Escolar Documentos
Profissional Documentos
Cultura Documentos
Applied Energy
journal homepage: www.elsevier.com/locate/apenergy
H I GH L IG H T S
A R T I C LE I N FO A B S T R A C T
Keywords: Improving efficiency is an enduring effort for all work-heat conversion cycles. Ideal regenerators working with
Real gas effects ideal gases bring about a lossless work and heat transfer over a temperature gradient, but real gases give rise to
Regenerator an “intrinsic” heat loss in regenerators because of the time-averaged enthalpy flow associated with the pressure
Refrigeration cycle dependence. Real gas effects play a vital role on the coefficient of performance (COP) of regenerators of re-
Heat input
frigeration cycles working at the temperatures close to or below the critical point. The “intrinsic” heat loss of real
Carnot efficiency
gases degrades the theoretical COP of regenerators to as low as 1% of the Carnot efficiency. In this paper, an
approach of heat input or removal aiming to improve the COP is proposed. The theoretical analysis of this
approach reveals the underlying mechanism. It is shown that the theoretical COP of an ideal regenerator working
with a real gas applying this approach is identical to the Carnot efficiency. A simplified approach of heat input is
further analyzed. The Carnot efficiency can be attained under certain circumstances, and it is possible to obtain
over 90% of the Carnot efficiency with a discrete method. The theory of improving the COP with the approach of
heat input in discrete regenerator locations is supported by the experiment results found in the relevant lit-
erature. This new approach provides a potential way to significantly improve the efficiency of the regenerator of
the refrigeration cycle working at the temperatures close to or below the critical point. This approach may
further provide a reference for studies of the heat pump cycle and the engine cycle working with real gases.
1. Introduction the temperature gradient from the hot end to the cold end with the least
work and heat losses possible in the oscillation flow. Ideal regenerators
Improving efficiency is an enduring effort for all work-heat con- working with ideal gases bring about a lossless work and heat transfer
version cycles [1–7]. The derivation for the theoretical efficiency re- over a temperature gradient. The refrigeration cycle applying an ideal
veals the working mechanism and sets a target for real devices, so it is regenerator and ideal pistons at the hot end and cold end, such as the
prior work for a new cycle and for a new condition [8–11]. The re- Stirling refrigeration cycle, reaches the Carnot efficiency [22]. The
frigeration cycle is a reverse work-heat conversion cycle that removes pulse tube refrigerator works without pistons at the cold end, and it
the heat from a low-temperature reservoir with a work input. Efficient dissipates the work in the phase shifter, thus its theoretical COP is equal
regenerative refrigerators are important for fulfilling the refrigeration to (Tc/Th) [22–24]. Recent investigations on the mass-spring (or gas
requirement of industrial gas liquefaction [12–15], superconducting spring) feedback [25–30], step piston [31], cascade style [32–34], etc.,
devices [16–20], infrared detectors [18,21], and so on. The function of realize a recovery of the dissipated work, thus the COP is improved
the regenerator in the refrigeration cycle is to transport the work over significantly. In fact, the theoretical COP of these improved
⁎
At: Institute of Refrigeration and Cryogenics, School of Mechanical Engineering, Tongji University, Shanghai 201804, PR China.
E-mail address: qiangcao@tongji.edu.cn.
https://doi.org/10.1016/j.apenergy.2018.03.102
Received 4 September 2017; Received in revised form 24 March 2018; Accepted 26 March 2018
Available online 03 April 2018
0306-2619/ © 2018 Elsevier Ltd. All rights reserved.
Q. Cao Applied Energy 220 (2018) 705–712
refrigerators is equal to the Carnot efficiency. “intermediate refrigeration power” can be used to increase the lique-
However, the theoretical COP is challenged if a real gas is used as faction rate [44,47], and to increase the cold-end cooling power of the
the working fluid. The enthalpy flow associated with the pressure de- cryocooler itself [43] by precooling the pulse tube, the radiation shield,
pendence, abbreviated as the pressure-induced enthalpy flow, which is and through other means. It has also been pointed out that the re-
a significant property of real gases, gives rise to an energy imbalance. generator loss could be decreased and the efficiency of the cryocooler
This imbalance is generally solved by generating an enthalpy flow as- could be improved by the “intermediate refrigeration power” method
sociated with imperfect heat transfer and limited matrix heat capacity [43,46].
in the regenerator, abbreviated as the heat-associated enthalpy flow An approach of heat input or removal is proposed to decrease the
[35,36]. But the heat-associated enthalpy flow is actually an “intrinsic” real gas losses and to improve the COP of refrigerators in this manu-
heat loss of the regenerator, so the COP of the cycles working with such script. The heat input or removal is the heat inputted or removed from
regenerators cannot reach the Carnot efficiency [37–40]. The theore- an external system to the refrigeration system at intermediate tem-
tical COP of such regenerators that work close to or below the critical peratures [49]. The working gas of the refrigeration system absorbs the
point is usually lower than 30% of the Carnot efficiency, and it gets external heat and it is turned to the enthalpy of the gas in the case of
even smaller if the refrigeration temperature goes lower [37–40]. This heat input, and the working gas turns its enthalpy to heat and the heat
means that over 70% of the availability gets lost in such situations. The is transferred to the external heat sink in the case of heat removal. In
loss mechanism caused by real gas effects has been analyzed in previous brief, the heat input or removal superimposes a finite time-averaged
studies [35,40–42], and the expression of the COP of the regenerator enthalpy flow in the regenerator. The heat input is actually one kind of
and the pulse tube cryocooler has been derived. refrigeration power (or heat sink) at intermediate temperatures, but the
However, the investigations dealing with the real gas effects are not heat removal is one kind of heat load (or heat source) at intermediate
sufficient. Systematical methods of increasing the COP of the re- temperatures. Because the heat input or removal takes place in a tem-
frigerators working with real gases have not been proposed. Although a perature range rather than a temperature point, a series of micro re-
practical method of benefiting from the “intrinsic” losses, i.e. gathering frigeration cycles should be applied to realize the process.
the “intermediate refrigeration power” (described as “excess cooling The ideal regenerator is one-dimensional, as assumed in the theo-
power” [43,44] or “free cooling power” [45,46]), has been proposed retical thermodynamic derivation, so the heat input or removal at a
and investigated in previous studies [43–46], the theoretical value of discrete location or at a certain temperature affects the corresponding
the “intermediate refrigeration power” and the theoretical COP of the section evenly.
regenerator taking account of the “intermediate refrigeration power” Although the theory of heat input or removal has been proposed in
have not been analyzed yet. The term “intermediate refrigeration the previous study, the real gas effects have not been taken into account
power” is used in this manuscript because it can explain the physical [49]. The working fluid was assumed to be an ideal gas in that im-
essence and describe the temperature range better. It has been found portant pioneering work. The conclusion was that the regenerator
that an “intermediate refrigeration power” can be extracted in the working with a finite enthalpy loss is able to absorb or remove a certain
middle section of a regenerator, leading to none or little degradation on amount of external heat. The influence of the heat input or removal to
the refrigeration performance at the cold end [43–48]. This “inter- the temperature profile and the energy flow was analyzed and simu-
mediate refrigeration power” method has been numerically and ex- lated.
perimentally studied in cryocoolers working down to 4 K [43–48]. Such This paper will systematically investigate the attainability of the
Fig. 1. Schematics of two typical regenerative refrigerators, a Stirling refrigerator (a) and a pulse tube refrigerator with a displacer (b).
706
Q. Cao Applied Energy 220 (2018) 705–712
Carnot efficiency by applying the approach of heat input or removal heat input decreases the total heat flow along the regenerator. The dash
over the whole temperature range of the regenerator. Furthermore, the red line represents the total heat flow above the point that the heat
simplified method of gathering the “intermediate refrigeration power” input is applied. Since the heat-associated enthalpy flow is eliminated
will be investigated. The theoretical value of the “intermediate re- in an ideal regenerator working with a real gas, it is only the pressure-
frigeration power” and the theoretical COP of the regenerator taking induced enthalpy flow that contributes to the total energy variation of
account of the “intermediate refrigeration power” will be analyzed. the regenerator [36], so Eq. (1) is reformed as:
Finally, experiment results found in the relevant literature will be cited
〈dHṗ 〉
to support the analysis of this approach. δQ̇ = dT .
dT (2)
2. Analysis on the attainability of the Carnot efficiency Since the status at the hot end of the regenerator is generally fixed in
one case, it becomes more meaningful to get a dimensionless parameter
Fig. 1 shows the schematics of two typical regenerative re- by setting that status as a reference. Divided by the PV power at the hot
frigerators, a Stirling refrigerator and a pulse tube refrigerator with a end, the dimensionless heat input per temperature range is expressed
displacer (mass-spring feedback of the PV power). It is assumed that as:
time-averaged isothermal processes proceed in the compressor and the
aftercooler. The regenerator transports the PV power over the tem- δQ̇/ 〈pV̇ 〉h 1 d [(1−Tβ ) ZT ]
= .
perature gradient from the hot end to the cold end in the oscillation δT (ZT )h dT (3)
flow. The time-averaged isothermal expansion in the Stirling re- Both the parameters (1 − βT) and (ZT) vary in almost the whole tem-
frigerator (Fig. 1(a)) takes place in the expansion space, which is lo- perature range [51], especially at the temperatures close to or below
cated between the displacer and the cold heat exchanger, but the time- the critical temperature, which means that the heat input is seldom is
averaged isothermal expansion in the pulse tube refrigerator (Fig. 1(b)) equal to zero. According to the principle of corresponding states for
takes place in the pulse tube, which transfers the PV power to the hot general gases, there is a minimum of the combined parameter ((1 − βT)
end over a temperature gradient. The PV power feeds back to the ZT) for a certain pressure, and the value of this parameter increases
compressor by the displacer both in these two refrigerators. monotonically at both sides [51]. It is clear that such a point is the
These typical examples show that the regenerator functions the separatrix point. The temperature at that point is generally close to the
same in regenerative refrigerators. The regenerator absorbs and re- critical temperature. The heat input is always negative (i.e. heat re-
leases the heat of the working gas in the matrix in the period of flowing moval) at temperatures higher than this point, and it is positive at
toward cold end and in the period of flowing toward the hot end [50], temperatures lower than this point. The temperature at the separatrix
respectively, as illustrated in Fig. 2. The yellow dash lines represent the point is exactly the critical temperature at the critical pressure. This
gas, and the thin red lines represent the heat transferred between the point is the vapor-liquid saturation point in the subcritical region, and
gas and the matrix. Ideal regenerators work with infinite heat transfer its temperature increases positively with the pressure both in the sub-
rate and infinite matrix heat capacity, and the heat capacity of an ideal critical region and supercritical region.
gas is only a function of the temperature, thus no heat loss is generated. One example of a typical gas, 4He [52], is given here, shown as solid
However, the heat capacity of a real gas is also a function of the pres- lines in Fig. 4. The unit of the temperature range for calculating the
sure besides the temperature, which means that a time-averaged en- dimensionless heat input is set as one kelvin, since the kelvin is a base
thalpy is generated because the average pressure in the half cycle of unit of temperature in the International System of Units. This unit is
absorbing heat is higher than that in the other half cycle of releasing also used for calculating the availability in the following section. The
heat. This time-averaged pressure-induced enthalpy flow must be ba- reduced pressure (the pressure normalized by the critical pressure) and
lanced by other enthalpy (heat) flows, according to the first law of reduced temperature (the temperature normalized by the critical tem-
thermodynamics. Generally, it is the heat-associated enthalpy flow in perature) are used because they demonstrate a more general property of
the regenerator that keeps the energy balance [35,36]. In order to gases. Comparing with the separatrix point of the critical point (shown
minimize the heat-associated enthalpy flow, the heat input or removal as the red line), the temperature of this point increases to a reduced
is proposed. These energy terms are shown in Fig. 2. In the calculation, temperature of 1.71 (8.9 K) at a reduced pressure of 4.0 (0.91 MPa)
the heat removal means a negative value of heat input, so it is ranked as (shown as the cyan line). There are a peak and a bottom for the heat
one kind of heat input in this manuscript. If we define the direction input and the heat removal, respectively, which are close to the se-
from the cold end to the hot end of the regenerator as the positive paratrix point. This causes a sharp change from heat input to heat re-
direction of the axis (noted as “x”), the energy flow (including the PV moval. The peak value and bottom value accelerate as the pressure gets
power, the heat flow and the enthalpy flow) is transferred in the same closer to the critical pressure, and the absolute values of the peak and
direction of the temperature gradient (dT/dx).
In the following derivation, the regenerator is assumed to be fric-
tionless, without axial conduction, and there can be perfect heat
transfer and infinite matrix heat capacity when needed [36]. So the
regenerator is lossless in an ideal case.
The real gas effects challenge the energy balance of the regenerator
[36]. In order to keep the energy balance, the local differential heat
input of a finite temperature range is responsible for the variation of the
total heat flow (sum of the pressure-induced enthalpy flow and the
heat-associated enthalpy flow) [36] of the regenerator, which is ex-
pressed as:
δQ̇ ̇
δQtot
= .
δx δx (1) Fig. 2. Illustration for the heat transfer process and the time-averaged energy
This process is illustrated in Fig. 3. The integration of the differential flow in the regenerator of the refrigeration cycle.
707
Q. Cao Applied Energy 220 (2018) 705–712
heat input refers to the availability divided by the PV power at the hot
end, and it is expressed as:
The lines of the availability regarding the temperature share the same
trend with the heat input. The separatrix points are actually that of the
heat input, but the peak and bottom points generally shift a bit lower.
The lines of availability become flatter at higher temperatures, and
become more rugged at lower temperatures, because the temperature
multiplier enhances the temperature effect.
The dimensionless availability can be integrated over the tempera-
ture from the hot end. The peak of the integrated availability is gen-
erally in the same magnitude as the PV power at that point, which
means that a significant availability can be utilized to improve the COP.
The peak at the critical temperature is even much greater than the PV
Fig. 4. The dimensionless heat input per kelvin for 4He and 3He corresponding power at the hot end. As the temperature gets farther from the critical
to various reduced pressures. The solid lines represent for 4He and the dash temperature, the peak becomes smaller.
lines 3He. The black, red, blue, and cyan lines represent for a reduced pressure The example of 4He in the previous section is analyzed continually,
of 0.5, 1.0, 2.0, and 4.0, respectively. (For interpretation of the references to as shown in Fig. 5. The peak and bottom points of the dimensionless
colour in this figure legend, the reader is referred to the web version of this
availability shift by about 0.3 K and 0.1 K comparing with that of the
article.)
heat input at a reduced pressure of 4.0, and the peak value and the
bottom value increase by about 1.7 and 3.3 times than that of the heat
the bottom reach over 40 1/K at the critical pressure, which agrees with input at this pressure. The integrated availability is shown in Fig. 6. The
the general principle that the real gas effects aggregate when the gas availability is integrated from the cold end. The peak point at the cri-
gets closer to the critical point. tical temperature is more than 10 times higher than the PV power at the
Another example of 3He [53] is shown as the dash lines in Fig. 4. Its hot end. The peak point decreases to below 1.0 when the reduced
tendency is the same as that of 4He. The value of the dimensionless heat temperature is lower than 0.5 or higher than 4.0. The integrated
is larger than that of 4He in most area. The reduced temperature of the availability over the whole temperature range is larger than zero in
peak and bottom points is a bit smaller than that of 4He when the these cases, which means that the net availability is related to the heat
pressure is smaller than the critical pressure, but it is a bit larger than input rather than the heat removal.
that of 4He when the pressure is larger than the critical pressure. The lines of 3He quite resemble those of 4He in most area, but there
However, since the critical temperature and critical pressure of 3He are some differences in specific points. For example, although the peak
(3.3157 K, 0.1146 MPa) are significantly smaller than that of 4He value of the dimensionless availability of 3He is 15% larger than that of
(5.1953 K, 0.2276 MPa), it is reasonable that the real gas losses of 3He the heat input, the bottom value is 4% smaller than that of the heat
are smaller than that of 4He at a given temperature and pressure. input at a reduced pressure of 4.0. The integrated availability of 3He is a
bit larger than that of 4He in most area, except when the pressure is
2.2. The availability of the heat input close to the critical pressure.
The availability of the heat means the work extracted in a reversible 2.3. The expression of the COP
process at a reference pressure and a reference temperature, i.e. the
average pressure and the hot-end temperature in a refrigeration cycle. Since the heat input keeps the energy balance in the ideal re-
The dimensionless distributed availability per temperature range of the generator working with real gases, the heat-associated enthalpy flow is
708
Q. Cao Applied Energy 220 (2018) 705–712
0.30 for 3He) when the reduced pressure is smaller than 1.0.
According to our calculation, the heat input can always improve the
COP to the value of the Carnot efficiency. The heat input effect is more
significant as the temperature gets closer to the critical point. For ex-
ample, it improves the COP by 1.4 times at the critical point of 4He,
comparing with an improvement of about 10% at a reduced tempera-
ture of 2.0. Case of 3He is in the same magnitude as that of 4He.
709
Q. Cao Applied Energy 220 (2018) 705–712
upper limit of the temperature range is the separatrix temperature of (ZT )c (βT )c
COP = N
.
the heat input, as discussed in Section 2.1. The upper limit is a reduced (ZT )h (ZT ) Th − Ti T 1 d [(1 − Tβ ) ZT ]
temperature of 1.73 at a reduced pressure of 4.0 for 4He, and it is a [1− (ZT )c −
h
∑ Ti
∫Ti−i 1 (ZT )h dT
dT ]
i=1 (9)
reduced temperature of 1.96 at a reduced pressure of 4.0 for 3He. The
cases of the temperature range between the lowest temperature and the In this equation, the number of the discrete regenerator locations is
upper limit at a reduced pressure of 4.0 for 4He and 3He are put into noted as N. It is clear that increasing the number will lower the dif-
investigation in the following, as shown in Fig. 8. The black line shows ference between the discrete method and the distributed method. The
the variation of the relative Carnot COP (rCOP) as a function of tem- temperature at the discrete regenerator locations is distributed evenly
perature, to which the “intermediate refrigeration power” is integrated in the temperature range in the following calculation for 4He and 3He.
from the cold-end temperature. The COP reaches the Carnot efficiency The rCOP increases from below 10% to a magnitude of 40% for both
when the distributed “intermediate refrigeration power” is integrated in gases if there is one discrete regenerator location in the middle. The
the whole temperature range. rCOP increases to a magnitude between 90% and 95% for both gases if
Under the circumstances when the regenerator works in a tem- there are 7 regenerator locations. The trend of the rCOP is that more
perature range in part of which the pressure-induced enthalpy flow discrete regenerator locations results in a closer trend of the distributed
decreases synchronously as the temperature decreases, the “inter- case, as shown in Fig. 8, which improves the COP to a value that ap-
mediate refrigeration power” can be obtained only in part of the tem- proaches the Carnot efficiency.
perature range. But the “intermediate refrigeration power” can still It is complicated to do the experiment of the heat input, because
significantly improve the COP. The reason is that the pressure-induced there are various heat losses and flow losses, and the temperature
enthalpy flow usually reaches a value of over 50% of the PV power at profile of the regenerator varies when the heat input is applied.
the cold end when the temperature is below the critical temperature, Fortunately, the previous experiment study [46] shows that, the more
and the “intermediate refrigeration power” method recovers a large discrete regenerator locations of heat input there are, the higher the
portion of the lost cooling power. COP of the refrigerator is. A heat input of 0.5 W was applied for one
The examples of 4He and 3He in the previous section are analyzed location, at a temperature of 6.4 K. But a heat input of 0.25 W and 0.5 W
for the “intermediate refrigeration power,” with a fixed hot-end tem- was able to be applied for two locations, at a temperature of 5.1 K and
perature (reduced temperature) of 3.85, and a fixed cold end tem- 7.1 K, respectively. The heat input did not degrade any performance of
perature (reduced temperature) of 0.42 and 0.30 for 4He and 3He, re- the origin stages, which was a refrigeration power of 0.5 W and 10 W at
spectively. The “intermediate refrigeration power” is integrated from a temperature of 3.8 K and 50 K, respectively. The input power was
the cold end, and it is divided by the PV power at the hot end. The 5.5 kW. If the refrigeration power at the temperature above the cold-
“intermediate refrigeration power” accounts for a maximum of 30% of end temperature is converted to the refrigeration power with a multi-
this PV power, as shown in Fig. 9. plier of the temperature ratio (Tc/T), which corresponds to the influ-
The amount of the “intermediate refrigeration power” increases ence of the temperature in an ideal process of the pulse tube refrigerator
synchronously with the pressure. As the reduced pressure increases that the expansion work is not recovered, the rCOP for the case without
from 0.5 to 4, the dimensionless “intermediate refrigeration power” at a heat input is equal to 1.79%. And the heat input in one location and in
reduced temperature of 0.5 increases by about 3 times both for 4He and two locations realizes an increase of the COP by 24% and 36%, re-
3
He, and the final integrated “intermediate refrigeration power” in- spectively.
creases from 0.022 to 0.26 for 4He and from 0.025 to 0.31 for 3He, both A previous measurement of the liquefaction rate of the helium gas
increasing by over 10 times. found in the literature [47] demonstrates how the “intermediate re-
The temperature range of the “intermediate refrigeration power” frigeration power” is applied to precool the gas with the distributed
gets broader as the pressure increases. The available temperature al- method. The numerical calculation gave a result that the “intermediate
ways starts from the cold end if the pressure-induced enthalpy flow refrigeration power” was equal to 2.04 W, which is over 3 times larger
reaches a maximum at the cold end. The upper limit increases from
about a reduced temperature of 0.8 to 1.4 and from 0.8 to 1.5 for 4He
and 3He, respectively, as the reduced pressure increases from 0.5 to 4.
The range of the available temperature almost doubles for both gases.
The rCOP for 4He and 3He is shown in Fig. 10. The rCOP without
“intermediate refrigeration power” is in the magnitude of 1%, but the
“intermediate refrigeration power” improves the COP by over 10 times.
The rCOP increases to over 80% when the reduced pressure is 4.0,
which means that the COP is close to the Carnot efficiency.
710
Q. Cao Applied Energy 220 (2018) 705–712
711
Q. Cao Applied Energy 220 (2018) 705–712
712