Category Archives: EuroSun2008-7

Results of most promising cases

In table 1 the selected building applications and energy efficiency standards are described.

[7] Coefficient of Performance (COP) = cooling or heating output (kW) divided by

electrical input. The only electrical input is 4 small circulation pumps and internal controls. COP in conventional compressor-based chillers and packaged air conditioning units is usually stated as cooling capacity (kW) divided by compressor electrical input. Since the Millennium MSS air conditioning doesn’t have a refrigeration compressor, COP is stated here as annual cooling/heating energy delivered divided by total electrical input of the pumps and internal controls.

Fig 1. Millennium MSS Air Conditioning system heating and cooling device The example below shows one of the two barrels discharging cooling.

[8] chillii® Solar Cooling System

The analysed solar driven chillii® Solar Cooling System of the Solar Next AG has been set up and installed as a test facility to cool their office building in Rimsting. This system includes a market available 15 kW LiBr absorption chiller (ACM), two 1 m3 hot water storage tanks, one 1 m3 cold storage tank, 37 m2 CS-100F flat plate collectors and 34 m2 TH SLU1500/16 solar vacuum tube collectors all facing south with an inclination of 30°, a 35 kW EWK wet recooling tower (Figure 1). For the distribution of the cooling energy chilled ceilings and fan coils are used with 16°C supply and 18°C return temperature and an automated supply temperature increase for dew point protection. An auxiliary heater is integrated in the system, but not considered in the present analyses.

[10] Introduction

The use of renewable energy in buildings is a very important challenge in order to decrease their primary energy consumption. In South of Europe, most of buildings, especially tertiary ones, need active cooling in summer; thus the set up of the solar cooling technology can often represents an important way to save fossil energy. Solar cooling technologies already exists and have shown their effectiveness at the demonstration stage. Nevertheless, a certain number of technical and economical barriers actually exist and prevent a larger set up of the systems.

[11] From the two groups in Fig. 2 the reverse result is obtained. But this is an effect of the scatter (e. g. the AAtmin value is lower than zero for AtACE=29K, which is physically not possible). The ‘real’ characteristic equation (derived from ideal data without any scatter) should be steeper, leading to positive values of AAtmin and an increased slope.

[12] Conventional AHU, whose layout and thermal cycle are shown in [1].

• Standard open-cycle desiccant system (DEC), according to the layout shown in Figure 1.

• Standard DEC provided with a desuperheater.

• Standard DEC modified with the use of an enthalpy recovery wheel, which operates a pre-dehumidification of the process air by releasing a certain fraction of its vapour content to the exhaust air.

For all the thermal cycles related to the DEC systems, the use of solar energy to assist regeneration will also be investigated (solar assisted DEC): the AHU is the same as in the previous points, but a solar section with a back-up heater is adopted to produce hot water for regeneration purposes. In order to test the sensitivity of the results to the size of the solar section, different values of the collector surface will be considered, corresponding to an annual solar fraction F ranging from 0 to 1. In order to compare the energy and exergy performance of the systems, we have considered a case study represented by the ventilation and the air-conditioning of an enclosed space with a latent

[13] Governing principles and parameters

So as to establish appropriate dimensioning guidelines, we will start by characterizing the governing principles and their link to amplitude-dampening along the pipes. We therefore will base on the simplified and analytically resolved case of a constant airflow subject to sinusoidal temperature input, with explicit treatment of diffusive heat storage into the soil [2], which yields following main insights:

• As long as the soil layer around each pipe is at least that thick, heat charge and discharge

around the pipes naturally extends over a penetration depth 8 which depends on the oscillation period:

[14] Koller, Zetzsche, Brendel, Muller-Steinhagen: Design and Operation of a small Ice Storage, 2nd International Conference Solar Air-Conditioning, Oct. 18th/19th 2007, Ostbayerisches Technologie- Transfer-Institut e. V., Regensburg, Germany, p. 359-364


[16] y is the response of the phenomena • xi is a factor or parameter influencing the phenomena • a0 is a constant effect,

• ai is the effect of single parameter

• aij is the effect of double interaction,

• aijk is the effect of triple interaction

It is evident that the number of effect to be determined will need the same number of experiments. In order to estimate the effects, each parameter varies between an upper and lower limit, so each parameter has two levels [2]. In our case we have 4 parameters; each one varying between 2 limits which mean we have 24 effects to be determined and 16 combinations (experiments) are then needed.

The following range of the parameters was considered:

• Outside temperature T1: [25, 35] [°C]

• Outside humidity ratio w1: [11, 14.5] [g/kg]

• Regeneration temperature T8 [55, 75] [°C]

• Regeneration humidity ratio w8 [10; 15] [g/kg]

Outside temperature and humidity ranges correspond to the most of the European climates (except some humid Mediterranean climates) and match with the domain of application of desiccant cooling (e. g. moderately hot and moderately humid climates). The upper limit of regeneration temperature domain is consistent with solar application with temperature not exceeding 75°C while

Heating and Cooling with a Small Scale Solar Driven Adsorption. Chiller Combined with a Borehole System

Dr. Tomas Nunez*; Bjorn Nienborg; York Tiedtke

Fraunhofer-Institut for Solar Energy Systems ISE, Heidenhofstrafte 2, 79106 Freiburg, Germany
* Corresponding Author, tomas. nunez@ise. fraunhofer. de


The performance of a solar driven adsorption cooling system is presented in this paper. The system consists of a reversible 5.5kW adsorption machine from the German company SorTech, a 20m2 flat plate collector field with a 2m3 buffer storage and a borehole array of three 80m boreholes. In the summer period, the adsorption machine is operated as a chiller driven by solar energy while the boreholes are used for heat rejection. In winter it is operated as a heat pump driven by the heat from a heating network and using the boreholes as low temperature heat source. The operation results presented here correspond to the period from June 4th to December 1st, 2007. The machine as well as the whole system operated reliably and as expected during the whole monitoring period. An overall thermal COP of 0.57 and an hourly mean chilling power of around 4.4kW were obtained during the cooling season. The electricity consumption was about 10% of the produced cold. During the heating season a thermal COP of 1.43 and hourly mean heating powers of 9.4kW were obtained. Frequency distributions of the registered driving, heat rejection and chilled water temperatures give a picture of operation conditions in a real application.

Keywords: adsorption, solar cooling, ground source heat exchanger, system performance

1. Introduction

In the frame of the finished EU project MODESTORE a solar driven cooling system with a 5.5kW reversible adsorption chiller was installed at the Fraunhofer Institute for Solar Energy Systems. In a previous publication [1] first simulation results which resulted in the present design and dimensioning of the system were published. In this publication the system performance and operation results from the first operation period in both modes (cooling and heating) is presented.


G. Gutierrez*, P. Rodriguez, A. Lecuona, M. Venegas, J. Nogueira

Universidad Carlos III de Madrid (UC3M). Departamento de Ingenieria Termica y de Fluidos. Avda.
Universidad 30, 28911 Leganes, Madrid, Spain

* Corresponding Author, glgutier@ing. uc3m. es


An experimental facility applying the concept of adiabatic absorption has been designed and built in Universidad Carlos III de Madrid. Plate heat exchangers are incorporated in the design functioning as generator, condenser, sub-cooler and solution heat exchanger. Other components include a separator and two fin-coiled tubes as evaporators. Trials were carried out in order to characterize components and performance. The range of controlled hot fluid temperature corresponds to a solar thermal energy source (below 100°C). Performance parameters, cooling capacity and COP, are expressed in terms of an ideal absorption model and compared with experimental results. The differences observed between ideal and experimental results help to identify the influence of components performances on the overall performance of the facility. The evaluation of the ideal and experimental cooling powers allows detecting a low performance of evaporators as both dry operation and overflow. Other influence factors are described and their effect is included in the thermodynamic analysis of the absorption cycle.

Keywords: Adiabatic absorption, Lithium bromide, Chillers.




coefficient of performance


specific heat capacity (kJ/kgK)


enthalpy (kJ/h)


lithium bromide


mass flow rate (kg/h)


plate heat exchanger


thermal power (kW)


temperature, external (°C)


temperature, internal (°C)


pump work (kJ/h)


mass concentration





evaporator, evaporation


generator, generation

















Greek letter:

П efficiency

1. Introduction

Absorption machines offer the possibility of amortizing thermal solar installations during summer, at the same time avoiding polluting emissions to atmosphere and increasing sustainability.

The adiabatic process is being investigated as a method for improving absorption in a chiller. It consists on dispersing the solution inside an adiabatic chamber. The resulting heat is extracted downstream using a compact heat exchanger. Thus, the mass and heat transfer processes are split in separated apparatus. Ref. [1-3] summarizes state of the art, including experimental and theoretical studies concerning adiabatic absorption. Further works on adiabatic absorption using aqueous LiBr, focus on experimental work [4] and simulation or theoretical studies [5, 6].

Interpretation of experimental data must include the particular features, in both design and operation, of the facility here presented. For this duty, diagnostic versions of the models are of much help. A basic thermodynamic model, including such features is presented and compared with experimental results, resulting from this the detection of some operational difficulties.

2. Experimental facility

The highly flexible experimental facility forms a single-effect water — lithium bromide absorption cooling system. The test facility design incorporates compact PHEs, due to the well-known benefits regarding to high heat transfer capacity in a reduced volume.

The absorption process takes place in an adiabatic chamber and the heat is extracted in an external PHE (subcooler). Because the solution absorbs less vapour as its temperature is higher, it is necessary to re-circulate part of the solution to the absorber after it has been cooled in the PHE (mr). Both the solution coming from the generator mstrong and the re-circulated solution mr flow

inside the absorber as free falling drops. The experimental setup configuration, the data acquisition system and the experimental procedure were described in detail in a previous work [7]. Fig. 1 shows a diagram of the experimental facility.

The facility is highly instrumented, such that individual component performance can de evaluated.

Data of temperature, pressure and mass flow rates in every component were recorded in intervals of 0.5 seconds in order to accurately determine steady state periods. The experimental setup was configured to allow determining COP and thermal powers exchanged in the different components.


Fig. 1. Diagram of the experimental facility 2.1. Experimental uncertainty analysis

A complete calibration process is periodically carried out for all instruments showed in Fig.1. This way, measurement errors can be reduced as far as possible. The uncertainty for an experimental result R, which is function of n different parameters x, is calculated as:


R = f(X1, X2,k xn )


Ur =





The uncertainty of performance parameters reached 16% at the worst case, see Fig. 4.

Description of installed system

1.1. System and operation concept

The system concept is shown in Fig. 1. The core components are the reversible adsorption machine which can be operated as a thermally driven chiller or heat pump and the borehole system that is used as heat rejection system for the chiller mode as well as the low temperature heat source for the heat pump mode.

In the summer operation mode (left schematic in Fig. 1) the adsorption machines works as a thermally driven chiller. It is driven by the heat from the solar collectors. A connection to the heating network of the building works as a heat backup in case of not sufficient solar driving heat. The system provides cooling to a cooling coil in the air handling unit, which is installed in the inlet air duct to the institutes’ canteen kitchen (3000m3/h). The waste heat from the chiller is rejected through the borehole system.


In the winter operation mode (right schematic in Fig. 1) the adsorption machine is working as a thermally driven heat pump. It is driven by heat from the heating network of the institute, which comes form a CHP unit, and lifts low temperature heat from the boreholes to the useful temperature level. This useful heat is used in a heating coil to pre-heat the air in the main duct of the ventilation system (9000m3/h).

Fig. 1. System schematic with summer (left) and winter operation mode (right).

The system installed at the Fraunhofer Institute consists of a solar collector array of 20m2 with a 2m3 buffer storage and three boreholes of 80m each. An adsorption chiller ACS05 from the German company SorTech AG with a rated nominal cooling power of 5.5kW is used. This chiller is a pre-series development based on a previous prototype presented in [2].

For a general application in the residential sector this system layout offers the following advantages:

• The solar system is used through the whole year. In winter it is used for solar assisted heating and in summer it provides the driving heat for the cooling system.

• The reversible adsorption machine is also used during the whole year: as a heat pump in winter and as a thermally driven chiller in the summer period.

• In the heat pumping mode the adsorption system enhances the energy output of the driving heat source through the use of ambient heat from the boreholes.

• The borehole system also serves two purposes: first as a low temperature heat source in winter and second as a heat rejection system for the chiller in summer. Thus the boreholes are not only used year round but are also regenerated in the summer season.

• In favourable conditions and periods of low cooling power requirements, the installed boreholes can be used for direct cooling.

Nevertheless, neither the direct solar heating nor the direct cooling with the boreholes has been implemented in the present system.

2.2. Control procedures

The operational concept foresees an operation only during weekdays. The operation conditions are as follows:

Cooling operation is carried out when:

• the inlet air temperature exceeds 20°C (2K hysteresis),

• the air temperature in the kitchen is above 23°C (2K hysteresis),

• the time is between 6:45 and 16:00 o’clock.

Heating operation is carried out when:

• the inlet temperature in the main air duct is below 14.5°C (3K hysteresis),

• the inlet air temperature is above 3°C (freeze protection of the machine),

• the time is between 6:45 and 19:00 o’clock.

Solar heat is used whenever the mean temperature in the upper part of the storage is above 73°C with a 5K hysteresis for turning of solar heat supply.

In spring and autumn it may happen that the air temperature falls below 14.5°C in the mornings and thus the heating mode is activated, but later during the day temperatures in the canteen kitchen rise above the threshold for cooling operation. In these cases the system is operated in the heating mode first and later in the cooling mode. This operation is called ‘alternate mode’.

The volume flows in the three circuits are kept constant and correspond to the nominal flows required by the chiller. Energy efficient pumps have been installed and the flow rate is set via the three power steps of the pumps.

3. Data acquisition and evaluation

A new prototype of the reversible adsorption machine was installed in spring of 2007 and is in operation since then. In the summer season of 2007 about 282 hours of operation in the cooling mode have been monitored. Evaluation is carried out at three levels:

1. performance of the chiller as a component

2. evaluation of the operation conditions of the whole system for system optimisation

3. evaluation of the overall systems energetic performance

These three evaluation levels address each a different group: while the first is mainly important for the chiller manufacturer in order to decide if the machine is working as expected, the second level targets the system developers which are interested in the optimisation and smooth operation of all system components in order to optimise the energy efficiency and thus produce the highest possible savings for the end-user. This result is covered with the third evaluation level.

The basic absorption model

Fig. 2 represents the thermodynamic states of the ideal simple effect absorption cycle, assuming mechanical, thermal and chemical equilibrium and assuming energy degradation exclusively concentrated on expansion valves. Steady state mass and energy balance in components yield the following equations:

Evaporator (no superheating):

Qm= m ref, K — h1); h1 = К


m ref, i X strong X weak

Considering =


weak strong

strong x weak














(a) (b)

Fig. 2. (a) LiBr/water absorption cycle. (b) Diagram P(T)-T LiBr/water

Generator (no superheating):

Mass balance:

Energy balance:

Qoi = ™strong • h2s — mweak * К + ™^ ,г * h2v ; T2s = T2v

Подпись: (5)

image643 Подпись: (6) (7) (8)

™ weak = ™ strong + ™ ref, г

Solution heat exchanger:

Подпись: (9)T — T T — T

n = Js _4s = 2s 3s

hex, i

T1’s — T4s T2s — T 3’s






h5s — h4s = wpi







h6s = К, К (h4s + Wpi)








The above expressions will be used next as a reference to evaluate the performance of the absorption facility.

Monitoring equipment and evaluation of raw data

The data acquisition system consists of internally integrating heat meters with matched Pt100 type temperature sensors. The integrator has a sampling rate of 1s and calculates cumulated energy amounts and mean temperatures and powers. This internal sampling rate assures a correct collection of energy data for the highly dynamic temperature patterns characteristic of adsorption systems. The integrator and further temperature sensors are read out by a computer with a sampling rate of 15s. The monitoring software further reduces these values to cumulated energies and mean temperatures which are stored with an interval of 5 minutes in the raw data measurement file. The storage interval can be set by the system operator and thus allows a flexible data management. The post processing of the raw data further reduces the values to hourly accumulated and mean values — depending on the quantity considered. For the hourly mean temperature values also a standard deviation is calculated in order to judge the stability of the temperature within the evaluated hour.

Experimental performance parameters

Experimental cooling capacity and COP is obtained from the external fluid temperatures in evaporators as follows:


QE, exp mchw Cw (tchw, in Phw, o )

From external fluid temperatures in generator:

QG, exp moil Coil (toil, in ^ oil, o )

Corresponding values per unit mass flow rate are:

Qe, exp















5. Results

Overview of operation period

The results presented here correspond to the operation period from June 4th, 2007 to December 1st, 2007. From these 180 days the system was in operation during 108 day. The days without operation were either weekends or days without monitoring data due to a lightning stroke into the data acquisition system. Within these 108 days 1034 hours of operation were registered. In some hours the machine was in operation only a few minutes, but in most of the registered hours a

continuous operation was observed. In Table 1 the number of hours with an operation time within given limits is shown.



Number of hours with an operation time t (in minutes) within the given limits

<15 min

15<t<30 min


45<t<60 min




















Table 1.Statistics of the operation hours

4. Operation results

Basic absorption model vs. experimental results. Factors affecting the facility performance

Подпись: to [°C] Fig. 3. Comparison of basic model and experimental results. (a) Ideal and experimental cooling capacity vs. tG (b) Ideal and. experimental COP vs. tG image655

Fig. 3 illustrates the comparison of experimental performance parameters with those obtained through the basic model, Eqs. (14) and (15), for 220 test points. The differences observed help identifying the influence of components performance, serving as a diagnosis. In the following, the causes of deviation from the ideal conditions are explained, evaluated and included in the model.

5.1. Influence of components performance. Modified basic model.

The evaporators in this facility layout show low performance [8]. Because of this, part of m ref

supplied to them is not evaporated, but overflows. Therefore, in order to measure the performance of evaporators (fraction of evaporated refrigerant), the cooling capacity obtained if all refrigerant is used in the evaporators Qref is compared with QE exp:

Пе = ; Qref = mref • (hEin — hEo) (21)


Energy balance in the generator yields:

m = m — m

ref weak strong (22)

The obtained average nE is 50% which is considerably low. With the current design, the excess of refrigerant is not recirculated to evaporators, and therefore it goes to the solution reservoir inside the absorber. Beside this, it has been detected that the distribution of mref (m f and m f) is not symmetrical in some cases, and therefore one of the evaporators runs dry. This situation also explains the low value of qexp and its tendency is not consistent with rise in temperature (Fig. 3). Corresponding values of COPexp show obviously the same behaviour.

Taking into account the evaporators limitations explained above, the modified cooling capacity will be:

4E, mod 4Ei ‘ ПE (23)

The facility performance is also affected by the solution heat exchanger efficiency phex and consequently is considered in the analysis. Beside this, a noticeable difference was identified between solution temperature at generator outlet and separator outlet, as the generation process continued and the refrigerant separation kept on in the path followed by the solution. The temperature drop associated ranged from 4°C to 10°C. The heat transfer associated is called qG.

Then, considering the real efficiency of the solution heat exchanger and the heat transfer losses in the system generator-separator, the modified generation power will be:

qG, mod = qG, i + qhex. i (1 — Vhex ) + qG-sep (24)

As a result, the modified COP becomes:

COPmod = ^mo1 (25)

qG, mod

Results obtained incorporating the specific behaviour of components to the basic model can be appreciated in Fig. 4. The predicted cooling capacity and COP shows good agreement with experimental results. This indicates that the main causes of deviation have been included in the model, thus verifying the preliminary diagnosis.

(a) (b)

6. Conclusions

A facility representing a single effect adiabatic LiBr/water chiller implementing adiabatic absorption facility has been operated under different conditions, corresponding to both design and off-design operational conditions. Performance parameters have been experimentally determined for every test condition.

The differences observed between ideal and experimental results helped to diagnose and identify the influence of components performances on the overall performance of the facility. The

evaluation of the ideal and real cooling powers allowed detecting a low performance of evaporators caused by both dry operation and overflow. Another influence factors which causes the deviation from ideal behaviour are the solution heat exchanger efficiency and the heat transfer losses in the system generator — separator.

Taking into account the particular layout and operation features tested, a good agreement with experimental performance parameters and those obtained through a modified basic absorption model has been achieved. This model incorporates a quantified loss and/or efficiency separating from ideal and has been fitted to experimental data. Experimental results demonstrate the operational possibilities and flexibility of the design, showing a great potential for further work.


The financial support of this study by the Ministry of Education, Science and Technology through CLIMABCAR project DPI 2003-01567, TRANSMACA project DPI 2002-02439 and MINICOM project (FIT 0204-2004-68 and FIT 020100-2003-233), is greatly appreciated. The authors express their gratitude to the technicians of Universidad Carlos III de Madrid Mr Manuel Santos and Mr Carlos Cobos for their invaluable help in this work.


[1] Venegas M., Izquierdo M., Rodriguez P., Lecuona A. Heat and mass transfer during absorption of ammonia vapour by LiNO3-NH3 solution droplets, Int. J. Heat Mass Transfer, 47 (12-13) (2004) 2653­2667.

[2] Venegas M., Rodriguez P., Lecuona A., Izquierdo M. Spray absorbers in absorption systems using lithium nitrate-ammonia solution, Int. J. Refrigeration, 28 (4) (2005) 554-564.

[3] Arzoz D., Rodriguez P., Izquierdo M. Experimental study on the adiabatic absorption of water vapor into LiBr-H2O solutions, Applied Thermal Engineering, 25 (5-6) (2005) 797-811.

[4] Warnakulasuriya F. S.K., Worek W. M. Adiabatic water absorption properties of an aqueous absorbent at very low pressures in a spray absorber, Int. J. Heat Mass Transfer, 49 (9-10) (2006) 1592-1602.

[5] Elperin T., Fominykh A., Orenbakh Z. Coupled heat and mass transfer during nonisothermal absorption by falling droplet with internal circulation, Int. J. Refrigeration, 30 (2) (2007) 274-281

[6] Wang L., Chen G. M., Wang Q., Zhong M. Thermodynamic performance analysis of gas-fired air-cooled adiabatic absorption refrigeration systems. Applied Thermal Engineering. 27 (8-9) (2007) 1642-1652.

[7] Gutierrez G., Venegas M., Rodriguez P., Izquierdo M., Lecuona A. Experimental characterization of a single stage LiBr-H2O absorption test rig. In Proc. ECOS 2006, Vol. 3, Crete, Greece, July 12-14 (2006) p. 1331-1316.

[8] Gutierrez G., Zacarias A., Venegas M., Rodriguez P. Cooling power evaluation of a water lithium bromide absorption test rig. In Proc. ECOS 2007, Vol. 2, Padova, Italy, June 25 -28 (2007) p. 1183-1190.

Operation of the adsorption chiller

Performance of the chiller

In order to evaluate the performance of the chiller a comparison of experimentally measured cooling capacities and COPs with expected values from calculations is shown in Fig. 2. The experimental values were selected according to the following criteria:

1. a continuous operation during the whole hour as well as in the hour before and after,

2. stable temperatures with a small standard deviation has been measures in all three circuits,

3. only cooling operation was considered.

TCooling operation

For the evaluation of the cooling operation only hours with constant and steady operation were considered. Operation periods less than 60 minutes within one hour were not considered in this evaluation in order to avoid effects of transient states. From Table 1 it can be seen that 282 hours could be considered.