Parametrical analysis of latent heat and cold storage for heating and cooling of rooms

Parametrical analysis of latent heat and cold storage for heating and cooling of rooms

Applied Thermal Engineering 84 (2015) 138e149 Contents lists available at ScienceDirect Applied Thermal Engineering journal homepage: www.elsevier.c...

3MB Sizes 0 Downloads 74 Views

Applied Thermal Engineering 84 (2015) 138e149

Contents lists available at ScienceDirect

Applied Thermal Engineering journal homepage:

Research paper

Parametrical analysis of latent heat and cold storage for heating and cooling of rooms* E. Osterman a, *, K. Hagel b, C. Rathgeber b, V. Butala a, U. Stritih a a b

Faculty of Mechanical Engineering, University of Ljubljana, Askerceva 6, 1000 Ljubljana, Slovenia ZAE Bayern, Walther-Meißner-Str. 6, 85748 Garching, Germany

h i g h l i g h t s  Thermal properties of paraffin RT22HC were measured.  Flow visualization was carried out and velocity between plates was measured.  Thermal and pressure drop analysis were performed.  Melting times are too long however, use of storage tank for heating and cooling looks promising.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 15 July 2014 Accepted 20 February 2015 Available online 31 March 2015

One of the problems we are facing today is the energy consumption minimization, while maintaining the indoor thermal comfort in buildings. A potential solution to this issue is use of phase change materials (PCMs) in thermal energy storage (TES), where cold gets accumulated during the summer nights in order to reduce cooling load during the day. In winter, on the other hand, heat from solar air collector is stored for evening and morning hours when solar radiation is not available. The main objective of the paper is to examine experimentally whether it is possible to use such a storage unit for heating as well as for cooling. For this purpose 30 plates filled with paraffin (melting point around 22 C) were positioned into TES and applied with the same initial and boundary conditions as they are expected in reality. Experimental work covered flow visualization, measurements of air velocity in the channels between the plates, parametric analysis in conjunction with TES thermal response and measurements of the pressure drops. The results indicate that this type of storage technology could be advantageously used in real conditions. For optimized thermal behavior, only plate thickness should be reduced. © 2015 Published by Elsevier Ltd.

Keywords: PCM Thermal energy storage Plates Free cooling Heating Solar air collector

1. Introduction In the last few years there has been a growing interest in reducing energy use in buildings, which is also stated in the European directive [1]. The literature on this topic shows a variety of approaches towards achieving this goal. This article concentrates on energy storage using phase change materials (PCM). Principle of operation of all systems with PCM is that the energy is stored when it is available and is released later when the energy demand arises. In order to achieve thermal comfort in buildings it is necessary to deliver a certain amount of heat or cold. The source can be *

This document is a collaborative effort. * Corresponding author. Tel.: þ386 1 4771 408. E-mail address: [email protected] (E. Osterman). 1359-4311/© 2015 Published by Elsevier Ltd.

conventional systems or more advanced technologies, proposed by researchers in recent years. Promising are two principles: one for cooling and the other one for heating. For cooling needs in summer, the idea is to store outdoor cold during the night and supply it to the indoor environment during the day. This type of cooling is suitable for climates where the diurnal temperature range is at least 15 K [2]. For the heating requirement, energy from the sun can be exploited in conjunction with solar air collectors [3]. It happens often that the heating during the day is not needed, because of relatively high ambient temperature and solar gains. Heating load increases after sunset, so in such cases day-cycle TES can be advantageously used. When the available energy exceeds the energy demand, energy is stored, later it is released and so it completely or partially replaces conventional systems. The advantage of such systems is that they can operate in a relatively small temperature

E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149


Roman symbols c specific heat [kJ/(kg K)] D hydraulic diameter [m] f friction factor [/] h specific enthalpy [kJ/(kg] K loss factor [/] L length of pipe [m] m mass flow [kg/s] P thermal power [W] Q heat [J] T temperature [ C] U the overall heat transfer coeffiecient [W/(m2 K)] v mean velocity of flow [m/s] Abbreviations EPS expanded polystyrene

range and the amount of energy is greater than in the case of sensible heat storage. Several publications have appeared in recent years documenting different types of TES such as capsules [4e6], tube banks [7], hybrid systems with fins [8,9], heat pipes [10], shell and tube [11], plates [12e14]. All of them use air as HTF. Experimental studies, where plates were used, can be divided into two groups, namely for cooling and for heating. Research work on former will be presented first. One of the first examples was presented in paper by Zalba et al. [15]. They outlined the development of an installation for free cooling that allows testing the performance of PCMs. Experiments were performed using 3 kg of paraffin (RT25, Rubitherm GmbH) which means a storage density of 28 kWh/m3. The main focus of the study was on: the ratio of energy/volume in capsules, load/unload rate of the storage, and cost of the installation. The results obtained suggest that the thickness of the encapsulation, the inlet temperature of the air and the air flow have significant influence on the solidification and melting process. Study carried out by Waqas and Kumar [16] was conducted to investigate thermal performance of the latent heat storage for free cooling of buildings in a dry and hot climate. 13 kg of PCM (SP29, Rubitherm GmbH) was encapsulated in the containers of the galvanized steel. The authors concentrated on the influence of air flow and the air inlet temperature on cold accumulation and it was shown that solidification of PCM was more sensitive to the charging air temperature compared to the air flow rate. Heat exchanger with two different geometries was investigated by Lazaro et al. [17,18] who developed empirical and numerical models. They proposed a modular structure and identified the required number of modules and the melting temperature to meet the specific cooling demand over time. In experiment 135 kg of paraffin (RT27, Rubitherm GmbH) was encapsulated in 10 mm thick aluminum plates. Second group are applications for space heating where Saman et al. [19] used experimental data for validation. They analyzed effects of different parameters (such as the air flow rate, the PCM slab thickness or the air gap) on TES performance. Melting temperature of calcium chloride hexahydrate (PCM29) was around 29  C. Labat et al. [20] presented a prototype which provided a 1 kW heating power during 2 h. It consisted of 34 aluminum containers filled with 28 kg of paraffin (Microtek 37D). Its melting range was between 31  C and 34  C. Their results revealed that



heat transfer fluid polymethylmethacrylat thermal energy storage

Subscripts a air al aluminum amb ambient avg average ch channel cs cross-section end end of cycle EPS expanded polystyrene h holder in inlet init initial out outlet PMMA polymethylmethacrylat

enough energy was stored but the heating power was lower than 1 kW in the first 2 h. In another study carried out by Charvat et al. [21] larger heat storage unit was investigated. They used 100 aluminum plates filled with a paraffin (RT42, Rubitherm GmbH) with the melting temperature around 40  C. Experimental results served as a means of validation. Data from each of the above mentioned experiments are collected in Table 1. Compactness refers to the ratio of plate's surface area to plate's volume, whereas energy density refers to the ratio of stored energy to volume of storage unit. The novelty of the paper is local storage unit for heating and cooling. In Table 1 different studies are presented, but none of them unites these two options. However, some studies [23] address both of them, but those are larger systems that are meant to operate more or less as central systems. In these systems more plates were positioned one after another along the flow. In this manner the area for heat transfer increased significantly and, consequently, the characteristics and thermal response of larger storage tank differed from the smaller one. With proposed system low exergy sources could be utilized throughout the year. Overall delivered energy for cooling and heating demands would increase, and consequently the energy demand from conventional systems would decrease. TES would operate in summer and winter, and thus would not be constrained to one season as are other systems in relevant literature. The objectives of the paper are to study the effects of inlet air temperature, air flow and air gap on the melting and solidification processes, to check the total energy exchanged between air and PCM for each case, to assess the required time to perform the phase-change process, to carry out flow analysis in the tank, since none of the previous studies addressed problems with uniform flow and nobody presented distribution of flow between channels and to determine thermal properties of paraffin and investigate if hysteresis and subcooling are present. 2. Experimental setup 2.1. Description of TES unit In order to achieve the target set in the introduction, an experimental rig for testing TES's thermal response was set up.


E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149

Table 1 Data from previous experimental work. Author



Zalba, Marin [15,22] Outer dimension of plates Dimension of TES Used PCM Melting range Mass of PCM # of plates Air gap Compactness Energy density Air flow Inlet temperature Maximum power Charge/discharge time

Waqas and Kumar [16] Lazaro, Dolado [17,18]

Saman et al. [19]


0.385  0.145  0.029 0.50  0.50  0.01

0.45  0.30  0.01

m /  C kg / m m2/m3 kWh/m3 m3/h  C W h

n. a. RT25 18e24 3 3 n. a. 100e150 28 100, 150 16, 18 28, 30 125 3.5

n. a. n. a. RT27 PCM29 25e28 29 135 6.1 216 28 0.01 0.005 200 400 / 24 675e1550 345 18, 20, 22, 25 40, 45, 50 20 40 4500 n.a. 2 1

1.5 (length) SP29 28e29 13 3 n. a. 200 26 52, 65 20, 22, 24 36, 38, 40 50 6

Casing was made of 8 mm thick PMMA and the external dimensions were 0.77 m  0.67 m x 0.42 m. Storage unit was insulated with 8 cm thick EPS. The unit contained 15 or 30 CSM plates (depending on the case) filled with paraffin RT22HC. PCM and melting point, respectively, was chosen in such a way that it ensured maximum solidification or melting of PCM during charging process. According to Yanbing et al. [4] the melting point of the PCM in case of cooling should be close to the designed room temperature or according to Medved and Arkar [24] it should be 2 K above average ambient temperature in the summer months. TES was in our case devoted also to heating, which was the reason for higher melting point. On the other hand, it should not be too high, because it might not get sufficiently charged. Proposed melting point for Slovenian climate is therefore between 22  C and 23  C. Plates (0.30 m  0.45 m x 0.015 m) were vertically or horizontally positioned in the storage tank (the longer side parallel to the flow) as in Fig. 1. In configuration with 15 plates, they were located in the middle of storage unit and the remaining space (on the sides) was filled with EPS. This brought additional mass into TES, but as the density of EPS is low, it did not significantly affect thermal capacity. Distance between plates was either 8 mm or 16 mm. Heat storage unit was a kind of PCM-air heat exchanger and its

Labat et al. [20]

Charvat et al. [21]

0.20  0.13  0.005 1.00  0.20  0.018 0.45  0.30  0.01 0.2  0.63  1 Microtek 37D 31-34 28 34 0.004 111 19 300e900 20 45 2700 1,25

0.44  0.62  1.8 RT42 40 55 100 0.02 200 7 230 25 58 1950 7.5

compactness (surface area/volume of storage) was 133 m2/m3and the energy density 16 kWh/m3. Average mass of filled plate was 1361 g, weight of paraffin in the plate was 1003 g and volume of each plate was 1.42 l. Approximately 9% of the plate's volume is empty in order to compensate the volume expansion of the liquid PCM and to avoid deformation of the plate due to higher pressure. Knowing the actual volume (weight) of the plate is very important when applying data for validation. If manufacturer's dimensions were taken into account, its thermal capacity would be overestimated. Consequently significant deviations in calculations would arise. 2.2. Test rig The experiments were carried out in a lab environment. Schematic representation of the entire test rig is shown in Fig. 2. Air was taken from the exterior of a building and conditioned through a finned tube heat exchanger (which was connected to a thermostatic bath) substituting an air collector or representing temperature of ambient air. Temperature sensors were installed at the inlet and outlet of the tank (thermocouple type K, ± 0.25  C, k ¼ 2), preceded by a static mixer which ensured a uniform

Fig. 1. Vertical and horizontal configuration of the plates installed in the mounting (holder).

E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149


Fig. 2. Scheme of the test rig.

temperature profile in the pipe. Consequently, it was possible to measure temperature at one point only. Perforated plate was installed before the plates (Fig. 3) which ensured uniform distribution of flow between plates. The design of the perforated plate has been carried out by means of CFD analysis. Pressure difference between the inlet and outlet was measured with differential pressure transducer (DMU2, ±1.5 mbar), volume flow with RCI FV 9060 (±0.06 m/s). Mass flow rates were chosen so as to resemble real conditions as best as possible. Minimal flow rate mirrors conditions in a room when a minimum air change rate needs to be ensured, whereas maximal flow rate simulates conditions when TES needs to be charged with heat/cold in a desired time. Targeted flow rate was reached with two parallel side channels fans and a valve. The fans allowed for the maximum air flow rate of 120 m3/ h. Every potential leakage spot was sealed to prevent incorrect flow rate readings. Thermal response of the TES to inlet temperature step function under different conditions was observed during the measurements. 10 thermocouples were used to measure temperature at various locations (Fig. 1), but for the sake of clarity, only inlet, outlet and ambient temperature are presented. Flow meter, pressure transducer and all thermocouples were connected to the data acquisition system Agilent 34901A. Commercial software Labview was used to acquire data from data acquisition system, and to record them in a database format on a server for further processing. All data were recorded at time intervals of 2 s.

One of the goals of this study was to test whether the distribution of flow between the plates is really uniform. For this purpose two methods were used. First one was visual observation of smoke flowing between plates, from which velocity was calculated on the basis of traveled distance and needed time. Pictures of smoke were taken just after it was blown into TES, so that it was possible to determine at what time it reached beginning and end of the plate. Visualization and calculation of flow time were carried out for vertical and horizontal configuration for 22 m3/h and 44 m3/h. In vertical configuration plates were positioned one by one (as in Fig. 1) with 8 mm air gap. In horizontal configuration plates were positioned in two parallel stacks with 10 mm air gap. In both configurations 30 plates were used. Second method comprised measurements of air velocity with anemometer, which was inserted between the plates. With this method only measurements for vertical configuration were performed. 2.3. Analysis TES is designed for cooling as well as for heating, hence boundary conditions (temperature) were chosen so that they simulate real conditions in full measure. In summer it can be expected that the inlet air temperature will be approximately between 16  C (night air) and 30  C (air temperature during the day). In winter temperature difference will be higher, approximately

Fig. 3. Experimental setup.


E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149

between 5  C (heating of external air) and 40  C (available warm air from the solar collector). These experiments were primarily carried out to determine the amount of heat that can be stored in the tank and thermal power. Latter is calculated by Eq. (1).

dQ ¼ m_ a cp;a ðTout  Tin Þ dt


As a rule, enthalpy of humid air should be considered, but as the deviation is negligible (0.6%), power can be calculated with a constant specific heat (1.006 kJ/kgK). Mass flow rate was calculated from volume flow rate and density. Eq. (1) assumes perfectly insulated storage tank, but in reality there are always thermal losses that need to be taken into account. They can be determined from the steady state as described by Eq. (2).

P ¼ m_ a cp;a ðTout  Tin Þ þ UAðTamb  Tst Þ ¼ 0


Tst is the unknown temperature of the storage unit. It can be replaced by the average temperature difference between inlet and outlet calculated with Tavg¼TinþTout/2. Definition of UA value follows as:

UA ¼

m_ a cp;a ðTout  Tin Þ Tamb  Tavg


Theoretical value of UA can be calculated if we consider forced convection inside, conduction through the insulation and natural convection outside. Since the dominant thermal resistance is due to conduction in the shell, convection has negligible effect on the UA value. Calculated UA value is 1.8 W/K. With steady state experiments UA value was determined as depicted in Fig. 4. Once UA was determined one could estimate thermal losses as UA(TambTavg) and add them to Eq. (1) for transient experiments. It is advisable to test whether different values of UA significantly impact thermal losses. Therefore a sensitivity analysis was carried

out, where estimation of losses in relation to thermal power was given as UA(TambTavg)/P. This ratio was determined for three different UA values (1.8 W/K, 2.7 W/K and 3.6 W/K) inside the phase change temperature range (between 20  C and 25  C) in which such a storage unit operates most of the time. It turns out that this ratio is, in most of the cases, below 10%, which means that deviations (possible wrong estimation) in UA value for a given range are not that influential. Still losses have to be taken into account when evaluating accumulated or rejected heat within one measurement (Eq. (4)).

Z Q¼

   m_ a cp;a ðTout  Tin Þ þ UA Tamb  Tavg dt


On the other hand accumulated or rejected heat can also be verified theoretically as in Eq. (5), since mass and thermal properties of used materials are known.

   Q ¼ mh cp;h þ mal cp;al þ mPMMA cp;PMMA Ta;out;end  Ta;out;init   Ta;out;end  Ta;out;init þ mPCM DhPCM þ mEPS cp;EPS 2 (5) Temperature difference is lower for expanded polystyrene, because within, the largest temperature gradient is generated. To calculate corresponding temperature, following is assumed: initial and final internal temperature (interior of storage tank) are the same as initial and final outlet temperature (Ta,out,init, Ta,out,end). In addition, temperature at the beginning of measurement is average between initial (Ta,out,init) and ambient temperature (Tamb); at the end (Ta,out,end) it is between end and ambient temperature (Tamb). In this way temperature of EPS is obtained as:

Fig. 4. UA value in relation to mass flow.

E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149

Ta;out;end þ Tamb Ta;out;init þ Tamb  2 2 Ta;out;end þ Tamb  Ta;out;init  Tamb Ta;out;end þ Ta;out;init ¼ ¼ 2 2 3. Results This section addresses influence of different inlet and boundary conditions on TES performance. Besides that, it presents measurements of thermal properties, analyzes flow as well as temperature conditions and examines pressure drop in the system.

3.1. DSC and T-History measurements Temperature dependence of PCM's specific heat, its thermal conductivity and density had to be determined. Data for the enthalpy changes upon melting and solidification between 14  C and 29  C were available from the manufacturer. To verify these data and to determine the influence of the sample size on the enthalpy curve, DSC step measurements were performed with a TA Q2000 heat-flux DSC and T-History measurements with a self-built T-History calorimeter available at ZAE Bayern. Further details about the DSC and the T-History calorimeter are given in Ref. [25]. A comparison of our data with the manufacturer's data is shown in Fig. 5. The enthalpy change upon melting between 15  C and 25  C measured via DSC and T-History deviates by less than 1%. Considering the enthalpy curves, RT22HC shows a volume-independent melting and solidification behavior without subcooling or hysteresis. For homogeneous materials, such as this one, DSC step measurements are favorable, because of negligible temperature gradient inside the sample. In the case of non-homogeneous materials (e.g. salt hydrates) it is more advisable to measure the enthalpy curves via T-History. In this way they can be larger and thus representative for selected material. Similar guidance give authors in Refs. [26,27], and [28]. Thermal conductivity in the solid state was measured with Isomet 2104 and density of the liquid with


a DMA 4500M from Anton Paar that was calibrated with pure water. Thermal properties are given in Table 2. 3.2. Flow analysis Measurements with anemometer were carried out at the beginning, in the middle and at the end of the plates on three different vertical positions (5 cm, 15 cm and 25 cm from the top of the plate) and for two different volume flow rates (22 m3/h and 44 m3/h) (Fig. 6). Air gap in this configuration was 8 mm. In the case of 22 m3/h, distribution of velocity is homogeneous, but with 44 m3/h it can be observed that at the beginning and end, the spread of values gets larger because of the entrance and exit effects. However, for the middle position, which is actually the most important, agreement is acceptable. Fig. 7 depicts results from flow visualization, namely flow distribution for the first 5 s in horizontal configuration with 22 m3/h (air gap was 10 mm). It seems that velocity is the highest in the middle, but it is just the matter of illumination (in this case lights were in the middle of the height). Based on the Reynolds number with the maximum value of 650 (Eq. (6)), laminar flow regime is expected, which can be confirmed from the figures. It can be clearly seen from Fig. 7c (channel number 7) that smoke forms shape of a parabola which is characteristic of a laminar flow. Flow visualization was repeated several times for channel numbers 2, 5, 9, 11, 15 and the results can be seen in Fig. 8. Deviations between measurements for one channel as well as between individual channels are small, which confirms satisfactory design of the perforated plate. Theoretical velocities are 0.07 m/s and 0.13 m/s for 22 m3/h and 44 m3/h, respectively. Results from the anemometer and from the flow visualization deviate by a factor of about 2. One reason could be air gap, which was 8 mm in anemometer measurement, whereas in flow analysis it was 10 mm, hence values measured with anemometer are higher. Another argument for deviations is also, that velocity measured with anemometer takes place in the channel's center. Velocity there is higher than near the plate. Besides it reduces air gap, which increases velocity. On the other hand values obtained via flow visualization are average values over the entire channel cross section. 3.3. Thermal analysis This subsection evaluates impacts of: number of plates, distance between plates, mass flow, initial and inlet temperatures on thermal power and accumulated heat. Results are presented for melting and freezing for vertical configuration with 15 or 30 plates. In all subsequent graphs thermal power curves are those where thermal losses are already subtracted and negative thermal power means freezing process.

Fig. 5. Enthalpy curves of RT22HC measured via DSC (solid line) and T-History (dashed line) compared with manufacturer's data (dot-dashed line).

3.3.1. Inlet temperature Time to melt or freeze PCM depends on the air inlet temperature; the higher the inlet temperature in case of melting, the larger the heat exchange, thus shorter time to melt. However, there is always limitation due to boundary conditions, e.g. outdoor temperature or outlet temperature from the solar collector. In Fig. 9a a comparison between three different inlet temperatures (30  C, 35  C and 45  C) is given. 30  C and 35  C are inlet temperatures for

Table 2 Thermal properties of RT22HC.



l (W/mK)

Dh (kJ/kg)

772.9 ± 0.1 @ 30  C

0.19± 0.01 @ 10  C

181 ± 9 (between 16  C and 30  C)


E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149

Fig. 6. Velocity measured with anemometer for vertical configuration.

Fig. 7. Flow visualization between channels for horizontal configuration.

E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149


Fig. 8. Velocity in individual channels from flow visualization for different flow rates.

two cases that would occur in the summer time during the day, whereas 45  C is the air temperature from the solar collector in wintertime. Initial temperature is 16  C, the same as in the summer time in the morning after night solidification. This temperature is not the most representative for the case which occurs in wintertime, but it is done so for the sake of comparison. End of melting is defined in inflection points for Tout just before it reaches steady state. Melting times are therefore 8.0 h, 12.5 h and 18.5 h for inlet temperatures 45  C, 35  C and 30  C. Similar analysis can be carried out for freezing (Fig. 9b) with 26  C being initial temperature. Solidification times are 13.5 h and 11.0 h for the inlet temperature 9  C and 4  C, respectively.

published in Ref. [30] are different, because in their case change of flow rate only affected melting/solidification time. The reason behind it is presumably very large flow rate, since heat transfer between air and PCM reached its limit. Same observation can be found in Ref. [20]. Time to reach end state in our experiment is not two times shorter, because also power is not two times larger all the time. Beside that, blue curve doesn't reach zero at the end (but it should, because losses were taken into account), which implies slight underestimation of UA value and results in longer melting time (in web version). Another reason was slightly lower initial temperature for larger mass flow. Stored heat for given conditions was approximately 6.5 MJ.

3.3.2. Mass flow rate Previous section reveals that for such application, solidification and melting times are too long. There are two reasons behind it, namely heat transfer in PCM itself and heat transfer between air and plate. Former cannot be influenced (except with some enhancement inside plates), whereas latter can be with larger mass flow. Results from now on are rather presented in terms of thermal power instead of temperature, as delivered heat or cold is of the main interest. Curves for melting with Tin¼[30] C and Tinit¼[16] C are presented in Fig. 10. Tin and Tinit are chosen to resemble summer time conditions during discharging. From the figure it can be seen, that if mass flow increases two times, then also thermal power in the melting area increases by the same factor (from 125 W to 220 W). Similar observation was found in Refs. [29], yet results

3.3.3. Distance between plates When designing such a storage unit, distance between plates needs to be determined. One objective is to store as much as possible heat/cold in a given volume, which leads to small distances, insufficient heat transfer and increased pressure drop. In Fig. 11 curves for different distances between plates are presented; in Fig. 11a four cases for melting; two for conditions in wintertime (Tin¼[40] C and Tinit¼[20] C, red and blue curve) and two for summer time (Tin¼[30] C and Tinit¼[17] C, orange and cyan curve). Mass flow rates and number of plates are the same, only distance between plates varies between either 8 mm or 16 mm. Thermal power for winter conditions is larger, because of the larger temperature difference at the beginning and because of larger difference TinTPCM in the phase change range. The effect of greater

Fig. 9. Outlet temperature over time.


E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149

Fig. 10. Thermal power for different mass flow rates (Tin¼30 C and Tinit¼16 C).

distance is lower thermal power in the initial period and longer melting time. One reason for this could be larger thermal entrance region, which results in smaller heat transfer in the first part of storage unit. Another reason may be built-up of boundary layer, leaving bulk of the flow experiencing almost no heat transfer. So for bigger air gap, thermal power represents approximately 80% of smaller air gap thermal power. Simulations carried out by Halawa and Saman [29] also indicate similar responses. Stored heat in all four cases is approximately 3.2 MJ. Analysis was also carried out for solidification imitating winter (Tin¼[16] C and Tinit¼[25] C, orange and cyan curve) and summer (Tin¼[10] C and Tinit¼[35] C, red and blue curve) conditions. Curves resemble those from melting (Fig. 11b), except for thermal power, which is lower because of smaller temperature difference. 3.3.4. Number of plates In real conditions space limit could be one of the impediments, therefore thermal response of TES with 15 plates was investigated. Thermal power decreases if velocity between channels remains constant because of reduced mass flow. Melting/freezing time does not change compared to experiment with 30 plates. On the other hand, velocity and heat transfer increase, if mass flow remains constant, which gives larger thermal power as in the case with constant velocity. Curves for melting (Tin¼[30] C and Tinit¼[16] C) and freezing (Tin¼[16] C and Tinit¼[25] C) are presented in Fig. 12. Under given conditions melting/solidification time reduces for more than 40%. Heat transfer is higher in the case with 15 plates (blue line) as there are 2 times less plates. This results in equal initial thermal power for both pairs of curves. In continuation of experiment blue curve presents 80% of thermal power for case with 30 plates (red curve) (in the web version).

3.3.5. Temperature profile As already mentioned, all measurements were carried out with constant inlet temperature except for two of them. The idea was to observe thermal response of a storage tank when the inlet temperature varies with time. Two temperature profiles were established for this purpose that imitated ambient temperature in summer time (Fig. 13). This is particularly important in the case of charging TES with cold, because of the smallest temperature difference that exists at that time. It is therefore questionable whether the tank will be charged or not. Entire storage tank was heated to 26  C before the measurement. Inlet temperature was gradually decreased from 26  C so that it reached 20  C in 1 h. After that it was decreased again, so that it reached 16  C in 6 h. This temperature represents the lowest temperature at night, after that it begins to rise again. This was illustrated by a temperature rise to 18  C over the next 2 h. In the second profile, external conditions were simulated also for day, so the temperature rose to 30  C in 6 h, representing discharging period. Profiles have been selected so that they correspond to temperatures from the test reference year for Ljubljana, Slovenia. Two measurements were carried out; one with 59 kg/h, 30 plates and 8 cm air gap and the other one with 95 kg/h, 15 plates and 8 mm. Problem with solidification times has already been highlighted in preceding subsections, yet it is obvious now. PCM was half solidified in the first case, resulting in insufficient amount of stored cold for the next day. Second case looks more promising in spite of smaller thermal power (around 100 W). PCM is solidified after 8 h when outlet temperature equals 20  C. At the end of experiment outlet temperature reaches 24  C. Temperature would be in reality even lower, because of reduced mass flow after solidification part.

3.3.6. Accumulated heat Correctness of measurements may be verified with calculation of stored/released heat as is done in Eq. (5). Amount of stored/ released heat from experiment and from calculation for different cases is presented in Fig. 14 and in Table 3. ‘Losses subtracted’ stands for calculation where heat flow due to losses was subtracted from measured thermal power (Eq. (4)). ‘Raw’ stands for calculation where second term in Eq. (4) is omitted. Experimental values, when plates were in TES and with losses subtracted, are slightly higher than calculated values. Contrary, they are slightly lower for empty storage (cases 1, 2, 3). The deviation between experimental  losses subtracted (Qexp) and theoretical heat (Qcalc) is calculated as (QexpQcalc)/Qcalc and is less than 10% in all cases.

Fig. 11. Thermal power for different distance between plates (Tin¼[40] C and Tinit¼[20] C, red and blue curve; Tin¼[30] C and Tinit¼[17] C, orange and cyan curve). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149


Fig. 12. Thermal power for different number of plates.

Fig. 13. Results for measurements with varying inlet temperature.

Thermal capacity of empty storage is just below 40 kJ/K and this amount, multiplied by temperature difference for a specific case, is subtracted from the amount of accumulated/released heat. In Table 4 are therefore presented values only for plates, either 30 or 15 plates for different temperature ranges. Given are mean values if more measurements were carried out for the same temperature span. As may be seen the amount of accumulated/released heat for 30 plates is twice as big as for 15 plates.

To illustrate applicability of proposed storage unit, following is assumed: heat storage goes through full thermal cycle every day; the heating period lasts 7 months and cooling period 5 months. Based on these assumptions, estimations project that annual energy consumption will be reduced for 690 kWh, of which 420 kWh will be at the expense of heating (assuming daily energy reduction of 7.2 MJ as in Table 4) and 270 kWh at the expense of cooling (assuming daily energy reduction of 6.0 MJ). Presumed Table 3 Data for different cases.

Fig. 14. Amount of accumulated/released heat for different cases.


# of plates

Tin ( C)

Tinit ( C)

Deviation (%)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

0 0 0 15 15 15 15 15 15 15 15 30 30 30 30 30 30 30 30 30

6 21 11 16 30 10 40 30 16 40 10 30 5 45 10 35 16 30 16 30

24 42 28 25 18 34 20 18 24 22 32 16 25 8 35 16 23 15 24 17

6 5 2 2 6 2 4 5 5 4 1 3 8 4 6 3 1 9 3 10


E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149

Table 4 Accumulated/released heat without capacity of casing. Number of plates

Accumulated/released heat (MJ)

Tin¼16 C, Tinit¼25 C Tin¼30  C, Tinit¼16  C Tin¼10  C, Tinit¼35  C



5.1 6.0 7.2

2.7 2.8 3.5

quadratically as in Eq. (8). An important implication of these findings is that in the numerical models, which are designed to optimize PCM storage units, it is necessary to consider pressure drop of the entire system. For this reason, it is advisable to add another term into 2D numerical models, which captures pressure drops of all components except plates (which are governed by conservation equations).

Re ¼ temperatures are valid when storage unit presents one part of the ventilation system. In that case, air is taken from ambient and inlet temperature equals ambient temperature (30  C in case of cooling and 10  C in case of heating). Another example of use would be exclusively for heating or cooling of indoor air (no air from outside). The temperature difference in this case would be much smaller, thus smaller stored energy. Assuming an operation of storage tank in the range between 16  C and 23  C in case of cooling and between 22  C and 40  C in case of heating this corresponds to 110 kJ/ kg and 95 kJ/kg of stored energy from the DSC measurement. It follows that daily energy consumption reduces for approximately 0.8 kWh for heating and 0.9 kWh for cooling. On a yearly basis this corresponds to 166 kWh for heating and 138 kWh for cooling. For a passive house (for example) the upper limit in annual energy consumption for heating is 15 kWh/(m2a), same applies to cooling. For a room with 10 m2, this results in 300 kWh/a. In theory, this means that the heating and cooling demand for a passive house can be covered in its entirety. Above estimations apply to passive building, but if the loads were higher, then it would be necessary to add another storage tank. Since it works as a modular unit, this would not constitute major problem. 3.4. Pressure analysis In evaluation of TES power for air circulation needs to be taken into account, therefore pressure drop between the inlet and outlet was measured. It is in theory affected by mass flow rate, distance between the plates and number of plates, but in practice, influence of the latter two is negligible. Therefore, pressure drop depends more or less on mass flow rate as presented in Table 5. It varies between 2.1 mbar for the lowest mass flow and 13.3 mbar for the largest. The above observation is very important for the numerical simulations. In most numerical models authors assume 2D geometry and calculate pressure drop only for the length of plates. This could be justified by the fact that the rest of the system remains the same, therefore only change of pressure drop between the plates would be of interest. It turns out that velocity through the entire system is of great concern. If only pressure drop between the beginning and the end of plates was measured, then a linear relationship between pressure drop and flow rate would be expected, because of small Reynolds number (see Eq. (6)). However, due to additional air ducts components (bends, valves, etc.; minor losses) pressure drop increases quadratically as in Eq. (7). Moreover, flow in pipes is turbulent, consequently pressure drop increases

Table 5 Approximate values of pressure drop for different mass flows. Mass flow (kg/h)

Pressure drop (mbar)

38 43 60 92 100 107

2.1 2.8 4.4 12.6 12.8 13.3

rvD 1:1$0:3$2$0:016 ¼ ¼ 650 m 1:85$105 1 2 rv 2


L 1 2 rv D2


Dp ¼ K

Dp ¼ f


4. Conclusions Experimental investigation of a full thermal cycling of latent heat storage unit for space heating and cooling was carried out. 30 plates filled with paraffin, with a melting temperature of about 22  C, were positioned into storage unit. Velocity measurements in different channels and flow visualization were performed. The results confirm that the perforated plate was well designed and has provided uniform flow. Analysis of the thermal response to various boundary conditions was executed. Measurements revealed significant impact of inlet temperature on the duration of melting or freezing. With the same initial temperature and for inlet temperatures 45  C, 35  C and 30  C, melting times resulted in 8 h, 12.5 h and 18.5 h. Time decreased for 3.5 h in case of solidification, if inlet temperature decreased from 9  C to 4  C. The data obtained indicate significant impact of mass flow. If it was increased by two times, then also thermal power in the melting area increased by the same factor (from 125 W to 220 W). Particular attention was paid to the total exchanged energy between air and PCM. Thermal storage was well insulated, nevertheless thermal losses needed to be taken into account. Otherwise, accumulated or released heat would be too large and not consistent with calculated heat. Deviation between accumulated and calculated heat was in average 30% if losses were not included. However, if they were, then deviation was in average 4%. The experiments demonstrate that in temperature range between 16  C and 30  C 1.8 kWh of cold is stored and between 35  C and 10  C 2 kWh of heat. Annual energy savings are 670 kWh in case of cooling/ heating of ambient air and 300 kWh in case of indoor air. One advantage of storage unit is its modular structure, which allows adding another storage tank if heating or cooling load exceeds supply. Measured was also pressure drop which varied from 2.1 mbar to 13.3 mbar. Its analysis demonstrates that the distance between the plates does not have significant impact on the electricity use for a fan, as most of the pressure drop is at the expense of additional air ducts components between the measuring points. The existence of these findings implies that the distance between plates could be smaller which would result in more compact TES. In addition, it is advisable to use pipes with larger diameter and fans with variable speed control. General conclusion from these results is that proposed local unit satisfactory operates under heating and cooling conditions. However, in most cases phase change takes place too slowly for given arrangement. TES would be further improved if plates’ thickness was reduced to 10 mm.

E. Osterman et al. / Applied Thermal Engineering 84 (2015) 138e149

Acknowledgements Financial support from the Slovenian Research Agency is gratefully acknowledged. The authors would like to extend their thanks to ZAE Bayern for its support and help in carrying out measurements.

References [1] EU, Directive 2010/31/EU of the European Parliament and of the Council, June 2010. [2] E. Osterman, V.V. Tyagi, V. Butala, N.A. Rahim, and U. Stritih. Rev. PCM Based Cool. Technol. Build.. 49:37e49. [3] U. Stritih, P. Novak, Thermal storage of solar energy in the wall for building ventilation, in: International IEA Workshop (Annex 17), Advanced Thermal Energy Storage Techniques, International Energy Agency (IEA), Lju, 2002, pp. 3e5 number april. [4] Kang Yanbing, Jiang Yi, Zhang Yinping, Modeling and experimental study on an innovative passive cooling system-NVP system, Energy Build. 35 (4) (May 2003) 417e425. [5] S. Takeda, K. Nagano, T. Mochida, K. Shimakura, Development of a ventilation system utilizing thermal energy storage for granules containing phase change material, Sol. Energy 77 (3) (September 2004) 329e338. [6] Zouhair Ait Hammou, Marcel Lacroix, A hybrid thermal energy storage system for managing simultaneously solar and electric energy, Energy Convers. Manag. 47 (3) (February 2006) 273e288. [7] Vadim Dubovsky, Gennady Ziskind, Ruth Letan, Analytical model of a PCM-air heat exchanger, Appl. Therm. Eng. 31 (16) (November 2011) 3453e3462. [8] U. Stritih, V. Butala, Experimental investigation of energy saving in buildings with PCM cold storage, Int. J. Refrig. 33 (8) (December 2010) 1676e1683. [9] Yoram Kozak, Boris Abramzon, Gennady Ziskind, Experimental and numerical investigation of a hybrid PCM-air heat sink, Appl. Therm. Eng. 59 (1e2) (September 2013) 142e152. [10] J.R. Turnpenny, D.W. Etheridge, D.A. Reay, Novel ventilation system for reducing air conditioning in buildings. Part II: testing of prototype, Appl. Therm. Eng. 21 (12) (August 2001) 1203e1217. [11] V. Antony Aroul Raj, R. Velraj, Heat transfer and pressure drop studies on a PCM-heat exchanger module for free cooling applications, Int. J. Therm. Sci. 50 (8) (August 2011) 1573e1582. [12] Pedro D. Silva, L.C. Gonçalves, L. Pires, Transient behaviour of a latent-heat thermal-energy store: numerical and experimental studies, Appl. Energy 73 (1) (September 2002) 83e98. [13] Miroslaw Zukowski, Mathematical modeling and numerical simulation of a short term thermal energy storage system using phase change material for heating applications, Energy Convers. Manag. 48 (1) (January 2007) 155e165. [14] A.H. Mosaffa, L. Garousi Farshi, C.A. Infante Ferreira, M.A. Rosen, Energy and exergy evaluation of a multiple-PCM thermal storage unit for free cooling applications, Renew. Energy 68 (August 2014) 452e458.


[15] Belen Zalba, Jose M. Marin, Luisa F. Cabeza, Harald Mehling, Free-cooling of buildings with phase change materials, Int. J. Refrig. 27 (8) (December 2004) 839e849. [16] Adeel Waqas, S. Kumar, Thermal performance of latent heat storage for free cooling of buildings in a dry and hot climate: an experimental study, Energy Build. 43 (10) (October 2011) 2621e2630. [17] Ana Lazaro, Pablo Dolado, Jose M. Marin, Belen Zalba, PCM-air heat exchangers for free-cooling applications in buildings: experimental results of two realscale prototypes, Energy Convers. Manag. 50 (3) (March 2009) 439e443. [18] Pablo Dolado, Ana Lazaro, Jose M. Marin, Belen Zalba, Characterization of melting and solidification in a real-scale PCM-air heat exchanger: experimental results and empirical model, Renew. Energy 36 (11) (November 2011) 2906e2917. [19] W. Saman, F. Bruno, E. Halawa, Thermal performance of PCM thermal storage unit for a roof integrated solar heating system, Sol. Energy 78 (2) (February 2005) 341e349. [20] M. Labat, J. Virgone, D. David, F. Kuznik, Experimental assessment of a PCM to air heat exchanger storage system for building ventilation application, Appl. Therm. Eng. 66 (1e2) (May 2014) 375e382. [21] Pavel Charv at, Lubomír Klimes, Milan Ostrý, Numerical and experimental investigation of a PCM-based thermal storage unit for solar air systems, Energy Build. 68 (January 2014) 488e497. [22] Jose M. Marin, Belen Zalba, Luisa F. Cabeza, Harald Mehling, Improvement of a thermal energy storage using plates with paraffin-graphite composite, Int. J. Heat Mass Transf. 48 (12) (June 2005) 2561e2570. [23] Thomas Haussmann, Hannah Neumann, Bagdat Oral, and Peter Schossig. Central PCM storage in HVAC systems. In Innostock 2012, The 12th International Conference on Energy Storage, pages INNOeSPe119. [24] Saso Medved, Ciril Arkar, Correlation between the local climate and the freecooling potential of latent heat storage, Energy Build. 40 (4) (2008) 429e437. [25] Christoph Rathgeber, Laia Miro, Luisa F. Cabeza, and Stefan Hiebler. Measurement of enthalpy curves of phase change materials via DSC and t-history: when are both methods needed to estimate the behaviour of the bulk material in applications? 596:79e88. [26] Michal Pomianowski, Per Heiselberg, Yinping Zhang, Review of thermal energy storage technologies based on PCM application in buildings, Energy Build. 67 (December 2013) 56e69. [27] Eva Günther, Stefan Hiebler, Harald Mehling, Robert Redlich, Enthalpy of phase change materials as a function of temperature: required accuracy and suitable measurement methods, Int. J. Thermophys. 30 (4) (August 2009) 1257e1269. [28] Christoph Rathgeber, Henri Schmit, Peter Hennemann, Stefan Hiebler, Calibration of a T-history calorimeter to measure enthalpy curves of phase change materials in the temperature range from 40 to 200 C, Meas. Sci. Technol. 25 (3) (March 2014) 035011. [29] E. Halawa, W. Saman, Thermal performance analysis of a phase change thermal storage unit for space heating, Renew. Energy 36 (1) (January 2011) 259e264. [30] Pablo Dolado, Ana Lazaro, Jose M. Marin, Belen Zalba, Characterization of melting and solidification in a real scale PCM-air heat exchanger: numerical model and experimental validation, Energy Convers. Manag. 52 (4) (April 2011) 1890e1907.