Investigation of a Flow Coefficient for Predicting a Natural Circulation of Water in a Built-in-storage Solar Water Heater

Pachern Jansa, Supachart Chungpaibulpatana, and Bundit Limmeechokchai Abstract - This paper describes an investigation of an overall flow coefficient Kf for predicting thermosyphon flow rate of water circulating in a built-in-storage (BIS) solar water heater. Firstly, a set of mathematical equations for solving storage tank temperatures of the BIS solar water heater has been developed based on the energy balances on three main components: absorber plate, collector channel and storage tank. The thermosyphon flow rate of water in the system is then modeled and an overall flow coefficient Kf is introduced in the flow model as a system performance parameter, which must be determined from experimental tests on the actual system. Procedure to determine the value of Kf has also been established. Consistent results have been satisfactorily obtained from several tests conducted on the system with different conditions of flow restrictions and under prevailing sky conditions. The investigation has shown that the overall flow coefficient Kf introduced in the flow model can be considered to be a parameter used for simulating the performance of BIS solar water heater. Keywords - Solar water heater, Built-in-storage, Thermosyphon, Thermal diode, Overall flow coefficient, System modeling, Simulation.

1. INTRODUCTION A passive solar water heater operated with natural circulation is often called the thermosyphon solar water heater. Especially when a collector and storage is integrated in the same unit, it is known in a name as a built-in-storage (BIS) solar water heater. The system mainly consists of an inclined tank, in which the walls of the tank are thermally insulated except front wall, which is painted with black matte from outside and covered with glass sheet [6]. Different developments are suggested by researchers to improve the performance of the BIS solar water heater such as using an insulation cover during night-time [5] and using a transparent insulation [1]. Another modification of BIS solar water heater which utilizes a reflector to minimize radiation losses is also presented [2]. The BIS solar water heater with a baffle plate placed between collector and storage tank is also proposed to reduce heat losses through the aperture area [4]. For a problem of convection losses during the night, a light plastic as a thermal diode is used to fix on an insulated baffle plate of a trapezoidal cross section tank to prevent reverse circulation [3]. As far as manufacturing is concerned, the trapezoidal cross section tank might be difficult to construct. Also, a light plastic plate using as a thermal diode might not be durable and not easy to change when damaged. A simple rectangular tank with a commercial check-valve as a thermal diode was therefore designed and constructed in order to conduct experimental tests [11]. A flow rate of water circulation is important to determine the thermal performance of a solar water heater. A thermosyphon flow rate occurred by a natural circulation of water in the BIS solar water heater is very difficult to measure. Thus, the main purpose of this study is to investigate an overall flow coefficient for predicting the thermosyphon flow rate of the constructed BIS solar water heater. A simulation model for predicting the storage tank temperature of the constructed BIS solar water heater has been developed. To verify the developed BIS system simulation model, the predicted results from the simulation are compared with those obtained from the experimental tests. __________________________ Pachern Jansa is with the mechanical engineering department, Sripatum University, 61 Phaholythin Rd., Jatujak, Bangkok 10900, Thailand. Phone: (+662) 579-1111 ext. 1202; Fax: (+662) 579-1111 ext. 1202 e-mail: p_jansa@hotmail.com, corresponding author Supachart Chungpaibulpatana is with Sirindhorn International Institute of Technology, Thammasat University,P.O.Box 22, Thammasat Rangsit Post Office, Klongluang, Patumthani 12121, Thailand. e-mail; supachar@siit.tu.ac.th Bundit Limmeechokchai is with Sirindhorn International Institute of Technology, Thammasat University,P.O.Box 22, Thammasat Rangsit Post Office, Klongluang, Patumthani 12121, Thailand. e-mail; bundit@siit.tu.ac.th

2. MATHEMATICAL MODEL FORMULATION The thermal performance of the developed built-instorage solar water heater can be described by the energy balances on three main components, i.e. the absorber plate, the water in the collector channel and the water in the storage tank as shown in Fig. 1 and 2.
Absorbed solar radiation, ()eIT

Glass cover Absorber plate

d
Ambient, Ta

n-1

j-1 3

j+1

Portion = 1

p T
T
f

Ts
Collector channel

y L

Storage

Fig.1 Schematic diagram of the three main components and the position of temperature measurement
Ut IT
Glass cover Absorber plate

Flow

Tp

hf

T f

Water in collector channel Insulated partition

Flow

Ts

Water in storage

Storage casing with insulation

(a) Enlarged schematic diagram of the three main components for an elementary portion y
IT

Ta
Tp
1 Ut
1 hf

Ta
1 U of

Qu Qc

Tf
1 Usf

Qu +Qu,y

Qu +Qu,y
Ta

Ts
1 Uos

Qu Qs

(b) Heat transferred network connecting the three main components Fig. 2 Main components of the built-in-storage solar water heater for model formulation.

2.1 ABSORBER PLATE The energy balance on the absorber plate, the heat capacity of which is assumed to be negligible, can be described by QA QTL = Qp-c or

2.2 COLLECTOR CHANNEL The energy balance of the water in the collector channel for each portion dy at any instant is described by the following equation.
Qc = Q p c Qc s Qc a Qu

( ) e I T U t (Tp Ta ) = h f (Tp T f )

(1)
f dC pf
dTf dt

or

where QA = absorbed total solar radiation on the absorber plate (W/m2), QTL = total heat losses from the top of the absorber plate to the ambient air (W/m2), Qp-c = total heat transfer from the absorber plate to the water in collector channel (W/m2), ()e = effective transmittance-absorptance product of the collector, IT = total solar irradiance falling on the collector surface (W/m2), Ut = overall top loss coefficient between the absorber plate and ambient temperature taken into account the effect of glass cover (W/m2K), Tp = average temperature of the absorber plate (C), Ta = ambient temperature (C), hf = convective heat transfer coefficient between the absorber and water in the channel (W/m2K), and Tf = average temperature of water in the channel (C). From eq. (1), one gets
Tp =
( ) e I t + h f T f + U tTa hf + Ut

= h f (Tp Tf ) U sf (Tf Ts )
U of ( 2d B )(T f Ta )

& pf dT f mC
B dy

(6)

(2)

The effective transmittance-absorbtance product can be approximated by [7], [8]

( )e

= 1.01 g p

(3)

where g = transmittance of glass cover, and p = absorptance of absorber plate. The overall top loss coefficient can be calculated by the following equation [7], [8].

Qc = total energy stored in the water in the collector where channel at any instant (W/m2), Qc-s = energy transfer between water in the collector channel and the storage tank by passing through the insulated partition (W/m2), Qc-a = energy loss from the collector channel to the ambient air (W/m2), Qu = useful energy in the water flowing out from the collector channel (W/m2), f = density of the water in the collector channel (kg/m3), d = collector channel depth (m), B = width of collector channel (m), Cpf = specific heat of water in the collector channel (kJ/kgK), m & = mass flow rate of water flowing through the system (kg/s), Tf = water temperature in the collector channel (C), Ts = water temperature in the storage tank (C), and Usf = heat transfer coefficient between water in the channel and the storage tank (W/m2K) = ksf/lsf, where ksf = heat conductivity of the insulated partition and lsf = thickness of the insulated partition, Uof = heat losses coefficient from water in the collector channel through the side insulation wall to the ambient air (W/m2K) = kof/lof, where kof = heat conductivity of the wall insulation partition and lof = thickness of the wall insulation. After substituting an expression of Tp into eq. (6) one has
f dC pf
dT f dt

h f [( ) e I T U t (T f Ta )] hf + Ut

N Ut = T -T TC N + f
* p * p

( )
* a

+ + -1 hv ( p + 0.00591Nhv ) +

-1

(Tp* + Ta* )(Tp*2 + Ta*2 )

2 N + f -1 + 0.133 p

(4)
-N

U sf (T f Ts ) U of (

2d B

)(T f Ta )

& pf dT f mC
B dy

(7)

where N = number of glass cover, T = absolute absorber plate temperature (K), Ta* = absolute ambient temperature (K), g = emittance of glass cover, p = emittance of absorber plate, = Stefan-Boltzmann constant (5.6697 x 10-8 W/m2K4), hv = wind heat transfer coefficient = 5.7 + 3.8vwind (W/m2C, [Duffie and Beckman, 1980 [7], where vwind = wind speed across the front cover of the collector, (m/s), e = 0.43(1-100/ Tp* ), f = (1 + 0.089hv 0.1166hvp)(1 + 0.07866N), and C = 520(1-0.0000512) for 0 < < 70, = collector tilt angle (degree). Note that Ut is calculated based on projected area of the absorber plate [7], [8]. The convective heat transfer coefficient between the absorber and water in the collector channel can be calculated for the inclined plate facing downward with approximately constant heat flux as follows [9]:

* p

Since the partial differential form of eq. (7) is not easy to solve analytically, a finite difference numerical method is used [10]. If the collector channel is represented by n portions along its flow direction as shown in Fig. 1. The temperature of the jth portion at the next time step t can be written as follows:
t + t
j

Tf

= T ft +
j

t
t tf dC pf

{
t

h f [( ) e I T U t (T f Ta )]
t t t t t
j

hf + U t
t

U sf (T f Ts ) U of (
t
j j

2d B

)(T f Ta )
t t
j

t & t C pf m

By
t

(T f T f )}
t t
j j 1

(8)

where

Tf = water temperature in the collector channel at time t

(C), T ft + t = water temperature in the collector channel at time t+t (C), Tst = water temperature in the storage tank at time t (C), Tat = ambient temperature at time t (C), tf = density of the water in the
t collector channel at time t (kg/m3), C pf = specific heat of water in

hf = Nu e kw/y

(5)
1/4

Nu e = average Nusselt number = 0.56(Gre Pre cos ) where < 88; 105 < Gre Pre cos < 1011, Gre = g(Tp Tf)L3/, g = gravitational constant (m/s2), = density of water (kg/m3), = thermal expansion coefficient of water (K-1), = viscosity of water (kg/m-s), kw = thermal conductivity of water (W/mK), y = length of the portion (m). Note that all properties of water except , using to calculated the average Nusselt number, are evaluated at a reference temperature Tref1 = Tp - 0.25(Tp - Tf); is evaluated at a temperature of Tref2 = Tf + 0.5(Tp Tf).

the collector channel at time t (kJ/kgK), I Tt = total solar irradiance falling on the collector surface at time t (W/m2), htf = convection heat transfer coefficient between the absorber and water in the channel at time t (W/m2K), U tt = overall top loss coefficient between the absorber plate and ambient temperature taken into & t = mass account the effect of glass cover at time t (W/m2K), and m flow rate of water flowing through the system at time t (kg/s).

Note that, for the bottom and the top end portions of the collector channel, there is an additional heat loss from the water through the end side to the ambient air. Hence the 3rd heat loss term on the right hand side of eq. (8) must be replaced by
U of (

is assumed to be linearly dependent on its density i.e. T = Cg, where C is a constant. The thermosyphon head can then be expressed as: HT = C1(h2-h1)(Tmb-Tsb) + C1(h3-h2)(Tfm-Tsb) - C1(h4-h1)(Tsm-Tsb) C1(h3-h4)(Tmt-Tsb)

1 y

2
B

) d (T f Ta ) .
t t
j

(12)

2.3 STORAGE TANK The energy balance for an elementary portion dy of the water in the storage tank can be described by the following equation.
Qs = Qu + Qc s Qs a

or

s HC ps

dTs dt

& ps dTs mC

dy

+ U sf (T f - Ts )

(9)
2

where C1 = constant, Tfm = average temperature of the water in the collector channel (C), Tsm = average temperature of the water in the storage tank (C), Tmb = average temperature of the water at the bottom, (Tsb+Tfb)/2, (C), Tmt = average temperature of the water at the top, (Tft+Tst)/2, (C), Tsb = temperature of the water at the bottom of the storage (C), Tfb = temperature of the water at the bottom of the collector channel (C), Tft = temperature of the water at the top of the collector channel (C), Tst = temperature of the water at the top of the storage (C). b) Friction Head The total friction head that occurs in the built-in-storage solar water heater due to the friction of water flowing through the entrance, exit, distributor and check-valve in the flow circuit can be described by: &2 HF = C2V2 = C3 m (13)

where Qs = total energy stored in the storage at any instant (W/m ), Qs-a = energy loss from the storage to the ambient air (W/m2), s = density of the water in the storage tank (kg/m3), H = storage tank depth (m), Cps = specific heat of water in the storage tank (kJ/kgK), Uos = heat loss coefficient between the storage tank and ambient (W/m2K); Uos = kos/los, where kos = heat conductivity of tank wall insulation (W/mK) and los = thickness of tank wall insulation (m). As shown in Fig. 1, the temperature at the next period of time for the jth node can be written as
Ts
t + t
j

= Tst +
j

t st HC ps B y j

t & t C ps m

(Ts - Ts )
j -1 j

C2, C3 = constant, V = velocity of flow (m/s). Note that where the friction due to the viscosity of water along the collector channel is relatively small in comparison with the above-mentioned friction. However it can be assumed that this small effect is already included in eq (13).
c) Mass Flow Rate Equating eqs. (12) and (13), the mass flow rate in passing through the collector channel can be obtained as; &2 C3m = C1{(h2-h1)(Tmb-Tsb) + (h3-h2)(Tfm-Tsb) (h4-h1)(Tsm-Tsb) (h3-h4)(Tmt-Tsb)}

+U sf (T f - Ts ) - U os (1 +
t t
j j

2H

)(Ts - Ta )]
j

(10)

where Tst + t = water temperature in the storage tank at time t + t

t (C), st = density of water in of storage tank at time t (kg/m3), C ps =

specific heat of water in the storage tank at the time t (kJ/kgK). Note that, for the bottom and the top end portions of the storage, there is an additional heat loss from the water through the end side to the ambient air. Hence the 3rd heat loss term on the right hand side of eq. (10) must be replaced by
U os (1 + 2H B + H y )(Ts Ta ) .
t t
j

(14)

& = Kf{(h2-h1)(Tmb-Tsb) + (h3-h2)(Tfm-Tsb) or m (h4-h1)(Tsm-Tsb) (h3-h4)(Tmt-Tsb)}(1/2) (15) where Kf = the overall flow coefficient due to the friction along the water flow passage through distributor, inlet and outlet hole of the collector channel and check valve (kg s-1{m K}-1/2). 3. SYSTEM SIMULATION PROCEDURE Table 1 Simulation parameter. Parameter Inclination angle of the system, Latitude of test location, Lat Collector width, B (m) Collector length, L (m) Average channel depth of the collector passage, d (m) Wall insulation thickness, los (m) Insulated partition thickness, lsf (m) Wall insulation heat conductivity, kos (W/mK) Insulated partition heat conductivity, ksf (W/mK) Number of glass cover, N Transmittance of glass cover, g Emittance of glass cover, g Emittance of absorber plate, p Absorptance of absorber plate, p Water volume in storage tank, s (liters) Heights (h2 h1) and (h3 h4) in Fig. 3.4 (m) Heights (h3 h2) and (h4 h1) in Fig. 3.4 (m) * Obtained from measurement ** Obtained from manufacturers specifications

2.4 THERMOSYPHON FLOW RATE a) Thermosyphon Head The thermosyphon built-in-storage solar water heater and its density-head diagram can be shown in Fig. 3. The thermosyphon head at any instant is created by the difference between the total pressure head along the path 123 and that along the path 143.
2 1 h1 h2 3 4 h4 h3 2 1 Temperature Density Head, h 4 3

Value 15 * 15 (north) 1.1 * 1.7 * 0.0425 * 0.03 * 0.035 * 0.07 ** 0.15 ** 1 0.88 ** 0.88 ** 0.95 ** 0.8 ** 273 * 0.10* 0.46*

Fig. 3 The density head diagram of the system. The thermosyphon head (HT) is described by HT =

gdh + gdh - gdh - gdh

1 2 1 4

(11)

where h = the vertical height (m). Since the temperature variation of water inside the system is small, the water temperature

A finite difference method has been used to solve the mathematical model of the constructed BIS system by dividing the system into several portions along the length of the system as shown in Fig. 1. For the simulation the BIS system was divided into 17 portions along the length of the collector. The parameters of the system used in the simulation are listed in Table 1 and the simulation procedure can be described by Fig. 4.
Start
Input System Parameters

where

Tsm , cal = average temperature of the water in the storage

t resulted from simulation at any time t (C), Tsm = average ,exp

, B, L, d, los,lsf, ksf, kos, N, g, g, p, p,and s

Input initial system temperatures (Tfi, Tsi)

Input a data file the meteorological data of interest (IT, Ta, and wind)

Input No. of time period for simulation (ntime)

t=1
Calculate Usf, Uos, U t j by eq. (3.4), h f j by eq. (3.5) and Tp j by eq. (3.2)

& t by eq. (3.15) Calculate m

temperature of the water in the storage resulted from experiment at any time t (C), n = number of the data. Finally the value of Kf that produces the minimum value of SEE is therefore selected for the system under the considered condition. The water in the collector channel flows upward during the heating period and, if provision for preventing reverse thermosyphon flow is not installed in the system, the water would flow downward in the collector channel during the night or cooling down period. Hence the value of the overall flow coefficient Kf during the heating may differ from that during the cooling since different friction resistances can be expected from the opposite direction of flow through any passage. Hence Kf is considered from two parts of data of each experimental test run, i.e. Kf for the heating period and Kf for the cooling period. For the heating period, Kf is suggested to determine from the days with high solar radiation intensity, say about 20 MJ/m2. Note that with such a high insolation day it is ascertained that the water flow exists in the system with a rate high enough to give accurate results of Kf. The value of Kf for the heating period is determined using the experimental data over the period which the thermosyphon head, HT, calculated by eq. (12) is found to be positive value whereas Kf for the cooling period is determined from the period when the thermosyphon head is negative. As, in each test, the system cooling down period occurred in the evening after the daytime heating up period, Kf for the heating period is determined first and then the accepted value of Kf for the heating period will be used to determine Kf for the cooling period. 4.2 DETERMINATION RESULTS AND DISCUSSION

Calculate T f j

t +t

and Ts j

t +t

by eq. (3.8) and (3.10)

The three different operational cases can be described as follows: Case 1: The check-valve in the flow passage was forced open throughout the test i.e. reverse circulation may occur at night. Case 2: The check-valve was allowed to work freely throughout the test. The valve opening is dependent on the pressure difference between each side of the valve lid. Case 3: The check-valve was allowed to work freely during the day but forced close at night i.e. no reverse circulation at night may occur. The best values of Kf obtained using the above mentioned determination procedure for both heating and cooling periods of several experimental test runs are shown in Table 2 and sample results showing the simulated storage tank temperatures, Tsm,cal in comparison with those Tsm,exp are given in Fig. 5, Fig. 6, and Fig. 7. In Table 2, it is found that Kf for heating period of Case 1 is higher than those of Case 2 and Case 3. As Kf is directly related with the mass flow rate of water circulating in the system, this mean that the flow rate in Case 1 was higher than the other two cases. Note that with higher Kf, the resistance to flow is less resulting in higher flow rate. This can be explained by the fact that, in Case 1, the check-valve installed in the flow passage was forced to fully open throughout the day (and the night as well). The resistance to flow was therefore less than those of the other two cases in which the check-valve was allowed to work freely, i.e. the valve was usually partly opened depending on the driving forced created by differential pressures of water located at both sides of the valve. The valve was then very rare in the full opening position. Similar results were also found in the case of cooling at night in which relatively very high reverse flow was shown in Case 1 in comparison with those in the other two cases. Note that, in Case 3 in which the check-valve was forced close, the value of Kf obtained from the night was not zero as it must be in reality. However the obtained value was very small and actually very close to zero. The small discrepancy may be explained by the errors in experimental measurements. It should also be noted here that the value of Kf for night-time period was found quite low for Case 2 indicating that the installed check-valve could significantly stop the reverse circulation.

& , HT, T f ,bottom , T f ,middle , T f ,top , Ts ,bottom Store h f , m Ts ,middle , Ts ,top , T fm , and Tsm in Output File

Check n time If t >

No

Yes

Stop
Fig. 4 Flow chart of simulation program. 4. DETERMINATION OF OVERALL FLOW COEFFICIENT, KF 4.1 DETERMINATION PROCEDURE The overall flow coefficient Kf, which appears in eq. (15) & , and must be known so that the thermosyphon mass flow rate, m of the water circulating in the system can be predicted by the simulation. Kf is the overall effects of all frictions occurred along the water flowing passage throughout the system. A trial and error procedure is proposed to determine the value of Kf from the experimental data obtained from the tests on the system. Several values of Kf are guessed systematically and input to the simulation program. For each value of Kf, the program predicts the average temperature of the water in the storage tank, Tsm, at any time interval of interest in accordance with the given meteorological data observed from the experimental test. These simulated values of Tsm are compared with their corresponding values of Tsm observed from the experiment. Their errors are estimated and the standard errors of estimate (SEE) of Tsm is then calculated by
(Tsm , cal Tsm ,exp )
t t
i i

SEE =

i =1

n 1

(C)

(21)

1000

100

5. CONCLUSIONS A mathematical model has been formulated for simulating the transient storage tank temperature of the developed BIS solar water heater operated under varying weather conditions of solar irradiance, ambient temperature and wind speed. A submodel for estimating the thermosyphon flow rate along the flow circuit between the collector channel and storage tank has also been proposed. An overall flow coefficient Kf has been introduced in the submodel to reflect the overall effects of various frictions occurring along the water flow passage. Procedure for determining the value of Kf from the experimental test conducted on the BIS system has been described. Using experimental data obtained from various test runs on the BIS system under different operations of check-valve, the values of Kf determined according to the above-mentioned procedure have been successfully found to be consistent, that is higher Kf which results in higher flow rate is obtained when the check-valve is forced fully open. Hence the overall flow coefficient Kf can be used as a performance parameter for the BIS system. ACKNOWLEDGEMENTS The author is grateful to Sripatum University and the Energy Policy and Planing Office--EPPO (formally called the Thailand National Energy Policy Office--NEPO) for their kind contribution in providing partial financial support. REFERENCES M. Arulanantham, K. S. Reddy and N. D. Kaushika (1998), Solar gain characteristics of absorber parallel transparent insulation materials. Energy Conversion and Management 39, 1519-1527 [2] G.N. Tiwari Doung Duc Hong and R.K. Goyal (1998), A low cost solar water heater suitable for rural, as well as urban, areas of Vietnam. Energy Conversion and Management 39, 953-962 [3] A.A. Mohamad (1997), Integrated solar collector-storage tank system with thermal diode. Solar Energy 61, 211-218. [4] I.N. Kaptan and A. Kilic (1996), A theoretical and experimental investigation of a novel built-in-storage solar water heater. Solar Energy 57, 393-400 [5] S.C. Kaushik, Rakesh Kumar, H.P. Garg, and J. Prakash (1994), Transient analysis of a triangular built-in-storage solar water heater under winter conditions. Heat Recovery Systems & CHP 14, 337-341 [6] M.S. Sodha, J.K. Nayak, S.C. Kaushik, S.P. Sabberwal, and M.A.S. Malik (1979), Performance of a collector/storage solar water heater. Energy Conversion and Management 19, 41-47 [7] John A. Duffie and William A. Beckman (1980), Solar Engineering of Thermal Processes (First Edition). John Wiley & Sons, Inc. [8] John A. Duffie and William A. Beckman (1991), Solar Engineering of Thermal Processes (Second Edition). John Wiley & Sons, Inc. [9] J. P. Holman (1997), Heat Transfer. McGraw-Hill, Inc. [10] G.Y. Saunier, Ph. Bouix, Supachart Chungpaibulpatana, and Thada Vitagsabootr (1985), A test procedure for thermosyphon solar water heaters. Solar Thermal Component and System Testing: Proceedings of the Fourth Asian School on Solar Energy Harnessing 12-20 December 1985 at AIT, Bangkok, 163-177 [11] P. Jansa, S. Chungpaibulpatana, and B. Limmeechokchai (2003), Development of a built-in-storage solar water heater for remote areas. Proceedings of the 2nd Regional conference on Energy Technology Towards a Clean Environment, The Joint Graduate School of Energy and Environment on 12-14 February 2003 at Phuket, Thailand, 422-428 [1]

900 800 700 600

IT

90 80 70

Tsm,cal

60 50 40 30

IT

500 400 300 200 100 0 10:30 12:00 13:30 15:00 16:30 19:30 21:00 22:30 6:00 7:30 0:00 1:30 3:00 4:30

Ta

Tsm,exp

20

vwind
10 0

Time

Fig.5 Simulation results in comparison with experimental results on 9/10/2002

1000 900 800 700 600

18:00

6:00

9:00

100

IT

90 80 70

Tsm,cal

60 50 40 30

IT

500 400 300 200 100 0 12:00 13:30 15:00 16:30 18:00 19:30 22:30 6:00 7:30 9:00 0:00 1:30 3:00 10:30 21:00 6:00 4:30

Ta Tsm,exp

vwind

20 10 0

Time

Fig.6 Simulation results in comparison with experimental results on 15/11/2002

1000 900 800 700 600 100

IT

90 80 70

Tsm,cal

60 50 40 30

IT

500 400 300 200 100 0 10:30 12:00 13:30 15:00 16:30 19:30 21:00 22:30 6:00 7:30 0:00 1:30 3:00 4:30

Ta Tsm,exp vwind
6:00

20 10 0

Time

Fig.7 Simulation results in comparison with experimental results on 12/7/200

Table 2 Best values of the overall flow coefficient Kf obtained for three different operations of check-valve.

18:00

9:00

vwind (x10-1 m/s), Temperature (C)

vwind (x10 m/s), Temperature (C)

-1

vwind (x10-1 m/s), Temperature (C)