- Email: [email protected]
NaNO3 – Graphite materials for thermal energy storage at high temperature: Part II. – Phase transition properties
Applied Thermal Engineering 30 (2010) 1586e1593
Contents lists available at ScienceDirect
Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng
KNO3/NaNO3 e Graphite materials for thermal energy storage at high temperature: Part II. e Phase transition properties Jérôme Lopez*, Zoubir Acem, Elena Palomo Del Barrio Laboratoire TREFLE UMR CNRS 8508, Esplanade des Arts et Métiers, 33405 Talence Cedex, France
a r t i c l e i n f o
a b s t r a c t
Article history: Received 1 February 2010 Accepted 4 March 2010 Available online 12 March 2010
Composites graphite/salt for thermal energy storage at high temperature (w200 C) have been developed and tested. As at low temperature in the past, graphite has been used to enhance the thermal conductivity of the eutectic system KNO3/NaNO3. A new elaboration method has been proposed as an alternative to graphite-foams infiltration. It consists of compression at room temperature of a physical mixing of expanded natural graphite particles and salt powder. Two different compression routes have been investigated: uni-axial compression and isostatic compression. The first part of the paper shows that both uni-axial and isostatic cold-compression are simple and equally efficient techniques for improving the salt thermal conductivity. The second part of the paper is focused on the analysis of their phase transition properties. It is shown that graphite does not degrade the salt within the composites; that is, no changes are observed neither in the salts transition temperatures nor in its latent heat. On the contrary, some negative effects as pores over pressurization and salt leakage can appear if no void space enough is available within the composite for salt volume expansion when melting. Such negative effects are only observed in the composites obtained by isostatic cold-compression. Ó 2010 Elsevier Ltd. All rights reserved.
Keywords: Phase change materials Thermal properties Thermodynamic analysis Exfoliated graphite
1. Introduction This paper addresses the development and the analysis of new low cost materials with improved thermophysical properties for thermal energy storage at high temperature (w200 C). Important applications using steam as working fluid can be found at temperature close to 200 C in the industry (food, textile, manufacturing, etc.) [1e4] and in solar power generation sector [5,6]. Efficient storage systems for such applications usually demand transfer of energy during the charging/discharging process at constant temperature. This is the reason why our attention has been focused on phase change materials (PCM) as storage media. Taking into account cost constraints, salts have been preferred to metals. However, salts are characterized by low thermal conductivities (w1 W/m/K) that reduce heat exchanges rates during melting/crystallization. Thus, graphite has been used to enhance their thermal conductivity. The eutectic system KNO3/NaNO3 (NaNO3 50 mol%) has been selected as PCM. It has been successfully used in the past for high temperature energy storage purposes, mainly in applications concerning electricity generation by solar concentration technologies
* Corresponding author. Fax: þ33 556 845 436. E-mail address: [email protected] (J. Lopez). 1359-4311/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.applthermaleng.2010.03.014
[7]. The melting temperature and the latent heat of this eutectic system are 223 C and 106 J/g respectively. Moreover, KNO3/NaNO3 has other desirable characteristics such as negligible undercooling, chemical stability, no phase segregation, low corrosion and hygroscopicity, as well as commercial availability at low cost. As said before, graphite has been used to enhance the thermal conductivity of the KNO3/NaNO3 binary system. It is well-known that graphite has strong resistance to corrosion and chemical attack, which make it compatible with most PCM. Moreover, thermal conductivity of graphite particles is considerably high and their density is lower than the metals one. As shown in ref. [8], using graphite additives (dispersion of graphite particles in molten salt) leads to insignificant thermal conductivity improvement in the overall application. From another hand, graphite-foams infiltration with molten salt leads to highly conductive materials but with very low energy density because low saturation degree of salt is achieved [6]. Hence, a new elaboration method has been proposed as an alternative to foams infiltration. It consists of coldcompression (at room temperature) of a physical mixing of expanded natural graphite particles and salt powder. Two different compression routes have been investigated: uni-axial compression and isostatic compression. The first part of the paper [9] has been devoted to the detailed description of the resulting composites as well as to the analysis of their thermal properties. It has been shown that both uni-axial and isostatic cold-compression are
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
simple and equally efficient techniques for improving the salt thermal conductivity. For instance, graphite amounts between 15 and 20%wt lead to apparent thermal conductivities close to 20 W/ m/K (20 times greater than the thermal conductivity of the salt). Furthermore, some advantages as low cost and safety come from materials elaboration carried out at room temperature. This second part of the paper is devoted to the analysis of the phase change properties of the proposed graphite/salt composites. In Section 2, the main characteristics of the microstructure of these composites are highlighted. It includes also a short description of the series of composites that have been elaborated for testing purposes. Section 3 is focused in the presentation and the discussion of the results coming from DSC (Differential Scanning Calorimetry) testing. Last section illustrates the interest of these new materials in terms of storage systems investment cost reduction.
2. Graphite/salt composites As said before, the graphite/salt composites we are studying are obtained by cold-compression (at room temperature) of a physical mixing of expanded natural graphite particles and KNO3/NaNO3 (50 mol%) crystals. Two different compression routes have been investigated: uni-axial cold-compression and isostatic compression (see ref. [9] for details). We remind that expanded natural graphite (ENG) is a well-known material produced from exfoliation of natural graphite particles (i.e. disk-like platelets, usually 0.5 mm diameter and 0.05 mm thickness). Exfoliation leads to worm-like particles whose length is up to 300 times the thickness of the initial platelets, while their diameter is unchanged. The worms are characterized by an irregular honeycomb-like microstructure, a very low density (w3 kg/m3), a very high porosity (>99%) and a huge specific surface (>40 m2/g). The uni-axial compression route leads to a cohesive matrix made of graphite whose porosity is partially occupied by salt grains. At the macroscopic scale, the resulting composites show a parallel layered structure with alternate “graphite dominant” and “salt dominant” layers. Moreover, uni-axial compressing obviously induces some preferential orientation of ENG particles in such a way they tend to lie within a plane perpendicular to the pressing stress. Consequently, “graphite dominant” layers show a parallel layered structure too (Fig. 1, left). Within the “salt dominant” layers, one observes a “random” mixture of salt grains and more or less deformed ENG particles (Fig. 1, center). As for porosity, we can roughly distinguish: (a) The porosity within the “salt dominant” layer, which mainly depends on salt crystals packing, and, consequently, on salt crystals-size distribution. This porosity is directly accessible to the liquid phase during salt melting. (b) The porosity within the “graphite dominant” layer, which depends on the graphite amount added as well as on the strength reached during
1587
compression. This porosity is rarely accessible to the salt when in liquid state. As shown in Fig. 1 (right), graphite/salt composites obtained by isostatic compression look like a random distribution of spheres (0.5e3 mm diameter) made of salt crystals within a continuous and homogeneous graphite-matrix at the macroscopic scale. Distance between salt particles ranges from some tenths of millimeter to some millimeters. As before, porosity within the salt spheres depends on salt crystals packing and the porosity within the graphite-matrix will not be accessible to the salt when in liquid state. We remind that for the elaboration of these composites ENG is first ground (w500 m mean size particles), then mixed with salt and compressed. As a consequence of that, the porosity of the resulting ENG matrix is expected to be less important than the porosity within the “graphite dominant” layers of the composites elaborated by uni-axial compression. Hence, a greater rigidity of the ENG matrix walls is also expected. For testing purposes, some graphite/salt composites have been elaborated. Graphite amount within those obtained by uni-axial compression ranges from 5 to 20%wt. As explained in ref. [9], the samples have been elaborated so that the total porosity is about 20e25% in all the cases. Samples apparent density ranges from 1500 kg/m3 to 1800 kg/m3. The samples of the graphite/salt composites obtained by isostatic cold-compression have been elaborated by SGL Technologies GmbH in the framework of the DISTOR project [6]. Graphite amount in all of them is about 20%wt. In all the cases, KNO3 and NaNO3 used within the KNO3/NaNO3 (50 mol%) eutectic system are 98% and 99.6% purity grade, respectively. We recall that the apparent thermal conductivity of the graphite/salt composites (uni-axial and isostatic compression routes) with 15e20%wt of graphite amount is close to 20 W/m/K [9]. 3. Phase transition properties Phase transition properties of the KNO3/NaNO3 (50 mol%), as well as those of the graphite/salt composites have been analyzed by Differential Scanning Calorimetry (Setaram DSC111). The testing protocol that has been used for estimating transition temperatures and latent heats is described in Section 3.1. Section 3.2 describes the results achieved when testing the salt, while next two sections are devoted to the presentation and the discussion of the results achieved when testing graphite/salt composites. 3.1. DSC testing protocol A small sample (w30 mg for salt, w100 mg for composites) of the material to be tested is placed in the DSC furnace and heated
Fig. 1. SEM images of graphite/salt composites. Right and center: composite obtained by uni-axial cold-compression. Right: composite elaborated by isostatic cold-compression.
1588
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
50 40
Compensation flux (mW)
30 20 10 0 -10 -20 -30 -40 -50 175
200
225
250
275
T (°C) Fig. 2. Measured compensation flux when heating (positive values) or cooling (negative values) the KNO3/NaNO3 sample at 5 C/min.
until it reaches 175 C. The sample is kept at this temperature during 10e15 min, and then it is heated again at constant temperature rate (5 C/min) up to 275 C. A new 10e15 min isothermal period follows heating. Then, the sample is cooled at constant temperature rate (5 C/min) up to 175 C. Cycles of melting/crystallization are repeated many times in order to test measurements repeatability as well as thermo chemical stability of the materials. The amount of energy absorbed or released (heat flux) by the sample as it is heated or cooled is measured by the DSC flux-meter. The energy related to sample phase transition (compensation flux in the following) is estimated by subtraction of the baseline from DSC flux measurements. The melting temperature is identified to the so called onset temperature on the compensation flux e temperature graphs, and the latent heat is determined from the scaled areas under the recorded signals. Calibration for heat flux and temperature was done with a standard tin metal at the same scan rates as in the experiments. The melting temperatures were reproducible to within 0.35 C, while latent heats were reproducible to within an accuracy of 2%.
3.2. Eutectic system KNO3/NaNO3 Twenty different samples of salt have been tested and cycled more than one hundred times. Fig. 2 is representative of compensation heat flux measured when heating/cooling samples at temperature rates of 5 C/min. Neither subcooling nor thermo chemical instability has been never observed. Estimated melting/ crystallization temperature and latent heat are 223 0.34 C and 104.5 1.6 J/g respectively. Such values are in very good agreement with those found in the literature [10e12]. Uncertainty on phase transition properties is calculated as the standard deviation of values recovered over the whole set of melting/crystallization cycles. 3.3. Materials obtained by uni-axial cold-compression As said before, different samples of graphite/salt composites have been elaborated by uni-axial cold-compression. Graphite amount within the samples ranges from 5 to 20%wt. Each sample has been cycled at least 10 times in the DSC. Fig. 3 shows the
50 40
Comp ens atio n fl ux(m W)
30 20 10 0 -10 -20 -30 -40 -50 175
200
225
250
275
T(°C) Fig. 3. Measured compensation flux when heating (positive values) or cooling (negative values) a sample with 15% wt of graphite amount (10 cycles at 5 C/min).
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
compensation heat flux measured when heating/cooling (5 C/min) one of the samples with 15%wt of graphite amount. Whatever may not be the graphite content, neither subcooling nor thermo chemical instability has been observed. Table 1 summarizes the results achieved in terms of onset temperatures and latent heats. Latent heat is reported to both the sample mass (LPCM in the table) and the mass of the salt within the composite (L in the table). As previously, uncertainty on phase transition properties is calculated as the standard deviation of values recovered over the whole set of melting/crystallization cycles carried out. Taking into account uncertainty values, it can be concluded that the graphite does not modify the phase transition properties (melting temperature and latent heat) of the salt.
compression route: a quite rigid and continuous conductive phase with spherical inclusions of salt and with not enough void space for salt volume expansion when melting. Hence, under melting, KNO3/ NaNO3 volume expansion (w10% [11]) will be constrained by the pore wall and the pressure within will thus increase. Main consequences of this over pressurization are a progressive augmentation of the salt melting temperature and a progressive reduction of its latent heat. To illustrate our purposes, the model proposed in refs. [13,14] for analysis of PCM melting within closed pores with homogeneous, isotropic and linear-elastic walls has been used. It can be demonstrated that the pressure P(t) within the pore at time t is a linear function of the fraction f*(t) of the initial volume of salt which has been already molten (molten salt fraction in the following):
3.4. Materials obtained by isostatic cold-compression
PðtÞ ¼ P0 þ Km
Fig. 4a shows a representative set of thermogramme coming from DSC tests carried out on isostatic cold-compressed materials (20%wt of graphite amount). They represent the heat flux supplied by the DSC (when heating, positive values) or delivered by the samples (when cooling, negative values) as a function of the prescribed temperature. Bold line corresponds to the first melting/crystallization cycle. Fig. 4b includes values of the energy amount supplied to the sample (circles) or delivered by it (squares) during successive cycles of melting/crystallization. Energy values are always reported to the mass of salt within the sample. First of all, it can be seen that one cycle is required before stabilization of the sample behavior. It is also observed that main features of the first melting/crystallization cycle are clearly different of those of next cycles: (a) the energy supplied by the DSC during first melting (Fig. 4b) is significantly higher than the energy required for melting the KNO3/NaNO3 (105 kJ/kg) within the sample at atmospheric pressure; (b) from time to time, the energy delivered by the sample during first crystallization is lower than the energy supplied by the DSC during the previous melting, and it could be lower than the energy required for complete crystallization of the amount KNO3/NaNO3 within the sample; (c) first melting takes place within a wide range of temperature. As shown in Fig. 4a, the classical peak that characterizes melting/ crystallization of eutectic substances in thermogramme looks now like a plateau. Similar melting/crystallization features are observed for tests carried out at 1 C/min heating/cooling rates instead of at 5 C/min.
E f* ¼
Zf * Vs0
rs rl Pdf þ rl
0 |fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl} mechanical
energy
Zf *
Z
T f*
rs Vs0 Dhf df þ
0 |fflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflffl} latent
heat
1589
rl rs rs
Vso * f ðtÞ Vmo
where P0 represents the initial pore pressure (atmospheric one), Km is the rigidity modulus of the pore wall, and Vm0 and Vs0 are respectively the initial volume of the pore wall (before wall deformation) and the initial volume of salt within the pore. Fig. 5 shows the variation of P with regard to f* corresponding to a pore with Km ¼ 8 109 Pa, (rsrl)/rl ¼ 4.6 102 and Vs0/Vm0 ¼ 1.37 (representative values for the materials we are studying [13]). In the same figure is represented the melting temperature evolution as well as the variation of the latent heat as predicted by the thermodynamical model proposed in ref. [14]. It can be seen that as a consequence of salt volume expansion pressure and melting point will be progressively increased. For melting progress, the pore have to be heated up to a melting point which is continuously going up and, consequently, melting will take place in a wide range of temperature instead of at constant temperature. Moreover, for getting complete melting in the example, the salt must be heated up to 300.9 C; otherwise, part of the salt within the pore remains in solid state. In such a case, the energy delivered by the sample during the first crystallization will be lower than expected because the preceding melting has not been completed. As show in Fig. 5, latent heat progressively decreases as melting goes on. This means that even for complete melting situations, a loss of latent heat storage capacity will be observed. On another hand, total energy stored within the pore during melting can be written as:
Vs0 rl cpl f þ rs cps ð1 f Þ dT
Tðf ¼ 0Þ
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} sensible
heat
where
Zf * Zf * rs rl * rs Vs0 Dhf df þ E f Vs0 Pdf þ ¼
rl
0 |fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl} mechanical
energy
0 |fflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflffl} latent
heat
Z
T f*
Vs0 rl cpl f þ rs cps ð1 f Þ dT
Tðf ¼ 0Þ
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} sensible
After stabilization the energy supplied/delivered by the salt within the samples is close to 105 kJ/kg (Fig. 4b), which correspond to the latent heat of KNO3/NaNO3 under normal pressure conditions. These observed behaviors are closely related to the microstructure of the graphite/salt composites obtained by the isostatic
heat
First term represents the energy used for wall pore deformation, the second one represents the energy required for melting a volume fraction f* of salt, and the third one is the term of sensible heat accumulation. Such three terms are represented in Fig. 6. One notices that mechanical energy is negligible compared to the total
1590
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
Table 1 Results from DSC tests: Tonset ¼ melting/crystallization temperature; L ¼ latent heat reported to the mass of salt in the sample; LPCM ¼ latent heat reported to the sample mass. Material
NaNO3/KNO3 5.4% ENG 8.0% ENG 11.1% ENG 15.2% ENG 20.0% ENG
Melting
Crystallization
Tonset ( C)
L (J/g)
LMCP (J/g)
Tonset ( C)
L (J/g)
LMCP (J/g)
220.6 0.3 220.9 0.2 221.3 0.6 221.3 0.5 221.1 0.2 220.2 0.9
104.3 1.9 103.2 0.5 97.5 1.5 96.9 0.7 92.3 0.3 86.8 0.7
104.3 1.9 106.0 0.5 103.2 1.7 105.2 0.8 103.7 0.4 102.8 0.8
223.0 0.34 223.3 1.68 222.8 2.13 223.2 2.18 222.9 0.3 223.1 0.3
104.6 1.5 102.2 0.6 97.0 1.6 95.8 1.1 91.8 0.4 86.4 0.6
104.6 1.5 104.99 0.58 102.6 1.9 104.0 1.2 103.1 0.5 102.3 0.8
stored energy. On the contrary, sensible energy is 30e40% of the total energy instead of zero as for eutectics melting at constant pressure. It represents the energy required for heating salt in order to follows a continuously going up melting point. That explains why the energy supplied by DSC during first heating of the graphite/salt materials is significantly higher than the energy required for KNO3/ NaNO3 melting at atmospheric pressure (105 kJ/kg). The behavior of the graphite/salt materials elaborated by isostatic compression during their first melting represents a serious drawback for using them in energy storage applications. The reason is that at the storage system scale significant salt leakage is expected due the in-pores pressurization.
4. Application The interest of the developed materials (mainly those obtained by uni-axial compression) is highlighted here through an example of application. It is shown that they could allow significant reduction of the storage system investment cost. A shell-and-tube storage system is considered. It consists of one block made of graphite/salt materials with a two-phase heat transfer fluid (water/vapor) passing through a set of parallel tubes crossing the block (Fig. 7). For modeling purposes, identical conditions are assumed in all the tubes. The storage unit can thus be assumed to be composed of parallel storage elements. In order to
a 100
first cycle
Compensation flux (mW)
80
following cycles
60 40 20 0 -20 -40 -60 -80
-100 150
175
200
225
250
275
300
325
350
375
400
T (°C)
b 150
melting crystallisation
140
Energy (kJ/kg)
130 120 110 100 90 80 0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
cycles Fig. 4. DSC results for one of the samples elaborated by the isostatic cold-compression route: (a) Thermogramme and (b) Energy supplied/delivered in successive melting/crystallization cycles.
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
1591
Fig. 5. Pressure (P), latent heat (Dh) and melting point (Tf) evolution as a function of the volume fraction of salt that have been molten within a closed pore.
1,6
total energy senible heat latent heat mechanical energy
1,4
Energy (J)
1,2 1,0 0,8 0,6 0,4 0,2 0,0 0,0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1,0
f* Fig. 6. Evolution of the stored energy within the pore during melting.
reduce the computation time, only a single pipe with surrounding storage material should be calculated (Fig. 8). The energy stored by the system is then the product of the number of tubes by the energy stored in only one of these units. The charging/discharging
processes is assumed to be completed when the solid/liquid phase front between adjacent tubes touches each other. Let Q and tc be, respectively, the energy storage capacity and the charge/discharge time required for the application. Calculation of the mass of PCM (salt and graphite) and the total tubes length required to fit Q and tc constraints is performed as follows: e Knowing the internal tubes diameter Di ¼ 2ri, the distance between tubes is calculated as Do ¼ 2ro ¼ 2rm(t ¼ tc), where rm(t) can be approached by ref. [15]:
Fig. 7. Sketch of a shell-and-tube storage system.
Fig. 8. System element for heat transfer modeling.
1592
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
# " rm ðtÞ 2 rm ðtÞ 2 rm ðtÞ 2 2 ln ¼ 1 1 þ 4 s ri ri Bi ri
Table 2 Values of the PCM physical properties used in the analysis. Salt properties
Graphite properties
Density ¼ 2220 kg/m3 Density ¼ 2100 kg/m3 Specific heat ¼ 1822 J/kg/K Specific heat ¼ 712 J/kg/K Melting temperature ¼ 223 C Latent heat ¼ 106 kJ/kg Thermal conductivity values as a function of the ENG amount According to the measurements carried out in ref. [9] 0%wt ENG: 1.2 W/m/K; 5% wt ENG: 2 W/m/K; 10% wt ENG: 10.5 W/m/K; 15% wt ENG: 20 W/m/K;
c Tfluid Tmelting s ¼ Ste:Fo; Ste ¼ ; LPCM
%wt ENG
20
2
0 5 10 15 0 5 10 15
121.5 90.1 36.3 27.8 44.1 31.5 11.2 8.1
0.03 0.08 0.12 0.14 0.10 0.13 0.21 0.26
0 5 10 15 0 5 10 15
607.5 450.5 181.4 138.9 220.5 157.5 55.8 40.5
0.06 0.08 0.12 0.14 0.10 0.13 0.21 0.26
0 5 10 15 0 5 10 15
3037.5 2252.3 907.0 694.5 1102.7 787.2 278.8 202.7
0.06 0.08 0.12 0.14 0.10 0.13 0.21 0.26
8
100
2
8
500
2
8
Ltubes (m)
Do (m)
Charging time (h)
Msalt (kg)
e The total length of tubes Ltubes of the heat exchanger is thus calculated as Q ¼ p(r20r2i )Ltubes r LPCM. e The mass of salt (active phase in the PCM) required is Msalt ¼ Q/ LPCM, and the mass of graphite is given by Mgraphite ¼ ((x/100) Msalt)/(1 x/100), where x represents the %wt of graphite amount within the PCM.
MENG (kg)
679.25 679.25 679.25 679.25 679.25 679.25 679.25 679.25 3396.3 3396.3 3396.3 3396.3 3396.3 3396.3 3396.3 3396.3 16 16 16 16 16 16 16 16
981.1 981.1 981.1 981.1 981.1 981.1 981.1 981.1
0 35.8 75.5 119.9 0 35.8 75.5 119.9 0 178.8 377.4 599.3 0 178.8 377.4 599.3 0 893.7 1886.8 2996.7 0 893.7 1886.8 2996.7
lt h ri ; Bi ¼ l r c ri2
where ri is the internal tube radius and h is the heat transfer coefficient between the tube wall and the fluid. LPCM represents the latent heat of the PCM, r and c are its density and its specific heat respectively, and l represents its thermal conductivity.
Table 3 Results from pre-dimensioning of the storage systems. Internal tubes diameter has been chosen equal to 20 mm. The heat transfer fluid temperature is assumed to be 10 C higher than the melting temperature of the salt. Convective heat transfer coefficient between the tubes wall and the fluid is 1000 W/m/K. Storage capacity (kWh)
Fo ¼
The part of the investment cost related to salt, graphite at tubes is:
C ¼ Msalt psalt þ Mgraphite pgraphite þ Ltubes ptubes where psalt, pgraphite and ptubes represent unitary prices. The PCM properties used in the analysis are given in Table 2 and results achieved are included in Table 3. It can be noticed that using graphite/salt composites instead of salt leads to significant reduction of the total tubes length. That is particularly interesting because the price of steel tubes represents a major part of the total cost of the storage system. Fig. 9 shows estimated investment cost reduction resulting from graphite/salt materials utilization. As expected, reduction cost is independent of the heat storage capacity; while it is strongly dependent on power requirements. As charging/discharging time decreases, the interest of using graphite/ salt composites increase. In all the cases, an optimal composition of the materials appears at 10%wt of graphite amount. It results from a balance between the cost of the graphite amount and the cost of the steel required for fitting to the design constraints.
60 Q = 20 kWh tc = 2 h
Q = 20 kWh tc = 8 h
Q = 100 kWh tc = 2 h
Q = 100 kWh tc = 8 h
Q = 500 kWh tc = 2 h
Q = 500 kWh tc = 8 h
Cost reduction (%)
50
40
30
20
10
0
0
5
10 15
0
5
10 15
0
5
10 15
0
5
10 15
0
5
10 15
0
5
10 15
%wt ENG Fig. 9. Investment cost reduction induced by thermal conductivity enhancement. Prices assumptions: 0.6 V/kg for the salt, 6 V/kg for ENG and 25 V/m for steel 316 L.
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
5. Conclusion New composites graphite/salt for thermal energy storage at high temperature (w200 C) have been developed and tested. They are elaborated by cold-compression of a physical mixing of expanded natural graphite particles and salt powder (eutectic KNO3/NaNO3). Two different compression routes have been investigated: uni-axial compression and isostatic compression. First one leads to a layered arrangement of the graphite within the composite material, while the second leads to a more or less homogeneous matrix including salt spheres. The thermal analyses carried out on in the first part of the paper show that both uni-axial and isostatic cold-compression are simple and equally efficient techniques for improving the salt thermal conductivity. For instance, graphite amounts between 15 and 20% wt lead to apparent thermal conductivities close to 20 W/m/K (20 times greater than the thermal conductivity of the salt). Furthermore, some advantages as low cost and safety come from materials elaboration carried out at room temperature. Phase transition properties of graphite/salt composites (melting/crystallization temperature and latent heat) have been studied by differential scanning calorimeter. Whatever may not be the graphite content, neither subcooling nor thermo chemical instability is observed. Moreover, it has been shown that graphite does not modify the phase transition properties (transition temperatures and latent heat) of KNO3/NaNO3. Composites obtained by uni-axial compression work correctly when cycling them. However, some problems have been identified when cycling composites obtained by isostatic compression. They are closely related to the microstructure of these materials: a quite rigid and continuous conductive phase with spherical inclusions of salt and with not enough void space for salt volume expansion. During melting (at least the first one), salt volume expansion will be constrained by graphite-matrix and pressure in-pores will thus increase. Main consequences of this over pressurization are a progressive augmentation of the salt melting temperature and a progressive reduction of its latent heat. For melting progress, materials have to be heated up to a melting point which is continuously going up. Hence, melting looks as spread melting. Moreover, a significant part of the energy supplied to the material will be used for heating it (sensible heat instead of latent heat). These behaviors represent a serious drawback for using graphite/ salt materials elaborated by isostatic compression in energy storage application. The reason is that at the storage system scale significant salt leakage is expected due to the in-pores pressurization. The interest of graphite/salt composites (uni-axial compression route) for high temperature thermal energy storage purposes has been illustrated through an example of application. Pre-dimensioning of shell-and-tube systems with storage capacity ranging
1593
20e500 kWh and charging/discharging times of 2 and 8 h, show that using graphite/salt materials instead of salt leads to significant reduction of the total tubes length required for fitting storage capacity and power constraints. As a consequence of that, significant reduction of the storage system investment cost is achieved (20e45%).
Acknowledgments The authors acknowledge the financial support of the CNRS Energy Program for subsidizing HTP-STOCK French Project, as well as the financial support of the European Commission for subsidizing DISTOR Project (6th framework program for research).
References [1] R.J. Wood, D.T. Baldwin, P.W. Ócallaghan, S.D Probert, Energy Cascading Combined with Thermal Energy Storage in Industry, Dordrecht, Boston, Lancaster, 1983, pp. 223e239. [2] F. Dinter, M. Geyer, R. Tamme, C. Richter, Thermal Energy Storage for Commercial Application (TESCA), a Feasibility Study on Economic Storage Systems. Springer Verlag, Berlin, 1991. [3] Enermodal Engineering LTD, cost reduction study for solar thermal power plants, final report, Prepared by Enermodal Engineering Limited with Marbek Resource Consultants Ltd., by contract of World Bank/GEF, Washington D.C., 1999. [4] Energy cascading combined with thermal energy storage in industry. In: sixth framework programme for research, European Commission, 2006. [5] R. Pitz-Paal, J. Dersch, B. Milow, ECOSTAR Roadmap Document. DLR Germany, SES6-CT-2003e502578, 2005. [6] R. Tammer, DISTOR: Energy Storage for Direct Steam Solar Power Plants. Final activity report, SES6-CT-2004e503526, 2008. [7] U. Hermann, B. Kelly, H. Price, Two-tank molten salt storage for parabolic trough solar power plants. Energy 29 (5e6) (2004) 883e893. [8] Z. Acem. Nouveaux composites graphite/sel destinés au stockage de l’énergie thermique à haute température: de l’élaboration au développement de méthodes de caractérisation thermique de matériaux conducteurs orthotropes. Université Bordeaux 1, PhD., 2007. [9] Z. Acem, J. Lopez, E. Palomo, KNO3/NaNO3-graphite composites for thermal energy storage at high temperature: Part I e Elaboration methods and thermal properties. Appl. Therm. Eng 30 (13) (2010) 1580e1585. [10] G.J. Janz, Molten Salts Handbook. Academic Press, New York, 1967. [11] M. Kamimoto, Thermodynamic properties of 50 mole% NaNO3 e 50% KNO3 (HTS2). Thermochimica Acta 49 (2e3) (1981) 319e331. [12] C.M. Kramer, Z.A. Munir, J.V. Volponi, Differential scanning calorimetry of sodium and potassium nitrates and nitrites. Thermochimica Acta 55 (1) (1982) 11e17. [13] L. Lopez, Nouveaux matériaux graphite/sel pour le stockage d’énergie thermique à haute température. Etude des propriétés de changement de phase. Université Bordeaux 1, PhD., 2007. [14] L. Lopez, G. Caceres, E. Palomo, W. Jomaa, Confined melting in deformable porous media: A first attempt to explain the graphite/salt composites behaviour. International Journal of Heat and Mass Transfer 53 (5-6) (2010) 1195e1207. [15] V. Morisson, M. Radi, E. Palomo del Barrio, E. Arquis, Thermal energy storage systems for electricity production using solar energy direct steam generation technology. Chemical Engineering and Processing 47 (3) (2008) 499e507.
Contents lists available at ScienceDirect
Applied Thermal Engineering journal homepage: www.elsevier.com/locate/apthermeng
KNO3/NaNO3 e Graphite materials for thermal energy storage at high temperature: Part II. e Phase transition properties Jérôme Lopez*, Zoubir Acem, Elena Palomo Del Barrio Laboratoire TREFLE UMR CNRS 8508, Esplanade des Arts et Métiers, 33405 Talence Cedex, France
a r t i c l e i n f o
a b s t r a c t
Article history: Received 1 February 2010 Accepted 4 March 2010 Available online 12 March 2010
Composites graphite/salt for thermal energy storage at high temperature (w200 C) have been developed and tested. As at low temperature in the past, graphite has been used to enhance the thermal conductivity of the eutectic system KNO3/NaNO3. A new elaboration method has been proposed as an alternative to graphite-foams infiltration. It consists of compression at room temperature of a physical mixing of expanded natural graphite particles and salt powder. Two different compression routes have been investigated: uni-axial compression and isostatic compression. The first part of the paper shows that both uni-axial and isostatic cold-compression are simple and equally efficient techniques for improving the salt thermal conductivity. The second part of the paper is focused on the analysis of their phase transition properties. It is shown that graphite does not degrade the salt within the composites; that is, no changes are observed neither in the salts transition temperatures nor in its latent heat. On the contrary, some negative effects as pores over pressurization and salt leakage can appear if no void space enough is available within the composite for salt volume expansion when melting. Such negative effects are only observed in the composites obtained by isostatic cold-compression. Ó 2010 Elsevier Ltd. All rights reserved.
Keywords: Phase change materials Thermal properties Thermodynamic analysis Exfoliated graphite
1. Introduction This paper addresses the development and the analysis of new low cost materials with improved thermophysical properties for thermal energy storage at high temperature (w200 C). Important applications using steam as working fluid can be found at temperature close to 200 C in the industry (food, textile, manufacturing, etc.) [1e4] and in solar power generation sector [5,6]. Efficient storage systems for such applications usually demand transfer of energy during the charging/discharging process at constant temperature. This is the reason why our attention has been focused on phase change materials (PCM) as storage media. Taking into account cost constraints, salts have been preferred to metals. However, salts are characterized by low thermal conductivities (w1 W/m/K) that reduce heat exchanges rates during melting/crystallization. Thus, graphite has been used to enhance their thermal conductivity. The eutectic system KNO3/NaNO3 (NaNO3 50 mol%) has been selected as PCM. It has been successfully used in the past for high temperature energy storage purposes, mainly in applications concerning electricity generation by solar concentration technologies
* Corresponding author. Fax: þ33 556 845 436. E-mail address: [email protected] (J. Lopez). 1359-4311/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.applthermaleng.2010.03.014
[7]. The melting temperature and the latent heat of this eutectic system are 223 C and 106 J/g respectively. Moreover, KNO3/NaNO3 has other desirable characteristics such as negligible undercooling, chemical stability, no phase segregation, low corrosion and hygroscopicity, as well as commercial availability at low cost. As said before, graphite has been used to enhance the thermal conductivity of the KNO3/NaNO3 binary system. It is well-known that graphite has strong resistance to corrosion and chemical attack, which make it compatible with most PCM. Moreover, thermal conductivity of graphite particles is considerably high and their density is lower than the metals one. As shown in ref. [8], using graphite additives (dispersion of graphite particles in molten salt) leads to insignificant thermal conductivity improvement in the overall application. From another hand, graphite-foams infiltration with molten salt leads to highly conductive materials but with very low energy density because low saturation degree of salt is achieved [6]. Hence, a new elaboration method has been proposed as an alternative to foams infiltration. It consists of coldcompression (at room temperature) of a physical mixing of expanded natural graphite particles and salt powder. Two different compression routes have been investigated: uni-axial compression and isostatic compression. The first part of the paper [9] has been devoted to the detailed description of the resulting composites as well as to the analysis of their thermal properties. It has been shown that both uni-axial and isostatic cold-compression are
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
simple and equally efficient techniques for improving the salt thermal conductivity. For instance, graphite amounts between 15 and 20%wt lead to apparent thermal conductivities close to 20 W/ m/K (20 times greater than the thermal conductivity of the salt). Furthermore, some advantages as low cost and safety come from materials elaboration carried out at room temperature. This second part of the paper is devoted to the analysis of the phase change properties of the proposed graphite/salt composites. In Section 2, the main characteristics of the microstructure of these composites are highlighted. It includes also a short description of the series of composites that have been elaborated for testing purposes. Section 3 is focused in the presentation and the discussion of the results coming from DSC (Differential Scanning Calorimetry) testing. Last section illustrates the interest of these new materials in terms of storage systems investment cost reduction.
2. Graphite/salt composites As said before, the graphite/salt composites we are studying are obtained by cold-compression (at room temperature) of a physical mixing of expanded natural graphite particles and KNO3/NaNO3 (50 mol%) crystals. Two different compression routes have been investigated: uni-axial cold-compression and isostatic compression (see ref. [9] for details). We remind that expanded natural graphite (ENG) is a well-known material produced from exfoliation of natural graphite particles (i.e. disk-like platelets, usually 0.5 mm diameter and 0.05 mm thickness). Exfoliation leads to worm-like particles whose length is up to 300 times the thickness of the initial platelets, while their diameter is unchanged. The worms are characterized by an irregular honeycomb-like microstructure, a very low density (w3 kg/m3), a very high porosity (>99%) and a huge specific surface (>40 m2/g). The uni-axial compression route leads to a cohesive matrix made of graphite whose porosity is partially occupied by salt grains. At the macroscopic scale, the resulting composites show a parallel layered structure with alternate “graphite dominant” and “salt dominant” layers. Moreover, uni-axial compressing obviously induces some preferential orientation of ENG particles in such a way they tend to lie within a plane perpendicular to the pressing stress. Consequently, “graphite dominant” layers show a parallel layered structure too (Fig. 1, left). Within the “salt dominant” layers, one observes a “random” mixture of salt grains and more or less deformed ENG particles (Fig. 1, center). As for porosity, we can roughly distinguish: (a) The porosity within the “salt dominant” layer, which mainly depends on salt crystals packing, and, consequently, on salt crystals-size distribution. This porosity is directly accessible to the liquid phase during salt melting. (b) The porosity within the “graphite dominant” layer, which depends on the graphite amount added as well as on the strength reached during
1587
compression. This porosity is rarely accessible to the salt when in liquid state. As shown in Fig. 1 (right), graphite/salt composites obtained by isostatic compression look like a random distribution of spheres (0.5e3 mm diameter) made of salt crystals within a continuous and homogeneous graphite-matrix at the macroscopic scale. Distance between salt particles ranges from some tenths of millimeter to some millimeters. As before, porosity within the salt spheres depends on salt crystals packing and the porosity within the graphite-matrix will not be accessible to the salt when in liquid state. We remind that for the elaboration of these composites ENG is first ground (w500 m mean size particles), then mixed with salt and compressed. As a consequence of that, the porosity of the resulting ENG matrix is expected to be less important than the porosity within the “graphite dominant” layers of the composites elaborated by uni-axial compression. Hence, a greater rigidity of the ENG matrix walls is also expected. For testing purposes, some graphite/salt composites have been elaborated. Graphite amount within those obtained by uni-axial compression ranges from 5 to 20%wt. As explained in ref. [9], the samples have been elaborated so that the total porosity is about 20e25% in all the cases. Samples apparent density ranges from 1500 kg/m3 to 1800 kg/m3. The samples of the graphite/salt composites obtained by isostatic cold-compression have been elaborated by SGL Technologies GmbH in the framework of the DISTOR project [6]. Graphite amount in all of them is about 20%wt. In all the cases, KNO3 and NaNO3 used within the KNO3/NaNO3 (50 mol%) eutectic system are 98% and 99.6% purity grade, respectively. We recall that the apparent thermal conductivity of the graphite/salt composites (uni-axial and isostatic compression routes) with 15e20%wt of graphite amount is close to 20 W/m/K [9]. 3. Phase transition properties Phase transition properties of the KNO3/NaNO3 (50 mol%), as well as those of the graphite/salt composites have been analyzed by Differential Scanning Calorimetry (Setaram DSC111). The testing protocol that has been used for estimating transition temperatures and latent heats is described in Section 3.1. Section 3.2 describes the results achieved when testing the salt, while next two sections are devoted to the presentation and the discussion of the results achieved when testing graphite/salt composites. 3.1. DSC testing protocol A small sample (w30 mg for salt, w100 mg for composites) of the material to be tested is placed in the DSC furnace and heated
Fig. 1. SEM images of graphite/salt composites. Right and center: composite obtained by uni-axial cold-compression. Right: composite elaborated by isostatic cold-compression.
1588
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
50 40
Compensation flux (mW)
30 20 10 0 -10 -20 -30 -40 -50 175
200
225
250
275
T (°C) Fig. 2. Measured compensation flux when heating (positive values) or cooling (negative values) the KNO3/NaNO3 sample at 5 C/min.
until it reaches 175 C. The sample is kept at this temperature during 10e15 min, and then it is heated again at constant temperature rate (5 C/min) up to 275 C. A new 10e15 min isothermal period follows heating. Then, the sample is cooled at constant temperature rate (5 C/min) up to 175 C. Cycles of melting/crystallization are repeated many times in order to test measurements repeatability as well as thermo chemical stability of the materials. The amount of energy absorbed or released (heat flux) by the sample as it is heated or cooled is measured by the DSC flux-meter. The energy related to sample phase transition (compensation flux in the following) is estimated by subtraction of the baseline from DSC flux measurements. The melting temperature is identified to the so called onset temperature on the compensation flux e temperature graphs, and the latent heat is determined from the scaled areas under the recorded signals. Calibration for heat flux and temperature was done with a standard tin metal at the same scan rates as in the experiments. The melting temperatures were reproducible to within 0.35 C, while latent heats were reproducible to within an accuracy of 2%.
3.2. Eutectic system KNO3/NaNO3 Twenty different samples of salt have been tested and cycled more than one hundred times. Fig. 2 is representative of compensation heat flux measured when heating/cooling samples at temperature rates of 5 C/min. Neither subcooling nor thermo chemical instability has been never observed. Estimated melting/ crystallization temperature and latent heat are 223 0.34 C and 104.5 1.6 J/g respectively. Such values are in very good agreement with those found in the literature [10e12]. Uncertainty on phase transition properties is calculated as the standard deviation of values recovered over the whole set of melting/crystallization cycles. 3.3. Materials obtained by uni-axial cold-compression As said before, different samples of graphite/salt composites have been elaborated by uni-axial cold-compression. Graphite amount within the samples ranges from 5 to 20%wt. Each sample has been cycled at least 10 times in the DSC. Fig. 3 shows the
50 40
Comp ens atio n fl ux(m W)
30 20 10 0 -10 -20 -30 -40 -50 175
200
225
250
275
T(°C) Fig. 3. Measured compensation flux when heating (positive values) or cooling (negative values) a sample with 15% wt of graphite amount (10 cycles at 5 C/min).
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
compensation heat flux measured when heating/cooling (5 C/min) one of the samples with 15%wt of graphite amount. Whatever may not be the graphite content, neither subcooling nor thermo chemical instability has been observed. Table 1 summarizes the results achieved in terms of onset temperatures and latent heats. Latent heat is reported to both the sample mass (LPCM in the table) and the mass of the salt within the composite (L in the table). As previously, uncertainty on phase transition properties is calculated as the standard deviation of values recovered over the whole set of melting/crystallization cycles carried out. Taking into account uncertainty values, it can be concluded that the graphite does not modify the phase transition properties (melting temperature and latent heat) of the salt.
compression route: a quite rigid and continuous conductive phase with spherical inclusions of salt and with not enough void space for salt volume expansion when melting. Hence, under melting, KNO3/ NaNO3 volume expansion (w10% [11]) will be constrained by the pore wall and the pressure within will thus increase. Main consequences of this over pressurization are a progressive augmentation of the salt melting temperature and a progressive reduction of its latent heat. To illustrate our purposes, the model proposed in refs. [13,14] for analysis of PCM melting within closed pores with homogeneous, isotropic and linear-elastic walls has been used. It can be demonstrated that the pressure P(t) within the pore at time t is a linear function of the fraction f*(t) of the initial volume of salt which has been already molten (molten salt fraction in the following):
3.4. Materials obtained by isostatic cold-compression
PðtÞ ¼ P0 þ Km
Fig. 4a shows a representative set of thermogramme coming from DSC tests carried out on isostatic cold-compressed materials (20%wt of graphite amount). They represent the heat flux supplied by the DSC (when heating, positive values) or delivered by the samples (when cooling, negative values) as a function of the prescribed temperature. Bold line corresponds to the first melting/crystallization cycle. Fig. 4b includes values of the energy amount supplied to the sample (circles) or delivered by it (squares) during successive cycles of melting/crystallization. Energy values are always reported to the mass of salt within the sample. First of all, it can be seen that one cycle is required before stabilization of the sample behavior. It is also observed that main features of the first melting/crystallization cycle are clearly different of those of next cycles: (a) the energy supplied by the DSC during first melting (Fig. 4b) is significantly higher than the energy required for melting the KNO3/NaNO3 (105 kJ/kg) within the sample at atmospheric pressure; (b) from time to time, the energy delivered by the sample during first crystallization is lower than the energy supplied by the DSC during the previous melting, and it could be lower than the energy required for complete crystallization of the amount KNO3/NaNO3 within the sample; (c) first melting takes place within a wide range of temperature. As shown in Fig. 4a, the classical peak that characterizes melting/ crystallization of eutectic substances in thermogramme looks now like a plateau. Similar melting/crystallization features are observed for tests carried out at 1 C/min heating/cooling rates instead of at 5 C/min.
E f* ¼
Zf * Vs0
rs rl Pdf þ rl
0 |fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl} mechanical
energy
Zf *
Z
T f*
rs Vs0 Dhf df þ
0 |fflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflffl} latent
heat
1589
rl rs rs
Vso * f ðtÞ Vmo
where P0 represents the initial pore pressure (atmospheric one), Km is the rigidity modulus of the pore wall, and Vm0 and Vs0 are respectively the initial volume of the pore wall (before wall deformation) and the initial volume of salt within the pore. Fig. 5 shows the variation of P with regard to f* corresponding to a pore with Km ¼ 8 109 Pa, (rsrl)/rl ¼ 4.6 102 and Vs0/Vm0 ¼ 1.37 (representative values for the materials we are studying [13]). In the same figure is represented the melting temperature evolution as well as the variation of the latent heat as predicted by the thermodynamical model proposed in ref. [14]. It can be seen that as a consequence of salt volume expansion pressure and melting point will be progressively increased. For melting progress, the pore have to be heated up to a melting point which is continuously going up and, consequently, melting will take place in a wide range of temperature instead of at constant temperature. Moreover, for getting complete melting in the example, the salt must be heated up to 300.9 C; otherwise, part of the salt within the pore remains in solid state. In such a case, the energy delivered by the sample during the first crystallization will be lower than expected because the preceding melting has not been completed. As show in Fig. 5, latent heat progressively decreases as melting goes on. This means that even for complete melting situations, a loss of latent heat storage capacity will be observed. On another hand, total energy stored within the pore during melting can be written as:
Vs0 rl cpl f þ rs cps ð1 f Þ dT
Tðf ¼ 0Þ
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} sensible
heat
where
Zf * Zf * rs rl * rs Vs0 Dhf df þ E f Vs0 Pdf þ ¼
rl
0 |fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl} mechanical
energy
0 |fflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflffl} latent
heat
Z
T f*
Vs0 rl cpl f þ rs cps ð1 f Þ dT
Tðf ¼ 0Þ
|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl} sensible
After stabilization the energy supplied/delivered by the salt within the samples is close to 105 kJ/kg (Fig. 4b), which correspond to the latent heat of KNO3/NaNO3 under normal pressure conditions. These observed behaviors are closely related to the microstructure of the graphite/salt composites obtained by the isostatic
heat
First term represents the energy used for wall pore deformation, the second one represents the energy required for melting a volume fraction f* of salt, and the third one is the term of sensible heat accumulation. Such three terms are represented in Fig. 6. One notices that mechanical energy is negligible compared to the total
1590
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
Table 1 Results from DSC tests: Tonset ¼ melting/crystallization temperature; L ¼ latent heat reported to the mass of salt in the sample; LPCM ¼ latent heat reported to the sample mass. Material
NaNO3/KNO3 5.4% ENG 8.0% ENG 11.1% ENG 15.2% ENG 20.0% ENG
Melting
Crystallization
Tonset ( C)
L (J/g)
LMCP (J/g)
Tonset ( C)
L (J/g)
LMCP (J/g)
220.6 0.3 220.9 0.2 221.3 0.6 221.3 0.5 221.1 0.2 220.2 0.9
104.3 1.9 103.2 0.5 97.5 1.5 96.9 0.7 92.3 0.3 86.8 0.7
104.3 1.9 106.0 0.5 103.2 1.7 105.2 0.8 103.7 0.4 102.8 0.8
223.0 0.34 223.3 1.68 222.8 2.13 223.2 2.18 222.9 0.3 223.1 0.3
104.6 1.5 102.2 0.6 97.0 1.6 95.8 1.1 91.8 0.4 86.4 0.6
104.6 1.5 104.99 0.58 102.6 1.9 104.0 1.2 103.1 0.5 102.3 0.8
stored energy. On the contrary, sensible energy is 30e40% of the total energy instead of zero as for eutectics melting at constant pressure. It represents the energy required for heating salt in order to follows a continuously going up melting point. That explains why the energy supplied by DSC during first heating of the graphite/salt materials is significantly higher than the energy required for KNO3/ NaNO3 melting at atmospheric pressure (105 kJ/kg). The behavior of the graphite/salt materials elaborated by isostatic compression during their first melting represents a serious drawback for using them in energy storage applications. The reason is that at the storage system scale significant salt leakage is expected due the in-pores pressurization.
4. Application The interest of the developed materials (mainly those obtained by uni-axial compression) is highlighted here through an example of application. It is shown that they could allow significant reduction of the storage system investment cost. A shell-and-tube storage system is considered. It consists of one block made of graphite/salt materials with a two-phase heat transfer fluid (water/vapor) passing through a set of parallel tubes crossing the block (Fig. 7). For modeling purposes, identical conditions are assumed in all the tubes. The storage unit can thus be assumed to be composed of parallel storage elements. In order to
a 100
first cycle
Compensation flux (mW)
80
following cycles
60 40 20 0 -20 -40 -60 -80
-100 150
175
200
225
250
275
300
325
350
375
400
T (°C)
b 150
melting crystallisation
140
Energy (kJ/kg)
130 120 110 100 90 80 0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
cycles Fig. 4. DSC results for one of the samples elaborated by the isostatic cold-compression route: (a) Thermogramme and (b) Energy supplied/delivered in successive melting/crystallization cycles.
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
1591
Fig. 5. Pressure (P), latent heat (Dh) and melting point (Tf) evolution as a function of the volume fraction of salt that have been molten within a closed pore.
1,6
total energy senible heat latent heat mechanical energy
1,4
Energy (J)
1,2 1,0 0,8 0,6 0,4 0,2 0,0 0,0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1,0
f* Fig. 6. Evolution of the stored energy within the pore during melting.
reduce the computation time, only a single pipe with surrounding storage material should be calculated (Fig. 8). The energy stored by the system is then the product of the number of tubes by the energy stored in only one of these units. The charging/discharging
processes is assumed to be completed when the solid/liquid phase front between adjacent tubes touches each other. Let Q and tc be, respectively, the energy storage capacity and the charge/discharge time required for the application. Calculation of the mass of PCM (salt and graphite) and the total tubes length required to fit Q and tc constraints is performed as follows: e Knowing the internal tubes diameter Di ¼ 2ri, the distance between tubes is calculated as Do ¼ 2ro ¼ 2rm(t ¼ tc), where rm(t) can be approached by ref. [15]:
Fig. 7. Sketch of a shell-and-tube storage system.
Fig. 8. System element for heat transfer modeling.
1592
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
# " rm ðtÞ 2 rm ðtÞ 2 rm ðtÞ 2 2 ln ¼ 1 1 þ 4 s ri ri Bi ri
Table 2 Values of the PCM physical properties used in the analysis. Salt properties
Graphite properties
Density ¼ 2220 kg/m3 Density ¼ 2100 kg/m3 Specific heat ¼ 1822 J/kg/K Specific heat ¼ 712 J/kg/K Melting temperature ¼ 223 C Latent heat ¼ 106 kJ/kg Thermal conductivity values as a function of the ENG amount According to the measurements carried out in ref. [9] 0%wt ENG: 1.2 W/m/K; 5% wt ENG: 2 W/m/K; 10% wt ENG: 10.5 W/m/K; 15% wt ENG: 20 W/m/K;
c Tfluid Tmelting s ¼ Ste:Fo; Ste ¼ ; LPCM
%wt ENG
20
2
0 5 10 15 0 5 10 15
121.5 90.1 36.3 27.8 44.1 31.5 11.2 8.1
0.03 0.08 0.12 0.14 0.10 0.13 0.21 0.26
0 5 10 15 0 5 10 15
607.5 450.5 181.4 138.9 220.5 157.5 55.8 40.5
0.06 0.08 0.12 0.14 0.10 0.13 0.21 0.26
0 5 10 15 0 5 10 15
3037.5 2252.3 907.0 694.5 1102.7 787.2 278.8 202.7
0.06 0.08 0.12 0.14 0.10 0.13 0.21 0.26
8
100
2
8
500
2
8
Ltubes (m)
Do (m)
Charging time (h)
Msalt (kg)
e The total length of tubes Ltubes of the heat exchanger is thus calculated as Q ¼ p(r20r2i )Ltubes r LPCM. e The mass of salt (active phase in the PCM) required is Msalt ¼ Q/ LPCM, and the mass of graphite is given by Mgraphite ¼ ((x/100) Msalt)/(1 x/100), where x represents the %wt of graphite amount within the PCM.
MENG (kg)
679.25 679.25 679.25 679.25 679.25 679.25 679.25 679.25 3396.3 3396.3 3396.3 3396.3 3396.3 3396.3 3396.3 3396.3 16 16 16 16 16 16 16 16
981.1 981.1 981.1 981.1 981.1 981.1 981.1 981.1
0 35.8 75.5 119.9 0 35.8 75.5 119.9 0 178.8 377.4 599.3 0 178.8 377.4 599.3 0 893.7 1886.8 2996.7 0 893.7 1886.8 2996.7
lt h ri ; Bi ¼ l r c ri2
where ri is the internal tube radius and h is the heat transfer coefficient between the tube wall and the fluid. LPCM represents the latent heat of the PCM, r and c are its density and its specific heat respectively, and l represents its thermal conductivity.
Table 3 Results from pre-dimensioning of the storage systems. Internal tubes diameter has been chosen equal to 20 mm. The heat transfer fluid temperature is assumed to be 10 C higher than the melting temperature of the salt. Convective heat transfer coefficient between the tubes wall and the fluid is 1000 W/m/K. Storage capacity (kWh)
Fo ¼
The part of the investment cost related to salt, graphite at tubes is:
C ¼ Msalt psalt þ Mgraphite pgraphite þ Ltubes ptubes where psalt, pgraphite and ptubes represent unitary prices. The PCM properties used in the analysis are given in Table 2 and results achieved are included in Table 3. It can be noticed that using graphite/salt composites instead of salt leads to significant reduction of the total tubes length. That is particularly interesting because the price of steel tubes represents a major part of the total cost of the storage system. Fig. 9 shows estimated investment cost reduction resulting from graphite/salt materials utilization. As expected, reduction cost is independent of the heat storage capacity; while it is strongly dependent on power requirements. As charging/discharging time decreases, the interest of using graphite/ salt composites increase. In all the cases, an optimal composition of the materials appears at 10%wt of graphite amount. It results from a balance between the cost of the graphite amount and the cost of the steel required for fitting to the design constraints.
60 Q = 20 kWh tc = 2 h
Q = 20 kWh tc = 8 h
Q = 100 kWh tc = 2 h
Q = 100 kWh tc = 8 h
Q = 500 kWh tc = 2 h
Q = 500 kWh tc = 8 h
Cost reduction (%)
50
40
30
20
10
0
0
5
10 15
0
5
10 15
0
5
10 15
0
5
10 15
0
5
10 15
0
5
10 15
%wt ENG Fig. 9. Investment cost reduction induced by thermal conductivity enhancement. Prices assumptions: 0.6 V/kg for the salt, 6 V/kg for ENG and 25 V/m for steel 316 L.
J. Lopez et al. / Applied Thermal Engineering 30 (2010) 1586e1593
5. Conclusion New composites graphite/salt for thermal energy storage at high temperature (w200 C) have been developed and tested. They are elaborated by cold-compression of a physical mixing of expanded natural graphite particles and salt powder (eutectic KNO3/NaNO3). Two different compression routes have been investigated: uni-axial compression and isostatic compression. First one leads to a layered arrangement of the graphite within the composite material, while the second leads to a more or less homogeneous matrix including salt spheres. The thermal analyses carried out on in the first part of the paper show that both uni-axial and isostatic cold-compression are simple and equally efficient techniques for improving the salt thermal conductivity. For instance, graphite amounts between 15 and 20% wt lead to apparent thermal conductivities close to 20 W/m/K (20 times greater than the thermal conductivity of the salt). Furthermore, some advantages as low cost and safety come from materials elaboration carried out at room temperature. Phase transition properties of graphite/salt composites (melting/crystallization temperature and latent heat) have been studied by differential scanning calorimeter. Whatever may not be the graphite content, neither subcooling nor thermo chemical instability is observed. Moreover, it has been shown that graphite does not modify the phase transition properties (transition temperatures and latent heat) of KNO3/NaNO3. Composites obtained by uni-axial compression work correctly when cycling them. However, some problems have been identified when cycling composites obtained by isostatic compression. They are closely related to the microstructure of these materials: a quite rigid and continuous conductive phase with spherical inclusions of salt and with not enough void space for salt volume expansion. During melting (at least the first one), salt volume expansion will be constrained by graphite-matrix and pressure in-pores will thus increase. Main consequences of this over pressurization are a progressive augmentation of the salt melting temperature and a progressive reduction of its latent heat. For melting progress, materials have to be heated up to a melting point which is continuously going up. Hence, melting looks as spread melting. Moreover, a significant part of the energy supplied to the material will be used for heating it (sensible heat instead of latent heat). These behaviors represent a serious drawback for using graphite/ salt materials elaborated by isostatic compression in energy storage application. The reason is that at the storage system scale significant salt leakage is expected due to the in-pores pressurization. The interest of graphite/salt composites (uni-axial compression route) for high temperature thermal energy storage purposes has been illustrated through an example of application. Pre-dimensioning of shell-and-tube systems with storage capacity ranging
1593
20e500 kWh and charging/discharging times of 2 and 8 h, show that using graphite/salt materials instead of salt leads to significant reduction of the total tubes length required for fitting storage capacity and power constraints. As a consequence of that, significant reduction of the storage system investment cost is achieved (20e45%).
Acknowledgments The authors acknowledge the financial support of the CNRS Energy Program for subsidizing HTP-STOCK French Project, as well as the financial support of the European Commission for subsidizing DISTOR Project (6th framework program for research).
References [1] R.J. Wood, D.T. Baldwin, P.W. Ócallaghan, S.D Probert, Energy Cascading Combined with Thermal Energy Storage in Industry, Dordrecht, Boston, Lancaster, 1983, pp. 223e239. [2] F. Dinter, M. Geyer, R. Tamme, C. Richter, Thermal Energy Storage for Commercial Application (TESCA), a Feasibility Study on Economic Storage Systems. Springer Verlag, Berlin, 1991. [3] Enermodal Engineering LTD, cost reduction study for solar thermal power plants, final report, Prepared by Enermodal Engineering Limited with Marbek Resource Consultants Ltd., by contract of World Bank/GEF, Washington D.C., 1999. [4] Energy cascading combined with thermal energy storage in industry. In: sixth framework programme for research, European Commission, 2006. [5] R. Pitz-Paal, J. Dersch, B. Milow, ECOSTAR Roadmap Document. DLR Germany, SES6-CT-2003e502578, 2005. [6] R. Tammer, DISTOR: Energy Storage for Direct Steam Solar Power Plants. Final activity report, SES6-CT-2004e503526, 2008. [7] U. Hermann, B. Kelly, H. Price, Two-tank molten salt storage for parabolic trough solar power plants. Energy 29 (5e6) (2004) 883e893. [8] Z. Acem. Nouveaux composites graphite/sel destinés au stockage de l’énergie thermique à haute température: de l’élaboration au développement de méthodes de caractérisation thermique de matériaux conducteurs orthotropes. Université Bordeaux 1, PhD., 2007. [9] Z. Acem, J. Lopez, E. Palomo, KNO3/NaNO3-graphite composites for thermal energy storage at high temperature: Part I e Elaboration methods and thermal properties. Appl. Therm. Eng 30 (13) (2010) 1580e1585. [10] G.J. Janz, Molten Salts Handbook. Academic Press, New York, 1967. [11] M. Kamimoto, Thermodynamic properties of 50 mole% NaNO3 e 50% KNO3 (HTS2). Thermochimica Acta 49 (2e3) (1981) 319e331. [12] C.M. Kramer, Z.A. Munir, J.V. Volponi, Differential scanning calorimetry of sodium and potassium nitrates and nitrites. Thermochimica Acta 55 (1) (1982) 11e17. [13] L. Lopez, Nouveaux matériaux graphite/sel pour le stockage d’énergie thermique à haute température. Etude des propriétés de changement de phase. Université Bordeaux 1, PhD., 2007. [14] L. Lopez, G. Caceres, E. Palomo, W. Jomaa, Confined melting in deformable porous media: A first attempt to explain the graphite/salt composites behaviour. International Journal of Heat and Mass Transfer 53 (5-6) (2010) 1195e1207. [15] V. Morisson, M. Radi, E. Palomo del Barrio, E. Arquis, Thermal energy storage systems for electricity production using solar energy direct steam generation technology. Chemical Engineering and Processing 47 (3) (2008) 499e507.