Six-Day System Test and Component test and System Simulation for Combitanks with Internal heat Exchangers

(1)

. .

Avdelningen för energi.

miljö och byggande Högskolan Dalarna

. 781 88 Borlänge

Tel: +46 23 778700 Tel: +46 23 778701

Besöksadress/Street adress:

Forskargatan 8 Borlänge

ISSN 1401 - 7555

ISRN HFB-SERC--61--SE June 1997

SERC

Högskolan Dalarna, S-781 88 Borlänge

for

Combitanks with Internal heat Exchangers and

Component test and System Simulation

*Solar Energy Research Center,

J.Dahm*, C.Bales**, K.Lorenz**, J-O.Dalenbäck*

**Department of building Services Engineering, Chalmers Univeristy of Technology

Solar Energy Research Center Centrum för solenergiforskning

Six-Day System Test

(2)

Six-Day System Test and

Component test and System Simulation

for

Combitanks with Internal heat Exchangers

J.Dahm*, C. Bales**, K. Lorenz**, J-O. Dalenbäck*

**Department of Building Services Engineering, Chalmers Unversity ofTechnology

*Solar Energy Research Center, Högskolan Dalarna, S-78188 Borlänge

ISSN 1401 -7555 ISRN DU-SERC--61--SE

(3)

Abstract

An international standard, ISO/DP 9459-4 has been proposed to establish a uniform standard of qualit y for small, factory-made solar heating systerns. In this proposal, system components are tested separatelyand total system performance is calculated using system simulations based on component model parameter values validated using the results from the component tests. Another approach is to test the whole system in operation under representative conditions, where the results can be used as a measure of the general system performance. The advantage of system testing of this form is that it is not dependent on simulations and the possible inaccuracies of the models. Its disadvantage is that it is restricted to the boundary conditions for the test. Component testing and system simulation is flexible, but requires an accurate and reliable simulation model.

The heat store is a key component conceming system performance. Thus, this work focuses on the storage system consisting store, electrical auxiliary heater, heat exchangers and tempering valve. Four different storage system configurations with a volume of 750 litre were tested in an indoor system test using a six -day test sequence. A store component test and system simulation was carried out on one of the four configurations, applying the proposed standard for stores, ISO/DP 9459-4A. Three newly developed test sequences for intemalload side heat exchangers, not in the proposed ISO standard, were also carried out. The MULTIPORT store model was used for this work. This paper discusses the results of the indoor system test, the store component test, the validation of the store model parameter values and the system simulations.

(4)

Sammanfattning

En internationell teststandard, ISO/DP 9459-4, har tagits fram för att försöka garantera en högkvalitet för fabrikstillverkade solvärmesystem. I denna föreslagna standard testas systemkomponenterna var för sig och systemprestanda beräknas med simuleringar av en systemmodell vars parametervärden har beräknats m.hj.a. data från komponentesterna. Ett annat sätt är att testa hela systemet i drift under realistiska förhållanden då resultaten från testet kan ge ett direkt mått på hela systemets prestanda. Fördelen med denna typ a systemtest är att den inte är beroende på simuleringar och de möjliga onnoggranheterna av modellerna.

N ackdelen är att resultaten är mycket beroende på testproceduren och förutsättningar . Komponenttest med systemsimulering är ett mycket flexibelt angreppssätt men är helt beroende på kvaliteten av simuleringsmodellerna som används.

Värmelagret (ackumulatortank eller förrådsberedare) är en nyckelkomponent vad gäller systemprestanda. Detta arbete koncentrerar därför kring värmelagret bestående ackumulatortank, elpatron, värmeväxlare för solfångarekretsen och varmvattenberedning samt blandningsventil för varmvattnet. Fyra olika tankkonfigurationer alla med en volym av 750 liter har testats inomhus med ett systemtest med en sex-dagars testsekvens. En komponenttest för ackumulatortanken tillämpades på en av dessa fyra tankkonfigurationer enligt den föreslagna teststandarden för värmelager, ISO/DP 9459-4A. Tre extra testsekvenser som inte är med denna ISO-standard användes också för att närmare studera varmvattenberdningen.

Simuleringsmodellen "MULTIPORT store model" användes med detta arbete.

Denna rapport beskriver resultaten av system- och komponenttesten samt redovisar hur resultaten användes för att validera modeIparametrar och för systemsimuleringar .En utvärdering av de fyra testa tankkonfigurationerna och av själva systemtest finns också med.

(5)

In Sweden there is a long tradition of using bioffiass, principally wood, for the heating needs of single family houses. In recent years it has become fairly common to install storage tanks together with wood fired boilers. One reason is increasing oil and electricity prices, another is the development of highly efficient boilers with strongly reduced emissions. This trend, together with an investment subsidy for solar heating systerns has created a rnarket for wood boiler systerns cornbined with a solar heating system for the summer period. These systerns provide both space heating and domestic hot water .

In order to offer reliable buyer information, quaIity control and appropriate investment subsidies, the systerns have to be tested by authorities to meet certain requirements. In Sweden there is aIready a standard quaIity labelling on solar collectors. This article deals with procedures to test storage systerns, that for the purposes of this article is taken as the subsystem comprising store, electricaI auxiliary heater, solar and load side heat exchangers together with the tempering vaIve for the domestic hot water (DHW). The indoor system test used the same collector loop for all tests, and can thus be considered as a storage system test for this work. This test is referred to as the six-day test in the rest of the article

A typical Swedish solar heating system comprises a storage volume of 500 -1000 litres with intemal heat exchangers for the production of domestic hot water as weIl as for the solar collector loop. In order to evaluate and improve the systerns' performance, a six -day indoor system test was designed and sixteen different store configurations were tested, four of which are discussed in this article. These have a storage volume of 750 litres. The test is designed to give result s that are representative for the weather conditions during the summer period in Borlänge, Sweden. The six-day test has a domestic hot water load only, as this was considered critical because of its high power. The inclusion of an extra boiler and space heating load was considered to give extra complexity that would make it difficult to analyse the causes leading to differences in performance.

Fig. Flow chart for the structure of this article.

Furtherrnore a store component test as specified in the proposed ISO standard was carried out for one of the storage system configurations tested with the six-day test. The proposed test sequences in the ISO standard were complemented with three new test sequences in order to describe the perforrnance of load side heat exchangers in more detail. The component parameters were identified with DF (Spirkl W., 1995) and the MULTIPORT store model (Drtick H. and Hahne E., 1996). The parameter values for one store configuration were then validated using the measurements of the six-day system test. The system simulation model was then modified in order to allow an evaluation of the other three store configurations.

(7)

Thus, all four storage system configurations could be simulated and evaluated for the six -day test.

In order to evaluate the designed six-day test all four investigated store configurations were then implemented in a system simulation model in order to predict the performance of the entire system for the summer period of Borlänge, Sweden. Measured weather data for five summers were used.

2. Test Apparatus and Measurement System

The test apparatus used in this study is shown in figure 2. It has been built up in order to test the performance of storage vessels and the heat exchangers connected to theffi (Lorenz et al., 1995). The solar collector is emulated by a simple collector hardware simulator (CHS) which can be used to emulate the heat transfer from a collector given a data-file containing weather information (insolation and ambient temperature) and the current temperatures in the test apparatus, or to generate a constant power .

FIXED TEMPERATURE

I

MIXING VALVE

~--y--;

T01 X

~LAUX. Too X :s=' .Too

X All cornponents with in the

dashed box are part of the

"storage system" ELECTRIC HEATER

The test cycle and all measurements are controIled by a PC-based control program. All measurements are made using a digital multimeter which is connected via a scanner to the sensors and to the PC. The control program converts the raw data into standard units, compensating for the individual sensors' calibration curves. Measurement scans take between

~--~-=-~-~ -10 and 15 seconds depending on the number of sensors.

The data are

FLOW SENSOR I averaged before I storing, the period for I averaging depending L on the current status I and type of test.

I Figure 2 shows the I placement of the J sensors in the test rig

-and tank.

temperature sensor

T 04X

T05 ~I

l-:::::w SEN~

GOLLEGTOR

!HARDWARE I

I T

: SIMULATOR ~ )( ..

FLOWSENSOR G

--1 )(T07

PUMP I ---

Fig 2. Schematic of the test apparatus. Valyes B, C and D are used to short circuit the heat exchangers in order to check the accuracy of the inlet and outlet temperature sensors .Valye A controls whether the discharge outlet temperature is constant (open) or yariable (closed).

Temperatures inside the tank are measured in order to assess the temperature stratification in the tank and also in order to calculate the energy in the tank at any given time point.

The instantaneous heat flows to and from the store via the heat exchangers are calculated from the volume flow rate, the temperature difference between inlet and outlet and the calculated fluid densityand heat capacity for those temperatures. The accumulated heat transfer (energy) is calculated using the instantaneous values for power and the relevant time period. The instantaneous electrical power supplied to the auxiliary heater is logged using a signal from the electrical meter that measures both power and supplied energy .

(8)

3. Studied Storage System Configurations

Configuration 1 Configuration 2

Configuration 3 Configuration 4

Fig. 3. Schematic of the four store configurations showing the locations of the intemal heat exchangers and electrical auxiliary heater (black circle). The numbers beside each heat exchanger refer to data for the size of the heat exchanger given in table I.

Figure 3 shows the four storage system configurations that were tested. All tests were carried out on the same physical storage yessel that had four heat exchangers mounted inside, t wo for the solar loop (not used at the same time) and t wo for the DHW production. All four store configurations used the same tempering yalye. Configuration 2 is the same as configuration 1 except that it has a second load side heat exchanger, placed just aboye the solar loop exchanger. Configuration 3 differs from configuration 2 in that it has a larger solar loop exchanger. Configuration 4 differs from configuration 3 in that the electrical auxiliary heater as a lower placement.

Tab. Sizes of the finned copper tube heat exchangers used in the tested storage systerns.

Size Nominal Inside Tube Surface

Type diameter Diameter Length Area

[mm] [mm] [m] [m1

1 22 16.9 11 2.74

2 28 22.5 9.5 3.01

3 22 16.9 15 3.73

The heat exchangers are made from finned copper tubes which are wound in a helix with a diameter of approximately 0.3 m. These are placed inside the tank, offset from the centreline of the tank, as is common practice in Sweden. Table 1 shows the sizes of the heat exchangers used in the four store configurations.

(9)

4. System Test (Six-Day Test)

4.1 lntroduction

The system test was designed to generate data from indoor tests that could be used for different purposes:

The basic requirement for the test was that the result could be used for direct comparisons, i.e. ranking of different system configurations. Furthermore, the solar fraction should be reasonably representative for Swedish summer conditions.

The measured data could be used to validate a simulation model to be used to calculate, among other things, yearly performance.

2.

It was also desirable to have a data set good enough to be able to detemline w hy certain systerns perform better than others. Furthermore, parameter identification requires a test that is performed under a wide variety of conditions.

At the time when the tests were planned and the first ones carried out (1994-5), there was no availahle simulation ffiodel that was validated for the storage systerns that were tested. We studied several "standard tests" (references), hut none of theffi could provide us with all the information that we required.

The test conditions have been created to be representative for the summer in central Sweden (taken as being April -September inclusive). Insolation data for Borlänge (60.48°N) were analysed. Several categories were defined to cover a range of daily insolation and short-term temporal variation, related to theoretical clear sky values. All days were allocated a category number according to the definition for the categories and the whole data set was analysed to find out which categories occurred most frequently, and the probability of one category being followed by another one.

Six days with the most frequently occurring categories were picked out from several years for the period at the beginning of August. These six days were then arranged in the most probable sequence. The days were chosen so that the average horizontal global irradiation for the six days was approximately the same as the average for the whole of the six month summer period, and in addition to give a wide variety of insolation conditions. The changeovers from day to day were smoothed out. Original six-minute data values were averaged to give 15 minute data values, a limit imposed by the function of the CHS at low flow rates. The weather data for the six-day test, created according to the procedure described above, is shown in figure 4.

(10)

The CHS was operated using TRNSYS collector model TYPE 1 together with parameters describing a typical Swedish solar collector, single cover and selective absorber, with an area of 10 m2 facing due south at a tilt of 40° from the horizontal. The energy transfer to the collector loop heat exchanger was the same as that generated by the collector, i.e. pipe losses were assumed to be zero.

Tab. 2. Data for the draw-off profile. Time of discharge, flow-rate, nominal volume (for 9-48°C) and energy content. This two-day profile is repeated twice to give the six-day profile.

The draw-off profile (table 2) was designed to give a daily discharge of 13 kWh. This is a relatively high value for Sweden (Aronsson S., 1996), but is a load consistent with the size of the tank and the solar collector. It represents a DHW load for a household with 5-6 occupants.

The load profile consists of four discharges each day at 7 a.m., noon, 6 p.m. and 8 p.m.. T wo different daily profiles were used in order to a large variety of load conditions. Flow rates in the range 0.1 -0.25 kg/s were used.

4.2 Test Procedure

For each test, the theflnostat for the electrical auxiliary heater was set to the lowest temperature possible so that, with 20°C in the rest of the tank, a 4.88 kWh discharge at a flow rate of 0.25 kg/s could just be drawn off without the outlet temperature declining to a value below 40°C. This setting assures that a standard Swedish bath can be drawn off even if there is no solar energy stored in the tank.

(11)

The test sequence starts with a uniform 20°C in the tank. The electrical auxiliary heater is switched on at the start of the test, and then the whole test is conducted automatically, using the fixed weather data, the CHS unit and the load profile discussed in the previous section.

The solar loop pump and the CHS is switched on if the control program calculates that the collector would deliver more than 250 W. Theyare switched off if the calculated heat flow is less than this threshold. This control method was chosen because the standard differential controller was relatively difficult to emulate in the control program. No real controller could be used as no realistic collector temperature was available. The draw-off loads occur at fixed times, and each is terminated as soon as the desired load energy for that draw-off is achieved.

The inlet temperature is constant at 9°C and the outlet is tempered to 48°C. The auxiliary heater is not controiled by the computer program. It turns on and off through the action of its own thermostat.

The energy balance equation for the six-day test is as follows:

~ol/ + QETOI = QLoad + Q SE + Q Loss ( 1 )

(2)

Where,

QE* = QETOt -Q El ( 3 )

This definition incorporates a compensation for the fact that the tank has a net gain in stored energy for the test {third term) and that the Iosses are apportioned to both solar and auxiliary heating {second term).

4.3 Test Results

The results of the six-day test for the four store configurations described earlier is summarised in table 3. Configuration 1 has much worse performance than the other three. This is due to having only one load side heat exchanger as opposed to t wo which the other three configurations have. Configuration 4 shows the best results. This is assumed to be due to the substantially lower thermostat setting for the auxiliary heater (nearly 10°C lower).

Tab. 3. Summary of results from six-day tests of the four store configurations.

f Auxiliary Set Temp

QE* ^QCol/ ^QLoad Q Loss Q SE Q El Q SE -Q El

System

[%]

52.2 68.9 71.4 76.7

[CC]

76.5 72 72 62.5

[MJ]

220 132 123 90

[MJ]

99 81 86 81

[MJ]

79 56 56 22

1 2 3 4

[MJ]

244 291 306 300

[MJ]

287 285 287 287

[MJ]

137 109 110 92

[MJ]

58 53 54 70

(12)

5. Store Component Test and Simulation

5.1 1ntroduction to ISOIDP 9459-4A

In this chapter a store component test, carried out according to ISO/DP 9459-4A on store configuration 3 is described. The proposed standard is not providing a specific test sequence for intemal load side heat exchangers. Therefore three additional tests to determine the performance of these heat exchangers were designed and carried out.

To be able to accurately simulate the performance of the storage system, the store model parameter values must be validated. Here the MULTIPORT Store Model (Drlick H. and Hahne E., 1996) for TRNSYS (Klein S.A., et al., 1996) was used. This store model is believed to be the most suitable and offers flexibility to model a variety of store types and configurations. It provides five direct connection ports for charge and discharge and three ports to model intemal heat exchangers. The modelling of mantIe heat exchangers is optional.

The MULTIPORT store model is bas ed on a 16PORT store model further developed from a 4PORT store model initially introduced by Marshall and Li at the European Solar Storage Testing Group (Visser H., editors, 1991 and 1992).

Tab. 4. Short description of the tests performed in the store component test. Those marked with * are not part of ISO/DP 9459-4A.

To detertnine the simulation model parameters of a store several tests sequences have to be carried out. The ones used for this work are summarised in table 4. The proposed standard ISO/DP 9459-4A provides test sequences to determine the heat loss coefficient(s) of the store, the position of the auxiliary heater, the effective storage tank volume, the dead volume (water volume above the highest and below the lowest intemal heat exchanger), the degradation of thefInal stratification during standby and the heat transfer capacity rate of the intemal solar loop heat exchanger. Three additional tests (SA, SB and SC), described later, were perfortned to detertnine the heat transfer capacity rate of intemal load side heat exchangers. The store

(13)

model parameters were identified with Dynamic Pitting and are summarised in table 5 at the end of this section.

The component tests were carried out on the same test rig as used for the system test. Direct ports at the bottom and top of the storage yessel were added to be able to discharge the tank directly after each test. Since the sampling rate of the measured data is yariable, the measured data were conyerted to a fixed time step for simulation purposes. A time step of 36 sec was used for tests with high energy transfer rates and 180 sec for tests with low energy transfer rates, e.g. to determine the heat losses of the tank. Furthermore, the number of store nodes in the model was chosen to be 100, referring to earlier inyestigations (Dahm J., et al., 1996).

For parameter identification, olle single fit, is proposed in ISO/DP 9459-4. This was not done in this case as parameters for the upper load side heat exchanger on its own were required for simulation of store configuration 1. Bach test sequence was thus treated separately.

5.2 Solar Loop Heat Exchanger

The solar loop heat exchanger is an intemal fin tube helical coil positioned at the bottom of the tank. Parameters which can be identified in Test C are the heat transfer capacity rate (UA- value) and the outlet height. The outlet height determines the size of the dead volume below the heat exchanger which can not be accessed for storage purposes. Test C makes it also possible to identify the tank volume and it was analysed first since the resulting parameters are a premise for all following tests.

At this point the heat losses of the store are still unknown; theyare negligible since the transferred energies are comparably high during a short test interval. Therefore, an estimated parameter for store heat losses can be used.

3 2.5

0.5 o

10 20 30 40 50 60 70 80 90 100

tm [OC]

Fig.5. UA-value for the solar heat exchanger as a function of the mean temperature for the heat transfer, tro , at a constant flow rate of 0.13 Us

Figure 5 shows the UA-value of the solar heat exchanger as a function of the mean temperature tro. Since the flow rate for the solar loop is usually relatively constant in a real

(14)

system, it was considered unnecessary to identify the flow dependency of the UA-value for this heat exchanger .

5.3 Store parameters

The heat losses of the store during standby can be detenDined from Test L. Theyare not measured but can be detenDined indirectly with parameter identification by comparing the charged and discharged energy. The MULTIPORT store modeIoffers the possibility to define the heat losses as a UA-value to ambient in six different zones. Here the UA-value to ambient was identified for t wo zones, an upper and a lower zone of equallength.

To provide the consumer with a certain amount of hot water at any time a part of the store has to be heated by the auxiliary heater at times of no or low solar input. This auxiliary volume is fixed to the volume ilbove the auxiliary heater. The height of the heater can be determined with test sequence EiA, ISOIDP 9459-4A.

The effective vertical heat conductivity describes the vertical destratification of the temperature field in the storage tank due to heat conduction of the store mantIe, intemal installations as weIl as the heat conductivity of water. ISO/DP 9459-4A provides test sequence EiB to determine this value.

5.4 Load Side Heat Exchangers

The transferred power in the load side heat exchangers during operation is about 4 to lOtimes the transferred power in the solar heat exchanger. Earlier investigations (Dahm J. et al., 1996) showed that the heat transfer capacity rate of intemal heat exchangers is time dependent. A convection plume develops slowly after starting to discharge (Van Berkel J ., 1995). The time until the plume reaches steady state depends on the kinematic viscosity of water, its heat capacityand the temperature differences between inversion layers.

For the investigated systerns the convection plume of the load side heat exchangers is marked- Here high temperature differences between the finned outside and the surrounding storage water occur. A high discharge power causes large convective currents inside the tank and it takes a certain time until the convection plume inside the tank reaches steady state.

Consequently the UA-value of the heat exchanger rises until steady state convection is reached. A basic assumption here is that this effect is too large to be neglected.

The MULTIPORT Store Model originally described the UA-value of a heat exchanger with an empirical equation:

t bhx,3

VA. = C.rhbbx.l .dt~bX.2 m,1

hX.l hx 1 (4

To take a time delay by calculating the UA- value in the beginning of a discharge sequence into account, this equation was modified, together with the ITW , Stuttgart, to the following:

bhx,3 'm,i

VAhx,i = ^1::rhx,t ..mhx C ^.bbxl ' .m,idtbhX2

5 )

(15)

where

Fhx.t = I=U Shx

(6)

Store configuration 3 is equipped with t wo serial coupled heat exchanger modules. These were tested using three test sequences, called SA, SB, SC, which are described in the following sections. To obtain information about the ffiass flow dependence, each test sequence was carried out three times with different constant ffiass flow rates (0.08 kg/s, 0.17 kg/s, 0.25 kg/s).

5.4.1 Test sequenceSA

The identified parameters from this test sequence are the outlet height of the upper heat exchanger module, to deterrnine the inaccessible volume above the heat exchanger in the top of the tank, and the UA-value as a function of the ffiass flow and mean temperature of the heat transfer. Figure 6 shows the power vs. time chart during the test for the ffiass flow rate of 0.17 kg/s.

o 1 2 3 4 5 6 7

time [h]

Fig. 6. Test SA; calculated and measured power and the difference between them during the test: l. conditioning, 2. heating the auxiliary volume of the tank to 80°C, 3. Discharging through the upper heat exchanger module with a constant flow rate.

(16)

The identified outlet height of the heat exchanger is 0.91. Thus the dead volume in the top of the tank is calculated to 66 l. Since this test is not explicitly designed to identify the capacitance factor for the calculation of the time dependent factor, the capacitance factor is strongly correlated to allother identified parameters. The UA-value for the upper heat exchanger module, as identified from this test, is shown in figure 7.

Fig.7 .UA-value of the upper load side heat exchanger as a function of ffiass flow and mean temperature of the heat transfer.

5.4.2 Test sequence SR

The parameters which can be identified from this test are the UA-values of the lower and upper heat exchanger module as a function of the ffieall temperature of the heat transfer and the ffiass flow. The dead volume in the top of the store, identified in test SA, is given as an input. Figure 8 shows the power vs. time chart during the test for the identified parameters at a constant ffiass flow rate of 0.17 kg/s.

(17)

~

Gi

~ o o.

o 1 2 3 4 5 6 7 8 9 10 11 12

time [h]

Fig.8. Test SB; calculated and ffieasured power and the difference between theffi during the test: t. conditioning;

2. heating the whole voluffie of the tank to 80°C; 3. discharging the store through both heat exchangers.

To exclude the influence of the time dependency on the UA-value during the parameter identification, the first 10 minutes after starting the discharge were not used.

Fig. 9. UA-value of the upper and lower load side heat exchanger as a function of the ffiass flow rate and mean temperature of the heat transfer .

(18)

It can be seen that the UA-value of the lower module is higher than the value for the upper module for the same operating point. This must be due to a better heat transfer on the outside of the tubes, the dominant factor for the UA-value of these heat exchangers. The difference between the modules for the d I capacitance factor and the UA- upper mo u e value may both be due to t wo

factors; there is a smaller driving load side temperature difference for the

upper module, and part of the lower module convection plume generated by the

.--upper module is passed

downwards to the lower module which itself generates a plume downwards (figure 10).

--.

solar module

Fig. 10. Convection plume inside the store during discharge through both load side heat exchangers.

5.4.3 Test sequence SC

Test sequence SC was designed to identify the time dependent factor for the upper and lower load side heat exchanger. The parameters to calculate the UA-values derived from test SB are used as input. Test sequence SC is similar to test SB except that the store is discharged in

10

o

-10

~ ...

al

~ o o.

-20

-30

-40

-50

o 1 2 3 4 5 6 7 8 9 10 11 12 13 14

time [h]

Fig. 11. Test SC; calculated and measured power and the difference between them during the test: 1.

conditioning; 2. heating the whole contents of the store to uniform 80°C; 3. discharging the store in intervals through the load side heat exchangers.

short intervals with a constant flow rate. Here the start up phase of the draw-off is of most interest, since a real withdraw is of short duration. Thus, the first 10 to 15 minutes after

(19)

starting a discharge are of major interest. Figure 11 shows the power -time chart during the test for a flow rate of 0.17 kg/s.

0.8

::I:: 0.6 -

><

.c LL 0.4

-Fhx,t -lower module -Fhx,t- upper module

0.2

ma ss flow = 0.17 kg/s

O T I I I I I I I I I

O 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 time [h]

Fig. 12. Test SC: variation in time of the capacitance factor for the upper and lower load side heat exchanger for the ffiass flow rate 0.17 kg/s.

The time dependent factor Fhx,t increases slower for the upper module. Consequently the capacitance factor for the upper heat exchanger module is greater than the one for the lower module.

5.5 ldentified Parameters

The identified store parameters are listed in table 5

Tab. 5. Identified parameters for the MULTIPORT store model for configuration 3; values in brackets are parameters for the upper load side heat exchanger from test SA.

""L.- -L- --,~~~~~~i~~j,()Ln .~n~-,~~i

effective store volume (height = 1.58 m) m- Q.73

effective vertical thermal conductivity W/(m.K) 1.73

Number of nodes in the store 1 QQ

WIK 2.27

heat loss of the lower half of the store

heat loss of the upper half of the store W IK 1.54 relative auxiliary heater and controller position -0.69

Lower DHW Heat Exchaneer Module

relative outlet position of the lower module -0.5

constant C in equation WIK 315.6

ffiass flow exponent bhxl -0.15 .

mean temperature exponent bhx3 -0.55

capacitance factor Shx kg 8628

Upper DHW Heat Exchaneer Module

relative outlet position of the upper module -0.91 (0.91)

(20)

mass flow exponent bhxl -0.18 (0.17)

mean temperature exponent bhx3 -0.75 (1.03)

capacitance factor Shx kg 15360

(1890) Solar Heat Exchane:er

relative outlet position of the morlule -0.01

constant C in equation WIK 10.9

mean temperature exponent bhx3 -1.1

There is one remaining parameter to identify after analysing all tests, which is the setting of the auxiliary heater thermostat. An evaluation of the function of the auxiliary heater thermostat, as weIl as a validation of the storage model parameter values is carried out by resimulating a dynamic test sequence for the whole store configuration.

5.6 Validatian af Parameter Values

Here, the obtained parameters were verified with a resimulation of a full scale test. The measured six-day system test data are used for this purpose.

Since the controller settings of the electrical heater are not identified, the measured auxiliary power and the switch on/off signal from the six-day test is initially provided as an input for the calculation ffiodel. Aiso the input temperatures and ffiass flows to each connection port have to be provided as inputs from the ffieasurement. Thus, it is possible to compare the calculated power and the total transferred energy through each connection port of the store with the measured values.

The relative error in transferred energy (ISO/DP 9459-4, equation 8.1)

Qx,c -Qx,m .100%

e =

Q

x,Q x,m 7

was calculated for each connection port and listed in table 6.

Tab. 6. Relative error and daily measured transferred energy for each connection port of system 3 for the resimulation of the six-day test; Auxiliary heat is input.

Rel. error [% ] / Qmeasured [kWh]

day 1 day 2 day 3 day 4 day S dav6

Solar hx sum load Auxiliary hXOHWl +hXOHW2 heater -0.5/34.1 -0.1/-13.2 0.0/24.0 -11.7/20.3 -2.9/-13.1 0.0/0.0

-13.1/8.4 -2.8/-13.2 0.0/1.2 -127.1/0.3 0.2/-13.1 0.0/7.2

2.5/6.1 0.0/-13.0 0.0/9.6

0.4/15.2 -4.1/-13.1 0.0/7.3 SUM -4.4 / 84.4 -1.6/-78.7 0.0/49.3

Using equation (2) the solar fraction for the measurements is calculated to be 62.6%, while the simulation gives 63.3%. However, since the relative error is exceeding i:3% for the solar heat

(21)

exchanger (table 6), the deterrnined pararneters are not validated according to ISO/DP 9459- 4A. Especially on day t wo and three the solar loop charges the tank less than measured and the calculated load is higher .

The auxiliary heater is a conventiona! imrnersion electrica! heater with an integrated temperature sensor insta!led at the same height of the tank. The rea! controller was set to 72°C (switch-off). Its dead band temperature range was unknown. The temperature in the node of the auxiliary heater was not measured. To identify the heater settings to be applied in the model (the switch on temperature and its dead band temperature range) the simulated temperature of the height where the auxiliary heater is placed was investigated. During the resimulation of the six-day sequence the measured auxiliary power and the switch on/off signal was provided as an input. Comparing the measured operation mode with the simulated temperature, the heater settings were determined (see figure 13) to be 65°C (switch-on) with a dead band of 5K. The first day of the simulation was neglecting because of inaccuracy in the results for that day, as is discussed later in the article.

>-

~

"x

~ IV 01 :5 -O .E C) 'Qi.c 01 :5

"=

01

~

~ E01 1-

Fig. 13. Identification of the heater's controller settings

Applying these settings, and with measured values as input to the heat exchangers, the six-day simulation was repeated.

Tab. 7. Relative and absolute error in transferred energy for each connection port of system 3 for the resimulation of the six-day test; Auxiliary settings are applied.

rel. error [% ] / abs. Error fkWhl

Solar hx SUM load hx

day 1 day 2 day 3 day 4 day 5 day 6 SUM

6.4 / 2.2 -7.4/-1.5 -12.9/-1.1 -128.7/-0.3

2.2 / 0.1 0.3 / 0.04

-5.4/0.7 -2.2/-0.3 -2.8 /-0.4 -1.1/-0.1 0.2 / 0.02

0.0/ 0.0

Auxiliary heater

(22)

Now measured and calculated energy transfer for solar and load heat exchangers are acceptable (table 7), but the solar fraction is -5% higher compared to providing auxiliary heat as input. The parameter values are still not validated with regard to ISO/DP 9459-4A as the energy transferred to the tank via the electrical auxiliary heater shows a relative error of - 10.6%. Laoking at the absolute values it can be seen, that the greatest error occurs on day one, where the heater is providing much less energy than measured. This is the principal cause of the large difference between measured and calculated solar energy for days one to three. If the auxiliary is provided as input (see figure 13), the model calculates consistently higher temperatures than the actual heater set temperature on the first day.

W hy is the model giving such erroneous results on day I ? This can be answered as follows: a closer look at the measurements and the auxiliary heater operation during the first day of the six-day test shows that the electrical heater switches on at times when it should not. This occurs during periods of charging with high heat flows through the solar heat exchanger, even though no discharge is occurring. The reason for this is a decrease in temperature at the height of the auxiliary heater' s thermostat during solar charging, caused by the upwardly moving plume from the solar heat exchanger .

The impulse of the upward ffioving water occasions the plume to rise to nodes with a higher temperature than in the plume. In these nodes mixing occurs between the plume and the water in the nodes, causing a decrease in the temperature in the nodes. This effect will be strongest when the stratification inside the tank is marked and the solar power is great. Typical conditions are a sunny day for the boundary node between the cold part and the Warffi auxiliary part of the tank. The six -day test sequence has extreme conditions on day one with 20°C in the cold part of the tank and a sharp boundary to the hot part at over 70°C. These conditions do not usually occur in real conditions. This effect is also noticeable on the last day of the validation sequence. The auxiliary heater and its thermostat sensor are placed at the boundary between cold and hot and, as soon as the plume (the solar power) is marked enough to mix the boundary node to a lower temperature, the heater is switched on. The higher the difference is between auxiliary set temperature and the temperature in the plume, the greater the influence of the upward ffioving plume on the heater .

In the ffiodel, in cases where an inversion is calculated, a mixing routine transfers energy upwards/downwards until the inversion is removed. In practice this ffieans that, for example during solar heating, all nodes above the heat exchanger are heated evenly up to the first node where there is a higher temperature. No energy is transferred to a node with a higher temperature in this routine. The one dimensional simulation ffiodel, as it is now, is thus not capable of modelling the above mentioned effect.

(23)

Fig. 14. Measured (black, thin) and simulated (grey, thick) temperatures and transferred powers for day six-day test using simulated auxiliary theunostat settings.

of the

Figure 14 shows the measured and simulated temperatures and transferred powers for day 1.

Up until 9 hours from the start, the measured temperature at the top of the upper load side heat exchanger is higher than the simulated one. Measured data shows a 2-3 K increase in temperature with height above the auxiliary whereas the simulation gives constant temperature above the auxiliary due to the mixing routine described earlier. From 9 until 12 the measured auxiliary energy is much greater than the simulated leading to much higher measured temperatures above the auxiliary .At approximately 13 the upper temperature decreases sharply. This is due to mixing with the relatively cool plume from the solar heat exchanger. This the same process that causes the auxiliary to tum on between 9 and 12.

5.6 Summary

Several tests to identify parameters for the MULTIPORT store model, system configuration 3, were performed according to the proposed standard ISO/DP 9459-4A. The store provided t wo serial coupled load heat exchanger modules. With focus on this type of heat exchanger three addition al test sequences were designed and carried out: Test SA concemed only the upper module; test SB focused on both modules; test SC was designed to identify a time dependent factor for the heat exchanger UA-value. Based on the experiences from test SA and SB, it can be stated that serial coupled heat exchangers mounted in the same storage vessel influence each others performance. Therefore the UA-value parameters for both heat exchangers have to be identified in one test sequence. However, the results of test SA can be used for the simulation of configuration 1 which only has one load side heat exchanger .

(24)

Furthermore, the model is not capable of modelling the influence of solar charge on the auxiliary heater during certain conditions, as the store model is not able to model the influence of the convection plume generated by the intemal heat exchanger in detail. The result is that the validation of the store model parameters does not comply 100% to the requirements in ISO/DP 9459-4A.

6. Evaluation of the Tested Storage Systerns

The following simulations are for systems where the auxiliary heater is not affected by the plume from the solar heat exchanger. The derived model for store configuration 3 is used as the basis for evaluating configurations 1, 2 and 4. Except for minor differences, i.e. electrical heater location and settings, number of load side heat exchanger modules and solar heat exchanger size, all parameters are the same as for configuration 3.

Configurations 1 and 2 are equipped with a smaller solar heat exchanger. Since the heat exchangers are of the same type and are mounted in the same way, the heat transfer capacity rate in the model is changed according to the difference in area ratio.

Store configurations I and 4 have different auxiliary heater settings compared to configurations t wo and three, and in configuration four the heater is placed at a different height. In the ffiodels the auxiliary heater settings and auxiliary volume are changed by the same absolute amount as for the measured storage systerns, referred to configuration three.

The dead band temperature range was kept the same. Changes compared to configuration 3 are shown in table 8.

Tab. 8. Parameter va1ues for store configuration 3 compared to those for configurations 1,2 and 4.

Svstem 1 System 2 Svstem 3 ~~stem 4

model parameter

model Darameter

model parameter

model parameter Test result

from test SA used

Test results from test SB and SC used

measured Store

One heat exchanger

module

3.01 m2 Pararneter 71 : 8.7 WIK

measured Store

T wo heat exchanger

modules

3.01 m2 Parameter 71 : 8.7W/K

measured Store

T wo heat

exchanger modules

3.73 mi Pararneter 71 : 10.9 WIK

measured Store T wo heat

exchanger modules

3.73 m2 Pararneter 71 : 10.9 WIK Load side

heat exchanger Area solar

loop heat exchanger

Auxiliary volume Controller settings Heater

DOWer

2501 Parameter 40:

0.685

2501 Parameter 40:

0.685

2501 Pararneter 40:

0.685

4071

76.5 °C Tset = 74.5 DC

dbT = -5 K

72°C Tset = 70 DC dbT = -5 K

72°C Tset = 70 °C dbT = -5 K

62.5 °C

3kW 3kW 3kW 3kW 3kW JkW 3.3kW

Pararneter 40:

0.5*

Tset = 60.5°C

dbT = -5 K

3.3kW

T wo evaluation studies were carried out using the measured values for temperature and flow as input to the solar and lower DHW heat exchangers in the models. First the auxiliary power was provided as an input for the simulation, since it is expected (as learned from system 3) that the control function of the heater is not the same in the model as in the measurements.

Second, the auxiliary heater settings from table 8 were used to simulate the heater function during the six-day test. Table 9, columns t wo and three respectively, shows the solar fraction of the systems for these t wo simulation rons.

(25)

Tab. 9. Solar fraction [ % ] of the configurations by resimulating the six-day system test, using the measured inlet temperatures and flows for the DHW and solar heat exchangers as input for the simulation.

With auxiliary power as input all system models are calculating the solar fraction within a range of +1- 2.3% compared to the measurements. This is considered accurate enough to state that all system models are able to be used for long term simulations, assuming that the function of the thermostat for the auxiliary is modelled perfectly.

Simulating the auxiliary heater, on the other hand, shows a different result. The solar fraction is higher for configurations 1, 2 and 3, and there is no major difference in performance between configuration 2, 3 and 4. A higher set temperature of the auxiliary heater causes a greater temperature difference between auxiliary volume and the rest of the store. Thereby the upwards ffioving convection plume, generated by the solar heat exchanger, has a greater impact on the controller by mixing (cooling) the node of the auxiliary heater. The function of the auxiliary heater in the ffiodel is therefore not validated.

One conclusion is that, within limits, store configurations of the same kind as shown here can be evaluated without addition al component tests. Further, it can be concluded that the main difference in performance between configuration 2, 3 and 4 in the six -day system test is due to the auxiliary heater, and the difference between configuration 2 and 3 in size of the solar heat exchanger has a minor influence on the performance.

7. Evaluation of the System Test

The previously described results show clearly that the designed system test could be used for parameter value validation purposes. One precondition for the design of the test sequence was, besides enabling a system ranking, that the solar fraction should be representative for the weather conditions during the summer period of Borlänge, Sweden.

To check, if the six-day test is representative conceming real weather conditions, the solar heating system, including all units, e.g. solar collector, store, controllers, load profile, has to be built up and simulated with TRNSYS using weather data for the region. The main components of the simulation model are the store model with identified parameter values and the solar collector model, i.e. TRNSYS TYPE t. The collector model is assumed to be temperature controIled and starts operation at 5°C and switches off at 3°C above the store's bottom temperature. The load profile for the simulation is the same two-day profile as applied in the six-day test.

All store configurations were simulated using the weather data from '88 to '92 for Borlänge.

The simulation for the summer period is starting on the 1 st of April and ending on the 31 st of

(26)

September. Additionally one simulation run was done by using the weather data from the six day test sequence. A time step of 36 seconds had to be applied due to the short and varied time intervals of draw-offs in the load profile.

The solar fraction, using Borlänge weather data, is calculated:

Q Q Q ColI

ColI -LoSS (n Q )

f B = ~oll + Etot ( 8 )

QLoad

Here the electrical energy supplied at the start and the net energy gain of the store at the end of the simulation are negligible compared to the length of the simulation and the total transferred energles.

Tab. 10. Solar fraction [% ] of the configurations by resimulating the six-day system test using the test insolation (figure 4) and using measured weather data for Borlänge and Göteborg.

Calculated 6-day test

(Paux is calculated}

Borlänge Göteborg

38.0 61.7 62.0 63.1

System 1 System 2 System 3 System 4

The five year averaged solar fraction for the summer half-year for each configuration is shown in table 1 Q. It is obvious that the absolute solar fraction is about 8 % less for each configuration than calculated using the six-day test weather data and about 10% less than measured in the six-day test sequence. However, the qualitative ranking among the four configurations remains the same as by resimulating the six-day test sequence with the simulated auxiliary heater (table 9) and using the six-day test weather data. Additionally the solar fraction calculated using the weather data for Göteborg 1984 (close to average) shows a solar fraction in the same order of magnitude as for Borlänge.

With the result from the simulation of the summer period using real weather data it can be concluded, that the designed six-day system test does not result in a particularly representative solar fraction for the weather conditions of Borlänge.

8. Discussion

A six -day system test was designed and four storage system configurations were tested.

Characteristic for the design of this test was that the test should result in a solar fraction ( calculated directly from the measured energy balance) representative for the location of Borlänge and that the test should be suitable for model (parameter values) validation. The major results are that the six-day test is suitable for system ranking and model validation.

However, the test sequence and/or the test procedure has to be modified if it is to result in a solar fraction better representative for real conditions. The result here is that, using real weather data for the summer period (6 months), the solar fraction derived is about 10% less than measured during the six-day system test and about 8% less than using the system test weather data for the simulation.

summer weather data '88- '92

summer weather data '84

47.6 71.0 71.3 72.7

37.3 63.1 63.4 64.2

Six-Day System Test and Component test and System Simulation for Combitanks with Internal heat Exchangers

SERC

for

Combitanks with Internal heat Exchangers and

Component test and System Simulation

Solar Energy Research Center Centrum för solenergiforskning

Six-Day System Test

CONTENTS