L. H. Alva S.

Department of Mechanical Engineering, University of Puerto Rico-Mayagez, Mayagez, Puerto Rico 00681-9045

Initial Analysis of PCM Integrated Solar Collectors

This paper investigates the technical feasibility of innovative solar collectors. The proposed collectors have a phase change material (PCM) integrated into them as the storage mechanism. The PCM-integrated solar collector eliminates the need of conventional storage tanks, thus reducing cost and space. The present work uses a parafn-graphite composite as the PCM to increase the conductivity of the PCM matrix. The parafns melting point is around 89 C, which is appropriate for use in single-effect absorption systems. The mathematical model that describes the thermal process in the PCM is presented and differs from the analysis of conventional at plate solar collectors making use of the lumped capacitance method which neglects spatial variations. The proposed model is calibrated favorably with a more detailed mathematical model that uses nite differences and considers temporal and spatial variations. Results for the collectors thermal performance are presented along with the effects of the composition of the PCM. The results for the PCM integrated collector proposed here are very encouraging. Therefore, there is an indication that conventional storage tanks may be replaced for the PCM integrated in the solar collector. DOI: 10.1115/1.2188532 Keywords: PCM, solar, thermal, collectors, storage, lumped, enthalpy

J. E. Gonzlez1
Department of Mechanical Engineering, Santa Clara University, CA 95053 e-mail: igonzalezcruz@scu.edu

N. Dukhan
Department of Mechanical Engineering, University of Puerto Rico-Mayagez, Mayagez, Puerto Rico 00681-9045

Introduction
The residential market in the Caribbean and other tropical locations represents a great opportunity for solar thermal applications for air conditioning systems. The typical cooling load in these buildings ranges between 10.5 and 35 kW. However, for solar driven absorption machines to be considered to satisfy cooling needs in the residential and commercial sectors in these regions, compact systems must be designed that eliminate the use of cooling towers and possibly the thermal storage tank. Recently we proposed a compact absorption machine that somehow addresses this need of downsizing solar absorption machines 1 . The system proposed in that work is an air-cooled closed absorption machine where the water cooling is replaced by air cooling through the use of compact fans, for both the condenser and the absorber. We concluded in that work that the penalty for using air cooling is minimum if extremely compact and efcient heat exchangers are used. The proposed machine was suggested for 10.5, 14.0, and 17.5 kW cooling loads. We now suggest in this work a mechanism to eliminate the use of storage tanks. We propose a at plate solar collector that incorporates a phase change material PCM as a mechanism for thermal storage. We suggest here an analysis for the proposed PCMsolar collector and investigate its thermal performance and technical feasibility. The proposed method of analysis is the lumped capacitance method. A computer program is used to simulate the thermal performance of the PCM under different working conditions.

al. 3 considered an arrangement of tubes over the PCM. Sokolov and Keizman 4 considered a PCM conned between two concentric tubes solar pipe .

The Parafn-Graphite Composite

The PCM considered is the composite proposed and reported by Py et al. 5 , which has a high thermal conductivity and a stable thermal capacitance. The composite is formed of two components: parafn as the PCM and compressed expanded natural graphite CENG . To form the composite, graphite powders are poured into a mold of aluminum and then pressed to obtain the porous graphite matrix with the desired bulk density 5 . Then the matrix is soaked into melted parafn and regularly weighted until maximum load is reached. In terms of capacity, and depending on bulk graphite density, the CENG/parafn composites present a weight percentage ranging from 65% to 95% 5 , which is the range used in this work. Figure 3 shows the maximum parafn weight loading versus bulk graphite matrix density, considering the loaded total porosity. Figure 3 also presents the thermal storage capacity ratio versus the CENG matrix density. It can be noted the heat storage capacity decreases with the CENG matrix density. The nal effective material properties used for the parafn at 89 C were 6 : density 900 kg/ m3; latent heat L 179 kJ/ kg; specic heat c 1770 J / kg K; and thermal conductivity k 0.2 W / m K.

Thermal Analysis for the PCM

For the solution of the phase change problem, the basic assumption is that the sensible heat is negligible compared to the latent heat Stefan number is assumed zero , and consequently all the heat is assumed to drive the phase-change 7 . This lumped approach is compared with a second method in which the energy equation is discretized by nite differences and that considers spatial and temporal variations. This second method is commonly referred as the enthalpy method 7 , and the results are shown later in this document.

The PCM-Solar Collector

In this work the PCM-solar collector considered is constructed, as shown in Fig. 1, with a group of tubes immersed inside the PCM. The PCM is conned in separated and insulated rectangular containers as shown in Fig. 2. No similar characteristics for a solar collector have been taken in consideration in the previous works reported in PCM collectors. Bansal et al. 2 and Rabin et
1 Corresponding author. Present address: Department of Mechanical Engineering, Santa Clara University, 500 El Camino Real, Santa Clara, CA, 95051. Contributed by the Solar Energy Division of ASME for publication in the JOURNAL OF SOLAR ENERGY ENGINEERING. Manuscript received November 4, 2003; nal manuscript received December 6, 2005. Reviewed conducted by Karen DenBraven.

Lumped Capacitance Method

A lumped approach was used to investigate the PCM collector along the following assumptions: MAY 2006, Vol. 128 / 173

Journal of Solar Energy Engineering

Copyright 2006 by ASME

Downloaded 08 Nov 2011 to 14.139.34.2. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Fig. 1 Schematic of the PCM solar collector

1 Phase transition occurs at a single xed temperature. 2 Heat transfer inside the PCM matrix is restricted to the conduction mode. 3 No density change occurs while phase change is occurring. 4 Physical properties are temperature independent. 5 The local interface is assumed planar and sharp a surface separating the phases , at the phase-change temperature. 6 Surface tension and curvature effects at the interface are assumed insignicant. Energy and mass balances were established for the PCM using the lumped capacitance method and the resulting equations are presented here. For the absorber plate, conduction in a surface with negligible thermal mass is considered as the heat transport mechanism to the PCM. The expression for the heat conduction through the absorber plate is S UT T P Ta k T P PCM = 0 T 1

Fig. 3 Maximum parafn weight in percentage and reduced storage thermal capacity versus density of the graphite matrix loaded total porosity 6

No Collector Flow, Sensible Heating Case. For the analysis of the PCM during sensible heating all the mass of the PCM is assumed at the same temperature and energy is balanced over the entire control volume. Note that at this stage no melting of the PCM or uid ow through the tubes is considered. The energy balance for this case is mc A dTPCM k = S UT T P Ta dt PCM TPCM Ta
INS

The above expression is solved at each time interval for the plate temperature using the secant method simultaneously with the PCM temperature whose solution is described below. The value of UT is found using the relationship suggested by Dufe and Beckman 8 for mean plate temperatures between ambient and 200 C to within 0.3 W / m2 C: UT = N C T P Ta TP N + f +
e

No Collector Flow, Melting the PCM Case. A melting stage for the PCM can be considered with no uid ow through the tubes. During this stage the liquid and the solid phases are considered both to be at the melting temperature and the interface is considered planar. At this stage the position of the interface into the PCM is an important parameter and is represented by x t . The energy balance during melting is L
PCM

1 hw

dx k = S UT T P Ta dt

T M Ta
INS

T P + Ta T2 + T2 P a 2N + f 1 + 0.133 1 + P + 0.00591Nhw
g

2
P

where: N is the number of glass covers, f is 1 + 0.089hw 0.1166hw P 1 + 0.07866N , 70, C is 520 1 0.000051 2 for 0 e is 0.43 1 100/ T P , is the collector tilt degrees ; 20 in this work , g is the emittance of glass; 0.9 in this work , and P is the emittance of plate; 0.09 in this work .

Collector Flow, Melting of PCM Case. A simultaneous stage can be simulated in which solar radiation is available and a uid is circulating through the solar collector, such that energy is being removed and added simultaneously from and to the PCM. Equation 5 below is used for the nal temperature of the PCM, which results from the application of an energy balance only when temperature of the PCM is above the melting temperature , dTPCM A = dt mc S UT T P Ta
PCM

k
INS

TPCM Ta 5

QCONV mc PCM

where QCONV is the heat transfer rate to the collector. For this simultaneous stage when the temperature falls to the melting temperature and solidication begins around the collector tube, a radial conguration is considered similar to that shown in Fig. 4, and Eq. 6 is used for the position of the interface at each time interval, dR = dt 2 1 L l 1 k QCONV + A UT T P Ta + R

PCM

INS

TPCM Ta S

Fig. 2 PCM element conguration

For the charging stage all the PCM is considered to be solidied, solar energy is transmitted to the composite by the absorbing plate, and no energy is removed inside the tubes, thus energy is transmitted from the upper side to the bottom. For this case a at interface is assumed. During the simultaneous stage all the PCM is considered to be melted, heat is being removed inside the tubes, and solar radiation is available. Therefore solidication is exTransactions of the ASME

174 / Vol. 128, MAY 2006

Downloaded 08 Nov 2011 to 14.139.34.2. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Fig. 4 Schematic representation of the PCM material for the simultaneous stage

pected to occur around the tubes and a radial conguration can be assumed for the interface. There is no conict in this approach because charging and simultaneous stages are assumed to take place at different times. Collector Fluid/Tube Analyses. The output uid temperature and the energy removed from the solar collectors by forced convection are calculated under the assumption of a constant PCM temperature along the collector tubes. For the collectors outlet uid temperature a differential element is considered inside the tube. The integration along the axis z direction of the collector tube gives TOUT z = Ti + e where 2h = RINV c The expression for the convective heat ux is given by 9 QCONV = ASh Ti TOUT Ti TIN Ti TOUT ln Ti TIN 8
z

Fig. 5 Node arrangement used in the enthalpy method

tn+1 tn

x j+1/2

tn+1

x j+1/2

A
x j1/2

E x,t dx dt =
tn

A
x j1/2

qx x,t dxdt 12

TIN Ti

Here the enthalpy E is the sum of sensible and latent heats of the PCM. Proceeding with the discretization of the control volume in a nite number of nodes, a backward-in-time and central-in-space difference formulation is used. The following expressions are obtained for the enthalpy of each node: En+1 = En + j j tn n q + qn j+1/2 x j j1/2 Tn Tn j j1 R j1/2 En j cS 13

The heat ux between nodes is qn = j1/2 The temperature of the nodes is Solid En j 0 . . . T n = TM + j En j 15 16 17 14

The inside wall temperature of the copper tube can be expressed as Ti = and RO RIN e l 1 = A Sh 2 lkc l ln TPCM + TIN 1+ 9

10

Interface 0 Liquid En j

L . . . T n = TM j En L j cL

The Biot number used in this work represents the ratio of the convective effects to the spatial conduction and is an important parameter to observe when using the lumped capacitance method. Its value should be less than 0.1 for the lumped capacitance method to be valid 9 . It is dened as Bi = hrc k 11

L . . . T n = TM + j

where rc is the characteristic length see Fig. 4 , h is the convection coefcient inside the tube, and k is the thermal conductivity of the composite.

The PCM element was divided vertically into ve sections Fig. 5 with the node at the center of each section. The rst node is considered to be placed in the section under the absorber plate and the fth node in the section immediately over the bottom insulation. Results obtained from the enthalpy method are presented in the next sections and are compared with those obtained with the lumped capacitance method.

Results
Data of the PCM conductivity for materials with a parafn content of less than 65% are not available 6 and are not considered in this work. Figure 6 shows the variation of the thermal conductivity of the 65% composite material with parafn weight percentage. The overall thermal conductivity of the composite is similar to that of the sole porous graphite matrix 7 where the thermal conductivity is mainly in the perpendicular direction to the composite surface. In this work we consider uniform thermal conductivity along the thickness of the PCM. MAY 2006, Vol. 128 / 175

Validation: The Enthalpy Method

The method used to validate the results obtained with the lumped capacitance method is referred as the enthalpy method 7 . The enthalpy method is based on the conservation of energy expressed in terms of enthalpy and temperature. The energy conservation applied to a one-dimensional problem is expressed in the following: Journal of Solar Energy Engineering

Downloaded 08 Nov 2011 to 14.139.34.2. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Fig. 8 Comparison of the fraction of volume melted results given by the lumped capacitance method and the enthalpy method Fig. 6 Composite percentage conductivity versus parafn weight

To compare the lumped capacitance and the enthalpy methods described above, simulations were conducted for a constant ambient temperature of 30.74 C and an incident solar irradiation of 526 W / m2. The dimensions of the PCM element considered are 0.056 m in width and 3.583 m in length. Figure 7 presents the results for the warming process of the PCM from the ambient temperature to the PCMs melting temperature 89 C for both methods. As described, ve nodes through the depth of the PCM were considered for the case of the enthalpy method. As can be observed results from both methods are very similar. The PCM reaches the melting temperature in 196.3 min using the lumped capacitance method and in 160.5 min with the enthalpy method. The time both methods take to melt the PCM is shown in Fig. 8, where the fractional melted volume for both methods is presented as a function of time. Here the lumped capacitance method predicts that the PCM will melt completely in 524.3 min, while the enthalpy method predicts it will occur in 512.5 min. Thus, the remaining results presented here make use of the lumped capacitance method. To nd practical values of PCM volume and dimensions for the integrated collector a range between 0.06 and 0.04 m of PCM width was found to be acceptable to avoid excessive melting for given mass ow rates while maintaining a uniform melting as indicated by the small values of the Biot number. These values were found by optimization using the thermal efciency and a maximum solidication rate of 80% as controlling parameters. The corresponding mass ow rates inside the tube that can be allowed for this range of PCM material were between 0.001 and 0.002 kg/ s per element. Higher values of mass ow rates may induce faster solidication of the PCM. Figure 9 shows the minimum 65%-parafn composite area, cor-

responding to a maximum solidication allowed of approximately 80%, with respect to the maximum mass ow rate per tube for different constant ow inlet temperatures. The maximum solidication is considered when the interface diameter equals the depth of the composite. Composite areas lower than the values shown may result in PCM solidication rates in excess of 80%, and consequently the PCM-collector may not work appropriately. It can also be noted that the composite material has a linear trend with respect to the mass ow rate, with a zero intercept. The value for the minimum composite area corresponding to an assumed mass ow rate and inlet temperature is a design parameter to be considered when sizing the total area of PCM collectors needed for a specic application. The mass ow rate and the collector input temperature under actual working conditions may be different than those used here. Figure 10 shows the width and length of a composite element versus the mass ow rate per tube, corresponding to the minimum composite area and a maximum solidication of 80%, for the case of a constant inlet temperature of 80 C. It can be noted from the gure that as the mass ow rate per tube increases the width of the element decreases, and the length increases. The area of the composite is the product of the composites width and length. Since the widths used 0.04 0.06 m to nd practical dimensions of the proposed PCM integrated collector correspond to the maximum permitted, then the results for the lengths correspond to minimums, as can be observed from Fig. 10. Figure 11 shows the total collector area and the collector efciency versus the composites width, for an element with a constant mass ow rate of 0.00142 kg/ s. This mass ow rate was selected arbitrarily in order to work with a small collector area. In the case shown in Fig. 11, the maximum collector efciency and the minimum total collector area are reached for a composites width in the range of 0.049 and 0.056 m.

Fig. 7 Comparison of the warming process of the PCM element for results given by the lumped capacitance method and the enthalpy method using ve nodes

Fig. 9 Minimum composite area versus maximum mass ow rate for different inlet temperatures

176 / Vol. 128, MAY 2006

Transactions of the ASME

Downloaded 08 Nov 2011 to 14.139.34.2. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

E h k L l m Q q qt R rc S T t UT V x z Greek Symbols

Fig. 10 Width and length of the composite material as function of mass ow rate for constant inlet temperature of 80 C

enthalpy J / m3 heat transfer coefcient W / m2 K thermal conductivity W/m K latent heat J/kg tube longitude m mass kg heat ux W heat ux per area W / m2 top losses due convection and radiation W / m2 tube radius m , thermal resistance K m2 / W characteristic length m absorbed solar radiation W / m2 temperature K time s top loss coefcient from the absorber plate to ambient W / m2 K uid velocity m/s distance m horizontal axis, collector tube length change, insulation thickness m density kg/ m3 plate thickness m Stefan Boltzmann constant characteristic length

Fig. 11 Collector area and collector efciency versus width of the PCM material for a mass ow rate equal to 0.00142 kg/ s

Conclusions
The following specic conclusions can be drawn from this work: The lumped capacitance method was compared with a onedimensional enthalpy method that used nite differences. The results for warming and melting of the PCM were very similar from the two methods. Therefore, the results obtained using the lumped capacitance method are acceptable and representative of the process presented here. The parafn-graphite composite with 65% of parafn in weight was found to work best mainly because of the highest conductivity obtained. Data for higher parafn fractions is not available, but should be investigated further. A minimum composite-element area was found for which a maximum solidication of approximately 80% was permitted to occur. This area depended strongly on the element mass ow rate and the collector input temperature. The results for this PCM integrated collector proposed here are very encouraging. Therefore, there is an indication that conventional storage tanks may be replaced for the PCM integrated in the solar collector. The performance of a collector as described in this paper depends strongly on an operational design point. Variations from this design point other insulation rates, charging/ discharging durations may affect the storage performance.

Superscripts n a b CONV c i IN INS j L M O OUT PCM P S s w

time step ambient bottom losses convective copper inside inlet insulation node number liquid melting outside outlet phase change material absorber plate solid lateral area wind

References
1 Alva, L. H., and Gonzlez, J. E., 2002, Simulation of an Air-cooled Solarassisted Absorption Air Conditioning System, ASME J. Sol. Energy Eng., 124, pp. 276282. 2 Bansal, N. K., and Buddhi, D., 1992, Performance of a Cum Storage System, Sol. Energy, 48, pp. 185194. 3 Rabin, Y., Bar-Niv, I., Korin, E., and Mikic, B., 1995, Integrated Solar Collector Storage System Based on a Salt-hydrate Phase-change Material, Sol. Energy, 55, pp. 435444. 4 Sokolov, M., and Keizman, Y., 1991, Performance Indicators for Solar Pipes with Phase Change Storage, Sol. Energy, 47, pp. 339346. 5 Py, X., Olives, R., and Mauran, S., 2001, Parafn/Porous-Graphite-Matrix Composite as a High and Constant Power Thermal Storage Material, Int. J. Heat Mass Transfer, 44, pp. 27272737. 6 Hoogendoorn, C. J., and Bart, G. C. J., 1992, Performance and Modeling of Latent Heat Stores, Sol. Energy, 48, pp. 5358. 7 Alexiades, V., and Solomon, A. D., 1993, Mathematical Modeling of Melting and Freezing Processes, Hemisphere, Washington. 8 Dufe, J. A., and Beckman, W. A., 1991, Solar Engineering of Thermal Processes, Wiley, New York. 9 Incropera, F. P., and DeWitt, D. P., 1996, Fundamentals of Heat and Mass Transfer, Wiley, New York.

Acknowledgment
The authors recognize the valuable editorial suggestions by Prof. Jeffrey Morehouse of University of South Carolina, which resulted in signicant improvements of the nal manuscript.

Nomenclature
A c PCM total area m2 specic heat J/kg K

Journal of Solar Energy Engineering

MAY 2006, Vol. 128 / 177

Downloaded 08 Nov 2011 to 14.139.34.2. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm