- Email: [email protected]
PCM thermal storage system for ‘free’ heating and cooling of buildings
Accepted Manuscript Title: PCM thermal storage system for ’free’ heating and cooling of buildings Author: E. Osterman V. Butala U. Stritih PII: DOI: Reference:
S0378-7788(15)00311-4 http://dx.doi.org/doi:10.1016/j.enbuild.2015.04.012 ENB 5804
To appear in:
ENB
Received date: Revised date: Accepted date:
12-1-2015 10-4-2015 11-4-2015
Please cite this article as: E. Osterman, V. Butala, U. Stritih, PCM thermal storage system for ’free’ heating and cooling of buildings, Energy & Buildings (2015), http://dx.doi.org/10.1016/j.enbuild.2015.04.012 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
PCM thermal storage system for ’free’ heating and cooling of buildings E. Ostermana , V. Butalaa , U. Stritiha,∗ of Mechanical Engineering, University of Ljubljana, Askerceva 6, 1000 Ljubljana, Slovenia
ip t
a Faculty
Abstract
an
us
cr
In order to reduce energy consumption in buildings, a solution using phase change materials (PCMs) as thermal energy storage (TES) is presented. During summer nights, cold is stored and delivered during the day to reduce cooling load, whereas in winter, heat from solar air collector is stored for heating during morning and evening hours. Proposed is a stand-alone unit suitable for offices, which consists of plates filled with paraffin RT22HC. The objective of the paper is to examine the functioning of the suggested TES system on an annual basis and to explore the feasibility of using it for both, cooling and heating. First numerical model is set up, which assumes 2D geometry and evaluates thermal response, calculations are carried out in Fluent. Then validation of simulations with experimental results is established. Finally the feasibility of storage unit on an annual basis is demonstrated. Keywords: PCM, thermal energy storage, plates, free cooling, heating, solar air collector
M
1. Introduction
Ac ce
pt
ed
In developed countries buildings account for a 20–40% of the total final energy consumption, in the European Union this number is at 40% [1]. Therefore one of the priorities of the EU is to minimize the energy consumed by buildings. The EPB Directive states that it is necessary to choose alternative solutions for heating and cooling and one option is thermal energy storage (TES) using phase change materials (PCMs). Latent heat storage has been gaining importance in recent years and literature review shows a variety of approaches, which are summarized in the papers written by Farid et al. [2], Soares et al. [3], Zhang et al. [4] and Osterman et al. [5]. These storage systems can be integrated into the building’s envelope (i.e., walls, roofs, and floors) and are regarded as passive systems [6, 7], since no additional energy is required for their operation. On the other hand, active systems need additional energy and they can be integrated into HVAC or operate as stand alone units. The most common forms of PCM’s encapsulation are plates [8, 9, 10], capsules [11, 12], shell and tube [13, 14, 15]. The idea is to utilize natural sources within stand-alone active system. Outdoor cold during summer nights is stored and supplied to the indoor environment during the day when the cooling load increases. For the heating requirements, energy from the sun can be exploited in conjunction with solar air collectors as in [16]. It happens often that heating during the day is not needed, because of high energy gains. Heating load increases after sunset, so in such cases day-cycle TES can be advantageously used. When the temperature of air from the air solar collector (Fig.1 marked with 1) in the winter time exceeds the outside temperature, air is taken from collector and it flows partially into building (2) and partially into TES (3) where it releases heat to PCM plates. After that it flows out of TES (5) into the surroundings. ∗ Corresponding
author Email address: [email protected], +386 1 4771 408 (E. Osterman)
Preprint submitted to Energy and Buildings: Special issue
18. junij 2015
Page 1 of 15
us
Slika 1: Conceptual design of a storage unit.
cr
ip t
2
Ac ce
pt
ed
M
an
Energy is stored and later air is taken from outside (4), it flows through TES and is supplied to the room (6). Even though previous studies have dealt with this topic, a key limitation was that they mainly addressed operation of TES for either heating or cooling, but none of the studies proposed a system that would unite these two options. Nonetheless, with our concept (Fig.1) it would be possible to further improve output of TES unit and utilize low exergy sources throughout the year. Moreover, research carried out so far was mainly focused on larger, central systems, studies on smaller systems are still lacking. The main objective of this paper is to examine closely the functioning of the suggested system on an annual basis and to explore the feasibility of using it for office cooling as well as for heating. The focus of recent research has been on thermal response of TES, however, to the authors’ best knowledge, very few publications [17, 18, 19] are available that discuss operation over a longer period of time. Deeper investigation is very important because it is the only clear-cut way to realistically assess performance of TES in buildings. This is necessary because the functioning of PCM applications is dependent on the condition of the PCM, which is variable over time. The capacity of a PCM differs between different times of a day. Using simulations over a year, an evaluation can be made whether the PCM application functions effectively enough, and whether the energy savings in buildings are of any significance. In order to achieve the previously specified goal, we first present experimental setup, then introduce numerical CFD model, establish comparison with experimental results and finally demonstrate the feasibility of operation in winter and summer. 2. Experimental setup
In order to verify the validity of the numerical model, an experimental rig for testing storage unit’s thermal response was set up. Casing was made of 0,8 cm thick PMMA and the external dimensions were 77 cm x 67 cm x 42 cm. 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 [20]. Air gap between plates was 0,8 cm. Outer dimensions of plates were 30 cm x 45 cm x 1,5 cm and they were vertically positioned in the storage tank (the longer side parallel to the flow) as in Fig.2. The actual volume of PCM is smaller, because plates have edges and the surface is not flat. So the dimensions, that are also used in the numerical model, are: 27.0 cm x 42.5 cm x 1.29 cm. Measured average mass of filled plate was 1361 g,
Page 2 of 15
cr
ip t
3
us
Slika 2: Vertical configuration of the plates installed in the mounting (holder).
an
weight of paraffin in the plate was 1003 g and volume of each plate was 1.48 l. Approximately 12% 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. It follows that smaller density of PCM needs to be considered in the numerical model. Otherwise, mass of PCM in a domain would be too large, thus leading to larger heat accumulation and erroneous results.
Ac ce
pt
ed
M
Analysis of: PCM’s specific heat dependence on the temperature, thermal conductivity and density was performed in order to gain data needed for the numerical model. DSC step (with temperature steps of 1 °C) and T-History measurements were performed and details about the DSC and the T-History calorimeter are given in [21]. Results are shown in Fig.3a, where 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 no subcooling or hysteresis. Thermal conductivity was measured in the solid state and density in the liquid state. Results of measurements are given in Table 1.
(a) Enthalpy curves of RT22HC measured via DSC (solid line) and T-History (dashed line) compared with manufacturer’s data (dot-dashed line).
(b) Apparent specific heat.
Slika 3: Enthalpy and specific heat of RT22HC.
Page 3 of 15
4
RT22HC
ρ (kg/m3)
λ (W/mK)
∆h (kJ/kg)
772.9±0.1 @ 30 °C
0.19±0.01 @ 10 °C
181±9 (between 16 °C and 30 °C)
Tabela 1: Thermal properties of RT22HC
3. Numerical model
M
an
us
cr
ip t
In order to define thermal response of a storage unit, two important processes need to be considered, namely the heat transfer from air to paraffin, and the heat transfer within the paraffin. The first one can be calculated with heat transfer coefficient obtained by means of correlations [22, 23, 24]. The disadvantage of this approach is that these correlations are only applicable when either plate’s temperature or heat flux remains constant. Moreover, they are valid only for developed flow. Temperature of plate could be considered as relatively constant, but this is not the case, since the phase change takes place within a specified temperature range. Also undeveloped part of flow is not negligible, therefore the problem is tackled with finite volume method, where both, flow and temperature field, are being solved. Numerical model of TES presented above, can be reduced to a 2D problem due to negligible changes in z-direction. Moreover, if we assume that the distribution of flow between plates is uniform, only one plate and one air channel can be considered. Going further, plate and air channel are symmetric, so only half of each of them needs to be included in the domain. Dimensions and the schematic diagram of the 2D computational domain used in this investigation are depicted in Fig.4.
Slika 4: Schematics of a numerical domain.
pt
ed
3.1. Governing equations In order to mathematically describe the process inside the storage unit, additional assumptions have to be met for the air as well as for the PCM: Air: • The range of velocity considered in the analysis is such that the changes in density are negligible (Mach number is below 0.2) therefore the air domain is considered as incompressible.
Ac ce
• We are dealing with Newtonian fluid which has constant density, specific heat, thermal conductivity and viscosity. • Viscous dissipation is negligible. • Since, in comparison with their spacing and width, the plates extend for a very long distance in the z direction, all locations in this direction appear essentially identical to one another. PCM:
• The PCM is homogenous and isotropic. • No distinction between PCM properties in solid and liquid phase is present. Values are assumed to be constant. • Natural convection in the PCM domain is not accounted for.
Page 4 of 15
3.2
5
CFD modeling
The steady equations governing the flow are solved only for the air and two unsteady energy equations are applied to the entire computational domain (PCM and air). The two-dimensional equations for conservation of mass, momentum and energy for air are expressed in Eqs.1-4. (1)
2 ∂ 2 vx ∂vx ∂vx ∂p ∂ vx + ρ a vx + vy =− + µa ∂x ∂y ∂x ∂x2 ∂y 2
(2)
ρa cp,a
∂Ta + ~v · ∇Ta ∂t
= ka ∇2 Ta + S
cr
2 ∂vy ∂ 2 vy ∂vy ∂p ∂ vy ρa v x + + vy =− + µa ∂x ∂y ∂y ∂x2 ∂y 2
ip t
∂vx ∂vy + =0 ∂x ∂y
(3) (4)
∂TP CM + ~v · ∇TP CM ∂t
= kP CM ∇2 TP CM
an
ρP CM cp,P CM,app
us
As previously written, there is no convection in PCM, therefore only unsteady energy equation is being solved (Eq. 5). (5)
3.2. CFD modeling
dQ =m ˙ a cp,a (Ta,out − Ta,in ) dt
(6)
ed
P =
M
With these equations air temperature at the outlet can be determined and thus thermal response of TES. Of the main interest is thermal power and amount of stored heat/cold, respectively. Former is calculated by Eq. 6.
Ac ce
pt
For the selected geometry grid independence study was carried out by comparing outlet temperature for the selected velocity and inlet temperatures. Size of the final mesh was 1000 x 30 for air and 1000 x 20 for PCM. Boundary conditions were selected so, that they best represent experimental work or operation in real situation. Thus, air velocity and temperature were prescribed at the inlet, static pressure was set to 0 Pa at the outlet and adiabatic wall was assumed at the ’wall’ (as marked in Fig.4). Based on the Reynolds number with the maximum value of 325 (Eq.7), laminar flow regime is expected, therefore laminar model was used for heat transfer fluid. Uniform temperature of the entire TES unit was taken as initial condition. The system of equations has thus been closed and solving was done in Ansys Fluent software, which employs the finite volume method for solving the mass, momentum and energy conservation equations [25]. Re =
ρvD 1.1 · 0.3 · 2 · 0.008 = = 325 µ 1.85 · 10−5
(7)
Relationship between enthalpy of PCM and temperature is known from DSC, hence latent heat of fusion can be taken into account via apparent specific heat of PCM. Because melting and solidification curves do not differ much, only one curve, representing both of them, can be used in simulations (Fig.3). As was mentioned above, density of PCM needs to be smaller, therefore apparent PCM density was calculated as ρP CM,app = ρP CM · vol% + ρa · (1 − vol%). Thermal properties of air and PCM used in simulations are presented in Table 2.
Page 5 of 15
3.3
6
Validation ρ (kg/m3 )
k (W/mK)
cp (kJ/kgK)
RT22HC
677
0.18
from the curve
Air
1.1
0.024
1005
µ (kg/ms) /
1.789 · 10-5
Tabela 2: Thermal properties of RT22HC and air used in simulation
3.3. Validation
us
cr
ip t
In order to confirm the numerical model it was necessary to carry out simulations with the same initial and boundary conditions as in experiment. Because TES unit was not completely insulated in the experiment, thermal losses emerged. They were included in equation for air (Eq.4) through additional term S, which is defined in Eq. 8. U A value is determined from the experiment, Tlab is temperature in the laboratory where experiment was carried out, Vchannels is volume of air channels. Temperature of storage unit was defined as average temperature between inlet and outlet. T −T U A · Tlab − a,in 2 a.out (8) S= Vchannels
an
Additional thermal mass due to casing, holder and PMMA was accounted for with increased specific heat of PCM. This means that curve in Fig.3 shifts upwards. 3.4. Principle of yearly operation
Ac ce
pt
ed
M
As already mentioned, TES is intended for heating and cooling, therefore two modes of operation need to be distinguished, ie. in summer and winter. Ljubljana, Slovenia is taken as a reference city (but it could be done for any other location as well) for which it is assumed that the heating season lasts from October to April and in the rest of the year, office needs to be cooled. In summer, during the night (from midnight to 7 a.m.), cold is being accumulated in the TES, if the average temperature of PCM (TP CM,avg ) exceeds inlet air temperature (Ta,in ) for at least 2 K. At that time air flow amounts to maximal flow (qv = qv,max ) and inlet temperature equals ambient temperature (Ta,in = Tamb ), otherwise air flow equals zero (qv = 0). During the day (from 7 a.m. to 8 p.m.) presumption will be, that minimal ventilation needs to be ensured due to presence of people, so air flow will be minimal (qv = qv,min ). But this is only valid when Ta,in exceeds TP CM,avg for at least 2 K. If Ta,in is up to 2 K higher than TP CM,avg , then air flow equals 0, because the temperature difference for heat transfer is too small, so it would not be reasonable to run the fan. Also there is no air flow if Ta,in falls behind TP CM,avg . At that time air directly enters into office. Operation in winter is slightly more complicated, because air can be taken from either ambient or from solar air collector. Unlike summer, TES unit does not operate at night, only during the day (from 7 a.m. to 8 p.m.). If TP CM,avg exceeds Tamb , then ambient air flows through TES with minimal flow, so the following condition Ta,in = Tamb holds true. At the same time the temperature of air from collector (Ta,col ) is monitored and if it is at least 2 K higher than TP CM,avg and also than Tamb , then air from collector flows through TES and not air from ambient. In that case TES gets charged with heat, so air flow increases to maximum. But flow rate through collector (qv,col ) does not equal flow through TES (qv ), because part of air from collector goes directly into office, so the following relation exists: qv = qv,col − qv,min . On the other hand, if these two conditions Ta,col < TP CM,avg + 2K and Ta,col > TP CM,avg − 2K are met, then air directly flows into office and TES does not operate. The same applies in the case when Tamb exceeds TP CM,avg . Model of solar air collector is presented in subsection 3.5. We are interested in operation throughout the year, but from the perspective of simulations this would be time consuming. Therefore the following will be assumed: the first week of each month is
Page 6 of 15
3.5
7
Description of the solar collector
pt
ed
M
an
us
cr
ip t
representative for the whole month. Data for ambient temperature are taken from the test reference year for Ljubljana and combined into temperature profile, which consists of 12 weeks.
(a) Winter.
(b) Summer.
Ac ce
Slika 5: Flow charts for winter and summer.
3.5. Description of the solar collector Calculation of air temperature at the outlet of solar collector was accomplished with TRNSYS software. Test reference year for Ljubljana serves as an input to module 561-2, which represents the solar collector without any special coatings. Its size is 2 m2 , collector is facing towards south and it is tilted for 60° from vertical. Other parameters set in the simulation are: ground reflectance 0.2, top loss convection coefficient 5,6 W/(m2 K), back loss convection coefficient 4,2 W/(m2 K). Air flow through collector was 75 m3 /h. 3.6. Evaluation of pressure drop Not only accumulated energy, but also power for air circulation needs to be taken into account in evaluation of TES performance. The pressure drop between the beginning and end of plates is calculated
Page 7 of 15
3.7
8
Implementation into Fluent
ip t
within simulations, but this pressure drop is negligible compared to the total pressure drop through such a system. This finding has important implication for the numerical modeling. In most numerical models authors focus only on plates and calculate pressure drop only for their length. But there are additional air ducts components (bends, valves, etc.; minor losses), where pressure drop increases quadratically as in Eq.9. Moreover, velocity in pipes is higher, so the flow is turbulent, consequently pressure drop (major losses) increases quadratically as in Eq.10. To evaluate TES accurately, it is crucial to examine pressure drop of the entire system when evaluating TES performance. For this reason, we suggest accounting all pressure drops, not only pressure drop due to plates. The main limitation of the proposed method is that pressure drop and characteristic of system can only be estimated with regard to components constituting it. Thus, according to our estimates relationship between fan consumption and air flow is given in Eq.11.
L1 2 ρv D2
Pf an = 0.0092 · qv1.961
us
∆p = f
cr
1 ∆p = K ρv 2 2
(9)
(10) (11)
4. Results
ed
M
an
3.7. Implementation into Fluent In the simulations, which were carried out for validation, additional term presented in Eq.8 was implemented into Fluent via UDF source term applied to air domain. Time step in these simulation was 30 s. Procedures described in subsection 3.4 were implemented into Fluent through script written in Scheme, which actually commanded Fluent. First, velocity field for minimum air flow was calculated and written to a file, then velocity field for maximum air flow was calculated and written to another file. Both of these were steady state calculations and they were succeeded by transient simulation, which embodied only energy equation. Data regarding flow field were read from the previously created two files according to the conditions presented previously. Used time step in yearly calculations was 300 s.
Ac ce
pt
4.1. Validation of a numerical model Figs. 6a and 6b give an example of outlet temperatures from experiment and simulation results, in the case with constant inlet temperature. In reality, inlet temperature varies with time, but for the purposes of validation, constant value is completely acceptable. In Fig.6a are depicted temperatures for solidification process, with inlet temperature set to 16 °C, initial temperature set to 24 °C and flow rate set to 78 m3 /h (conditions in experiment). Simulations were performed for different flow rates and specific heats (cp,shif t and Tm,shif t mean additional amount to the base value). Beside solidification, also melting was simulated and compared with experimental results (Fig.6b). Initial temperature was just below 16.5 °C, inlet temperature was 30 °C and flow rate was 78 m3 /h. Simulation results are in good agreement with experiment, however there is a slight difference in the end of the process. Curves indicate that melting or solidification ends faster in simulation than in experiment. This can be contributed to the non-uniform flow distribution between plates. Apparently some plates get less air, which hinders heat transfer. As a consequence, thermal response gets slightly distorted. Let us define relative temperature difference (∆Trel ) at every time step as a measure of agreement T −Ta,exp,out between simulations and experiment. It is characterized as : ∆Trel = a,out,sim . Accordingly, Ta,exp,out average value of relative temperature difference for solidification in the time interval 0 h - 18 h and for melting in the time interval 0 h - 6 h is less than 1.0%.
Page 8 of 15
9
Yearly performance
cr
ip t
4.2
(b) Melting.
us
(a) Solidification.
Slika 6: Curves from experiment and simulations.
an
4.2. Yearly performance
M
This subsection presents the results of TES unit in winter and summer, and in the end, energy savings on an annual basis are evaluated. In all simulations minimal volume flow equaled 40 m3 /h, whereas maximal flow equaled 75 m3 /h.
pt
ed
4.2.1. Winter Fig.7 depicts operation of TES in the first week of March. Temperatures (Ta,avg is the average air temperature in TES unit) are illustrated in upper subfigure and volume flow in lower subfigure. TES unit is not in operation in the beginning, so the temperature is constant and the flow is zero. Minimum ventilation starts at 7 a.m., so Ta,in = Tamb , TP CM,avg starts to decrease, the flow rate equals 40 m3 /h. When the temperature of the solar collector (Ta.col ) exceeds the average temperature of PCM (TP CM,avg ), operation is switched to charging mode, so flow rate changes to 35 m3 /h (75 m3 /h - 40 m3 /h). In the moment when TP CM,avg exceeds Ta,col , flow in TES unit goes back to 40 m3 /h. At 8 p.m. it shuts down and temperatures are again more or less constant, except Tin , which rises due to warmer PCM.
Ac ce
4.2.2. Summer For visual representation of temperatures and flow rate in August, the reader is referred to Fig.8. Air flow through TES at night is 75 m3 /h, then in the morning it shuts down, because of lower ambient air temperature than average PCM temperature. But when Tamb exceeds TP CM,avg TES is activated and air flows through it with 40 m3 /h. Such regime is maintained until evening or until TES gets discharged. At night charging process starts again if Tamb drops below TP CM,avg . 4.2.3. Yearly energy savings In addition to the annual thermal response of TES, also energy savings are of interest. Let us assume that ambient air needs to be heated up or cooled down to the desired room temperature. If this air is at least partly heated or cooled in the TES, then this part, presumably, represents energy savings Eaccu due to TES as defined in Eq. 12. Energy use for fan operation also needs to be considered, which is described by Eq.13. If the actual energy savings Esaving have to be determined, then energy consumption needs to be subtracted from accumulated energy as in Eq.14. It is necessary to point out, that these two (heat,
Page 9 of 15
10
Yearly performance
an
us
cr
ip t
4.2
Slika 7: Temperatures and volume flow in TES unit in the first week of March.
M
electricity) are, from exergy point of view, two different quantities despite the same units. But for the purposes of estimation of yearly savings, this premise will be good enough. ˆ Eaccu = qm,a cp,a (Ta,out − Ta,in ) dt (12)
ed
ˆ
Econsum =
qv,a ∆p dt η
pt
Esaving = Eaccu − Econsum
(13) (14)
Ac ce
Fig.9 depicts accumulated heat and electricity use in winter (Fig.9a) and in summer (Fig.9b). From Fig.9a it may be first observed that electricity consumption for such a small flow is negligible. It is worth nothing again that, in this case, the flow through collector amounts to 75 m3 /h, whereas through TES to 35 m3 /h, since one part of air flows directly into the office. As expected, the smallest accumulated heat is in November and December (7 kWh), due to the smallest solar radiation. Accumulated heat increases in spring and it reaches its peak in March, when approximately 20 kWh is stored. In April, accumulated heat is reduced, as the ambient temperature increases, consequently TES does not go through a complete cycles. Let us compare March’s result with the theoretical maximum stored heat. Assuming that about 50 Wh is stored in one plate in the temperature range from 15 °C to 30 °C, then for 30 plates this gives 1.5 kWh in one cycle. This would result in 45 kWh if TES was daily completely charged and discharged. In this manner the reasonableness of the simulations is confirmed. In the case of summer, the flow through the storage tank is actually 75 m3 /h, therefore electricity use for fan operation exceeds the one in winter. Electricity represents a bit more than 10% of the accumulated cold. The least cold is accumulated during the transition period (September, 7 kWh), as the outdoor temperature is low enough so that ambient air directly flows into the building. Ambient
Page 10 of 15
11
Yearly performance
an
us
cr
ip t
4.2
Slika 8: Temperatures and volume flow in TES unit in the first week of August.
M
temperature in the warmest months is much higher, and as a result, storage tank goes through complete cycles. Thus, the amount of accumulated cold in July and August equals 17 kWh.
Ac ce
pt
ed
4.2.4. Savings in building To illustrate presented results, a simulation of energy use in buildings was performed. As a reference, office with 4 m x 3 m x 2.8 m was chosen and the following was assumed: it has one wall adjacent to the ambient, facing south, size of windows is 3 m2 , shading (external shading factor = 0.9) turns on when total radiation exceeds 150 W/m2 on a vertical wall, on average one person is present form 7 a.m. to 8 p.m., computer and lights are on in the same time period, ventilation amounts to 1.2 h-1 , infiltration to 0.2 h-1 , maximum temperature should not exceed 26 °C in summer and in winter temperature in office should be kept above 20 °C during the day and above 17 °C during the night, internal accumulation is estimated to be 500 kJ/K. In order to calculate hourly heating and cooling load we embedded these data into Trnsys and for the results, see Fig.10a. The bar graph represents accumulated heat, the use of electricity, heat loads and ventilation losses. The results show that the biggest part of the heat load is at the expense of ventilation losses. Comparing accumulated heat to heat load, then it can be seen that the former accounts 8% of the heat load during winter months. In early autumn and in spring TES seems to be more promising, as almost all (March) or all (April, October) heat load can be covered with stored heat. Energy savings in entire winter will be approximately 84 kWh under given circumstances. Similar analysis can be carried out for summer (Fig.10b). It can be seen that the ventilation does not represent marginal part of cooling load anymore, as the most gains are at the expense of solar radiation. In all cases the cooling load exceeds accumulated cold, which means that energy use will be only partially reduced. Still, energy savings in summer will be approximately 58 kWh. The investment cost for the presented system is approximately 1140€, allocation of costs is given in table 3. The largest share of the costs represent plates and air solar collector, whose costs are 850€. Conventional system (for example split system air conditioning) suitable for office would cost around 500€. From these numbers it follows that TES is twice as expensive as a conventional system. However,
Page 11 of 15
(b) Monthly accumulated cold and electricity use for summer, qv,max = 75 m3 /h, 15, cm thick plates, 0,8 cm air gap, no shift in melting point.
us
(a) Monthly accumulated heat and electricity use for winter, qv,max = 35 m3 /h, 1,5 cm thick plates, 0,8 cm air gap, no shift in melting point.
cr
ip t
12
Slika 9: Monthly operation of TES
Ac ce
pt
ed
M
an
costs during the operation stage in the case of TES are lower, thus making TES promising alternative. Moreover, the costs would decrease, if TES was mass produced or if the state provided subsidies.
(a) Monthly accumulated heat and electricity use for winter, building’s heating load and ventilation losses, qv,max = 35 m3 /h, 1,5 cm thick plates, 0,8 cm air gap, no shift in melting point.
(b) Monthly accumulated heat and electricity use for summer, building’s cooling load and ventilation gains, qv,max = 75 m3 /h, 1,5 cm thick plates, 0,8 cm air gap, no shift in melting point.
Slika 10: Temperatures and volume flow in TES unit.
5. Conclusions Experimental and numerical investigations of a latent heat storage unit for space heating and cooling was carried out. It consisted of 30 plates filled with paraffin RT22HC. Thermal response of TES to the temperature step function was monitored in the experimental part. In the numerical part 2D model was
Page 12 of 15
Equipment
Price (€)
30 plates
600
Fan
60
Ducts
20
Insulation
10
Motor unit
80
Solar air collector
250
Speed regulator for a fan
80
Programmable clock
40
Together
1140
cr
Tabela 3: Breakdown of the investment costs.
ip t
13
M
an
us
set up in Fluent and calculations were executed for such conditions as they appeared in the experiment. A comparison of the results demonstrates very good agreement between experiment and simulations, which confirms adequateness of the numerical model. The developed model was used for evaluation of TES on a yearly basis. Calculations reveal that in winter the largest savings are in March, as larger quantities of heat are available at that time. In summer, the maximum amount of accumulated cold is in July and August, because of larger temperature fluctuations between day and night. Consequently, TES goes through complete cycles and stored energy increases. In comparison to heating and cooling loads in an office it could be seen, that accumulated heat/cold is relatively small. Nevertheless, energy consumption in an office can be annually reduced for approximately 142 kWh. Acknowledgements
References
ed
Financial support from the Slovenian Research Agency is gratefully acknowledged.
pt
[1] EU. Directive 2010/31/EU of the European Parliament and of the Council, June 2010.
Ac ce
[2] Mohammed M Farid, Amar M Khudhair, Siddique Ali K Razack, and Said Al-Hallaj. A review on phase change energy storage materials and applications. Energy Conversion and Management, 45(9-10):1597–1615, June 2004. [3] N. Soares, J.J. Costa, A.R. Gaspar, and P. Santos. Review of passive PCM latent heat thermal energy storage systems towards buildings energy efficiency. Energy and Buildings, 59:82–103, April 2013. [4] Yinping Zhang, Guobing Zhou, Kunping Lin, Qunli Zhang, and Hongfa Di. Application of latent heat thermal energy storage in buildings: State-of-the-art and outlook. Building and Environment, 42(6):2197–2209, June 2007. [5] E. Osterman, V.V. Tyagi, V. Butala, N.A. Rahim, and U. Stritih. Review of PCM based cooling technologies for buildings. Energy and Buildings, 49:37–49, June 2012. [6] Tiago Silva, Romeu Vicente, Nelson Soares, and Victor Ferreira. Experimental testing and numerical modelling of masonry wall solution with PCM incorporation: A passive construction solution. Energy and Buildings, 49:235–245, June 2012.
Page 13 of 15
14 [7] Shazim Ali Memon. Phase change materials integrated in building walls: A state of the art review. Renewable and Sustainable Energy Reviews, 31:870–906, March 2014. [8] Belen Zalba, Jose M. Marin, Luisa F. Cabeza, and Harald Mehling. Free-cooling of buildings with phase change materials. International Journal of Refrigeration, 27(8):839–849, December 2004.
ip t
[9] Pablo Dolado, Ana Lazaro, Jose M. Marin, and Belen Zalba. Characterization of melting and solidification in a real-scale PCM-air heat exchanger: Experimental results and empirical model. Renewable Energy, 36(11):2906–2917, November 2011.
cr
[10] Pavel Charvát, Lubomír Klimeš, and Milan Ostrý. Numerical and experimental investigation of a PCM-based thermal storage unit for solar air systems. Energy and Buildings, 68:488–497, January 2014. [11] Taha K. Aldoss and Muhammad M. Rahman. Comparison between the single-PCM and multi-PCM thermal energy storage design. Energy Conversion and Management, 83:79–87, July 2014.
us
[12] R. Parameshwaran and S. Kalaiselvam. Energy conservative air conditioning system using silver nano-based PCM thermal storage for modern buildings. Energy and Buildings, 69:202–212, February 2014.
an
[13] Vadim Dubovsky, Gennady Ziskind, and Ruth Letan. Analytical model of a PCM-air heat exchanger. Applied Thermal Engineering, 31(16):3453–3462, November 2011.
M
[14] N.H.S. Tay, M. Belusko, and F. Bruno. Designing a PCM storage system using the effectivenessnumber of transfer units method in low energy cooling of buildings. Energy and Buildings, 50:234–242, July 2012.
ed
[15] Trp Anica. An experimental and numerical investigation of heat transfer during technical grade paraffin melting and solidification in a shell-and-tube latent thermal energy storage unit. Solar Energy, 79(6):648–660, December 2005.
pt
[16] U Stritih and P Novak. Thermal Storage Of Solar Energy In The Wall For Building Ventilation. In International IEA Workshop (Annex 17), Advanced Thermal Energy Storage Techniques, number april, pages 3–5, Lju, 2002. International Energy Agency (IEA).
Ac ce
[17] B.L. Gowreesunker, S.A. Tassou, and M. Kolokotroni. Coupled TRNSYS-CFD simulations evaluating the performance of PCM plate heat exchangers in an airport terminal building displacement conditioning system. Building and Environment, 65:132–145, July 2013. [18] Maciej Jaworski. Thermal performance of building element containing phase change material (PCM) integrated with ventilation system - an experimental study. Applied Thermal Engineering, 70(1):665– 674, September 2014. [19] Fabien Rouault, Denis Bruneau, Patrick Sebastian, and Jerome Lopez. Experimental investigation and modelling of a low temperature PCM thermal energy exchange and storage system. Energy and Buildings, 83:96–107, November 2014. [20] RUBITHERM GmbH, 2013.
[21] 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? Thermochimica Acta, 596:79–88, November 2014.
Page 14 of 15
15 [22] Ming Liu, Frank Bruno, and Wasim Saman. Thermal Performance Of A Pcm Thermal Storage Unit. pages 2766–2771. Springer Berlin Heidelberg, January 2009. [23] Hed and R. Bellander. Mathematical modelling of PCM air heat exchanger. Energy and Buildings, 38(2):82–89, 2006.
Ac ce
pt
ed
M
an
us
cr
[25] ANSYS. ANSYS FLUENT Theroy Guide 14.0, November 2011.
ip t
[24] Jacques Bony and Stéphane Citherlet. Numerical model and experimental validation of heat storage with phase change materials. Energy and Buildings, 39(10):1065–1072, October 2007.
Page 15 of 15
S0378-7788(15)00311-4 http://dx.doi.org/doi:10.1016/j.enbuild.2015.04.012 ENB 5804
To appear in:
ENB
Received date: Revised date: Accepted date:
12-1-2015 10-4-2015 11-4-2015
Please cite this article as: E. Osterman, V. Butala, U. Stritih, PCM thermal storage system for ’free’ heating and cooling of buildings, Energy & Buildings (2015), http://dx.doi.org/10.1016/j.enbuild.2015.04.012 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
PCM thermal storage system for ’free’ heating and cooling of buildings E. Ostermana , V. Butalaa , U. Stritiha,∗ of Mechanical Engineering, University of Ljubljana, Askerceva 6, 1000 Ljubljana, Slovenia
ip t
a Faculty
Abstract
an
us
cr
In order to reduce energy consumption in buildings, a solution using phase change materials (PCMs) as thermal energy storage (TES) is presented. During summer nights, cold is stored and delivered during the day to reduce cooling load, whereas in winter, heat from solar air collector is stored for heating during morning and evening hours. Proposed is a stand-alone unit suitable for offices, which consists of plates filled with paraffin RT22HC. The objective of the paper is to examine the functioning of the suggested TES system on an annual basis and to explore the feasibility of using it for both, cooling and heating. First numerical model is set up, which assumes 2D geometry and evaluates thermal response, calculations are carried out in Fluent. Then validation of simulations with experimental results is established. Finally the feasibility of storage unit on an annual basis is demonstrated. Keywords: PCM, thermal energy storage, plates, free cooling, heating, solar air collector
M
1. Introduction
Ac ce
pt
ed
In developed countries buildings account for a 20–40% of the total final energy consumption, in the European Union this number is at 40% [1]. Therefore one of the priorities of the EU is to minimize the energy consumed by buildings. The EPB Directive states that it is necessary to choose alternative solutions for heating and cooling and one option is thermal energy storage (TES) using phase change materials (PCMs). Latent heat storage has been gaining importance in recent years and literature review shows a variety of approaches, which are summarized in the papers written by Farid et al. [2], Soares et al. [3], Zhang et al. [4] and Osterman et al. [5]. These storage systems can be integrated into the building’s envelope (i.e., walls, roofs, and floors) and are regarded as passive systems [6, 7], since no additional energy is required for their operation. On the other hand, active systems need additional energy and they can be integrated into HVAC or operate as stand alone units. The most common forms of PCM’s encapsulation are plates [8, 9, 10], capsules [11, 12], shell and tube [13, 14, 15]. The idea is to utilize natural sources within stand-alone active system. Outdoor cold during summer nights is stored and supplied to the indoor environment during the day when the cooling load increases. For the heating requirements, energy from the sun can be exploited in conjunction with solar air collectors as in [16]. It happens often that heating during the day is not needed, because of high energy gains. Heating load increases after sunset, so in such cases day-cycle TES can be advantageously used. When the temperature of air from the air solar collector (Fig.1 marked with 1) in the winter time exceeds the outside temperature, air is taken from collector and it flows partially into building (2) and partially into TES (3) where it releases heat to PCM plates. After that it flows out of TES (5) into the surroundings. ∗ Corresponding
author Email address: [email protected], +386 1 4771 408 (E. Osterman)
Preprint submitted to Energy and Buildings: Special issue
18. junij 2015
Page 1 of 15
us
Slika 1: Conceptual design of a storage unit.
cr
ip t
2
Ac ce
pt
ed
M
an
Energy is stored and later air is taken from outside (4), it flows through TES and is supplied to the room (6). Even though previous studies have dealt with this topic, a key limitation was that they mainly addressed operation of TES for either heating or cooling, but none of the studies proposed a system that would unite these two options. Nonetheless, with our concept (Fig.1) it would be possible to further improve output of TES unit and utilize low exergy sources throughout the year. Moreover, research carried out so far was mainly focused on larger, central systems, studies on smaller systems are still lacking. The main objective of this paper is to examine closely the functioning of the suggested system on an annual basis and to explore the feasibility of using it for office cooling as well as for heating. The focus of recent research has been on thermal response of TES, however, to the authors’ best knowledge, very few publications [17, 18, 19] are available that discuss operation over a longer period of time. Deeper investigation is very important because it is the only clear-cut way to realistically assess performance of TES in buildings. This is necessary because the functioning of PCM applications is dependent on the condition of the PCM, which is variable over time. The capacity of a PCM differs between different times of a day. Using simulations over a year, an evaluation can be made whether the PCM application functions effectively enough, and whether the energy savings in buildings are of any significance. In order to achieve the previously specified goal, we first present experimental setup, then introduce numerical CFD model, establish comparison with experimental results and finally demonstrate the feasibility of operation in winter and summer. 2. Experimental setup
In order to verify the validity of the numerical model, an experimental rig for testing storage unit’s thermal response was set up. Casing was made of 0,8 cm thick PMMA and the external dimensions were 77 cm x 67 cm x 42 cm. 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 [20]. Air gap between plates was 0,8 cm. Outer dimensions of plates were 30 cm x 45 cm x 1,5 cm and they were vertically positioned in the storage tank (the longer side parallel to the flow) as in Fig.2. The actual volume of PCM is smaller, because plates have edges and the surface is not flat. So the dimensions, that are also used in the numerical model, are: 27.0 cm x 42.5 cm x 1.29 cm. Measured average mass of filled plate was 1361 g,
Page 2 of 15
cr
ip t
3
us
Slika 2: Vertical configuration of the plates installed in the mounting (holder).
an
weight of paraffin in the plate was 1003 g and volume of each plate was 1.48 l. Approximately 12% 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. It follows that smaller density of PCM needs to be considered in the numerical model. Otherwise, mass of PCM in a domain would be too large, thus leading to larger heat accumulation and erroneous results.
Ac ce
pt
ed
M
Analysis of: PCM’s specific heat dependence on the temperature, thermal conductivity and density was performed in order to gain data needed for the numerical model. DSC step (with temperature steps of 1 °C) and T-History measurements were performed and details about the DSC and the T-History calorimeter are given in [21]. Results are shown in Fig.3a, where 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 no subcooling or hysteresis. Thermal conductivity was measured in the solid state and density in the liquid state. Results of measurements are given in Table 1.
(a) Enthalpy curves of RT22HC measured via DSC (solid line) and T-History (dashed line) compared with manufacturer’s data (dot-dashed line).
(b) Apparent specific heat.
Slika 3: Enthalpy and specific heat of RT22HC.
Page 3 of 15
4
RT22HC
ρ (kg/m3)
λ (W/mK)
∆h (kJ/kg)
772.9±0.1 @ 30 °C
0.19±0.01 @ 10 °C
181±9 (between 16 °C and 30 °C)
Tabela 1: Thermal properties of RT22HC
3. Numerical model
M
an
us
cr
ip t
In order to define thermal response of a storage unit, two important processes need to be considered, namely the heat transfer from air to paraffin, and the heat transfer within the paraffin. The first one can be calculated with heat transfer coefficient obtained by means of correlations [22, 23, 24]. The disadvantage of this approach is that these correlations are only applicable when either plate’s temperature or heat flux remains constant. Moreover, they are valid only for developed flow. Temperature of plate could be considered as relatively constant, but this is not the case, since the phase change takes place within a specified temperature range. Also undeveloped part of flow is not negligible, therefore the problem is tackled with finite volume method, where both, flow and temperature field, are being solved. Numerical model of TES presented above, can be reduced to a 2D problem due to negligible changes in z-direction. Moreover, if we assume that the distribution of flow between plates is uniform, only one plate and one air channel can be considered. Going further, plate and air channel are symmetric, so only half of each of them needs to be included in the domain. Dimensions and the schematic diagram of the 2D computational domain used in this investigation are depicted in Fig.4.
Slika 4: Schematics of a numerical domain.
pt
ed
3.1. Governing equations In order to mathematically describe the process inside the storage unit, additional assumptions have to be met for the air as well as for the PCM: Air: • The range of velocity considered in the analysis is such that the changes in density are negligible (Mach number is below 0.2) therefore the air domain is considered as incompressible.
Ac ce
• We are dealing with Newtonian fluid which has constant density, specific heat, thermal conductivity and viscosity. • Viscous dissipation is negligible. • Since, in comparison with their spacing and width, the plates extend for a very long distance in the z direction, all locations in this direction appear essentially identical to one another. PCM:
• The PCM is homogenous and isotropic. • No distinction between PCM properties in solid and liquid phase is present. Values are assumed to be constant. • Natural convection in the PCM domain is not accounted for.
Page 4 of 15
3.2
5
CFD modeling
The steady equations governing the flow are solved only for the air and two unsteady energy equations are applied to the entire computational domain (PCM and air). The two-dimensional equations for conservation of mass, momentum and energy for air are expressed in Eqs.1-4. (1)
2 ∂ 2 vx ∂vx ∂vx ∂p ∂ vx + ρ a vx + vy =− + µa ∂x ∂y ∂x ∂x2 ∂y 2
(2)
ρa cp,a
∂Ta + ~v · ∇Ta ∂t
= ka ∇2 Ta + S
cr
2 ∂vy ∂ 2 vy ∂vy ∂p ∂ vy ρa v x + + vy =− + µa ∂x ∂y ∂y ∂x2 ∂y 2
ip t
∂vx ∂vy + =0 ∂x ∂y
(3) (4)
∂TP CM + ~v · ∇TP CM ∂t
= kP CM ∇2 TP CM
an
ρP CM cp,P CM,app
us
As previously written, there is no convection in PCM, therefore only unsteady energy equation is being solved (Eq. 5). (5)
3.2. CFD modeling
dQ =m ˙ a cp,a (Ta,out − Ta,in ) dt
(6)
ed
P =
M
With these equations air temperature at the outlet can be determined and thus thermal response of TES. Of the main interest is thermal power and amount of stored heat/cold, respectively. Former is calculated by Eq. 6.
Ac ce
pt
For the selected geometry grid independence study was carried out by comparing outlet temperature for the selected velocity and inlet temperatures. Size of the final mesh was 1000 x 30 for air and 1000 x 20 for PCM. Boundary conditions were selected so, that they best represent experimental work or operation in real situation. Thus, air velocity and temperature were prescribed at the inlet, static pressure was set to 0 Pa at the outlet and adiabatic wall was assumed at the ’wall’ (as marked in Fig.4). Based on the Reynolds number with the maximum value of 325 (Eq.7), laminar flow regime is expected, therefore laminar model was used for heat transfer fluid. Uniform temperature of the entire TES unit was taken as initial condition. The system of equations has thus been closed and solving was done in Ansys Fluent software, which employs the finite volume method for solving the mass, momentum and energy conservation equations [25]. Re =
ρvD 1.1 · 0.3 · 2 · 0.008 = = 325 µ 1.85 · 10−5
(7)
Relationship between enthalpy of PCM and temperature is known from DSC, hence latent heat of fusion can be taken into account via apparent specific heat of PCM. Because melting and solidification curves do not differ much, only one curve, representing both of them, can be used in simulations (Fig.3). As was mentioned above, density of PCM needs to be smaller, therefore apparent PCM density was calculated as ρP CM,app = ρP CM · vol% + ρa · (1 − vol%). Thermal properties of air and PCM used in simulations are presented in Table 2.
Page 5 of 15
3.3
6
Validation ρ (kg/m3 )
k (W/mK)
cp (kJ/kgK)
RT22HC
677
0.18
from the curve
Air
1.1
0.024
1005
µ (kg/ms) /
1.789 · 10-5
Tabela 2: Thermal properties of RT22HC and air used in simulation
3.3. Validation
us
cr
ip t
In order to confirm the numerical model it was necessary to carry out simulations with the same initial and boundary conditions as in experiment. Because TES unit was not completely insulated in the experiment, thermal losses emerged. They were included in equation for air (Eq.4) through additional term S, which is defined in Eq. 8. U A value is determined from the experiment, Tlab is temperature in the laboratory where experiment was carried out, Vchannels is volume of air channels. Temperature of storage unit was defined as average temperature between inlet and outlet. T −T U A · Tlab − a,in 2 a.out (8) S= Vchannels
an
Additional thermal mass due to casing, holder and PMMA was accounted for with increased specific heat of PCM. This means that curve in Fig.3 shifts upwards. 3.4. Principle of yearly operation
Ac ce
pt
ed
M
As already mentioned, TES is intended for heating and cooling, therefore two modes of operation need to be distinguished, ie. in summer and winter. Ljubljana, Slovenia is taken as a reference city (but it could be done for any other location as well) for which it is assumed that the heating season lasts from October to April and in the rest of the year, office needs to be cooled. In summer, during the night (from midnight to 7 a.m.), cold is being accumulated in the TES, if the average temperature of PCM (TP CM,avg ) exceeds inlet air temperature (Ta,in ) for at least 2 K. At that time air flow amounts to maximal flow (qv = qv,max ) and inlet temperature equals ambient temperature (Ta,in = Tamb ), otherwise air flow equals zero (qv = 0). During the day (from 7 a.m. to 8 p.m.) presumption will be, that minimal ventilation needs to be ensured due to presence of people, so air flow will be minimal (qv = qv,min ). But this is only valid when Ta,in exceeds TP CM,avg for at least 2 K. If Ta,in is up to 2 K higher than TP CM,avg , then air flow equals 0, because the temperature difference for heat transfer is too small, so it would not be reasonable to run the fan. Also there is no air flow if Ta,in falls behind TP CM,avg . At that time air directly enters into office. Operation in winter is slightly more complicated, because air can be taken from either ambient or from solar air collector. Unlike summer, TES unit does not operate at night, only during the day (from 7 a.m. to 8 p.m.). If TP CM,avg exceeds Tamb , then ambient air flows through TES with minimal flow, so the following condition Ta,in = Tamb holds true. At the same time the temperature of air from collector (Ta,col ) is monitored and if it is at least 2 K higher than TP CM,avg and also than Tamb , then air from collector flows through TES and not air from ambient. In that case TES gets charged with heat, so air flow increases to maximum. But flow rate through collector (qv,col ) does not equal flow through TES (qv ), because part of air from collector goes directly into office, so the following relation exists: qv = qv,col − qv,min . On the other hand, if these two conditions Ta,col < TP CM,avg + 2K and Ta,col > TP CM,avg − 2K are met, then air directly flows into office and TES does not operate. The same applies in the case when Tamb exceeds TP CM,avg . Model of solar air collector is presented in subsection 3.5. We are interested in operation throughout the year, but from the perspective of simulations this would be time consuming. Therefore the following will be assumed: the first week of each month is
Page 6 of 15
3.5
7
Description of the solar collector
pt
ed
M
an
us
cr
ip t
representative for the whole month. Data for ambient temperature are taken from the test reference year for Ljubljana and combined into temperature profile, which consists of 12 weeks.
(a) Winter.
(b) Summer.
Ac ce
Slika 5: Flow charts for winter and summer.
3.5. Description of the solar collector Calculation of air temperature at the outlet of solar collector was accomplished with TRNSYS software. Test reference year for Ljubljana serves as an input to module 561-2, which represents the solar collector without any special coatings. Its size is 2 m2 , collector is facing towards south and it is tilted for 60° from vertical. Other parameters set in the simulation are: ground reflectance 0.2, top loss convection coefficient 5,6 W/(m2 K), back loss convection coefficient 4,2 W/(m2 K). Air flow through collector was 75 m3 /h. 3.6. Evaluation of pressure drop Not only accumulated energy, but also power for air circulation needs to be taken into account in evaluation of TES performance. The pressure drop between the beginning and end of plates is calculated
Page 7 of 15
3.7
8
Implementation into Fluent
ip t
within simulations, but this pressure drop is negligible compared to the total pressure drop through such a system. This finding has important implication for the numerical modeling. In most numerical models authors focus only on plates and calculate pressure drop only for their length. But there are additional air ducts components (bends, valves, etc.; minor losses), where pressure drop increases quadratically as in Eq.9. Moreover, velocity in pipes is higher, so the flow is turbulent, consequently pressure drop (major losses) increases quadratically as in Eq.10. To evaluate TES accurately, it is crucial to examine pressure drop of the entire system when evaluating TES performance. For this reason, we suggest accounting all pressure drops, not only pressure drop due to plates. The main limitation of the proposed method is that pressure drop and characteristic of system can only be estimated with regard to components constituting it. Thus, according to our estimates relationship between fan consumption and air flow is given in Eq.11.
L1 2 ρv D2
Pf an = 0.0092 · qv1.961
us
∆p = f
cr
1 ∆p = K ρv 2 2
(9)
(10) (11)
4. Results
ed
M
an
3.7. Implementation into Fluent In the simulations, which were carried out for validation, additional term presented in Eq.8 was implemented into Fluent via UDF source term applied to air domain. Time step in these simulation was 30 s. Procedures described in subsection 3.4 were implemented into Fluent through script written in Scheme, which actually commanded Fluent. First, velocity field for minimum air flow was calculated and written to a file, then velocity field for maximum air flow was calculated and written to another file. Both of these were steady state calculations and they were succeeded by transient simulation, which embodied only energy equation. Data regarding flow field were read from the previously created two files according to the conditions presented previously. Used time step in yearly calculations was 300 s.
Ac ce
pt
4.1. Validation of a numerical model Figs. 6a and 6b give an example of outlet temperatures from experiment and simulation results, in the case with constant inlet temperature. In reality, inlet temperature varies with time, but for the purposes of validation, constant value is completely acceptable. In Fig.6a are depicted temperatures for solidification process, with inlet temperature set to 16 °C, initial temperature set to 24 °C and flow rate set to 78 m3 /h (conditions in experiment). Simulations were performed for different flow rates and specific heats (cp,shif t and Tm,shif t mean additional amount to the base value). Beside solidification, also melting was simulated and compared with experimental results (Fig.6b). Initial temperature was just below 16.5 °C, inlet temperature was 30 °C and flow rate was 78 m3 /h. Simulation results are in good agreement with experiment, however there is a slight difference in the end of the process. Curves indicate that melting or solidification ends faster in simulation than in experiment. This can be contributed to the non-uniform flow distribution between plates. Apparently some plates get less air, which hinders heat transfer. As a consequence, thermal response gets slightly distorted. Let us define relative temperature difference (∆Trel ) at every time step as a measure of agreement T −Ta,exp,out between simulations and experiment. It is characterized as : ∆Trel = a,out,sim . Accordingly, Ta,exp,out average value of relative temperature difference for solidification in the time interval 0 h - 18 h and for melting in the time interval 0 h - 6 h is less than 1.0%.
Page 8 of 15
9
Yearly performance
cr
ip t
4.2
(b) Melting.
us
(a) Solidification.
Slika 6: Curves from experiment and simulations.
an
4.2. Yearly performance
M
This subsection presents the results of TES unit in winter and summer, and in the end, energy savings on an annual basis are evaluated. In all simulations minimal volume flow equaled 40 m3 /h, whereas maximal flow equaled 75 m3 /h.
pt
ed
4.2.1. Winter Fig.7 depicts operation of TES in the first week of March. Temperatures (Ta,avg is the average air temperature in TES unit) are illustrated in upper subfigure and volume flow in lower subfigure. TES unit is not in operation in the beginning, so the temperature is constant and the flow is zero. Minimum ventilation starts at 7 a.m., so Ta,in = Tamb , TP CM,avg starts to decrease, the flow rate equals 40 m3 /h. When the temperature of the solar collector (Ta.col ) exceeds the average temperature of PCM (TP CM,avg ), operation is switched to charging mode, so flow rate changes to 35 m3 /h (75 m3 /h - 40 m3 /h). In the moment when TP CM,avg exceeds Ta,col , flow in TES unit goes back to 40 m3 /h. At 8 p.m. it shuts down and temperatures are again more or less constant, except Tin , which rises due to warmer PCM.
Ac ce
4.2.2. Summer For visual representation of temperatures and flow rate in August, the reader is referred to Fig.8. Air flow through TES at night is 75 m3 /h, then in the morning it shuts down, because of lower ambient air temperature than average PCM temperature. But when Tamb exceeds TP CM,avg TES is activated and air flows through it with 40 m3 /h. Such regime is maintained until evening or until TES gets discharged. At night charging process starts again if Tamb drops below TP CM,avg . 4.2.3. Yearly energy savings In addition to the annual thermal response of TES, also energy savings are of interest. Let us assume that ambient air needs to be heated up or cooled down to the desired room temperature. If this air is at least partly heated or cooled in the TES, then this part, presumably, represents energy savings Eaccu due to TES as defined in Eq. 12. Energy use for fan operation also needs to be considered, which is described by Eq.13. If the actual energy savings Esaving have to be determined, then energy consumption needs to be subtracted from accumulated energy as in Eq.14. It is necessary to point out, that these two (heat,
Page 9 of 15
10
Yearly performance
an
us
cr
ip t
4.2
Slika 7: Temperatures and volume flow in TES unit in the first week of March.
M
electricity) are, from exergy point of view, two different quantities despite the same units. But for the purposes of estimation of yearly savings, this premise will be good enough. ˆ Eaccu = qm,a cp,a (Ta,out − Ta,in ) dt (12)
ed
ˆ
Econsum =
qv,a ∆p dt η
pt
Esaving = Eaccu − Econsum
(13) (14)
Ac ce
Fig.9 depicts accumulated heat and electricity use in winter (Fig.9a) and in summer (Fig.9b). From Fig.9a it may be first observed that electricity consumption for such a small flow is negligible. It is worth nothing again that, in this case, the flow through collector amounts to 75 m3 /h, whereas through TES to 35 m3 /h, since one part of air flows directly into the office. As expected, the smallest accumulated heat is in November and December (7 kWh), due to the smallest solar radiation. Accumulated heat increases in spring and it reaches its peak in March, when approximately 20 kWh is stored. In April, accumulated heat is reduced, as the ambient temperature increases, consequently TES does not go through a complete cycles. Let us compare March’s result with the theoretical maximum stored heat. Assuming that about 50 Wh is stored in one plate in the temperature range from 15 °C to 30 °C, then for 30 plates this gives 1.5 kWh in one cycle. This would result in 45 kWh if TES was daily completely charged and discharged. In this manner the reasonableness of the simulations is confirmed. In the case of summer, the flow through the storage tank is actually 75 m3 /h, therefore electricity use for fan operation exceeds the one in winter. Electricity represents a bit more than 10% of the accumulated cold. The least cold is accumulated during the transition period (September, 7 kWh), as the outdoor temperature is low enough so that ambient air directly flows into the building. Ambient
Page 10 of 15
11
Yearly performance
an
us
cr
ip t
4.2
Slika 8: Temperatures and volume flow in TES unit in the first week of August.
M
temperature in the warmest months is much higher, and as a result, storage tank goes through complete cycles. Thus, the amount of accumulated cold in July and August equals 17 kWh.
Ac ce
pt
ed
4.2.4. Savings in building To illustrate presented results, a simulation of energy use in buildings was performed. As a reference, office with 4 m x 3 m x 2.8 m was chosen and the following was assumed: it has one wall adjacent to the ambient, facing south, size of windows is 3 m2 , shading (external shading factor = 0.9) turns on when total radiation exceeds 150 W/m2 on a vertical wall, on average one person is present form 7 a.m. to 8 p.m., computer and lights are on in the same time period, ventilation amounts to 1.2 h-1 , infiltration to 0.2 h-1 , maximum temperature should not exceed 26 °C in summer and in winter temperature in office should be kept above 20 °C during the day and above 17 °C during the night, internal accumulation is estimated to be 500 kJ/K. In order to calculate hourly heating and cooling load we embedded these data into Trnsys and for the results, see Fig.10a. The bar graph represents accumulated heat, the use of electricity, heat loads and ventilation losses. The results show that the biggest part of the heat load is at the expense of ventilation losses. Comparing accumulated heat to heat load, then it can be seen that the former accounts 8% of the heat load during winter months. In early autumn and in spring TES seems to be more promising, as almost all (March) or all (April, October) heat load can be covered with stored heat. Energy savings in entire winter will be approximately 84 kWh under given circumstances. Similar analysis can be carried out for summer (Fig.10b). It can be seen that the ventilation does not represent marginal part of cooling load anymore, as the most gains are at the expense of solar radiation. In all cases the cooling load exceeds accumulated cold, which means that energy use will be only partially reduced. Still, energy savings in summer will be approximately 58 kWh. The investment cost for the presented system is approximately 1140€, allocation of costs is given in table 3. The largest share of the costs represent plates and air solar collector, whose costs are 850€. Conventional system (for example split system air conditioning) suitable for office would cost around 500€. From these numbers it follows that TES is twice as expensive as a conventional system. However,
Page 11 of 15
(b) Monthly accumulated cold and electricity use for summer, qv,max = 75 m3 /h, 15, cm thick plates, 0,8 cm air gap, no shift in melting point.
us
(a) Monthly accumulated heat and electricity use for winter, qv,max = 35 m3 /h, 1,5 cm thick plates, 0,8 cm air gap, no shift in melting point.
cr
ip t
12
Slika 9: Monthly operation of TES
Ac ce
pt
ed
M
an
costs during the operation stage in the case of TES are lower, thus making TES promising alternative. Moreover, the costs would decrease, if TES was mass produced or if the state provided subsidies.
(a) Monthly accumulated heat and electricity use for winter, building’s heating load and ventilation losses, qv,max = 35 m3 /h, 1,5 cm thick plates, 0,8 cm air gap, no shift in melting point.
(b) Monthly accumulated heat and electricity use for summer, building’s cooling load and ventilation gains, qv,max = 75 m3 /h, 1,5 cm thick plates, 0,8 cm air gap, no shift in melting point.
Slika 10: Temperatures and volume flow in TES unit.
5. Conclusions Experimental and numerical investigations of a latent heat storage unit for space heating and cooling was carried out. It consisted of 30 plates filled with paraffin RT22HC. Thermal response of TES to the temperature step function was monitored in the experimental part. In the numerical part 2D model was
Page 12 of 15
Equipment
Price (€)
30 plates
600
Fan
60
Ducts
20
Insulation
10
Motor unit
80
Solar air collector
250
Speed regulator for a fan
80
Programmable clock
40
Together
1140
cr
Tabela 3: Breakdown of the investment costs.
ip t
13
M
an
us
set up in Fluent and calculations were executed for such conditions as they appeared in the experiment. A comparison of the results demonstrates very good agreement between experiment and simulations, which confirms adequateness of the numerical model. The developed model was used for evaluation of TES on a yearly basis. Calculations reveal that in winter the largest savings are in March, as larger quantities of heat are available at that time. In summer, the maximum amount of accumulated cold is in July and August, because of larger temperature fluctuations between day and night. Consequently, TES goes through complete cycles and stored energy increases. In comparison to heating and cooling loads in an office it could be seen, that accumulated heat/cold is relatively small. Nevertheless, energy consumption in an office can be annually reduced for approximately 142 kWh. Acknowledgements
References
ed
Financial support from the Slovenian Research Agency is gratefully acknowledged.
pt
[1] EU. Directive 2010/31/EU of the European Parliament and of the Council, June 2010.
Ac ce
[2] Mohammed M Farid, Amar M Khudhair, Siddique Ali K Razack, and Said Al-Hallaj. A review on phase change energy storage materials and applications. Energy Conversion and Management, 45(9-10):1597–1615, June 2004. [3] N. Soares, J.J. Costa, A.R. Gaspar, and P. Santos. Review of passive PCM latent heat thermal energy storage systems towards buildings energy efficiency. Energy and Buildings, 59:82–103, April 2013. [4] Yinping Zhang, Guobing Zhou, Kunping Lin, Qunli Zhang, and Hongfa Di. Application of latent heat thermal energy storage in buildings: State-of-the-art and outlook. Building and Environment, 42(6):2197–2209, June 2007. [5] E. Osterman, V.V. Tyagi, V. Butala, N.A. Rahim, and U. Stritih. Review of PCM based cooling technologies for buildings. Energy and Buildings, 49:37–49, June 2012. [6] Tiago Silva, Romeu Vicente, Nelson Soares, and Victor Ferreira. Experimental testing and numerical modelling of masonry wall solution with PCM incorporation: A passive construction solution. Energy and Buildings, 49:235–245, June 2012.
Page 13 of 15
14 [7] Shazim Ali Memon. Phase change materials integrated in building walls: A state of the art review. Renewable and Sustainable Energy Reviews, 31:870–906, March 2014. [8] Belen Zalba, Jose M. Marin, Luisa F. Cabeza, and Harald Mehling. Free-cooling of buildings with phase change materials. International Journal of Refrigeration, 27(8):839–849, December 2004.
ip t
[9] Pablo Dolado, Ana Lazaro, Jose M. Marin, and Belen Zalba. Characterization of melting and solidification in a real-scale PCM-air heat exchanger: Experimental results and empirical model. Renewable Energy, 36(11):2906–2917, November 2011.
cr
[10] Pavel Charvát, Lubomír Klimeš, and Milan Ostrý. Numerical and experimental investigation of a PCM-based thermal storage unit for solar air systems. Energy and Buildings, 68:488–497, January 2014. [11] Taha K. Aldoss and Muhammad M. Rahman. Comparison between the single-PCM and multi-PCM thermal energy storage design. Energy Conversion and Management, 83:79–87, July 2014.
us
[12] R. Parameshwaran and S. Kalaiselvam. Energy conservative air conditioning system using silver nano-based PCM thermal storage for modern buildings. Energy and Buildings, 69:202–212, February 2014.
an
[13] Vadim Dubovsky, Gennady Ziskind, and Ruth Letan. Analytical model of a PCM-air heat exchanger. Applied Thermal Engineering, 31(16):3453–3462, November 2011.
M
[14] N.H.S. Tay, M. Belusko, and F. Bruno. Designing a PCM storage system using the effectivenessnumber of transfer units method in low energy cooling of buildings. Energy and Buildings, 50:234–242, July 2012.
ed
[15] Trp Anica. An experimental and numerical investigation of heat transfer during technical grade paraffin melting and solidification in a shell-and-tube latent thermal energy storage unit. Solar Energy, 79(6):648–660, December 2005.
pt
[16] U Stritih and P Novak. Thermal Storage Of Solar Energy In The Wall For Building Ventilation. In International IEA Workshop (Annex 17), Advanced Thermal Energy Storage Techniques, number april, pages 3–5, Lju, 2002. International Energy Agency (IEA).
Ac ce
[17] B.L. Gowreesunker, S.A. Tassou, and M. Kolokotroni. Coupled TRNSYS-CFD simulations evaluating the performance of PCM plate heat exchangers in an airport terminal building displacement conditioning system. Building and Environment, 65:132–145, July 2013. [18] Maciej Jaworski. Thermal performance of building element containing phase change material (PCM) integrated with ventilation system - an experimental study. Applied Thermal Engineering, 70(1):665– 674, September 2014. [19] Fabien Rouault, Denis Bruneau, Patrick Sebastian, and Jerome Lopez. Experimental investigation and modelling of a low temperature PCM thermal energy exchange and storage system. Energy and Buildings, 83:96–107, November 2014. [20] RUBITHERM GmbH, 2013.
[21] 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? Thermochimica Acta, 596:79–88, November 2014.
Page 14 of 15
15 [22] Ming Liu, Frank Bruno, and Wasim Saman. Thermal Performance Of A Pcm Thermal Storage Unit. pages 2766–2771. Springer Berlin Heidelberg, January 2009. [23] Hed and R. Bellander. Mathematical modelling of PCM air heat exchanger. Energy and Buildings, 38(2):82–89, 2006.
Ac ce
pt
ed
M
an
us
cr
[25] ANSYS. ANSYS FLUENT Theroy Guide 14.0, November 2011.
ip t
[24] Jacques Bony and Stéphane Citherlet. Numerical model and experimental validation of heat storage with phase change materials. Energy and Buildings, 39(10):1065–1072, October 2007.
Page 15 of 15