Category Archives: EuroSun2008-7

Presentation of the installation and the first experimental results

1.1. Presentation of the installation

image556

The aim of our solar cooling installation is to cool four classrooms (total area = 170 m2). We can notice that there are not backup systems and the average temperature in the classrooms is not steady depending of the outside conditions (Tout — Tin = 5 to 7 °C). We can see all main components on fig. 1:

1. an absorption chiller (frigorific power = 30 kW),

2. 36 solar collectors (total area = 90m2),

3. a cooling tower (cooling power = 80 kW),

4. a hot tank (V = 1500 L) and a cold tank (V = 1000 L),

5. 13 cold fans which provide frigorific power to the classrooms.

Cold storage TRNSYS type

A user-defined type was developed to describe the behaviour of an ice-making ice-encapsulated cold storage. The storage can be represented like a stack of layers (nodes), each one containing a certain number of capsules. The number of capsules in each layer is directly calculated as follows:

N = Vtank і1′ Z)

capsules

y capsule 1 v layers і

where Vtank is the tank volume, Vcapsule is the capsule volume and є is the void fraction around the capsule, typically in the range 0.4 to 0.5 depending on the capsule shape. Heat balance equations over each layer yields the following:

^refVrefCpref j. ■ refCpref Tj-1 Tj NcapsulesQref^capsule

dt when mref is downward (3)

^refVrefCpref dt m refCPref Tj+1 Tj NcapsulesQref ^capsule

dt when mref is upward (4)

image602 Подпись: 144 • kk, (4.00 • k„, (Tj image604 image605
image606

where Sref, Vref and Cpref are the refrigerant density, volume and specific heat respectively, mref is the refrigerant mass flow rate, Tj is the refrigerant temperature within the j layer and Qref^capsule is the heat transferred from refrigerant to capsule. The expression for Qref^capsule depends on the particular state of the PCM inside the capsule. Four conditions are possible: 1) PCM liquid, 2) PCM in transition from liquid to solid, 3) PCM in transition from solid to liquid and 4) PCM solid. When the PCM is completely liquid or solid, the lumped capacity model is used, assuming the PCM temperature is uniform in the capsule and equal to the temperature of the layer. During PCM phase transition inside the capsule, the following experimental equations, as described in [6], have been used:

where L is the latent heat of water, Vcapsule is the capsule volume, Ro is the capsule radius, 5wat and kwat are the density and thermal conductivity of water, 5ice and kice are the density and thermal conductivity of ice, Tpc is the phase change temperature, t is the time since phase change has started, and t* is the dimensionless time. For a detailed description see [8].

5 Simulation results

In a parametric study, the influences of the collector field area, size of the latent heat storage (commercially available nodule ice storage), the chilled mass flow that is to be cooled and the possible effect of dust on the reflector area, were investigated. Some results are presented below.

5.1 Evaluated Parameters

To rate the system two performance figures where evaluated: The extracted heat per incident beam radiation as total efficiency (ntotal) and the ratio between the cooling work carried out by the solar cooling system and the total cooling work required by the given process for the specified flow rate of the return flow as described in 5.3 (Solar fraction, SF).

Подпись: (7)Подпись: (8)extracted, solar

Подпись: total— |y°l

beam, incident

Qextracted, solar

Qtotal [%]

Simulation of the adsorption chiller

Since the discontinuous dynamics of adsorption/desorption cycle causes a variable cooling effect of the chiller, it is important to have a model that takes into account this transient behavior. The dynamic model proposed by Saha [5] has been implemented in INSEL. Given that only little information on the physical properties of the chiller and the nature of the silica-gel used is available, some modifications/simplifications in the model have been done. The silica-gel/water equilibrium model based on Henry’s law equation (1) developed by Ng and Chua [6] for the silica — gel Fuji-Davidson type RD was used to describe the adsorption/desorption process.

q* is amount of water adsorbed at equilibrium [kg/kg] (1)

Gat is the isosteric heat of adsorption (Gst=2510 kJ/kg)

KQ is a constant (if? = 2.1C-1S Pa"1)

P is the equilibrium pressure of the adsorbate in the gas phase [Pa]

image014

The model was compared with measurement data (in 10 seconds time interval) of one chiller for different operating conditions of temperatures and volume flow rates. Figures 5 shows the simulated and measured temperature values of the 3 water circuits during 3 cycles for a heating temperature of 78-80 °C and a volume flow rate of

image015

Figure 6 shows the measured and simulated temperatures for a heating temperature of 70-72°C and a volume flow rate of 90 m3/h.

Подпись: Figure 7: Simulated and measured values of the adsorption/desorption cycles (case 3) 5

The results for a hot water temperature of 70°C and a reduced hot water volume flow rate of 50 m3/h are shown in figure 7.

Figure 8 shows how the model behaves when the inlet cooling water temperature fluctuates, which happens often in reality when the cooling towers cannot cool down the water coming from the chillers.

image017

2. Discussion

In this part, the results of the simulations of both the solar plant and the adsorption chillers are discussed and different ways of using the model for optimization of the system via online simulation will be proposed.

Desiccant Cooling System Modelling and Optimisation

S. D. White[4]*, P. Kohlenbach1, C. Bongs[5]

1 CSIRO Energy Technology, PO Box 330, Newcastle, 2300, Australia
2 Fraunhofer Institute for Solar Energy Systems, Heidenhofstr.2 79110 Freiburg, Germany
* Corresponding Author, stephen. d.white@csiro. au

Abstract

The performance of a once-through solar desiccant cooling system, for airconditioning a commercial office space, was modelled using the TRNSYS computer simulation software package. The study particularly focused on the potential for designing and operating a desiccant cooling system, without any thermal backup provided to mitigate for intermittent solar availability in three Australian cities. Increasing (i) indirect evaporative cooler effectiveness, (ii) air flow to the office space, and (iii) solar collector area were all shown to reduce the frequency of high temperature events inside the building occupied space. In the warm temperate climate of Melbourne (and to a lesser extent Sydney), high ventilation rates enabled comfort conditions to be maintained at or near acceptable levels in the occupied space, with-out the use of a backup thermal source. The synergy between evaporative cooling and solar desiccant cooling, observed in the warm temperate climates, was not evident in the tropical Darwin climate, suggesting that the selected ventilation desiccant cooling cycle is not appropriate for tropical climates.

Keywords: desiccant cooling, solar cooling, comfort

image121This study further develops the results of the previous paper, particularly focussing on (i) the potential benefit of improving the performance of the indirect evaporative cooler, and (ii) the performance of the solar desiccant configuration in other climates.

Simulation environment

The base installation is made up off a solar installation with water collectors and an absorption chiller. The meteorological data used have been obtained of the program Meteonorm® with the format TMY. [1]. The simulations have been executed with the program TRNSYS v16 [2], that offers a simulation environment mainly focused to solar installations in particular and energetic ones in general.

2. Loads description

For the simulations, they have been modelled the loads of a real installation, the ones of the facilities that the Research Centre CARTIF has in Boecillo (Valladolid). On the fig. 1 can be seen inside the box in red the air-conditioned zone in which the working hours are from 7:00 to 15:00. The area of the air-conditioned zone is 200 m2, and has the distribution shown in the next figure:

image251

Fig. 1. Plan of the building with the air-conditioned zone.

Based on a physical model of the part of the building to be used and an estimation of the internal loads, it was developed a model with TRNBUILD ® of the zone that was going to be conditioned.

The maximum load obtained was 7 kW for refrigeration, being this the power of the chiller chosen, and for weekends the result obtained was 3 kW. Subsequently they were compared the data obtained by means of the simulation with those measured on the real facility and they were extremely similar.

Evaluating the maximum power on the surface, it’s obtained a value of 36 kW/m2. Note that this value corresponds with the ones obtained within the project SACE [3] for this kind of locales (30 kW/m2).

Daily performances

In this section, data and performance figures of the system in a typical day of operation are presented. Figure 2 shows on a psychrometric chart the measured air status in the typical cooling and dehumidification process of the system. It has to be noted that the measured values of humidity ratio are affected of errors probably due to the accuracy of the relative humidity transmitters used. For instance, the outlet humidity ratio of the pre-dehumidification coil CC1 is slightly different from the outside air value, even if the coil was set to off. Also the slope of the dehumidification line (1-3) is lower than the enthalpy line, contrarily to the well known thermodynamics of the process.

1 Подпись: 5 7 9 11 13 15 17 19 21 23 Humidity ratio [g/kg] Fig. 2. - Measured values of the desiccant cooling process in the T-x-diagram - date: 25.07.2008 time: 16.00 — 2 cooling coil — CC1

2 — 3 dehumidification — DW

3 — 4 HX and infiltration of return air

4 — 5 cooling coil — CC2

5 — 6 building

6 — 7 humidification of return air — HU

7 — 8 heat exchanger — HX

8 — 9 condensation heating coil HC1

9 — 10 solar heating coil HC2

10 -11 regeneration — DW

In addition, the humidity ratio at outlet of the heat exchanger (point 4) is higher than at outlet of the desiccant wheel (point 3), probably due to the air leakages from the return air side to the supply air side. For the considered operation conditions, it was estimated a leakage flow rate of about 18% of the whole return air flow rate. In order to reduce the relevance of leakages into the supply air it will be necessary to improve the air sealing system.

Подпись: 90 80 70 60 50 „ О о 40 “ 30 20 10 0 Подпись: O c-i rn ^ IT, 6 К CC C О c-i rn ^ IT, 6 К CC C О -2 cj Fig. 3. Solar radiation, collectors thermal power, heating coil power and various temperatures - date: 25.07.2008 I sol — Solar radiation

P_coll — Solar collectors P_HC 2 — Solar heating coil

T_coll out

Подпись: — — T coll in

T regeneration

4

Figure 3 shows the performance of the solar plant and the regeneration temperature achieved in the heating coil HC 2. It can be noted that the maximum regeneration temperature achieved was 63°C, providing a heating power of 6.4 kW to the regeneration air flow.

image329
In Figure 4 some results related to the regeneration process are presented. It can be noted that the operation of condensation heat recovery coil HC1 is not continuous but intermittent according to the switching on and off of the chiller. The temperature rise up achieved by means of the condensation coil HC1 is in the range of 12-13 °C. The same figure shows the distribution of the heating power for the regeneration by the heat exchanger, the condensation and solar coils (PHX PHCi P HC2 ) and the efficiency of the heat exchanger ranging from 54 to 70%.

Подпись: Figure 5 shows the results of the cooling power produced in the AHU. It can be noted that the supply temperature of 20°C is met by means of the second cooling coil CC2. No additional dehumidification is generally required to reach the desired supply humidity ratio, due to the moderate humidity ratio of the outside air of about 14 g/kg. The contributions of the cooling power produced by the desiccant cycle PDEC and the auxiliary cooling power PCC2 are shown for the considered sample day of operation.Подпись: OJ) O' image332P CC2 — Cooling coil P DEC — Desiccant cooling P CC1 — Pre-dehumidification coil

——— x set point supply

x supply —•— T supply

——— T set point supply

— — x ext — Outside humidity ratio T ext — Outside temperature

Fig. 5. Cooling power of desiccant cooling cycle and auxiliary cooling coils — date: 25.07.2008

Practice test with real time data

In a next test both models, i. e. Type 307 (PD2) and Type 177, are fed with two weeks of real time measuring data of the solar cooling plant at Ebner Solartechnik. Since both types do not account for transient effects only stationary operation periods have been considered. Fig. 6 and Fig. 8 show the simulated cooling capacity (QE, sim) compared to the measured capacity at the plant (QE, Ebner).

Regarding these curves, large deviations are observed for Type 307 (PD2). Here, the simulated cooling capacity is always higher than the measured data at Ebner. The reason for this is different operating conditions for the chiller at Ebner compared to the chiller measured at ILK. Especially the cooling water mass flow rate at Ebner is only the half of the nominal mass flow rate (i. e. 5000 kg/h). In addition, the chilled water flow rate differs from the nominal value (2000 kg/h) and

Подпись: 40 £ § 35 * 30 25 20 M a "" 15 Подпись: 2100 2300 2500 2700 2900 ME / (kg/h) Подпись: Fig. 7. Deviation of the cooling capacity as a function of the mass flow for Type 307 (PD2)image436Подпись: 1st International Congress on Heating, Cooling, and Buildings, 7th to 10th October, Lisbon - Portugal / 0 5 10 15 20 25 Measured capacity QE,Ebner / kW Fig. 6. Cooling capacity, real time versus simulated value for Type 307 (PD2) 25

5s

I 20

In contrast to the results of Type 307 the appli­cation of Type 177 to the measured data at Ebner show (surprisingly) good agreement for the cooling capacity (cf. Fig. 8). Due to high chilled water temperatures (e. g. tEo is often higher than 16°C) most of the Ebner data have a temperature lift in a range of 10 to 15 K which is considerably lower as included in the ILK test data (13 to 29 K). From the ILK data a decreasing slope with decreasing AtACE is derived[11]. If this decrease is too pronounced com­

Подпись: 20Подпись:Подпись:image441

Подпись: 25
Подпись: 0 5 10 15 20 25 Measured capacity QE,Ebner / kW Fig. 8. Cooling capacity, real time versus simulated value for Type 177

varies in a range of ~2200 to 2900 kg/h. Fig. 7 shows a decreasing deviation with increasing chilled water mass flow for Type 307 (PD2). This can be ascribed to an increasing heat transfer coefficient at the external side of the evaporator, which partially compensates the reduced heat transfer coefficient in the cooling water circuit.

pared to reality, the slope is calculated too low especially at low AtACE-values. Thus, the reduced cooling capacity at Ebner (having its physical cause in a reduced cooling water flow rate) is coincidentally calculated ‘correctly’ by the characteristic equation with variable parameters. As a consequence, due to the scatter in the measured data the variable parameters seem to be too uncertain for application at the lowest end of temperature lifts (e. g. AtACE < 13K).

On a first view the result is as expected: Better agreement of measured data to simulated values if variable parameters are used. Nevertheless, in this case the better agreement is just an artefact and even the method of characteristic equations has to be improved for non-nominal flow conditions.

3. Conclusion

Concluding the results it has been found that both TRNSYS-Types are able to predict laboratory test data correctly. For Type 107 based models care has to be taken when creating the performance data file. The opportunity to fill in a great many of supporting points describing the chillers’

performance easily entices the user to replicate each measuring point without taking the unavoidable measuring errors into account. Thus, if measured data are used in the performance file in combination with stepwise interpolation, numerical results are obtainable, were the physical background is questionable. Care has to be taken, that the same flow rates are used in the simulation as have been used to prepare the performance data file. Alternatively a further improvement of Type 107/307 would be necessary to consider mass flow deviations inside the performance data file by an enhanced look-up approach. Only apparently the simulation results with Type 177 show better agreement to the measurements of a solar cooling plant. However, also Type 177, which uses the method of characteristic equations, has to be improved for simulating non-nominal flow conditions.

Acknowledgements

The authors would like to thank Ebner Solartechnik for their collaboration and especially Stiftung Sudtiroler Sparkasse for the financial support.

Nomenclature

A

Area

m2

B

Duhring parameter

COP

Coefficient of performance

cp

Spec. heat capacity

kJ/(kgK)

f

Fraction

M

Mass flow rate

kg/s

PD

Performance data file

Q

Capacity

kW

r

Axis interval of characteristic equation

kW

s

Slope of characteristic equation

kW/K

t

External temperature

°C

AtACE

External temperature lift

K

AAt*

Modified characteristic temperature difference

K

Subscripts

A

Absorber

Ebner

Chiller at Ebner Solartechnik

o

Outlet

avg

Averaged

G

Generator (Desorber)

sim

Simulated

C

Condenser

i

Inlet

set

Set value

DEI

Design energy input

ILK

Chiller at ILK-Dresden

var

Variable

E

Evaporator

NCC

Nominal cooling capacity

X

Placeholder

References

[1] Henning, H.-M.: Solar-Assisted Air-Conditioning in Building — A Handbook for Planners. Springer-Verlag/Wien, 2004.

[2] Witte et al: Modelling of a Solar Combi-Plus System for residential and small commercial applications. estec 2007 3rd European Solar Thermal Energy Conference, June 2007, Freiburg i. Br., Germany; pp. 444-445.

[3] University of Wisconsin-Madison: TRNSYS 16: A TRaNsient SYstem Simulation program — Volume 5 Mathematical reference; Type 107: Single Effect Hot Water Fired Absorption Chiller; Page 155-158. Solar Energy Laboratory, University of Wisconsin-Madison. 2004.

[4] H.-M. Hellmann, C. Schweigler, F. Ziegler : The characteristic equation of sorption chillers.

Proc. Int. Sorption Heat Pump Conf., Munich, 24.-26. March 1999; pp. 169-172.

[5] Albers, J.: TRNSYS Type107 Part load simulation of single staged absorption chillers in quasi steady states — Contribution to a design tool for solar assisted air conditioning systems developed in IEA TASK25 Subtask B, IEMB-Report 2-67/2002.

[6] H.-M. Hellmann, F. Ziegler: Simple absorption heat pump modules for system simulation programs, ASHRAE Transactions 105 (1), (1999)

Task D: System technologies

Thermally driven heat pumps need three different circuits, two of them provide the machine with the energies that let it work, and the other one reject it closing the energy balance. All three circuits have to be controlled at the same time, taking into account that, except for some cases, the temperature levels of the energy (both supply and demand) vary. Overall control is indispensable if the maximum energy yields are wanted be reached.

In some cases the generator depends very much on the variability of the sun, or the amount of waste energy that can be used. Therefore the design of a system to store exceeding free energies, mitigating thus the fluctuations, is essential and also auxiliary power requirements must be evaluated to assure a constant energy to drive the machine.

Similar considerations have to be done also for the other two energy circuits. There is, however, an extra variable to be taken account of: depending on the final elements and the relation of sensible and latent load the fluid must be sent to the demanding zones or applications with different temperature levels that maintain the satisfaction of the users keeping the thermal/electrical yield of all the facility as high as possible:

For cooling applications, the fluid of the evaporator will be sent to a — usually variable — demand application while the absorber and condenser must be cooled with a heat rejection system.

For heating applications, the varying demand is connected to the medium temperature level (absorber and condenser) and the evaporator must provide with energy at low temperature level. With the — possibly “free” — driving energy in the generator circuit this low temperature energy can be lifted up to the temperature level needed, saving primary energy.

Since only little practical experience with thermally driven heat pumps for heating and cooling applications is available to date, it is necessary to monitor demonstration projects as case studies, providing information on performance under every day conditions and yields of the control strategies implemented on them. At the end of the task the documentation and evaluation of existing projects will help us to show and teach in a real way what kind of things can and cannot be done with TDHP’s depending on the applications.

Also, because of this lack of installed systems where new things can be tested, it is necessary to have simulation programs with trustable results. A list of the available software and a comparison among them must be produced.

Low-powered absorption machines range

Research and development in the field of absorption machines has traditionally be linked to industrial processes, where it is usual to have residual heat that can be used to feed the absorption machine. That

is one of the reasons of the shortage of low-powered absorption chillers models for domestic applications. Besides, the cheap price of electricity in the last decades, has allowed an incredible technical development of domestic vapour compression systems and the consequent widespread installation.

Today, the increasing price of electricity and the progressive growth of the social awareness about the environmental effects of climate change and its consequences have modified the requirements of the demand, which currently gives priority to the reduction of energy consumption and subsequently promotes the development of smaller machines.

The range of power of the most installed commercial absorption machines for domestic applications covers from 4,5kW to 32kW, stressing an interesting 10kW machine from a Swedish firm. The smallest does not need cooling tower, which is required for the other models to avoid lithium-bromide crystallization.

3.4. Complementary uses

In cooling demand periods, the required solar collector field is going be fully used to feed the absorption machine. Nevertheless, in non-cooling demand periods it could be used to pre-heat sanitary hot water and central heating water. Depending on the surface of the solar field, it is feasible to reach annual savings up to 90% in sanitary hot water and up to 55% in central heating. By this way, the use of the solar field would be continuous during all the year, profiting it 100%, and preventing the collectors from overheating because of water stagnation.

Discussion and look ahead

A composite carrier reflector based parabolic trough system was successfully developed.

The composite material reflective panel approach developed specifically for roof mounted parabolic trough applications, has use in all solar technologies employing mirrors. Depending on the application (small aperture parabolic troughs, heliostats, linear Fresnel reflectors, large parabolic troughs for CSP or parabolic dish elements) the structural support system and tracking mechanism will differ. The low mass, strength and stiffness of the composite material carrier however, which is common throughout, results in a simplified and cheaper support structure and tracking mechanism.

Acknowledgements

[1] The development of the collector was supported by the Renewable Energy Development Initiative of the Commonwealth Government of Australia.

[2] The Commonwealth Scientific and Industrial Research Organisation for the photogrammetry and ray tracing analysis.