Thermal performance of PCM thermal storage unit for ...

Solar Energy 78 (2005) 341–349 www.elsevier.com/locate/solener

Thermal performance of PCM thermal storage unit for a roof integrated solar heating system W. Saman *, F. Bruno, E. Halawa Sustainable Energy Centre, University of South Australia, Mawson Lakes SA 5095, Adelaide, Australia Received 12 September 2003; received in revised form 19 August 2004; accepted 25 August 2004 Available online 14 October 2004 Communicated by: Associate Editor Claudio Estrada-Gasca

Abstract The thermal performance of a phase change thermal storage unit is analysed and discussed. The storage unit is a component of a roof integrated solar heating system being developed for space heating of a home. The unit consists of several layers of phase change material (PCM) slabs with a melting temperature of 29 C. Warm air delivered by a roof integrated collector is passed through the spaces between the PCM layers to charge the storage unit. The stored heat is utilised to heat ambient air before being admitted to a living space. The study is based on both experimental results and a theoretical two dimensional mathematical model of the PCM employed to analyse the transient thermal behaviour of the storage unit during the charge and discharge periods. The analysis takes into account the effects of sensible heat which exists when the initial temperature of the PCM is well below or above the melting point during melting or freezing. The significance of natural convection occurring inside the PCM on the heat transfer rate during melting which was previously suspected as the cause of faster melting process in one of the experiments is discussed. The results are compared with a previous analysis based on a one dimensional model which neglected the effect of sensible heat. A comparison with experimental results for a specific geometry is also made. 2004 Elsevier Ltd. All rights reserved. Keywords: Phase change materials; Solar space heating; Thermal storage unit; Roof integrated heating systems

1. Introduction The work is currently underway at the Sustainable Energy Centre, University of South Australia, to design a roof integrated solar air heater for space heating. The

* Corresponding author. Tel.: +618 8302 3008; fax: +618 8302 3380. E-mail address: [email protected] (W. Saman).

main goal is producing a technically reliable system with economic attractiveness. This will be realised through two important strategies: (1) utilising existing roof construction as the thermal collector thereby minimising the capital cost, and (2) introducing a latent thermal storage unit which will store excess solar energy for use when there is no solar radiation and thereby minimising the auxiliary energy requirement. Belusko et al. (2001) have developed mathematical models for the roof integrated solar heating system (RISHS) and a full scale prototype system has recently

0038-092X/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.solener.2004.08.017

342

W. Saman et al. / Solar Energy 78 (2005) 341–349

Nomenclature A Af c f h H hc hs k m q T t U

surface area of a boundary node, m2 passage cross-sectional area, m2 specific heat, J/kg C liquid fraction of a node volumetric sensible enthalpy of a node, J/m3 total volumetric enthalpy, J/m3 local heat transfer coefficient, W/m2 C sensible enthalpy of a surface node, J/m3 thermal conductivity, W/m C mass flow rate, kg/m3 heat transfer rate, W temperature, C time, s velocity, m/s

density, kg/m3

Subscripts cond conduction E east node l liquid phase m melting N north node P current (centre) node s solid phase S south node W west node w wall Superscript o previous (old) value

Greek symbols k latent heat of fusion, J/kg a thermal diffusivity, m2/s

Aux Heat Storage Unit

q

Fan

Insulation Fig. 1. Schematic of the roof integrated solar heating system.

been installed in a house in Adelaide, Australia. The schematic of the RISHS is shown in Fig. 1 (Belusko et al., 2001). As shown, the system utilises the existing roof as a solar collector/absorber and incorporates a PCM thermal storage unit consisting of calcium chloride hexahydrate to store heat during the day and releases it to heat the living space during the night or when there is no sunshine. The inclusion of an energy storage unit in the system aims to ensure the maximum utilisation of solar energy absorbed by the system. By this, it maximises the solar contribution to serve the heating load and minimises the need for auxiliary energy. The installed RISHS is currently under monitoring and evaluation. In general, thermal energy storage offers the following advantages (Dincer and Rosen, 2001 and Dincer and Dost, 1997): reduced energy costs, reduced energy consumption, improved indoor air quality, increased flexibility of operation, reduced initial and maintenance costs, reduced equipment size, increased efficiency and effectiveness of equipment utilization, conservation of

fossil fuels, and reduced pollutant emissions (e.g. CO2 and CFCs). The proposed system aims to have all these advantages. A number of thermal energy storage systems using PCMs as the storage media have been designed and/or installed. Farid and Husian (1990) designed and tested an electrical storage heating system which utilised offpeak electrical energy. They showed that a PCM based storage unit has much less weight and could replace the conventional storage using bricks. They suggested further studies regarding the economic feasibility of the system for domestic purposes. Farid and Kong (2001) designed and tested an underfloor heating system using encapsulated PCM. The PCM (calcium chloride hexahydrate, CaCl2 Æ 6H2O) with a melting point of 28 C was placed in the concrete floor during the construction. The system was able to provide uniform heating throughout the day and kept the floor surface near the desired temperature of 24 C. Arkar and Medved (2002) designed and tested a latent heat storage system used to provide ventilation of a building. The spherical encapsulated polyethylene spheres were placed in a duct of a building ventilation system and acted as porous absorbing and storing media. The heat absorbed was used to preheat ambient air flowing into the living space of a building. The ‘‘solar wall’’ is another application of PCM for thermal storage. In this case the solar radiation that reaches the wall is absorbed by the PCM buried in the wall. Stritih and Novak (2002) designed an ‘‘experimental wall’’ which contained black paraffin wax as the PCM heat storage agent. The stored heat was used for


the validity of the model two sets of experimental results from similar and different PCM geometry are compared with the model predictions.

2. Mathematical model The schematic of the thermal storage unit under analysis is shown in Fig. 2. It consists of several layers of PCM slabs placed parallel to each other. Air flows through the passages between the PCM slabs. The most realistic model of the TSU under the study is a two dimensional model where a temperature gradient exists and therefore heat transfer occurs in both vertical and horizontal directions. In the vertical direction, heat transfer occurs due to the temperature difference between the air flow along the surface of the PCM and the layer of the PCM. In the horizontal direction, heat transfer occurs due to the variation in the temperature of the PCM along the horizontal plane. This scheme is shown in Fig. 3. In this scheme heat transfer between the PCM and the fluid occurs along the left, right, lower and upper surfaces of the PCM. The mathematical model employed in the current work is based on the enthalpy formulation where the dependent variable is enthalpy. For a phase change process involving either melting or freezing, energy conservation can be expressed in terms of total volumetric enthalpy and temperature as follows (Voller, 1990): oH þ rðuH Þ ¼ rðkðrT ÞÞ ot

ð1Þ

where H = total volumetric enthalpy, J/m3; t = time, s; u = velocity; k = thermal conductivity, W/m C; T = temperature, C.

L Air Flow

PCM slab Fig. 2. Schematic of the thermal storage unit (TSU).

y W

heating and ventilation of a house. The results of this work, according to the authors, were ‘‘very promising’’. Recently Vakilaltojjar (2000) and Vakilaltojjar and Saman (2001) developed a mathematical model and designed and tested a thermal storage unit using PCM for a space heating and cooling system integrated with a reverse cycle air conditioner. The model used by Vakilaltojjar was an extension of the one developed by Morrison and Abdel-Khalik (1978) for an air system. In his work, Vakilaltojjar treated the thermal energy storage system model as a forced convection problem with variable wall temperature. In the model the sensible heat was ignored and heat transfer was assumed to occur at a constant melting point temperature (Vakilaltojjar, 2000 and Vakilaltojjar and Saman, 2001). As a consequence of this simplification, in the simulation the initial temperature of PCM was set equal to the melting temperature. In the model, the air inlet temperature was 40 C and the melting temperature of the PCM was 29 C. It is expected that the largest portion of sensible heat transfer occur on the initial instant after complete melting of PCM; that is when the driving force (i.e. temperature difference) is the highest. As a rough estimate, when the PCM used in the model reaches the equilibrium temperature or steady-state conditions, given the temperature difference between the melting point and the air flowing outside the PCM containers, the maximum sensible heat convected by the flowing air to the PCM that can be stored is 23.43 kJ/kg (or 11% of the total) for CaCl2 Æ 6H2O and 40.48 kJ/kg kg (or 14.9% of the total) for KF Æ 4H2O. These figures may justify Vakilaltojjars claim for low differences between the air temperatures and melting/ freezing temperatures. The inclusion of the sensible heat term(s) in a model that describes the heat transfer between PCM and air, however, is useful when this term cannot be neglected. For example, Bruno and Saman (2001) have shown that for PCM fibreboards used as heat storage, the sensible heat is significant in the overall heat transfer; and the model involving only the latent heat is not valid. The effect of the sensible heat is also important in relation to thermal comfort consideration. Since the unit being analysed is used for delivering air to a living space to maintain the required level of thermal comfort, any factors that affect the thermal comfort need to be considered carefully. In the present work, the sensible heat term is included in the analysis while maintaining the effects of varying wall temperature on the air temperature. The model considers both the melting and freezing processes and the effects of operating temperature level on the melting/ freezing time. The model is used to analyse the thermal performance of the PCM thermal storage unit (TSU). To check

343

x L Fig. 3. Two dimensional model of the TSU.

344


In general, the thermophysical properties of materials undergoing phase change can be assumed constant. In such a case, Eq. (1) can be simplified as oH ¼ rðkðrT ÞÞ ot

ð2Þ

In the above equations, the total volumetric enthalpy H is the sum of the sensible and latent heats of the PCM and is related to the temperature of the PCM as follows: Z T H ðT Þ ¼ qc dT þ qfl ðkÞ ð3Þ Tm

The first term on the right side of Eq. (3) accounts for the sensible heat while the second term accounts for the latent heat of the PCM. In the above formulation, the latent heat of the PCM is related to the liquid fraction of the PCM, fl. To be able to calculate the latent heat, the liquid fraction fl needs to be defined. For the case of isothermal phase change (at T = Tm) the liquid fraction fl is calculated as follows: 1 if T > T m fl ¼ ð4Þ 0 if T < T m where Tm = melting temperature, C. Eq. (3) can also be written as ð5Þ

H ¼ h þ qfl k

where k = latent heat of fusion, J/kg. The first term on the right-hand side of Eq. (3) defines the sensible heat, h: Z T hðT Þ ¼ qc dT ð6Þ Tm

Following Eqs. (3) and (6) the enthalpy of the PCM is Z T qs cs dT ; T < T m ðsolidÞ ð7aÞ H¼ Tm

H ¼ ql fl k; H¼

Z

T ¼ T m ðmeltingÞ T > T m ðliquidÞ

Referring to Fig. 4, discretisation of Eq. (9) yields the following equations for internal nodes (Voller, 1990 and Costa et al., 1998): aE hE þ aW hW þ aP hP þ aN hN þ aS hS ¼ hoP þ q1 kðfPo fP Þ

ð10aÞ

where aE ¼ aW ¼ aN ¼ aS ¼ aR aP ¼ 1 aE aW aN aS

ð10bÞ

where hP = volumetric sensible enthalpy of the current node, P; hE, hW, hN and hS = sensible enthalpies of adjacent (East, West, North and South) nodes; ql = density of liquid PCM, kg/m3; hoP = volumetric sensible enthalpy of the current node from previous time step; fPo = liquid fraction of the current node from previous time step; fP = liquid fraction of the current node; a = thermal diffusivity of PCM, m2/s. The geometry of the current TSU has a high ratio of length to the thickness of the slabs. Direct application of the discretized Eq. (10) in the simulation will lead to unnecessarily long computing time. Therefore, in the present work a quantity / is introduced and defined as / = dx/dy, where dx is the length of a node in horizontal (x) direction and dy is the thickness of a node in vertical (y) direction. Inclusion of this quantity during the discretization of Eq. (9) results in the following modified equations of internal nodes coefficients: 9 > aE ¼ aW ¼ aR/2 = ð10cÞ aN ¼ aS ¼ aR > ; 2 aP ¼ 1 þ 2aRð/ þ 1Þ In Eqs. (10b) and (10c), R is defined as R ¼ Dt=ðDyÞ2

ð10dÞ

When the ratio / = 1, the coefficients of internal nodes, Eq. (10c), will be identical to Eq. (10b).

ð7cÞ

Initial condition For the melting process, the PCM is initially solid and its temperature is assumed at a certain value below the melting point. For winter operation, the initial temperature of the PCM, Tinit, is assumed to be 20 C. For freezing, the PCM is initially liquid and its temperature

Tm

where qs = density of solid PCM, kg/m3; cs = specific heat of liquid PCM, J/kg C; ql = density of liquid PCM, kg/m3; cl = specific heat of liquid PCM, J/kg C. Solving Eqs. (7) for the PCM temperature, we get: T ¼ Tm þ

H q s cs

T ¼ Tm

0 6 H 6 q1 k ðinterfaceÞ

ð8bÞ

T ¼ Tm þ

H q1 k ql cl

ð8cÞ

H < 0 ðsolidÞ

ð9Þ

ð7bÞ

T

ql cl dT þ ql k;

oh ofl ¼ r ðarhÞ qk ot ot

ð8aÞ

N P

H > q1 k ðliquidÞ

Substituting Eq. (5) into (2) we obtain:

E

W S

Fig. 4. Two dimensional domain of the PCM internal nodes notation.


is assumed at a certain value above the melting point. The value of this temperature depends on the final condition of melting during the charge period. These two situations can be used to calculate the initial sensible enthalpy of the PCM, hinit, as follows: qs cs ðT m T init Þ ð11Þ hinit ¼ ql cl ðT init T m Þ

345

used to determine the sensible enthalpy, hs, at the slab surface as follows: hs ¼ qs cs ðT PCM T Melt Þ T PCM < T Melt hs ¼ 0 T PCM ¼ T Melt hs ¼ ql cl ðT PCM T Melt Þ T PCM > T Melt

ð16Þ

3. Results and discussion

Qair ¼ mair cP DT air

ð12Þ

The air mass flow rate mair can be calculated from mair ¼ qair V air Af

ð13Þ

where Vair = air velocity inside the passage, m/s; Af = passage cross-sectional area, m2. Heat convected by the air, Qconv, to the PCM container wall (see Fig. 5) is given by Qcov ¼ hc AðT air T wall Þ

ð14Þ 2

where hc = local heat transfer coefficient, W/m C; A = surface area of a boundary node, m2; Tair = air temperature, C; Twall = wall temperature, C. Heat conducted to the PCM surface node is given by Qcond ¼ k w A

ðT wall T PCM Þ Dy wall

ð15Þ

Referring to Fig. 3, for a differential length Dx along the air flow direction, the constant wall temperature assumption is valid. A number of correlations for the Nusselt number, Nu, are available (Kays and Perkins, 1985) depending on the prevailing flow regime. Eqs. (12)–(15) are used to calculate the fluid, wall and PCM surface node temperatures at any distance from the entrance. The surface node temperature, TPCM, is

T air

q

T PCM

T Wall

3.1. TSU Specification The TSU under the numerical investigation consists of 45 flat PCM slabs arranged one on top of each other. This arrangement is contained in a rectangular duct through which air flows to release or remove heat from the PCM. The PCM (CaCl2 Æ 6H2O) has the following thermophysical properties: melting temperature = 29 C, densities: qsolid = 1710 kg/m3, qliquid = 1500 kg/m3, thermal diffusivities: asolid = 2.9684 · 107 m2/s, aliq7 2 m /s, thermal conductivities: uid = 1.5464 · 10 ksolid = 1.09 W/m K, kliquid = 0.54 W/m K. Air flows through the air gap between the slabs. The total mass of PCM is 600 kg and each slab has dimensions 1 m · 0.89 m · 0.01 m. Each slab is contained in a plastic container. For melting the PCM initial temperature was 20 C whilst for freezing the PCM initial temperature was 40 C. 3.2. Simulation results of the TSU 3.2.1. Effect of inlet temperature Fig. 6 shows how inlet temperature affects the heat transfer rate and the melting time at a given air flow rate. The higher the inlet temperature the shorter the melting time due to increased heat transfer rate. Further reduction in inlet air temperature results in a very long melting time which is not viable practically and economically.

7.5 Heat Transfer Rate, kW

Boundary conditions The PCMs storage being modelled is subjected to the following convection boundary conditions. Heat convected by the air to the PCM equals the heat conducted into the PCM. The heat conducted into the PCM is used to raise its internal energy and at the melting point this energy is used for changing its phase. A thermal circuit for the heat exchange between the air, the wall and the PCM is given in Fig. 5. Referring to Figs. 2 and 3 the heat transfer rate from the air from one node to the next in the +x direction is given by

Flow rate = 400 l/s 40°C 37.5°C

5.0

35°C

2.5

0.0

1 hc A

∆ywall

0

120

240

360 480 600 Time, minutes

720

840

k wA

Fig. 5. Thermal circuit of the air-wall-PCM heat transfer.

Fig. 6. Effect of the inlet temperature on the heat transfer rate and melting time (flat slabs).


3.2.2. Effect of air flow rate Fig. 7 shows the simulation results of the heat transfer rates and the melting time during a melting process for three different air flow rates. As seen, a higher flow rate increases heat transfer rate and shortens the melting time. The effect of the air flow rate is similar to that of inlet temperature. As the charge period is limited by the period of solar energy availability, this parameter is an important factor in determining the effectiveness of the TSU. Individual or combined effect of the inlet temperature and air flow rate may be the only practical factors that can control the range of the melting time. 3.2.3. Outlet air temperature profile For heating, the useful commodity of a TSU is the outlet air which should be warm enough to provide thermal comfort when it is delivered to a living space. Figs. 8 and 9 show the outlet air temperature resulting from a discharge (freezing) process. Air entering the TSU at a constant temperature of 20 C is heated as it passes the air passages in the TSU having initial temperature of 40 C. Initially, due to the high temperature difference between the air and the PCM, the outlet air temperature

Heat Transfer Rate, kW

10.0 Inlet temp. = 40°C 600 l/s

7.5

Flow rate = 400 l/s 35 15°C

30

17.5°C

20°

25 20 0

120 240 360 480 600 720 840 Time, minutes

Fig. 9. Effect of air inlet temperature on the air outlet temperature and freezing time (flat slabs).

approaches that of the PCM initial temperature and drops linearly and sharply during the period of the first hour. This is desirable as it delivers warming effect during the initial period of the system operation. When the PCM surface temperature reaches the melting point, the outlet temperature drops gradually because some of the heat is used to freeze the PCM. The outlet temperature suitable for heating must be limited by that necessary for achieving thermal comfort. The effect of the air flow rate on the outlet air temperature is also worth noting. As the flow rate increases the outlet air temperature curve shifts downwards which results in a lower air temperature supply to the living space (Fig. 8). Conversely, increasing the air inlet temperature results in higher air temperature of the air leaving the TSU (Fig. 9).

400 l/s

3.3. Comparison with experimental results

300 l/s

5.0

2.5

0.0 0

120 240

360 480 600 720 Time, minutes

840

Fig. 7. Effect of air flow rate on the heat transfer rate and melting time (flat slabs).

40 Temperature, °C

40 Temperature, °C

346

Inlet Temp. = 20°C

35

600 l/s

30

400 l/s

300 l/s

25 20 0

120

240

The results predicted by the model have been compared with two sets of experimental data. The first set consists of experimental data for PCM in flat containers while the second set came from the experiments with PCM encapsulated with conical capsules. The errors encountered in the experiments came from the measurements of air flow rates and the temperatures. The air flow rate was measured using a nozzle flowmeter having an accuracy of ±2%. The temperatures were measured with type T thermocouples calibrated against a resistance temperature detector (RTD) with an accuracy of ±0.5 C. Based on these values, the temperature measurements reported in the paper are within ±0.5 C whilst the heat transfer rates reported are within ±0.35 kW.

360 480 600 Time, minutes

720

840

Fig. 8. Effect of air flow rate on the air outlet temperature and freezing time (flat slabs).

3.3.1. Flat slabs Fig. 10 shows the inlet temperature of the air flowing into the TSU, the actual outlet temperature from the previous experiment (Vakilaltojjar, 2000) and the outlet temperatures predicted by Vakilaltojjars model and the present model. The PCM and the thermophysical properties are the same as in the simulation (Section 3.2). The data used in both the models and the experiment were as

W. Saman et al. / Solar Energy 78 (2005) 341–349 3.0

40 Heat Transfer Rate, kW

Inlet Temperature, °C

347

35 Vakilaltojjar's Present model

30

Experiment

2.5

Experiment

Present model

2.0 1.5 1.0 Vakilaltojjar's

0.5 0.0

25 0.25

0.5 0.75 Time, hrs

1

0

1.25

Fig. 10. Inlet and outlet air temperature profiles of the TSU from the two models and the experiment (flat slabs, melting).

follows. Weight of the PCM slabs = 6.05 kg consisting of 28 slabs, air flow rate = 96 l/s, slab thickness = 5 mm, width = 200 mm, length = 126 mm and air gap = 5 mm. The PCM initial temperature was 25 C. The figure shows a close agreement between the present model, Vakilaltojjars model and the experimental results. Similar results were found with the experiments involving freezing. Due to the small mass of PCM used in the experiments, the sensible heat effect only appears during the very short time at the beginning of the melting process. 3.3.2. Conical capsules The model also has been used to predict the heat transfer rate and the outlet temperature profile of a TSU containing PCM encapsulated in conical capsules. Due to the difference in the geometry, while keeping other parameters constant, the equivalent thickness of a slab and the air gap need to be estimated. Estimation was based on the actual heat transfer surface that results from the arrangement of the conical capsule in the TSU. The data of the experiment are as follows (Bruno and Saman, 2002): TSU dimensions = 1.3 m · 0.9 m · 0.6 m, PCM weight = 253.5 kg, container: conical capsule, air flow rate = 264 l/s with varying air inlet temperature. Fig. 11 shows the heat transfer rates predicted by both the models and the actual experimental results. As can be seen the prediction of the present model underestimates the heat transfer rates during half of the time, however it predicts accurately the melting time and the total heat stored during the melting. Fig. 12 shows the inlet temperature and the outlet temperatures from the experiment and from both models. As can be seen, in the initial period of melting where the sensible heat is predominant, the actual outlet temperature and that predicted by the present model rise sharply by some degrees above the initial temperature (about 22 C) to a value above the melting point (28 C). In the later period, where the whole surface of

120 240

360 480 600 720 Time, minutes

840

Fig. 11. Comparison of the experimental heat transfer rates with those predicted by the models (conical capsules, melting).

40 Inlet Temperature, °C

0

35 30 Vakilaltojjar's Present model

25

Experiment 20 0

120

240 360 480 Time. minutes

600

720

Fig. 12. Inlet and outlet air temperature from the experiment and those predicted by the models (conical capsules, melting).

the PCM is at the melting point, the outlet temperature increases gradually and finally approaches the inlet temperature at the end of melting. Since the heat transfer rate and air temperature are coupled, underestimate of heat transfer rate at the beginning of melting (Fig. 10) is reflected in the rather high outlet air temperature at the same period. Since Vakilaltojjars model ignored the sensible heat, it predicts higher outlet temperature at the initial period of melting. Finally, Figs. 13 and 14 show the heat transfer rate and air outlet temperature from the actual experimental results and the models prediction for freezing. As shown, the present model predicts accurately both the heat transfer rate and the air outlet temperature. 3.4. Natural convection Previously it was suspected that natural convection, occurring inside the PCM during the melting, speeds up the melting process (Bruno and Saman, 2002). In the present model, natural convection effect has not been included; yet it is able to predict quite accurately the melting time and the heat transfer rate during the melting. This suggests that for the geometry being considered, the effect of natural convection is rather


Heat Transfre Rate, kW

348 3.0 2.5

Present model

Vakilaltojjar's

2.0 1.5 1.0 Experiment

0.5 0.0 0

60

120 180 240 300 360 420 Time, minutes

Fig. 13. Comparison of the experimental heat transfer rates with those predicted by the models (conical capsules, freezing).

(2) A higher inlet air temperature increases the heat transfer rates and shortens the melting time. Conversely, during freezing, a lower inlet air temperature increases the heat transfer rates and shortens the freezing time. (3) Likewise, a higher air flow rate increases the heat transfer rate and shortens the melting time but increases the outlet air temperature. For freezing, a higher air flow rate increases the heat transfer rate and shortens the freezing time but reduces the outlet air temperature. (4) The model employed in this study has been validated using existing experimental data and the comparison is quite satisfactory.

References

35 Temperature, °C

Present model 30

Vakilaltojjar's

25 20 Inlet Temp.

Experimen t

15 0

60

120 180 240 300 360 420 Time, minutes

Fig. 14. Outlet air temperature from the experiment and those predicted by the models (conical capsules, freezing).

insignificant. This is understandable as the natural convection depends on the value of the Rayleigh number which is small for the geometry considered in this study. As observed by Laouadi and Lacroix (1999) for the TSU geometry similar to that in the present study, the effect of natural convection is significant only when the PCM slabs are in the vertical position and heated from the side.

4. Conclusions Analysis of the thermal performance of a phase change storage unit has been carried out. It demonstrates the need to include the effect of sensible heat in analysing this particular application. The following conclusions can be drawn from the analysis: (1) The effect of sensible heat is perceived in the initial periods of both melting and freezing. The effect are reflected in sharp increase in the outlet air temperature in the initial periods of melting and a sharp decrease in the initial periods of freezing. For heating purposes, this means a significant warming effect is perceived during the initial periods of delivering air to the living space. This is advantageous from the thermal comfort point of view.

Arkar, C., Medved, S., 2002. Enhanced solar assisted building ventilation system using sphere encapsulated PCM thermal heat storage. Second Workshop: IEA, ECES IA Annex 17, Advanced thermal energy storage techniques—feasibility studies and demonstration projects, 3–5 April 2002, Ljubljana, Slovenia. Belusko, M., Saman, W., Bruno, F., 2001. Experimental study of a roof integrated solar collector in a solar heating system, Solar World Congress, Adelaide, November. Bruno, F., Saman, W., 2001. Utilization of PCM Fibreboards for space heating. In: Proceedings of ISES Solar World Congress, Adelaide, Australia. Bruno, F., Saman, W., 2002. Testing of a PCM energy storage system for a space heating, World Renewable Energy Congress, 2002. Costa, M., Buddhi, D., Oliva, A., 1998. Numerical simulation of a latent heat thermal energy storage system with enhanced heat conduction. Energy Convers. Manage. 39 (3–4), 319–330. Dincer, I., Rosen, M.A., 2001. Energetic, environmental and economic aspects of thermal energy storage systems for cooling capacity. Appl. Therm. Eng. 21, 1105–1117. Dincer, I., Dost, S., Li., X., 1997. Performance analyses of sensible heat storage systems for thermal applications. Int. J. Energy Res. 21, 1150–1171. Farid, M., Kong, W.J., 2001. Underfloor heating with latent heat storage. Proc. Inst. Mech. Eng. 215 (Part A), 601–609. Farid, M.M., Husian, R.M., 1990. An electrical storage heater using the phase-change method of heat storage. Energy Convers. Manage. 30 (3), 219–230. Kays, W.M., Perkins, H.C., 1985. Forced convection, internal flow in ducts. In: Rohsenow, W.M., Hartnett, J.P., Ganic, E.N. (Eds.), Handbook of Heat Transfer. McGraw-Hill, New York. Laouadi, A., Lacroix, M., 1999. Thermal performance of a latent heat energy storage ventilated panel for electric load management. Int. J. Heat Mass Transfer 42, 275–286. Morrison, D.J., Abdel-Khalik, S.I., 1978. Effect of phasechange energy storage on the performance of air-based and liquid based solar heating system. Solar Energy 20 (1), 57– 67. Stritih, U., Novak, P., 2002. Thermal storage of solar energy in the wall for building ventilation. Second Workshop: IEA,

W. Saman et al. / Solar Energy 78 (2005) 341–349 ECES IA Annex 17, Advanced thermal energy storage techniques—feasibility studies and demonstration projects, 3–5 April 2002, Ljubljana, Slovenia. Vakilaltojjar, S.M., 2000. Phase change thermal storage system for space heating and cooling. Ph.D. Thesis, School AME, University of South Australia.

349

Vakilaltojjar, S.M., Saman, W., 2001. Analysis and modelling of a phase change storage system for air conditioning applications. Appl. Therm. Eng. 21, 249–263. Voller, V.R., 1990. Fast implicit finite difference method for the analysis of phase change problems. Numer. Heat Transfer, Part B 17, 155–169.