Imece2013-64947

Home , Enthalpy of vaporization, Specific energy

Proceedings of the ASME 2013 International Mechanical Engineering Congress and Exposition IMECE 2013 November 15-21, 2013, San Diego, CA, USA

IMECE2013-64947

ISOBARIC, ISOCHORIC AND SUPERCRITICAL THERMAL ENERGY STORAGE IN R134a

Benjamin I. Furst Adrienne S. Lavine Mechanical and Aerospace Engineering Dept. Mechanical and Aerospace Engineering Dept. University of California, Los Angeles University of California, Los Angeles Los Angeles, CA,USA Los Angeles, CA,USA

Reza Baghaei Lakeh Richard E. Wirz Mechanical and Aerospace Engineering Dept. Mechanical and Aerospace Engineering Dept. University of California, Los Angeles University of California, Los Angeles Los Angeles, CA,USA Los Angeles, CA,USA

ABSTRACT e enthalpy for CPTES or internal energy for CVTES The effective thermal energy density of R134a subjected to (kJ kg-1) an isobaric or isochoric process is determined and evaluated in the two-phase and supercritical regimes. The results are h enthalpy (kJ kg-1) qualitatively extended to other fluids via the principle of -1 corresponding states. It is shown that substantial increases in hfg enthalpy of vaporization (kJ kg ) volumetric energy density can be realized in the critical region -1 for isobaric processes. Also, for isobaric processes which utilize ufg,eff effective latent heat for an isochoric process (kJ kg ) the full enthalpy of vaporization at a given pressure, there exists a pressure at which the volumetric energy density is a T temperature (K) maximum. For isochoric processes (supercritical and two- phase), it is found that there is no appreciable increase in u internal energy (kJ kg-1) volumetric energy density over sensible liquid heat storage; the effective specific heat can be enhanced in the two-phase, ∆( ) signifies a change in quantity ( ) isochoric regime, but only with a significant reduction in volumetric energy density.  average fluid density (kg m-3)

NOMENCLATURE INTRODUCTION The goal of this paper is to explore and evaluate the CPTES constant pressure thermal energy storage Thermal Energy Storage (TES) potential of isobaric and isochoric processes, using R134a as an example fluid. The CVTES constant volume thermal energy storage focus is placed on the energy density that can be obtained in a TES system using these processes. The motivation for -1 -1 investigating isobaric and isochoric processes is twofold. cp specific heat at constant pressure (kJ kg K ) Firstly, by utilizing an isobaric or isochoric process, energy can -1 -1 be stored under the dome in the two-phase regime where the cv specific heat at constant volume (kJ kg K ) latent heat of vaporization is available. Secondly, these -1 -1 processes can both be used to store energy near the critical ceff effective specific heat (kJ kg K ) point, where large enhancements in specific heat have been -3 -1 measured [1]. Both of these regimes (under the dome and in the cvol,eff effective volumetric energy density (kJ m K )

1 Copyright © 2013 by ASME critical region) have been identified as having potentially high each. For example, in a CVTES system the pressure can energy densities [2], which may offset their disadvantages. increase substantially with temperature, while a CPTES system Isobaric and isochoric TES under the dome and in the would likely require some sort of robust, distensible volume. critical region appears to not have been thoroughly studied in The larger optimization problem has been investigated the literature. In TES reviews, the liquid-vapor phase change is elsewhere for the case of a high temperature (~400°C) CVTES mentioned only in passing (if at all) when discussing latent heat system with encouraging results [7]. Here, the focus is purely TES [3]. This is understandable considering that isobaric on the energy density potential of different thermodynamic processes require possibly impractical changes in volume—as a regimes. fluid changes from liquid to vapor at constant pressure its volume can change by several orders of magnitude. Additionally, the pressure in such a system could be high, depending on where the dome is crossed. In principle isochoric systems have been studied, but not in the same way proposed here. The case often studied in literature is the one where a liquid (e.g. water) is contained in a pressurized vessel in order to enable the liquid to store sensible heat at elevated temperatures without completely vaporizing [4]. In this paper a more general analysis is done to see if the energy density of an isochoric TES system can be increased by using the latent heat of vaporization. The TES characteristics of R134a were investigated using the highly accessible and accurate data from NIST REFPROP [5]. This program facilitated a systematic graphical and numerical evaluation of pertinent thermodynamic data. All Figure 1 – P-v diagram for R134a with isotherms graphs and thermodynamic values in this paper were derived [5]. from NIST REFPROP unless otherwise noted. R134a has a critical temperature, pressure and density of 101°C, 4.06 MPa and 512 kg/m3 respectively. RESULTS The evaluation of isobaric and isochoric TES with a There are a few subtleties involved in quantifying TES. particular fluid (R134a) has been done with the idea that using Generally the quantities of interest are the energy stored per real properties for an actual fluid would elucidate the TES degree temperature change per unit mass or unit volume. potential of these thermodynamic regimes in general, via the Temperature change is not a pertinent parameter for a purely well-established principle of corresponding states, which latent heat process. When there is a change in temperature (∆T) somewhat unifies the thermodynamic behavior of all fluids [6]. involved, it is useful to use an effective specific heat value (ceff) The trends seen for this particular fluid would then qualitatively defined as the change in specific energy (∆e) that occurs over apply to other fluids as well. the change in temperature divided by that change in In a Constant Volume Thermal Energy Storage (CVTES) temperature: or isochoric system, a fixed mass of fluid would be loaded into a fixed volume (container). With the addition of heat, the ∆푒 푘퐽 푐 = [ ] temperature and pressure of the fluid would increase while the 푒푓푓 ∆푇 푘푔 퐾 average density would remain constant and would depend on the initial quantity of mass and container size. Such a system Note that ∆e is enthalpy for a CPTES and internal energy for would follow a vertical line on a P-v diagram, and fluid could CVTES. Another useful quantity is the effective volumetric exist as a liquid, vapor or two-phase mixture (Figure 1). To pass energy density, defined as: through the critical point the fluid needs to be loaded at its critical density. ∆푒 ∙ 휌푚푖푛 푘퐽 In a Constant Pressure Thermal Energy Storage (CPTES) 푐푣표푙,푒푓푓 = [ ] ∆푇 푚3 퐾 or isobaric system, a fixed mass of fluid would be loaded into a distensible volume. As heat was added to the fluid, the where 휌푚푖푛 is the minimum density that occurs in the process-- container would adjust its size such that the pressure would this is appropriate when dealing with CPTES systems where the remain constant. Perhaps the simplest version of such a system volume (and thus average density) changes, as it corresponds to would be the prototypical, vertical piston/cylinder configuration the largest volume of the system. where the pressure exerted on the fluid in the cylinder is due to These quantities facilitate convenient comparisons between the force acting on the piston (its weight and the surrounding TES in different regimes and other materials, but must be used pressure). The piston would be allowed to move and thus with care. The effective values are only valid for the specific accommodate the changes in volume of the fluid while conditions under which they were derived, namely the specified maintaining a constant pressure. Such a system would traverse change in energy and temperature. They should not be a horizontal line on the P-v diagram (Figure 1). extrapolated to use with another ∆T without great care. For The overall optimization and feasibility of both a CPTES example, in a CPTES system an extremely high effective and CVTES system are beyond the scope of this paper, specific heat can be calculated along the critical isobar if ∆T is although there are clearly important implementation issues for made small enough, however this value of ceff is only valid for 2 Copyright © 2013 by ASME the given ∆T and the given enthalpy change, and using a different ∆T with that ceff would not make sense. (See the section on supercritical CPTES). This is emphasized since it contrasts with the familiar case of the specific heat being relatively independent of temperature (e.g. for liquids and solids).

CPTES An overview of CPTES in R134a is given by the h-T diagram in Figure 2. This graph suggests that there are 3 main CPTES regions: 1) under the two-phase dome where a large change in enthalpy occurs at a fixed temperature (the enthalpy of vaporization); 2) outside the dome far from the critical region where enthalpy is approximately linearly dependent on temperature; and 3) in the critical region where enthalpy is non- Figure 3 – The enthalpy of vaporization, saturated vapor linearly related to temperature. From a TES perspective the density and volumetric energy density versus pressure for regions of interest are the critical region and under the dome, R134a. where there are relatively large changes in enthalpy over small (or no) changes in temperature. These two regions are discussed SUPERCRITICAL CPTES in more detail below. A CPTES system has the potential to exploit the large

increase in cp that occurs near the critical point. Figure 4 shows these spikes in cp for several pressures in the supercritical region. The critical temperature and pressure of R134a are 101°C and 4.06 MPa. The energy storage potential of this region depends on what pressure is selected (which spike), and the size and location of the ∆T at this pressure. It is clarifying to compare Figures 2 and 4: pressures at which the spike is more pronounced in Figure 4 correspond to the enthalpy versus temperature isobar being more vertical in the critical region in Figure 2. Clearly the greatest TES is obtained for a given spike when the ∆T is centered on the maximum. It is also clear from Figure 2 that for a given pressure, a larger ∆T yields a larger ∆h, and that the largest ∆h per ∆T occurs at the critical pressure.

Figure 2 – An enthalpy versus temperature diagram for R134a [5]. The two-phase dome and lines of constant pressure are shown.

SUBCRITICAL CPTES Below the critical pressure the most energetically interesting region is under the dome where large excursions in h exist for isothermal processes. This is the realm of the familiar latent heat of vaporization at constant pressure. The TES characteristics of this region are succinctly contained in the enthalpy of vaporization at a given pressure. These values are plotted in Figure 3. The enthalpy of vaporization monotonically decreases with increasing pressure. The saturated vapor density Figure 4 – The specific heat at constant pressure versus (ρmin in this case) is also included in Figure 3 for the corresponding pressure; it increases as the enthalpy of temperature for several isobars [5]. At the critical point (4.06 vaporization decreases. The product of density and enthalpy of MPa) this quantity dramatically increases; for pressures above and close to the critical point the spike diminishes. A vaporization is the volumetric energy density (Figure 3). There subcritical pressure (2 MPa) is included for reference--the is a maximum volumetric energy density at about 3.31 MPa. discontinuity occurs at the liquid/vapor phase change.

These observations are quantified in Figures 5 and 6. Figure 5 shows the effective specific heat and effective volumetric energy density at different supercritical pressures for a ∆T of 5 °C centered on the location of the cp maximum for each pressure. As expected ceff and cvol,eff increase as the pressure is reduced towards the critical pressure. Note that the ∆T range was chosen so that it captured most of the elevated cp

3 Copyright © 2013 by ASME along the critical isobar. This choice is somewhat arbitrary, but CVTES also unimportant since the same basic trend would be seen An overview of CVTES is provided by Figure 7. In this whatever the chosen ∆T was. Figure 6 shows the effective graph the internal energy is plotted against temperature for a specific heat and effective volumetric energy density as a characteristic sample of isochores. The dome (saturated liquid function of ∆T at the critical pressure. The ∆T is centered on the and vapor values) is also shown. Outside of the dome all of the maximum in each case. As expected, the effective specific heat isochores have approximately the same slope implying that the diverges as the ∆T decreases; in the limit of zero ∆T, ceff effective specific heat (and cv) in this region is practically approaches cp. Note that while ceff and cvol,eff increases independent of density and temperature. Inside the two-phase dramatically as ∆T is reduced, the total energy that can be dome there is a non-linear relation between u and T along the stored at this high ceff decreases. This trend can also be seen in isochores. In particular there are larger slopes under the dome, Figure 2. The values for cvol,eff were obtained by multiplying ceff indicating enhanced ceff. Furthermore, as the density gets by the density at the upper value of the temperature interval, lower, portions of the u versus T slope are greater, indicating which is the minimum density for the process. local regions of enhanced ceff. This improved effective specific heat under the dome should be expected since latent heat effects are present.

Figure 5 – The effective specific heat and effective volumetric Figure 7 – Internal energy versus temperature for lines of energy density at different supercritical pressures for a ∆T of 5 constant density [5]. The two-phase dome is also shown. The 3 °C centered on the cp maxima. critical density is 512 kg/m .

Note that there is no special TES phenomenon near the critical point. Unlike lines of constant pressure in an h versus T diagram, here the lines of constant density do not approach being vertical near the critical point. This is reinforced by looking at a c versus T diagram (Figure 8)—the increase in v specific heat at constant volume near the critical point is modest, increasing about 25% over the nearby saturated liquid values.

Figure 6 – The effective specific heat and effective volumetric energy density at the critical pressure as a function of decreasing ∆T. The associated enthalpy change available at this Figure 8 – Specific heat at constant volume versus temperature for lines of constant density [5]. Note that the ceff/cvol,eff is also included. The ∆T is centered on the location of increase near the critical point is very modest. the maximum in cp. As the effective specific heat increases, the available enthalpy decreases.

4 Copyright © 2013 by ASME Figure 7 establishes that under the dome is the most The volumetric effective specific isochoric latent heat is valuable region from the perspective of effective specific heat also of interest. This quantity can be determined by multiplying (best ∆푢/∆푇). In order to further explore the potential of this the effective specific isochoric latent heat by the associated region, the change in internal energy that occurs under the density. These effective volumetric densities are shown in dome was looked at for a range of densities. Note that an Figure 11; they monotonically increase with density. analog to the latent heat of vaporization does not have an obvious definition here, as it does in the case of constant pressure, since for an isochoric process the fluid can start out being two-phase (imagine a P-v diagram with lines of constant volume). However, Figure 7 suggests that all the densities do converge to the same value of internal energy at low temperatures (about -100°C). Then a metric for quantifying the latent heat changes in a constant volume process for R134a can be defined as the difference in internal energy between the reference temperature of -100°C and the point where the isochores leave the two-phase region. These effective isochoric latent heat values (ufg,eff) are shown in Figure 9 for a range of densities. This metric shows the same trend that is seen in

Figure 7: smaller densities under the dome have larger changes Figure 11 – The effective volumetric energy density for ufg,eff in internal energy per unit mass. and the optimal case (∆T of 5 °C) as a function of density.

It is important to keep in mind that the ∆T associated with each isochoric latent heat is different and that even though the ∆T may be larger for a given 휌 it may have a smaller ufg,eff than another 휌 with a smaller ∆T. This can be seen in Figure 7 where the lowest density clearly has the smallest ∆T and a relatively large ufg,eff. While the definition adopted for ufg,eff provides a convenient metric for CVTES under the dome, it does not provide a comprehensive picture. As in the case of CPTES near the critical point the values of ceff and cvol,eff are strongly dependent on the ∆T chosen. In particular if a ∆T is judiciously chosen, higher values of ceff can be realized. In Figure 7 it can Figure 9 – The effective isochoric latent heat (ufg,eff) for be seen that the highest values of ceff along a given isochore different densities. ufg,eff is defined as the change in internal occur just before the isochore intersects the saturation dome energy along an isochore between -100°C and the point (the slope is steepest). Then the best achievable (optimal) c where the isochore leaves the dome. eff and c for a given isochore occurs when the ∆T is chosen to vol,eff be just inside the dome. These values are shown in Figures 10 Unlike in a constant pressure process, the temperature and 11 using a ∆T of 5°C, where the upper value of the ∆T changes during an isochoric liquid/vapor phase change. This interval is the saturation point for the given isochore. In accord suggests that an effective specific heat can be defined for this with Figure 7, the values for c and c are larger for the process by dividing u by the ∆T that is associated with it. eff vol,eff fg,eff small ∆T just inside the dome at all densities shown except the This quantity is shown in Figure 10. As expected from Figure 7, highest, where the values converge with those calculated for lower densities have higher effective specific heats. ufg,eff. The hump-type feature that occurs for the case of a 5°C ∆T is due to the shape of the dome, and it is only visible for a small ∆T. The upper values provide an upper bound to the achievable ceff and cvol,eff in this domain

DISCUSSION TES can be divided into two classes: that involving a ∆T, and that not involving a ∆T (pure phase change process). For TES processes that involve a ∆T, the metrics of ceff and cvol,eff provide a convenient basis for comparison. In fact these parameters can be seen as figures of merit since they lump the important TES parameters together in the appropriate way: the ideal TES system has a large energy capacity (numerator) over a small ∆T and in a small volume/mass (denominator). Note that as far as practical TES is concerned, volumetric energy Figure 10 – The effective specific heat for ufg,eff and the optimal case (∆T of 5 °C) as a function of density. density (represented by cvol,eff) is probably the most important metric—it is generally of greatest interest to have a small

5 Copyright © 2013 by ASME volume with a large capacity. Another important aspect to keep Table 1—Summary of results in mind while evaluating TES is the ∆T, since ceff and cvol,eff generally depend on ∆T. c c ∆T ∆e ∆e*ρ Modality Region eff v ol,eff min Table 1 compares the TES potential of the thermodynamic (kJ/kg/K) (MJ/m^3/K) (K) (kJ/kg) (MJ/m^3) regions explored in the previous section for processes involving along critical a ∆T: CPTES in the critical region and CVTES in the two- CPTES 12.8 3.85 5 63.4 19.2 phase region. To facilitate the comparison, the largest values of isobar c and c available in each thermodynamic region are listed eff vol,eff along critical CPTES 65.8 25.2 0.5 32.9 12.6 for the case of a ∆T of 5°C; the quantity of energy stored in this isobar interval is also included. For CPTES, energy is equal to under dome -- CVTES enthalpy; for CVTES, energy is internal energy. Values for the lowest 3.8 0.38 5 18.8 1.9 (optimal) sensible heat TES of R134a and water at 20°C have been density included for reference. (Note that isobaric and isochoric under dome -- CVTES sensible heat changes in this interval differ by about 1%). highest 1.2 1.8 5 6.03 9.0 (optimal) The most prominent features of Table 1 are that the density sensible heat of water is hard to beat (which is well known), sensible saturated 1.4 1.7 5 7.1 8.5 and that CPTES in the critical region can perform very well. (R134a) liquid at 20 C For a ∆T of 5°C, R134a CPTES along the critical isobar can sensible saturated 4.2 4.2 5 20.9 20.8 have an effective specific heat about three times larger than (water) water at 20 C water and an effective volumetric energy density rivaling water

(only 8% smaller). These values depend on the ∆T, and as the ∆T is reduced, critical CPTES only improves; the trade-off is CONCLUSION A preliminary evaluation of isobaric and isochoric TES in that less and less energy can be stored at the high values of c eff R134a has been conducted. The TES potential of the liquid, and cvol,eff. The second entry in Table 1 (critical CPTES with a ∆T of 0.5°C) illustrates this. two-phase and supercritical regions were compared using An evaluation of CVTES is slightly more complex. In thermal energy density metrics. The investigated TES processes were divided into two classes: those involving a ∆T, and those order to evaluate the potential for TES in this region, values not involving a ∆T. Only CPTES in the two-phase regime fell from the “optimal” case were used, where the ∆T is taken to be under the latter category. In this regime it was found that there just inside the two-phase dome. These represent the best possible performance in this region (see previous section). As is an optimum pressure of about 3.31 MPa where the enthalpy stored per unit volume reaches a maximum of 18 MJ/m3. can be seen from Figures 15 and 16, ceff and cvol,eff have opposite trends: c decreases with increasing density while c For processes involving a ∆T, the TES capacity of different eff vol,eff regimes was compared to the TES capacity of saturated liquid increases with increasing density. To bound the TES potential R134a at 20°C with a 5°C ∆T. The metrics of comparison were of this region, values for the lowest and highest densities are the effective specific heat (c ) and the effective volumetric used in rows three and four of Table 1. As can be seen, for low eff energy density (cvol,eff) over a 5°C ∆T. The greatest values from density an appreciable increase in ceff over saturated liquid can be obtained (170% larger). However, this is accompanied by a each regime were selected for comparison. 79% reduction in the volumetric storage capacity (c ) It was found that a CVTES system was unable to achieve vol,eff substantial gains in c over the reference sensible TES liquid compared to sensible heat TES in R134a at 20°C. At the other vol,eff values. Values of c could be increased by 170% over the end of the spectrum (high density CVTES), the values of c eff eff sensible heat TES values, but only at the expense of a dramatic and cvol,eff approach the saturated liquid values at 20°C. This is expected from Figure 10, where the high density isochores reduction in volumetric energy density (down 79% from the practically follow the saturated liquid line. The c of sensible sensible heat TES reference). This implies that a CVTES vol,eff system is unable to practically improve upon the energy density TES in liquid cannot be substantially improved upon by a of sensible TES in saturated liquid. CVTES system. In the supercritical regime it was found that substantial Evaluating the latent enthalpy of vaporization is somewhat benefits over sensible liquid TES can be realized in a CPTES more elusive since it cannot be directly compared to sensible 3 modes of TES due to there being no ∆T involved. It would system. For a ∆T of 5°C, a cvol,eff of 19.2 MJ/m can be obtained probably be most appropriate to compare this mode of TES to (1.3 times larger than the sensible liquid TES reference); the effective specific heat in this region is 12.8 kJ kg-1 K-1 (8.1 other latent heat storage (melting/freezing); however data for times larger than the saturated liquid sensible TES value). the latent heat of fusion for R134a could not be found in the These values can be improved by reducing the ∆T; increasing literature. The most noteworthy aspect of this regime found here is the maximum in volumetric energy storage that occurs the ∆T diminishes the advantage. No significant advantages for a pressure of about 3.31 MPa. To compensate for the lack of were apparent for CVTES in the supercritical region. data for R134a, the optimal volumetric energy density of water While the quantitative conclusions above solely apply to R134a, the general trends observed can be expected to apply to and ammonia in isobaric processes were compared to their other fluids as well via thermodynamic similarity (the principle latent heat of fusion. It was found that the volumetric energy of corresponding states [6]). These more general trends are density of the latent heat of fusion is 3.2 times greater than the optimal isobaric process for water, and 5.2 times greater for summarized below: ammonia.

6 Copyright © 2013 by ASME 1) For a CVTES system operating in the two-phase [4] Medrano, M., et al. "State of the art on high-temperature regime, the volumetric energy density (cvol,eff) cannot thermal energy storage for power generation. Part 2—Case be substantially improved over values available in studies." Renewable and Sustainable Energy Reviews, 14.1 sensible liquid TES. (2010): 56-72. 2) A CVTES system in the two-phase regime can significantly increase the effective specific heat over [5] NIST Reference Fluid Thermodynamic and Transport sensible heat storage in liquid, however volumetric Properties—REFPROP, Version 9.0. Thermophysical energy density decreases substantially. Properties Division National Institute of Standards and 3) There is practically no TES benefit in the supercritical Technology. Boulder, Colorado 80305. region for a CVTES system. 4) A CPTES system operating in the critical region can [6] Leland, T. W., Chappelear, P. S. "The corresponding states substantially increase the volumetric energy density principle—a review of current theory and practice." Industrial and effective specific heat over sensible liquid TES & Engineering Chemistry 60.7 (1968): 15-43. values. The increase in ceff and cvol,eff decreases with increasing ∆T. [7] Tse, L. A., Ganapathi, G. B., Wirz, R. E., Lavine, A. S. 5) For a CPTES system exploiting the entire enthalpy of 2012. “System modeling for a supercritical thermal energy vaporization, there is a pressure at which the storage system”. Proceedings of the ASME 2012 6th volumetric energy density is a maximum. International Conference on Energy Sustainability & 10th Fuel Cell Science, Engineering and Technology Conference. It should be emphasized that the conclusions drawn here only pertain to energy density. Energy density is only one [8] Ganapathi, G. B., Berisford, D., Furst, B., Bame, D., aspect of a practical TES system, and other considerations Pauken, M., Wirz, R.”A 5 kWh Lab Scale Demonstration of a could be important such as pressure, cost and containment. For Novel Thermal Energy Storage Concept with Supercritical example, although a CVTES system cannot yield substantially Fluids”. ASME 7th International Conference on Energy higher energy densities than sensible liquid TES, it does have Sustainability, Jul 14-19, 2013, Minneapolis, MN, USA. other advantages. Given a specified ∆T, a CVTES system will always incur less of a pressure increase than a pure liquid sensible heat TES system traversing the same ∆T (compare lines of constant density on a P-v diagram inside the dome and in sub-cooled liquid). In fact CVTES systems are currently being explored [2, 7, 8], and have shown promise. Conversely, regions that have been shown here to have a very high energy density (e.g. the critical region) may not be practically accessible in many situations. The main value of this study is the trends highlighted above that are expected to hold for other fluids. These provide a guide of what relative energy densities are to be expected in a fluid undergoing an isobaric or isochoric processes.

ACKNOWLEDGEMENTS This effort was supported by ARPA-E Award DE-AR0000140 and Grant No. 5660021607 from the Southern California Gas Company.

REFERENCES [1] Sengers, J. M. H. Levelt. “Supercritical Fluids: Their Properties and Applications.” Supercritical Fluids: Fundamentals and Applications. Springer Science+Business Media, 2000.

[2] Ganapathi, G. B., Wirz, R. E. “High density Thermal Energy Storage with Supercritical Fluids,” ASME 6th International Conference on Energy Sustainability, Jul 23-26, 2012, San Diego, CA, USA

[3] Hasnain, S. M. "Review on sustainable thermal energy storage technologies, part I: heat storage materials and techniques." Energy Conversion and Management, 39.11 (1998): 1127-1138.