, triple pass unit. Based on laboratory as had been anticipated at the design stage. Testing was
tests and the heat exchanger model, the anticipated concluded in case the bearing failed completely causing the
effectiveness of both the recuperator and aftercooler units destruction of the entire CAU. There was no evidence of any
was 80%. gearbox deterioration during testing.

6. Instrumentation 8. Discussion of measured performance

A range of conventional pressure and temperature From the calorimeter performance measurements, the
instrumentation was installed on the air-cycle demonstrator primary objective of the project had been achieved. A
plant. Air temperature and pressure was logged at inlet and unique air-cycle refrigeration system had been developed
outlet from each heat exchanger, compressor and the within the same physical envelope as the existing Thermo
turbine. The speed of the primary compressor was King SL200 refrigeration unit, which delivered the
determined from the speed measurement on the diesel equivalent cooling capacity using air as the working fluid.
engine control unit, while the cold air unit was equipped Indeed, the 7,8 kW full load capacity of the air-cycle unit
 S.W.T. Spence et al. / International Journal of Refrigeration 27 (2004) 503–510 509

Table 2
Measured performance for the air-cycle demonstrator plant

Full-load 0 8C Full-load 220 8C Part-load 220 8C

Cooling capacity (W) 9500 7800 3400

Ambient temperature (8C) 30,6 29,3 29,9
Trailer temperature (8C) 0,0 220,0 220,0
Discharge air temp. (8C) 229,4 246,4 235,8
Engine speed (rev/min) 2210 2210 1760

exceeded the 7,2 kW rated capacity of the Thermo King consumption indicates that the air-cycle system maintained
SL200 unit (Table 1) at 220 8C by 8%. However, it was its efficiency over a wide range of duty. By comparison, the
clear that the diesel engine was heavily loaded and that fuel consumption of the standard SL200 at an equivalent
considerably more power was necessary to drive the air- part-load condition was still as high as 73% of the fuel
cycle demonstrator plant. At full load at 2 20 8C the fuel consumption at full-load. The fuel consumption of the air-
consumption of the air-cycle plant was approximately three cycle demonstrator plant was around 80% more than the fuel
times greater than the standard SL200 refrigeration unit. The consumption of the standard SL200 at part-load operation,
air-cycle demonstrator plant with its compromised turbo- compared with approximately three times more at full-load;
machinery, prototype heat exchangers and crudely designed although potential for substantial improvements in fuel
gearbox was never expected to rival the efficiency of the consumption has been identified.
standard SL200 unit. The large disparity in fuel consump- The part-load performance was very encouraging, since
tion could clearly be improved through developments in the it demonstrated that a fully optimised air-cycle plant could
CAU bearing system. The estimated 2 – 3 kW power potentially rival the efficiency of the standard SL200 at part-
dissipation in the bearings would correspond to approxi- load, which is the condition at which working units spend
mately 2 – 3 kW of additional power input from the diesel most time operating. The efficiency of an air-cycle system is
engine. Improving the bearing system alone could reduce unlikely to ever match a well-designed vapour-cycle system
the power consumption of the demonstrator plant by at full-load operation, which is why vapour-cycle systems
approximately 10%. Significant potential for improvements have been preferred for decades. However, the good part-
also exists in the gearbox and through improved turboma- load performance of the air-cycle system may mean that the
chinery and heat exchanger performance. overall efficiency comparison becomes more acceptable and
At 0 8C the capacity of the air-cycle plant was 9,5 kW, that the environmental benefits of the air-cycle could be
21% less than the Thermo King SL200 unit. The COP of the realised in practice.
vapour-cycle improves much more rapidly with increasing
temperature than for the air-cycle. At the design stage, the
capacity of the air-cycle plant had not been expected to
match the vapour-cycle at 0 8C. The fuel consumption was 9. Conclusions
still approximately three times greater than the Thermo
King SL200 unit. The attractions of a refrigeration cycle using air as the
Approximately 3 kW is normally adequate to maintain a working fluid are considerable, making it worthy of close
temperature of 2 20 8C in a trailer in good condition. scrutiny for some applications. Following a thorough
Therefore, for long haul transport the refrigeration unit may thermodynamic design analysis, a unique air-cycle demon-
have a few hours operation at full load, followed by many strator plant was constructed to fit within the restricted
hours at less than 50% capacity. The operating costs of a physical envelope of existing trailer refrigeration units for
transport refrigeration unit are often dominated by part-load road transport, using modified or adapted commercially
performance rather than full-load. The previous discussion available components wherever possible. Detailed perform-
of Fig. 2 had highlighted that the air-cycle COP remained ance testing on an industry standard calorimeter test facility
relatively constant over a broad range of operating established that the objective had been achieved, which was
conditions. It was therefore anticipated that although the to deliver the equivalent cooling capacity using an air-cycle
air-cycle would be comparatively inefficient at full-load system within the same physical space as the standard
operation, it could potentially rival the performance of the vapour-cycle unit. The full load capacity of the air-cycle
vapour-cycle system at part-load operation. The air-cycle unit at 220 8C was 7,8 kW, 8% greater than the standard
demonstrator plant was also tested at a part-load condition vapour-cycle unit. At 0 8C, the full-load capacity of the air-
of 3,4 kW at 220 8C, which was 44% of it’s full load cycle unit was 9,5 kW, 21% less than the standard vapour-
capacity. The fuel consumption at part-load was 43% of the cycle unit. At both temperatures, the fuel consumption of the
full-load consumption. The proportional change of fuel air-cycle plant was excessively high at more than three times
510 S.W.T. Spence et al. / International Journal of Refrigeration 27 (2004) 503–510

the consumption of the standard vapour-cycle unit. John Gough for his enduring support. The authors would
However, part load operation was more encouraging, also like to thank Scania (Sweden) for their technical and
where the disparity in fuel consumption dropped from hardware support.
over 200 to 80%. Through development work, an air-cycle
system using optimised turbomachinery, heat exchangers,
transmission and bearings would realise much better References
efficiency than the demonstrator plant. Consequently the
efficiency could rival the efficiency of the standard vapour- [1] Boeck I, Keske K, Köhler W. Betriebserfahrungen mit einer
cycle system at part-load operation, which represents the Kaltluftklimaanlage (Operational experiences with a cold air
biggest proportion of operating time for most units. While conditioner in the intercity express train) ICE 3, KI Luft-und
there will be a fuel consumption penalty for delivering the Kältetechnik, 2/2002, p. 81–85.
environmental benefits of air-cycle refrigeration, in some [2] Verschoor MJE, editor. Guidelines for the application and
applications, such as road transport, the penalty would be design of air cycle systems for heating, ventilating and air
much less than comparisons of full-load performance would conditioning in buildings. Apeldoorn, The Netherlands: TNO-
suggest. MEP; 2001.
[3] Spence SWT, Doran WJ, Artt DW. Optimisation of turboma-
chinery performance for air-cycle refrigeration systems. Proc
Instn R 1997;94(4th December). ISBN 1 872719 12 0.
Acknowledgements [4] Spence SWT, Doran WJ, Artt DW. Experimental performance
evaluation of a 99,0 mm radial inflow nozzled turbine at larger
The authors would like to thank Enterprise Ireland and stator/rotor throat area ratios. Proc Instn Mech Eng, Part A, J
Thermo King for funding this project, and particularly Mr Power Energy 1999;213:205–18. ISSN 0957 6509.