A resorption cycle for the cogeneration of electricity and refrigeration

Applied Energy 106 (2013) 56–64

Contents lists available at SciVerse ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

A resorption cycle for the cogeneration of electricity and refrigeration Liwei Wang a,b,⇑, Felix Ziegler c, Anthony Paul Roskilly b,1, Ruzhu Wang a, Yaodong Wang b a

Institute of Refrigeration and Cryogenics, Shanghai Jiao Tong University, Shanghai 200240, China Sir Joseph Swan Centre for Energy Research, Newcastle University, Newcastle NE1 7RU, UK c Technische Universitat Berlin, KT2, Marchstrasse 18, 10587 Berlin, Germany b

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

" A resorption cogeneration cycle for

A novel resorption cycle driven by low grade heat for cogeneration of electricity and refrigeration is studied. The cycle features in high exergy efficiency, very little or no ammonia liquid inside and simple structure.

electricity and refrigeration is proposed. " The cycle improved refrigeration COP by 10 times compared with Goswami cycle. " The highest exergy efficiency of the cogeneration cycle is as high as 0.9. " The cycle also features in safety and simple structure.

a r t i c l e

i n f o

Article history: Received 23 October 2012 Received in revised form 11 January 2013 Accepted 12 January 2013

Keywords: Resorption Refrigeration Electricity generation Exergy efficiency

a b s t r a c t This paper describes a novel resorption cycle driven by the low grade heat for the cogeneration of electricity and refrigeration, which is based on ammonia adsorption refrigeration technology. The presented cycle features a variable endothermic process which stands for higher adaptability if compared with the traditional Rankine cycle, very little or no ammonia liquid in the system which is a safety feature, solid adsorbents inside the beds, and simple structure for the fact of no rectifying equipment and circulation pumps required by the working fluids. This cycle can be utilised for the heat source with the temperature higher than 100 °C, and it has an electricity generation exergy efficiency of up to 0.69 and a refrigeration coefficient of performance (COP) of up to 0.77. If compared with the Goswami cycle, which is established based on the absorption Kalina cycle for the cogeneration of electricity and refrigeration, the novel resorption cycle kept the merit of the high exergy efficiency for electricity generation, meanwhile, it overcame the limitation of the low refrigeration coefficient of performance (COP) of Goswami cycle, and improved the COP by 10 times. The optimum overall exergy efficiency is as high as 0.9, which is 40– 60% improved compared with the Goswami cycle under the same working conditions. Ó 2013 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Institute of Refrigeration and Cryogenics, Shanghai Jiao Tong University, Shanghai 200240, China. Tel.: +86 21 34206056; fax: +86 21 34206814. E-mail addresses: [email protected] (L.W. Wang), [email protected] (F. Ziegler), [email protected] (A.P. Roskilly), [email protected] (R.Z. Wang), [email protected] (Y.D. Wang). 1 Tel.: +44 191 222 5869. 0306-2619/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.apenergy.2013.01.041

L. Wang et al. / Applied Energy 106 (2013) 56–64

57

Nomenclature a, b, n, and k the stoichiometric factors of the reaction C specific heat (kJ/(kg K)) COP coefficient of refrigeration performance H enthalpy (kJ) m mass of ammoniate HTS (kg) M molar mass (mol) Q heat (kW) tc cycle time T temperature (K) W electricity (kW) DH enthalpy change (kJ) DU internal energy change (kJ) DS entropy change (kJ) E exergy (kW)  m mass flowrate (kg/s) g exergy efficiency

1. Introduction The utilisation of low grade heat is one of the main options to overcome the developing constraints which are related to the ever rising energy demand [1,2]. The conventional Rankine cycle is a well known power cycle for the utilisation of the low grade heat, and its primary challenge limiting performance is the isothermal endothermic and exothermic processes which often cannot match the temperature profiles of the heat source and sink perfectly. Using an ammonia–water mixture, as in the Kalina cycle [3], is attractive because it allows the exploitation of available exergy to be maximised due to the adjustable temperature variation during the liquid–vapour phase change by controlling the ammonia concentration. The Kalina cycle, under special circumstances [4], generally produces higher exergy efficiency when compared to the Rankine cycle [5,6]. A totally different approach for harvesting the waste heat for power production is the use of thermoelectric systems. This technology has great appeal in terms of its simplicity, but its energy efficiency is only about 0.1 for a heat source temperature lower than 350 °C [7]. Combining the electricity generation with refrigeration may improve the exergy efficiency. This is done in the so-called Goswami cycle [8], in which a heat exchanger (precooler) is added before the absorber in a Kalina cycle so as to recover the low temperature potential of the expanded cold gas stream for refrigeration. Of course, as this heat is only specific heat and not latent heat from a phase change, the Goswami cycle suffers from a fundamental limitation of a low refrigeration coefficient of performance (COP), which is only approximately 0.05 [9], although the exergy efficiency of the cogeneration cycle can reach 0.5 [10]. Chemisorption refrigeration technology is attractive because of the wide range of chloride salts that can absorb ammonia, which allows high flexibility in the application of different temperature heat sources. Cycles of this kind have been theoretical studied extensively by Alefeld and other researchers from the 1980s [11–13]. The working principle [14] is fundamentally the same as that of absorption refrigeration. Consequently, the cycle will have a comparable COP for refrigeration if it was used for the cogeneration cycle. Resorption [15–19] is a form of absorption or chemisorption with the advantage of improved safety. It also utilises the heat of decomposition in the refrigerating process [20], which provides the potential for performance improvement for the energy conversion [21]. This paper proposes a novel resorption cogeneration cycle for electricity and refrigeration that has high refrigeration COP,

Subscripts ad adsorption am ammonia c cooling de desorption el electricity env environmental ex exergy h heat, heating H high temperature salt hr heat recovery L low temperature salt Me metal element ref refrigeration

has a reduced operating pressure, possesses little or no ammonia liquid in the system, and has a great potential for a very simplified system for the recovery of the low grade heat. In order to have an overall understanding of this new type of cycle, the different working pairs are compared by analysing energy and exergy efficiency, and the performance is compared with the Goswami cycle operating under the same working conditions. 2. Resorption cycle for the cogeneration of electricity and refrigeration 2.1. The principle of the cycle The resorption cogeneration cycle presented in Fig. 1 utilises a turbine located between the high temperature salt (HTS) and low temperature salt (LTS) beds to generate electrical power. A superheater and a precooler are integrated in the system to optimise the performance, and two different working phases which take place one after the other are involved. The first phase is the power generation phase. The thermodynamic diagram is shown in Fig. 2a. In this process the endothermic reaction of the ammoniate HTS at temperature TdeH with heat input Qde (Figs. 1 and 2a) adds an internal energy difference of DUhH and provides the reaction enthalpy DHdeH (for decomposition process) to the ammoniate HTS, as well generates ammonia vapour with the enthalpy of Ham1, which is superheated to temperature Th by isobaric heat addition Qh. The ammonia vapour with the enthalpy of Ham2 is (in the ideal case) isentropically expanded to produce mechanical work W in the turbine. The expanded ammonia may additionally generate some refrigeration effect Qref1 in an isobaric process at temperature Tref by the enthalpy difference in the precooler if the temperature after expansion TLo is lower than the refrigeration temperature Tref. This is comparable to the situation in the Goswami cycle. Finally the ammonia vapour with enthalpy of Ham4 is adsorbed in Bed 1 by the LTS at temperature TadL and releases heat QadL to the environment. Consequently, the internal energy and enthalpy of ammoniate LTS are decreased by DUcL and DHadL, respectively. The second phase is the resorption refrigeration phase, and the thermodynamic diagram is shown in Fig. 2b. In this process the refrigeration Qref2 (Figs. 1 and 2b) is generated due to the sum of the decomposition heat DHdeL and internal energy difference DUhL of the ammoniate LTS in Bed 2. The generated vapour with the enthalpy of Ham5 has to be adsorbed in Bed 2 by the HTS and the heat of adsorption QadH is rejected to the environment at the temperature of TadH. Simultaneously, the internal energy and

58

L. Wang et al. / Applied Energy 106 (2013) 56–64

Fig. 1. The principle of the resorption cogeneration cycle.

Fig. 2. Thermodynamic diagram for the cogeneration cycle: (a) power generation process, (b) resorption refrigeration process, (c) T–S diagram for heat recovery of HTS and (d) T–S diagram for heat recovery of LTS.

enthalpy of ammoniate HTS are decreased by DUcL and DHadL, respectively. The working conditions of the four beds will be switched as soon as the adsorption and desorption in the beds are completed. Afterwards, the beds change their respective role and the beds have to interchange their temperature. The recovery of sensible heat [22] from Bed 1 of the HTS to Bed 2 of the HTS at the time of switching as shown in Fig. 2c will decrease the heat (Qde) required for the heating process of HTS bed and, consequently, make up for some of the loss in performance which is due to the batch operation. The heat recovery of sensible heat from Bed 2 of LTS to Bed 1 of LTS as shown in Fig. 2d will decrease the cooling requirement for the bed which has to be supplied from the external

cold carrier for the cooling process of LTS bed, i.e. provide part of the internal energy change of ammoniate LTS (DUhL), and consequently it will improve the refrigeration output. Theoretically, the temperatures after heat recovery for both beds are equal. 2.2. Energy balancing and exergy analysis equations The equilibrium reaction for different salts is:

Mea Clb  ðNH3 Þn þ ðk  nÞNH3 () Mea Clb  ðNH3 Þk

ð1Þ

where Me is the metal element, a, b, n, and k are the stoichiometric factors of the reaction.

59

L. Wang et al. / Applied Energy 106 (2013) 56–64

Due to the fact that no fundamental equation or enthalpy diagram for the used working substances exists approximate balances are being used. The power producing phase includes two heating steps, i.e. the desorption process (Qde) and the superheating process (Qh). In the desorption process the decomposition heat of  of desorbed ammonia vaDHdeH to produce the mass flowrate m pour has to be provided, as well as the internal energy difference of the ammoniate HTS. The specific heat capacity of the heat exchanger, the shell, the heat carrier in the exchanger etc. is neglected here. Depending on the design it will increase the required heat input and, consequently, bring down efficiency. The heat recovery process needs to be considered for the calculation of the internal energy difference of the ammoniate HTS. At the beginning of the heat recovery process, the ammoniate HTS in the bed after desorption (point C in Fig. 2c) is MeaClb(NH3)n shown in Eq. (1), and the ammoniate HTS in the bed after adsorption (point A in Fig. 2c) is MeaClb(NH3)k shown in Eq. (1). The temperature after heat recovery is:

T hrH ¼






MMea Clb ðNH3 Þn  C Mea Clb ðNH3 Þn  T deH
 þ M Mea Clb ðNH3 Þk  C Mea Clb ðNH3 Þk  T adH =
M Mea Clb ðNH3 Þn  C Mea Clb ðNH3 Þn þ M Mea Clb ðNH3 Þk  C Mea Clb ðNH3 Þk ð2Þ

where M is the molar mass, C is the specific heat.  the When the average mass flowrate of desorbed vapour is m, mass of ammoniate HTS after adsorption is calculated as follows:

mH ¼ m  t c

  ðk  nÞ  MNH3  MMea Clb ðNH3 Þk

ð3Þ

MeaClb(NH3)n shown in Eq. (1), and the ammoniate LTS in the bed after adsorption (point C in Fig. 2d) is MeaClb(NH3)k shown in Eq. (1). The temperature after heat recovery is:

T hrL ¼ ½ðM Mea Clb ðNH3 Þn  C Mea Clb ðNH3 Þn  T ref Þ þ ðMMea Clb ðNH3 Þk  C Mea Clb ðNH3 Þk  T adL Þ= ðMMea Clb ðNH3 Þn  C Mea Clb ðNH3 Þn þ M Mea Clb ðNH3 Þk  C Mea Clb ðNH3 Þk Þ ð8Þ  the mass of When the mass flowrate of desorbed vapour is m, ammoniate LTS after adsorption is:

 mL ¼ m  t c ½ðk  nÞ  MNH3   MMa Clb ðNH3 Þk

ð9Þ

Then:

Q ref1 ¼ m  ðHam4  Ham3 Þ

ð10Þ

Q ref2 ¼ m  DHdeL þ m  DU hL ¼ m  DHdeL þ ½mL  C Ma Clb ðNH3 Þk  ðT ref  T hrL Þ=tc ¼ m  DHdeH þ m  M Ma Clb ðNH3 Þk =ðk  nÞ  M NH3  C Ma Clb ðNH3 Þk  ðT ref  T hrL Þ

ð11Þ

The values of refrigeration exergy from the precooler and the resorption refrigeration process are:

Eref1 ¼ Q ref1  ðT env =T ref  1Þ;

Eref2 ¼ Q ref2  ðT env =T ref  1Þ

ð12Þ

where tc is the half cycle time. tc is equal to the cooling and adsorption time. Then the desorption heat is:

Because the electricity is a type of energy having different grade with the input heat, the exergy of heat is utilised and the exergy efficiency (gel,ex) is calculated for the evaluation of the system performance.

Q de ¼ m  DHdeH þ m  DU hH

gel;ex ¼ W Eh



¼ m  DHdeH þ ½mH  C Ma Clb ðNH3 Þk  ðT deH  T hrH Þ=tC

Coefficient of performance (COP) for refrigeration process is:

¼ m  DHdeH þ m  MMa Clb ðNH3 Þk =ðk  nÞ  M NH3  C Ma Clb ðNH3 Þk  ðT deH  T hrH Þ

ð4Þ

The heat provided for the enthalpy increment of the ammonia vapour by the superheating process is:

Q h ¼ m  ðHam2  Ham1 Þ

ð5Þ

The exergy for the heating process Eh is determined by the amount of heat, the heating temperature for desorption and superheating processes, i.e. TdeH and Th (K), and the environmental temperature of Tenv (K). Internal equilibrium temperatures are being used, so the irreversibility in heat exchange is not accounted for

Eh ¼ Q de  ð1  T env =T deH Þ þ Q h  ð1  T env =T h Þ

ð6Þ

For the electrical generating phase the power is calculated by the isentropic enthalpy decrease of the ammonia vapour through the turbine.

W ¼ m  ðHam3  Ham2 Þ

ð13Þ

ð7Þ

There are two processes which can produce refrigeration. This is to a small part the cooling Qref1 that is provided by the precooler, which is determined by the enthalpy difference of the ammonia vapour when it heats up after leaving the expansion turbine; this refrigeration is available in the power producing phase. The predominant part is the resorption refrigeration power Qref2, which is due to the desorption heat DHdeL of LTS in the refrigeration phase. In this process the internal energy change DUdL of the ammoniate LTS, which is used for cooling down the LTS after the heat recovery phase needs to be considered. Again, the thermal mass of additional components is neglected. At the beginning of the heat recovery process for LTS, the ammoniate LTS in the bed after desorption (point A in Fig. 2d) is

COP ¼ ðQ ref1 þ Q ref2 Þ=ðQ de þ Q h Þ

ð14Þ

The exergy efficiency for the refrigerating process (gref,ex) is:



gref;ex ¼ ðEref1 þ Eref2 Þ Eh

ð15Þ

The total exergy efficiency of the cycle is:

gtotal;ex ¼ gel;ex þ gref;ex

ð16Þ

In the calculation the values of desorption enthalpy for ammoniate HTS (DHdeH) and ammoniate LTS (DHdeL) compounds as well as pressures and specific heat are determined by the choice of adsorbents. 2.3. Analysis of adsorbents The equilibrium reaction lines and thermodynamic parameters of the salts are shown in Fig. 3 and Table 1 [19]. Five HTS (number 4–8) and three LTS (1–3) selections are presented in Table 1 and Fig. 3 (along with pure ammonia, 0), and two examples of P–T cycles are shown in Fig. 4. PbCl2(8–3.25) NH3, BaCl2(8–0) NH3, and CaCl2(8–4) NH3 were chosen as potential options for the ammoniate LTSs. In this notation, the numbers in brackets refer to m and n in Eq. (1). For the resorption refrigeration phase, the potential HTS selection is limited by the heat rejection temperature and the resorption refrigeration temperature. Therefore, with an environmental temperature of 30 °C, cooling the HTS bed limits the range of states of the HTS to the yellow area shown in Fig. 3. Using the principle of isobaric resorption between two types of beds under the condition of that Tref is equal to 10 °C, when the LTS is CaCl2 (point k), the possible HTS selections are MnCl2, FeCl3, MgCl2, and NiCl2 (points

60

L. Wang et al. / Applied Energy 106 (2013) 56–64

Fig. 3. Equilibrium P–T reaction lines of different salts.

Table 1 Thermodynamic parameters for different salts. Number shown in Fig. 3

Material

DHde/ad (kJ mol1)

DSde/ad (kJ kg1 K1)

C (kJ kg1 K1)

0 1

NH3 Pb83.25 Ba8-0 Ca8-4 Sr8-1 Mn6-2 Fe6-2 Mg6-2 Ni6-2

1375 2019

8.835 13.16

4.72 (liquid) 4.12

2216 2413 2437 2789 3016 3274 3483

13.37 13.55 13.46 13.42 13.41 13.57 13.40

4.42 4.27 4.44 4.29 4.50 4.19 4.21

2 3 4 5 6 7 8

the equilibrium reaction pressure of the ammoniate HTSs, indicated by the red dots in Fig. 3 for different working pressure. The pressure of the turbine outlet will be the equilibrium reaction pressure of the ammoniate LTSs, restricted by the environmental cooling media at 30 °C. As a result of the cooling temperature, the equilibrium reaction points of f, g, and h with pressure of 6 bar, 2 bar and 1 bar are achieved for PbCl2, BaCl2 and CaCl2 respectively. In this phase some refrigeration power is generated inside the precooler by the expanded ammonia from the turbine as in the Goswami cycle. The most important state points of the thermodynamic processes for the two working pairs, MnCl2–CaCl2–NH3 and MnCl2– BaCl2–NH3, with a refrigeration temperature of 10 °C are shown in Fig. 4. For the electricity generation phase, point A is the isothermal and isobaric desorbing process of MnCl26NH3, and A–B is the isobaric superheating process. B–C and B–G are isentropic electricity generation processes for the working pairs of MnCl2–BaCl2–NH3 and MnCl2–CaCl2–NH3 respectively. This shows that the pressure ratio in the expansion process for MnCl2–CaCl2–NH3 is larger than that of MnCl2–BaCl2–NH3, consequently the expanded ammonia leaves the turbine at a temperature below 10 °C for MnCl2–CaCl2– NH3 and higher than 10 °C for MnCl2–BaCl2–NH3. Only MnCl2– CaCl2–NH3 generates refrigeration power in the precooler below 10 °C by the isobaric heating process (G–H) of the expanded ammonia. C–D and H–I are isobaric heating processes of the expanded ammonia above 10 °C (not usable for refrigeration), while at D and I the isobaric and isothermal adsorption processes for the ammoniate LTSs take place for MnCl2–BaCl2–NH3 and MnCl2– CaCl2–NH3. For the resorption refrigeration phase, the isobaric and isothermal refrigerating processes at 10 °C take place at points E and J, with D–E and I–J as isochoric and isosteric cooling processes (switching time between the phases) for BaCl2-ammonia and CaCl2-ammonia compounds respectively. E–F and J–K are isobaric heating processes of the ammonia vapour desorbed from the LTS ammonia compounds. In points F and K the isobaric and isothermal adsorption processes of the ammoniate HTS compound take place for MnCl2–BaCl2–NH3 and MnCl2–CaCl2–NH3, respectively.

3. Energy and exergy efficiency analysis 3.1. Electricity generation phase

Fig. 4. P–T diagrams of typical resorption cogeneration cycles.

of a, b, c, d). For a LTS BaCl2 (point j) or PbCl2 (point i) the HTS selection options extend to include SrCl2. If the pressure in the resorption process is too low the mass transfer resistance will be high and, consequently, the reaction rate and power density will be poor. Therefore, a minimum pressure was chosen to be 0.2 bar and as a result only BaCl2 and PbCl2 serve as the LTSs for a refrigeration temperature of 0 °C (points m and n) and 10 °C (points o and p). The lowest refrigeration temperature is 26 °C with PbCl2 as the LTS (point q). For the electricity generation phase, a number of HTSs are suitable for different levels of desorbing temperature. As Fig. 3 shows, ammoniate SrCl2 has the lowest desorbing temperature of around 100 °C and ammoniate NiCl2 has the highest desorbing temperature of around 250 °C. The pressure of the turbine inlet will be

The output of specific work for electricity generation related to 1 kg of ammonia by the enthalpy difference ((Ham2  Ham3) in Fig. 2b) is shown in Fig. 5a–c. PbCl2 has the lowest electricity generation while CaCl2 has the highest. It is mainly because PbCl2 has the highest constraint pressure (6 bar) while the CaCl2 has the lowest (1 bar) at the exit of the expander. Two optimal LTSs, i.e. CaCl2 and BaCl2 are chosen for the exergy analysis. The exergy efficiency for the electrical generating process and refrigeration process is shown in Fig. 6. For the electrical generation phase (Fig. 6a and b) the exergy efficiency increases with the increasing pressure ratio and decreases with the increasing temperature for the desorption and superheating process. For the HTSs with higher desorption temperature the exergy efficiency increases while the superheating temperature increases because the increment of power is larger than the increment of exergy input. For the salt of SrCl2, which has the desorption temperature of about 100 °C, the increment of power is larger than the increment of exergy while the superheating temperature is lower, and the exergy efficiency increases. After that it will decrease because the increment of exergy is larger than the increment of power. The working pair with the LTS with the

61

b

900 800

Superheating temperature o of 400oC 250 C

700 600

200oC

500

300oC

400 300 200

150oC

100

Specific work output /(kJ/kg)

Specific work output /(kJ/kg)

a

100oC 0 0.5 1 1.5 2 2.5 3 3.5

900 800

Superheating 300oC temperature of 400oC

700 600 500 400 300

150oC

200

200oC

100oC

100 0 0.5

Superheating pressure/MPa

250oC

1

1.5

2

2.5

3

c

900

Specific power output /(kJ/kg)

L. Wang et al. / Applied Energy 106 (2013) 56–64

800

400oC

700

300oC

250oC

600 500 400 300 200 100

200oC Superheating o temperature 150 C o of 100 C

3.5

0.5 1 1.5 2 2.5 3 3.5

Superheating pressure/MPa

Superheating pressure/MPa

Fig. 5. Specific power output for (a) LTS of PbCl2, (b) BaCl2 and (c) CaCl2.

3.2. Refrigeration phase

b 0.8

0.7

0.7

0.6

0.6

0.5

0.5

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1

0.0 0

100 200 300 400 500

0.0

0

100 200 300 400 500

Superheating temperature/oC

Superheating temperature/oC

Fig. 6. Exergy efficiency for electricity generation process under the condition of 3 MPa superheating pressure and 10 °C refrigeration temperature, (a) with LTS of BaCl2 and (b) with LTS of CaCl2.

lower equilibrium reaction pressure at a given temperature and the HTS with the higher equilibrium reaction pressure at a given temperature showed better performance. Generally it holds that for a working pair with a desorption temperature for the HTS higher than 200 °C the exergy efficiency is lower than 0.3–0.4. For a HTS with desorption temperature lower than 200 °C the exergy efficiency generally is higher than 0.4. The highest exergy efficiency for the power generation is 0.69 for the working pair of BaCl2– SrCl2–NH3 with the superheating temperature of 250–300 °C.

0.9

SrCl2 as the HTS

0.8

MgCl2

0.6

NiCl2

SrCl2 as the HTS

0.8

c

0.9 0.8

MnCl2

MnCl2 FeCl3

0.7

COP

b 0.9

FeCl3

0.7

COP

a

For a superheating pressure of 3 MPa and a refrigeration temperature of 10 °C, the trends for the refrigeration coefficient of performance (COP) are analysed and plotted in Fig. 7a–c. These show that the LTS with the higher reaction temperature, i.e. BaCl2 and CaCl2, and the HTS with the lower generation temperature, such as SrCl2 and MnCl2, will have the higher COP, and the values decrease with increasing superheating temperature. The value of COP achievable ranges between 0.39 and 0.77. The highest COP of 0.77 is obtained with BaCl2–SrCl2–NH3 (Fig. 7b) under the condition of 100 °C superheating temperature. It can be observed that the change in COP with superheating temperature is stronger at lower temperatures than at higher temperatures. The reason for this behaviour is the additional refrigeration power gained at small superheating quantity due to very cold turbine outlet, as well as the additional requirement of input heat for high superheating temperatures. The contribution to the COP by the precooler and the resorption process is analysed to compare the contribution of both parts for the optimal working pair of BaCl2–SrCl2 with a superheating temperature of 150 °C, and the results are shown in Fig. 8. The contribution from the precooler is much less than that from the resorption process although it increases with the increasing pressure. It is mainly because of that the cooling capacity contributed by the precooler is the sensible heat of the refrigerant vapour, and it is much less than the desorption heat of LTS for generating the cooling capacity. The largest contribution from the precooler is only approximately 6.2%. Fig. 8 also indicates that for some occa-

MgCl2 NiCl2

0.6 0.5

0.5

0.4

0.4

0.4

0

100 200 300 400 500

Superheating

temperature/ oC

0.3

0

100 200 300 400 500

Superheating

temperature/ oC

MgCl2

FeCl3

NiCl2

0.6

0.5

0.3

MnCl2

0.7

COP

a 0.8

0.3

0

100 200 300 400 500

Superheating temperature/ oC

Fig. 7. COP for the LTSs of (a) PbCl2, (b) BaCl2 and (c) CaCl2 under the condition of 3 MPa superheating pressure and 10 °C refrigeration temperature.

62

L. Wang et al. / Applied Energy 106 (2013) 56–64

tion phases in one cycle were both taken into account. The refrigerating exergy efficiency is shown in Fig. 9a and b. It decreases with increasing temperature, and ranges between 0.064 and 0.29. The highest exergy efficiency is gotten from the working pair of BaCl2–SrCl2–NH3 at the superheating temperature of 100 °C.

0.8

COP

0.6

3.3. Total exergy efficiency for the cogeneration cycle

0.4

The total exergy efficiency of resorption cycle is analysed, and the results are shown in Fig. 10a and b. For the cogeneration of the electricity and refrigeration, the maximum exergy efficiency is as high as 0.9 for the working pair of BaCl2–SrCl2–NH3 at the superheating temperature of 150 °C.

0.2

0

1

1.5

2

2.5

3MPa

Superheating pressure

4. Performance comparisons with Goswami cycle

COPLTS contributed by the desorption heat of LTS COPprecooler contributed by t he precooler Fig. 8. Contribution comparison for COP by precooler and resorption for BaCl2– SrCl2 under the condition of 150 °C superheating temperature and 10 °C refrigeration temperature.

0.30

SrCl2 as the HTS

0.25

MnCl2 FeCl3 MgCl2

0.20

b 0.35

FeCl3 0.25

MgCl2

0.20

0.15

0.15

0.10

0.10

0.05

MnCl2

0.30

ref.ex

ref.ex

a 0.35

NiCl2

0.05

NiCl2

0.00

0.00 0

100 200 300 400 500

0

Superheating temperature/ oC

100 200 300 400 500

Superheating temperature/ oC

Fig. 9. Exergy efficiency for refrigeration process under the condition of 3 MPa superheating pressure and 10 °C refrigeration temperature, (a) with LTS of BaCl2 and (b) with LTS of CaCl2.

b 1.0

0.8

0.8

SrCl2 as the HTS

0.6

total.ex

total.ex

a 1.0

0.4 MnCl2 MgCl2

0.2 0.0

0

FeCl3 NiCl2

100 200 300 400 500

Superheating

temperature/ oC

MnCl2 FeCl3

0.6 0.4 MgCl2

0.2

NiCl2

0.0

0

100 200 300 400 500

Superheating temperature/ oC

Fig. 10. The total exergy efficiency for cogeneration under the condition of 3 MPa superheating pressure and 10 °C refrigeration temperature, (a) with LTS of BaCl2 and (b) with LTS of CaCl2.

sions if the cooler is removed from the system, the performance of the system would not be influenced very much but the structure of the system will be much simplified. Again two optimal LTSs for refrigeration phase, i.e. CaCl2 and BaCl2 are chosen for the exergy analysis, for which two refrigera-

In order to have an overall understanding for the resorption cogeneration cycle, another thermally driven cogeneration cycle, which is the Goswami cycle [23], is analysed for the same superheating temperature, superheating pressure, refrigeration temperature, and environmental temperature. The performances of two types of cycles are compared. The diagram of Goswami cycle is shown in Fig. 11 [23,24]. It is a simple Rankine cycle with solution circuit [11] with the only specialty that the cold gas after expansion is utilised to produce refrigeration, as stated before. The working processes of Goswami cycle in Fig. 11a are as follows: The ammonia and water in the absorber is pumped to the generator. The generator is heated by the heat Qde at the temperature Th2, and the desorbed ammonia is superheated by the heat Qh in the super heater at the temperature Th1. After that the ammonia vapour enters the turbine and generates the power of W there. The expanded ammonia enters the precooler. If its temperature is lower than the refrigeration temperature Tref it will generate the refrigeration Qref by heating up there to the temperature of Tref. After that it will be absorbed by the absorber, which receives the weak solution from the generator, and rejects the heat to environment. A heat recovery process proceeds in the solution heat exchanger between generator and absorber in order to improve the efficiency. For the Goswami cycle (Fig. 11b) two equilibrium lines are chosen for the analysis, i.e. the absorption equilibrium line of 0.397 kg/kg, which has the same sorption pressure as that of BaCl2 while the cooling temperature is 30 °C, and the equilibrium line of 0.199 kg/kg, which has the same desorption pressure as MnCl2 while the desorption temperature is about 150 °C. In Fig. 11b the solid lines represent the ammonia fluid and the dotted lines represent the solution. For the ammonia vapour F is the end of the desorption process, F–G is the superheating process, G–H is the power generating process. H–K is the refrigeration process by the precooler. For the solution after leaving the generator the poor solution will be cooled in the solution heat exchanger (F–b) and depressurised to the absorber pressure by entering the absorber subcooled, with a flash, or in equilibrium (shown here, b–I) with I as the equilibrium state. Absorption of the ammonia vapour at the state K will take place in the solution between I and O with O being the final state. The solution then is pumped to the high pressure (O–a) and heated in the solution heat exchanger (a–J) with the boiling start state of J. J–F is the desorption process in the generator. The same working conditions as in the resorption cycle are chosen for the Goswami cycle, i.e. the superheating pressure is 3 MPa, the environmental temperature is 30 °C, and the refrigeration temperature is 10 °C, and the energy efficiency and exergy efficiency of

L. Wang et al. / Applied Energy 106 (2013) 56–64

63

0.8

0.20

0.6

0.15

Refrigeration exergy efficiency 0.4

Exergy efficiency for power generation

0.10

the highest value is 0.63 for 400 °C superheating temperature, while the refrigeration COP is much lower than that of resorption cycle. The total exergy efficiency for refrigeration and power generation of Goswami cycle is between 0.57 and 0.64, which is as well much lower than that of resorption cycle.

COP

Exergy efficiency

Fig. 11. Goswami cycle, (a) principle diagram and (b) P–T diagram.

5. Conclusions

Refrigeration COP 0.2

0.05

0 0

100

200

300

400

0.00 500

Superheating temperature / oC Fig. 12. The performances for the Goswami cycle under the condition of 3 MPa superheating pressure and 10 °C refrigeration temperature.

the cycle are calculated. The calculation is similar to that of the resorption cycle except for the pumping power in the Goswami cycle which has to be subtracted from the turbine power. The results are shown in Fig. 12. Fig. 12 shows that the highest value of COP for the refrigeration process is only 0.05. This, of course, is obvious because it produces refrigeration due to subcooled vapour only. The Goswami cycle only generates refrigeration when the superheating temperature is lower than 250 °C. Fig. 12 also shows that the exergy efficiency for refrigeration is very small, and the value only ranges between 0.006 and 0.014. The exergy efficiency for the power generating process of the Goswami cycle is similar with that of the resorption cycle, and

The resorption cogeneration cycle shows a very attractive performance. The highest exergy efficiency gotten for electrical generation and refrigeration are 0.69 and 0.29, while the highest COP for refrigeration is 0.77. The optimum total exergy efficiency is as high as 0.9. Compared with another type of low grade heat driven cogeneration cycle, i.e. the Goswami cycle that was developed from the Kalina cycle, the resorption cogeneration cycle has comparable electricity generation performance but features a refrigeration COP over than ten times higher. At the same time the total exergy efficiency for cogeneration is improved by 40%-60% compared with the Goswami cycle. The resorption cogeneration cycle can be used for recovering solar energy or waste heat in large scale or small scale applications. As both electricity and refrigeration for industrial processes or for air conditioning are produced independently it is possible to adjust, e.g., the electricity output to the actual demand and use the produced cold for instance in refrigerated warehouses as a buffer. The system can also be designed in such a way that it exhibits a large buffering capacity inherently. Sources of waste heat often are dispersed and come in a range of relatively small energy. Therefore, small scale applications have a

64

L. Wang et al. / Applied Energy 106 (2013) 56–64

significant market potential. However, small scale systems usually are required to be simple and have low first cost. To this end, there is also the possibility to eliminate the superheater because the desorbed vapour from the resorption system has no liquid ammonia inside. The possibility for excluding the precooler will further simplify the system. Moreover, sacrificing some efficiency due to leaving out the possibility of internal heat recovery, the plant can be reduced to a very simple system with one HTS bed, one LTS bed, and a turbine. Due to the use of consolidated solid sorbents the plant will be very compact, which is of great importance for small scale application. For such an application, approximately 1.42 kW electricity and 5–6 kW refrigeration can be achieved from 10 kW of waste heat at 200 °C for the ideal cycle.

Acknowledgements We gratefully acknowledge the support from the EU Marie Curie International Incoming Fellowship and the International Collaborating Project Funded by the Foundation of Science and Technology Commission of Shanghai Municipality, China (contract number 11160706000). L.W. Wang thanks the National Science Foundation of China for young excellent academics under the contract number of 51222601.

References [1] Landry BA. Utilization of waste heat. Science 1953;3:3. [2] Hifele W. A global and long-range picture of energy developments. Science 1980;209:174–82. [3] Kalina AI. Combined cycle and waste-heat recovery power systems based on a novel thermodynamic energy cycle utilising low temperature heat for power generation. ASME Paper; 1983, 83-JPGC-GT-3. [4] Köhler S. Geothermisch angetrieben Dampfkraftprozesse: Analyse und Prozessbergleich binärer Kraftwerke. Dissertation. Berlin: Institute of Energy Engineering, Technische Universität Berlin; 2005. [5] Zamfirescu C, Dincer I. Thermodynamic analysis of a novel ammonia–water trilateral Rankine cycle. Thermochim Acta 2008;477:7–15.

[6] Bombarda P, Invernizzi CM, Pietra C. Heat recovery from diesel engines: a thermodynamic comparison between Kalina and ORC cycles. Appl Therm Eng 2010;30:212–9. [7] Bell LE. Cooling, heating, generating power, and recovering waste heat with thermoelectric systems. Science 2008;321:1457–61. [8] Goswami DY. Solar thermal power—status of technologies and opportunities for research. In: Jaluria Y, editor. Proceedings of the second ISHMT—ASME Heat & Mass Transfer Conference, Suratkal, India: Tata McGraw Hill; 1995. p. 57–60. [9] Tamm G, Goswami DY, Lu S, Hasan AA. Theoretical and experimental investigation of an ammonia–water power and refrigeration thermodynamic cycle. Sol Energy 2004;76:217–28. [10] Vidala A, Best R, Rivero R, Cervantes J. Analysis of a combined power and refrigeration cycle by the exergy method. Energy 2006;31:3401–14. [11] Alefeld G, Radermacher R. Heat conversion systems. Bocca Raton, Florida: CRC Press; 1994. [12] Ziegler F. Novel cycle for power and refrigeration. In: 1st European conference on polygeneration. Tarragona, Spain; October 1995. p. 16–7. [13] Meunier F. Solid sorption heat powered cycles for cooling and heat pumping applications. Appl Therm Eng 1998;18:715–29. [14] Critoph RE, Tamainot-Telto Z. Solar sorption refrigerator. Renew. Energy 1997;12:409–17. [15] Wang LW, Bao HS, Wang RZ. A comparison of the performances of adsorption and resorption refrigeration systems powered by the low grade heat. Renew. Energy 2009;34:2373–9. [16] Vasiliev LL, Mishkinis DA, Antukh AA, Kulakov AG. Resorption heat pump. Appl Therm Eng 2004;24:1893–903. [17] Bao HS, Wang RZ, Oliveira RG, Li TX. Resorption system for cold storage and long-distance refrigeration. Appl. Energy 2012;93:479–87. [18] Bao HS, Oliveira RG, Wang RZ, Wang LW, Ma ZW. Working pairs for resorption refrigerators. Appl Therm Eng 2011;31:3015–21. [19] Choudhury B, Saha BB, Chatterjee PK, Sarkar JP. An overview of developments in adsorption refrigeration systems towards a sustainable way of cooling. Appl. Energy 2013;104:554–67. [20] Goetz V, Spinner B, Lepinasse E. A solid–gas thermochemical cooling system using BaCl2 and NiCl2. Energy 1997;22:49–58. [21] Powell JR, Salzano FJ, Yu WS, Milau JS. A high-efficiency power cycle in which hydrogen is compressed by absorption in metal hydrides. Science 1976;193:314–7. [22] Neveu P, Castaing J. Solid-gas chemical heat pumps: field of application and performance of the internal heat of reaction recovery process. Heat Recov. Syst. CHP 1993;13:233–51. [23] Goswami DY. Solar thermal power technology: present status and ideas. Energy Sources 1998;20:137–45. [24] Poolen LJV, Rainwater JC. Critical-region model for bubble curves of ammonia– water with extrapolation to low pressures. Fluid Phase Equilib 1998;150– 151:451–8.