Cooling Systeme

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n x x x ( 2 0 1 3 ) 1 e2 0
Available online at www.sciencedirect.com
w w w . i i fi i r . o r g
journal homepage: www.elsevier.com/locate/ijrefrig
Solar cooling with watereammonia absorption

chillers and concentrating solar collector e
Operational experience
Christine Weber a,*, Michael Berger b, Florian Mehling a,

Alexander Heinrich a, Tomas Núñez a,1
a
Fraunhofer Institute for Solar Energy Systems ISE, Department of Thermal Systems and Buildings,
Heidenhofstrasse 2, 79110 Freiburg, Germany
b
Industrial Solar GmbH, Emmy-Noether-Straße 2, 79110 Freiburg, Germany
article info abstract
Article history: Concentrating solar collectors provide high efficiency at high driving temperatures
Received 13 March 2013 favourable for thermally driven chillers. Therefore, they offer applications for hot climates
Received in revised form and industrial process integration, especially in combination with NH3eH2O chillers that
12 August 2013 provide refrigeration temperatures below 0 C. The presented solar cooling installation
Accepted 18 August 2013 comprises a linear concentrating Fresnel collector that provides the driving heat for two
Available online xxx NH3eH2O absorption chillers at temperatures up to 200 C. Chilled water temperatures are
This work is dedicated to Dr. Tomás produced in the range between 12 C and 0 C. Collector capacities reached up to 70 kW at
Núñez who passed away abruptly in peak times and the total cooling capacity of both chillers showed peak values up to 25 kW.
2011. This project was one of his last For good operating conditions, the thermal system EER was 0.8 and an electrical system
achievements as an extremely EER of 12 was easily achieved. The system showed a sound operating behaviour. The
committed, reliable and creative performance of different operation and control strategies was analysed, evaluated and
researcher. His vision was a future enhanced within the two year operation phase.
energy system based on renewable ª 2013 Elsevier Ltd and IIR. All rights reserved.
energies. Due to his work and
experiences, he became one of the
leading experts on solar cooling in
Germany and Europe. We will
always remember him as a very
warm-hearted, inspiring and
motivating person.
Keywords:
Solar cooling
Fresnel collector
NH3eH2O absorption
Ice storage
Process heat
Solar steam generation
* Corresponding author. Tel.: þ49 176 62016966; fax: þ49 761 458 0.
E-mail addresses: christine.weber.de@gmail.com (C. Weber), peter.schossig@ise.fraunhofer.de. (P. Schossig)
0140-7007/$ e see front matter ª 2013 Elsevier Ltd and IIR. All rights reserved.
http://dx.doi.org/10.1016/j.ijrefrig.2013.08.022
Please cite this article in press as: Weber, C., et al., Solar cooling with watereammonia absorption chillers and concentrating
solar collector e Operational experience, International Journal of Refrigeration (2013), http://dx.doi.org/10.1016/
j.ijrefrig.2013.08.022
2 i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n x x x ( 2 0 1 3 ) 1 e2 0
Nomenclature u0; u1 linear and quadratic heat loss coefficients

[W m2K1; W m2K2]
CSP concentrating solar power
DHW domestic hot water Greek letters
EC European Commission a Azimuth angle
EERel electric energy efficiency ratio (cooling h efficiency
power/auxiliary electricity consumption) h0 optical efficiency for sun in zenith position
¼ cooling power to electricity consumption ratio q zenith angle
EERth thermal energy efficiency ratio (cooling d angle between collector’s absorber pipe and
power/driving power) (¼ COPth) ¼ cooling power to northesouth direction
heat input ratio w temperature
IAM incidence angle modifier Dw temperature difference [T1 T0]
ISE Fraunhofer Institute for Solar Energy Systems ISE
Suffixes
ISG Industrial Solar GmbH
aperture collecting cross section surface
NH3 ammonia
col collector
PID proportional integral and derivative control
loss power loss
H2O water
HT high temperature, driving circuit
TDC thermally driven chiller
LT low temperature, cold water circuit
Nomenclature for performance data MT medium temperature, heat rejection circuit
A area [m2] out outlet
DNI direct normal irradiation [W m2] in inlet
G solar irradiation [W m2], in this context also DNI 1 chiller 1
Q_ power/heat flow [kW] 2 chiller 2
Q energy [kWh] l longitudinal
T temperature [ C] t transversal
V volume flow [l h1] sol solar
cp heat capacity at constant pressure [kJ kg1 K] tot total, both chillers
_ m
m, mass flow rate [kg s1] load thermal power measured in electric load circuit
Froid solaire avec des refroidisseurs à absorption

d’ammoniac-eau et un capteur solaire à
condensation e expérience de fonctionnement
Mots clés : Froid solaire ; Capteur solaire de Fresnel ; Absorption de NH3-H2O ; Accumulation de glace ; Chaleur industrielle ; Génération
solaire de vapeur
1. Introduction energy balance on annual level, energy saving and energy

efficiency measures have to be fully exploited. Even
Worldwide there is an increasing energy demand, especially though, a demand for active heating and cooling will
in industrialized countries and emerging countries with high remain in most buildings and under most climatic condi-
economic growth. The cooling for buildings and industrial tions (Henning and Döll, 2012).
processes holds a significant proportion of this demand. Next to the cooling demand of buildings, process heat and
At first, energy efficiency measures and renewable energy cooling for industrial applications is an increasingly important
usage need to be further developed and implemented. The topic, especially regarding energy efficiency and renewable
goal is to develop systems with a low energy demand by sources to produce a carbon-neutral and environmentally
energy-efficient measures coupled with a large proportion of harmless goods (Henning and Döll, 2012). Here, solar energy is
renewably generated cooling power. Solar cooling provides a the best and most available renewable energy source on the
carbon-neutral and environmental protecting option to cover spot.
this demand and reduce greenhouse gas emissions. Regarding refrigeration, perishable goods, like food or
Passive house standards and net zero energy buildings pharmaceutics, require constant cooling. In this case, the
have become an important topic in the R&D work on fluctuating production of cooling power needs to be stored to
buildings for the present future. In order to achieve a zero maintain the product quality.
i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n x x x ( 2 0 1 3 ) 1 e2 0 3
1.1. Solar cooling e benefits developments of cleaning robots are initiated by various col-
lector manufacturers. Because the usage of diffuse irradiation
The use of solar energy for cooling has the advantage of a is not possible with concentrating collectors, they have to be
high correlation between solar irradiation and cold demand installed in southern countries with high amount of direct
in most cases. The two main components of a solar irradiation.
cooling plant are the solar collector and the chiller. The Compared to other concentrating systems, linear concen-
overall system efficiency depends on the coupling and trating Fresnel collectors are advantageous regarding rooftop
control strategy between these two components. Systems viability because of very low wind loads (Häberle et al., 2007).
for solar air-conditioning vary in a wide range of applica- Parabolic trough collectors are not viable at very high wind
tions as different system sizes and designs enable diverse speeds as operation must be interrupted.
uses. Chemisana et al. (2013) examined the performance of
About one thousand solar cooling systems were estimated building integrated concentrating collectors for solar cool-
to be installed worldwide by the end of 2011 (Henning, 2011). ing with simulative work that looks at the benefits of a
Of these, less than thirty systems use linear concentrating high driving temperatures that are interesting for double-
collectors to supply the heat required by the thermally driven effect absorption technologies. Their results show that
cooling processes. for the concentrating system the aperture area can be
Eicker and Pietruschka (2009) showed in their cost analysis significantly reduced compared to a standard solar thermal
of solar cooling systems that the operation in southern Euro- installation.
pean locations leads to significantly lower costs compared to
locations in northern Europe. This is caused by a higher 1.4. Absorption and refrigeration with NH3eH2O
cooling demand. For a southern German climate, the costs systems
resulted to be more than double.
Several studies in the past years underlined the high potential
1.2. Solar cooling e drawbacks for refrigeration with solar cooling systems e especially con-
cerning industrial applications e in different areas worldwide
According to Chemisana et al. (2013), the main technical (Ayadi et al., 2012).
drawbacks of solar cooling systems are the low efficiency of To reach freezing temperatures below 0 C, the only avail-
the commonly used single effect absorption chillers and the able thermally driven chillers use NH3eH2O as working pair.
large collector areas needed to produce the thermal energy. These chillers operate with a high temperature lift compared
However, they are referring to flat plate collectors and evac- to lithium bromide absorption chillers, so that dry heat rejec-
uated tube collectors that have been used in most of the solar tion with cold production below 0 C is possible even in hot arid
cooling systems. Furthermore, solar cooling systems are climates. Fresh water consuming wet cooling towers along
complex and operation control as well as system design needs with costly hydraulic circuits are not necessary. In dry climates
skilled implementation. Economically speaking, the high in- where the main source of water is desalination this is a major
vestment costs are one of the main obstacles, requiring advantage. However, the process requires high driving tem-
further effort to produce economical solar cooling technolo- peratures (140e200 C) for the generator that are easily pro-
gies (Otanicar et al., 2012). This is why in the recent past vided by concentrating solar collectors. Furthermore, several
research activities were initiated to look at solar cooling sys- industrial applications require both cooling and steam for
tems composed of PV panels with electrically driven processes, and some chiller manufacturers use steam as
compression chillers. preferred heat transfer medium. Here also, concentrating solar
collectors can be very advantageous as they can provide both.
1.3. Concentrating solar technology Regarding economically aspects, the costs for NH3 systems
can be lower than for LiBr systems according to Otanicar et al.
Concentrating solar systems can improve the energy effi- (2012).
ciency of solar cooling systems when high driving tempera- The use of ice storages is a cheap and compact way to store
tures above 120 C are required. They offer advantages cold. Heat losses of a cold storage are low compared to a hot
compared to the solar collectors traditionally used in solar air- storage, because of the smaller temperature difference to the
conditioning and refrigeration facilities. ambience. Using the latent heat of fusion of water provides a
Those advantages of concentrating technologies are lower high storage density. The general advantage of using storages
thermal losses and thus higher temperature of the fluid, high is the possibility of peak shifting for applications, where the
ground usage and adequate recycling of component materials. daily cooling load profile does not coincide with the solar
A higher overall system efficiency at higher working tem- irradiation. As a result, industrial cold applications with ice
peratures is reached and the aperture area for a given power production become feasible.
requirement is reduced. The risk of reaching stagnation
temperatures is reduced as the safety control enables the 1.5. Operation experience
collector to move into stow position.
The disadvantages of concentrating technologies are Publications on solar cooling with concentrating collector
increased investment and installation due to the tracking technologies are either summarizing papers or simulation
system, higher effort for hydraulics and balance of plant. As work. Published works on testing and operation data for
surface cleaning is an issue for all solar collector types, NH3eH2O absorption in combination with concentrating
collectors are scarce. A few existing publications are confer- since then according to the manufacturer. Also, the general set-
ence contributions or project reports that typically provide up and several components were less elaborated at the time.
general findings rather than detailed results. Bermejo et al. (2010) examined the solar fraction, solar
A thorough general overview on solar cooling installations, cooling ratio and performed a brief economic analysis of a
research projects and IEA Task 38/48 work was elaborated by system that comprises of a concentrating Fresnel collector
Cabrera et al. (2013) in their literature survey. It focusses on with a double-effect H2OeLiBr absorption chiller in Spain.
double-effect absorption cycles and parabolic trough collec- They conclude to use the system at places with high irradia-
tors. They also conclude that publications on testing data and tion and a high quantity of sunny hours.
standardisation are scarce. Saulich et al. (2012) presented system enhancements of a
For performance assessment of solar cooling installations, solar cooling installation consisting of a single-effect H2OeLiBr
one important figure is the primary energy saving potential in system with evacuated tube and flat plate collectors. Main
comparison to a conventional reference system (Napolitano enhancements were carried out by changing hydraulic in-
et al., 2011). Therefore, a high yearly value is the goal for terconnections and the control strategy that improved the
R&D work. usable solar energy for the cooling process by 53%. This leads
Ayadi et al. (2012) give an overview on solar cooling in- to an increase of 18% for the sorption process. Still, the cooling
stallations utilizing concentrating collectors. They state that capacity accounted 15 kW, the thermal COP was around 0.65.
for higher primary energy savings, double and triple effect No information about electricity consumption was given.
absorption chillers are required which require high driving
temperatures that can only be provided by concentrating col- 1.6. The project “SOLERA” and the solar cooling test
lectors with a reasonable efficiency. For urban buildings, the plant
area limitations on roofs restricts feasibility of single effect
chillers with large flat plate or evacuated tube collectors. They In the EC co-funded project “SOLERA” (SOLERA, 2007e2011),
conclude that in future developments, it is essential to the general objective was to develop highly integrated solar
decrease the parasitic consumption of the solar cooling heating and cooling systems in the medium capacity range.
installation, especially for the heating-up phase and chiller The principal goals of the project were to examine the
operation. Feasible potential are: variable speed control of technical feasibility of solar heating and cooling systems for
solar pumps, advanced design of the generator heat exchanger small and medium scale applications by installing and oper-
and direct steam generation for industrial applications. ating three different demonstration systems. Also to analyse
Although several systems have been implemented world- the hydraulic design, review suitable components and to un-
wide, each installation presents a unique set-up. Despite derstand the characteristics and potential of the demon-
advancing developments of thermally driven chillers in the past strated systems in the two years monitoring phase. The
years, the systems are still no standardized technical plants projects results were to be disseminated to stimulate market
(Henning, 2011). To further increase economical balance and interest and further developments.
the system efficiency and in order to ensure an environmentally As the systems design requires extensive planning due to
friendly technology, further R & D activities need to be accom- the innovative character, the focus was set on novel compo-
plished. According to Schicktanz et al. (2011), Henning (2007) nent matching, system integration and a diverse range of
and experience from related research projects (PolySMART, operating strategies. The coupling of the components and
2006e2010), (SOLERA, 2007e2011), (MEDISCO, 2006e2009), sen- managing efficient operation by good functioning control
sitive parameters are the parasitic power consumption of the strategies is still challenging.
installation, a elaborated control strategy and a thorough sys- The presented solar cooling installation in Freiburg/Ger-
tem design. If those parameters are considered in a corre- many is one of three systems that were installed, monitored
sponding way, solar cooling systems utilizing concentrating and evaluated in the course of the project.
collectors become feasible. In the current work, the solar cool- The main components of the system, a linear concen-
ing system design was created from the scratch without influ- trating Fresnel collector and two NH3eH2O absorption chillers,
ence by existing structures, except for the collector itself. are market-available. With NH3 as refrigerant, freezing tem-
An assessment of solar cooling technologies based on cost peratures below 0 C can be generated. To store this energy at
and performance parameters as well as boundary conditions a low temperature level, a set of ice storages was installed.
of weather and cooling demand was presented by Mokhtar Additionally, as cooling load, electric heaters substitute a
et al. (2010). They simulated 25 solar cooling systems with real application, thus providing a flexible experimental po-
concentrating collector technologies and PV cells and their tential. Cascading of several small chillers offers the possi-
results showed that on a smaller scale the Fresnel collector bility of an optimised operation of the several devices and an
option is the most economical one. The two most influencing inefficient part load operation with low performance can be
parameters on the cost economy were the investment costs of avoided. No backup devices were part of this experimental
the collector and the chillers performance that is again set-up. The linear Fresnel collector provides high operation
influenced by the heat rejection technology. temperatures at low heat losses. This allows pressurised hot
One of the first installations with a similar set-up was erected water or steam production also for applications with direct
in Bergamo, Italy in 2006 (Häberle et al., 2007). For this system, steam driven chillers.
only one chiller and no heat storages were installed. The col- The performance of different operation strategies was
lector performance was already good, but has been enhanced analysed and evaluated. The results give valuable information
Fig. 1 e System schematic with main components and characteristics.
on application potential with focus on hot climates and in- required driving power is around 20 kW each. The chillers are
dustrial processes. Special attention is given to the overall modified versions of Robur’s standard gas fired product and
results and optimisation potential for system design and therefore need high driving temperatures to regenerate the
efficient operation and control strategies. absorption process. With this configuration, temperatures
down to 12 C are reached, allowing ice storage usage. An ice
1.7. Paper composition storage block with four single units of 300 l water each (1200 l
in total) is integrated for performance testing. The storage
This paper is structured the following way: after a compre- capacity is approximately 110 kWh. The ice storages are ice
hensive introduction with literature review that refers to the building devices. This means that ice is built up on a coil in
present developments and activities, the system description open water tanks. Those four storage tanks are installed in a
about the components, operation regimes and control strat- joint insulation box. The cooling load is simulated by three
egy is given in chapter 2. Chapter 3 provides an overview on electric flow heating units, providing defined testing suit-
the monitoring periods, the data acquisition and analysis and ability for fixed load profiles of real applications, e.g. load
concludes with the optimisation activities for both hardware patterns in food industry. The system schematic is shown in
and control strategy. Chapter 4 summarises the overall sys- Fig. 1. On both the driving and chilled water side of the
tem and component behaviour with representative operation chillers, the hydraulics was planned for serial and parallel
days. Further analysis of typical example days for steam and operation mode. The modes are switched by manual valves.
pressurised water operation are presented. The general con- The main operation tested was parallel interconnection of the
clusions and an outlook is given in chapter 5. chillers. In this configuration, one of the chillers can be shut
off, so that a cascasing operation of the chillers can be tested.
This was a further focus for evaluation of full and part load
2. System description behaviour, especially for the start-up and shut-down phases.
As the chiller’s heat exchangers can only be operated with
The solar thermal system is a linear concentrating Fresnel liquid heat transfer medium, for steam operation, the chillers
collector of 132 m2 of the company Industrial Solar GmbH. The are connected via a secondary circuit and are not connected to
collector has a nominal peak power of 74.2 kW.1 For the given the steam flow directly. The circuitry on the driving side can
conditions in Freiburg, a thermal output up to 70 kW was be altered by manual valves from a single-circuit-system with
measured. According to the collector power and for testing pressurised water to an indirect three-circuit-system with
part load operation, two units of the Robur NH3eH2O ab- direct steam generation. Thus, different operation strategies
sorption chiller for cooling temperatures between 10 C and are tested and evaluated.
0 C (Model ACF60-00 LB) were installed. They have a nominal
cooling power of 12 kW each and a nominal COP of 0.6. The
2.1. System and operation goals
1 2
According to the manufacturer: DNI ¼ 900 W m , longitudinal
angle ¼ 0 , transversal angle ¼ 30 and 170 C operating tem- For the system design and operation, various objectives were
perature at 30 C ambient temperature. determined and tested:
- Cold production with temperatures below 0 C.

Table 1 e Technical data of the IS-LF11.
- Integrated dry heat rejection without the need of water for a
wet cooling tower. Length of one module 13.32 ft 4.06 m
Height 14.76 ft 4.5 m
- Combine the absorption chillers with a concentrating
Width 24.61 ft 7.5 m
Fresnel collector for process heat at temperatures between Aperture width of one module 18.04 ft 5.5 m
140 and 200 C. Total primary mirror area of 72.18 ft2 22 m2
- Analyse interdependency of component characteristics. one module
- Demonstrate the feasibility of the configuration and obtain Width of one mirror row 1.64 ft 0.5 m
reliable performance data. Number of primary mirror rows 11 11
Pipe connection (outer diameter) 2.76 in 70 mm
- Operation of two chillers for evaluation of full and part
Specific weight (related to 88.56 kg ft2 27 kg m2
load behaviour, analyse their performance for different
installation surface area)
operation modes (parallel or serial configuration; cascaded Heat loss coefficient y 0.00043 W m2K2
or synchronised operation, direct load cooling, storage Maximum operating temperature 752 F 400 C
charge/discharge, both). Life expectancy >20 years
- Use ice storages for cold storage and evaluate its
performance.
- Test and evaluate the steam generation of the collector
The total width is 7.5 m, whereas the aperture width, being the
regarding feasibility and further development. cumulated width of the 11 primary mirror rows, is 5.5 m. The
total height of the collector is around 4.5 m, depending on the
Specific operation goals and corresponding test runs were type of ground and the foundations. The 11 mirror rows are
the following: tracking the sun in one axis, while the receiver, consisting of
absorber tube, secondary mirror and housing, stays fixed.
- Behaviour of chillers with constant driving temperatures at The efficiency of these collector systems depends on
different levels (140e200 C). various factors: geographic location, operation conditions (e.g.
- Constant steam generation at different pressure levels
operation temperature and corresponding heat losses), field
(3 bare16 bar). layout, orientation of the collector field and others. In Table 1
- Ice storages, e.g. storage capacity, charging and discharging
the technical data of the Fresnel collector is summarised.
behaviour and time, suitability for peak load shifting or The optical efficiency of the collector is dependent on the
night usage, heat losses. incidence angle of direct solar irradiation. By definition h0 is
- Keeping provided temperatures at chillers constant:
the optical efficiency at normal incidence (i.e. sun in zenith)
different temperature levels (e.g. 140 Ce200 C), automatic and under ideal conditions (e.g. with 100% clean primary and
control of active collector area. secondary reflectors and receiver glass tube). For the here
- Collector maximum power (all mirrors focussed/all mirrors
described collector it is given by:
defocussed) and efficiency measurement.
- Cooling curve of collector overnight (test of heat losses and h0 ¼ 0:635
vacua).
To calculate the efficiency for any other incidence angle, an
incidence angle modifier (IAM) is defined as a correction factor
to be multiplied with h0. The transversal IAM is independent
2.2. The linear concentrating Fresnel collector of the field design, whereas due to end losses the longitudinal
IAM depends on the number of modules being combined in
The collector is a boiler for industrial use. It is designed as a one string. Data on an representative IAM are shown in Fig. 3.
linear focussing solar thermal collector for industrial process
heat. The building block of the collector is a “module”. Its
length of 4.06 m is defined by the length of the SCHOTT PTR70
vacuum absorber tube. Three modules are depicted in Fig. 2.
Fig. 3 e IAM of a string of 4/8/16 IS Fresnel collector

Fig. 2 e Three modules of the Industrial Solar Linear modules (referred to DNI), obtained from raytracing
Fresnel Collector. simulation by the manufacturer.
Because of shading effects, the IAM is >1.0 for some sun

positions, therefore the optical performance is not at its
maximum, when the sun is in zenith position. The maximum
efficiency will be reached, when the sun is at a transversal
zenith angle of 5 and a longitudinal zenith angle of 0 .
hmax ¼ 0:663ðfor sun at 5 transversal zenith angle;

0 longitudinal zenith angleÞ
Due to the use of a vacuum absorber, thermal performance

is almost independent of wind speed. Conductive losses as
well as convective losses are very low and thermal losses are
dominated by radiative losses.
The thermal losses per m2 of primary reflector can be
approximated in second order by: Fig. 4 e Characteristic curves of the Industrial Solar linear
Fresnel collector LF-11.
q_ loss ¼ u0 Aaperture DT þ u1 Aaperture DT2 (1)
with DT being the temperature difference between absorber

and ambience. As a result from measurements and based on Robur NH3eH2O absorption chiller (Model ACF60-00 LB) were
the aperture area the heat loss coefficients of the IS-LF11 are: installed. They have a nominal cooling power of 12 kW each
and a nominal COP of 0.6. The required driving power is
u0 ¼ 0
around 20 kW each. The chillers are modified versions of
u1 ¼ 0:00043 Wm2 K2 Robur’s standard gas fired product and therefore need high
resulting in equation by (Lüpfert driving temperatures to regenerate the absorption process.
et al., 2008):

q_ loss ¼ 0:00043 Wm2 K2 Aaperture m2 DT2 K2 (2) Table 2 displays the main technical specifications of the
chillers.
By knowing the optical efficiency and the heat losses in the
The Robur chiller needs driving temperatures between
collector, the expected usable thermal power of the collector
140 C and 200 C and therefore easily provides cooling tem-
can now be calculated:
peratures below 0 C with NH3 as a refrigerant. The two
q_ ¼ hoptical ða; qÞd $A$G q_ loss ðDTÞzh0 $IAMl ðql;d Þ$IAMt ðqt;d Þ$A$G chillers are always flowed through pressurized hot water at
16 bar, also in steam mode. The higher temperature lift
u1 $A$ðDTÞ2
compared to lithium bromide chillers, allow a dry heat rejec-
(3) tion with the built-in-ventilation of the Robur chillers. The
where a and q are the sun’s azimuth and zenith angle, d is the
angle between the collector’s absorber pipe and northesouth
direction, A is the collector’s active aperture area, and G is the
direct normal irradiance (DNI). By an approximation the op-
tical efficiency can be expressed with the factorized biaxial
IAM as a product of functions of the projected longitudinal
and transversal zenith angles q1 and qt. The thermal efficiency
h of the collector now gives the ratio of the usable thermal
power to the collectable solar irradiance ADG:
q_
h¼
A$G
q_ loss ðDTÞ2
¼ hoptical ða; qÞd zh0 $IAMl ðql;d Þ$IAMt ðqt;d Þ u1 $ (4)
A$G G
For the sun being in zenith position, IAMl and IAMt each
become 1, and the display of h over DT2/G is then being called
the characteristic curve of the collector, which is shown in
Fig. 4 for different values off G.
The thermal efficiency is the basis for obtaining the expected
gross heat production of the collector. In order to calculate the
expected gross heat production, only the values for irradiation,
sun position and the actual collector size are needed. Fig. 5
shows the linear concentrating Fresnel collector on site.
2.3. The water ammonia absorption chillers

Fig. 5 e Photograph of the linear concentrating Fresnel
As connecting core of the whole plant, the absorption chillers collector with an aperture area of 132 m2 in Freiburg,
are the interface of the two main circuits. Two units of the Germany.
For both circuits, the driving and cold side, the hydraulic
Table 2 e Technical data of the Robur NH3eH2O
absorption chiller. layout is designed for both serial or parallel flow through the
absorption chillers. Furthermore, the driving side is also
Operational characteristics Unit Model
designed for the two options operation with pressurized water
ACF60-00 LB
directly or steam indirectly via a secondary circuit linked by a
Heat capacity at regenerator Nominal kW 24 heat exchanger (condenser). Both operational in-
Cooling capacity at the kW 12 terconnections are shown in Fig. 7.
following conditions: For operation with pressurized water, the chillers are part

T in water C 210 of the primary collector circuit and therefore are flowed

T amb C 35 through directly. An expansion tank is used to maintain

T in water C 0

constant pressure. At driving temperatures up to 200 C, an
T out water C 5
operation pressure of 16 barg was used. In a real application,
Chilled water flow Nominal l h1 2600 for 200 C driving temperature an operation pressure of at
min. 2300
least 20 barg is recommended.
max. 2900
For steam operation, the connection of the chillers is
Pressurized water flow Nominal l h1 3500 explained in Fig. 7. The primary collector circuit is displayed
Operating temperature (TAIR) min.
C 12 in red. The steam is extracted from the steam drum into the
max. þ45 steam network (green) and condensed in a heat exchanger. In
Electric power consumption max. kW 0.84 order to keep a high condensating temperature in the
condenser, a control valve is used to maintain the pressure in
this part of the steam net. The condensate therefore is still far
advantage is that no wet cooling tower with additional hy- above 100 C at the heat exchanger outlet. An air handling
draulic circuit is needed. The disadvantage is an increased unit (ventilation) cools down the condensate below 100 C,
electricity demand of the chiller itself. For the process of an before the liquid is being expanded in the control valve to
NH3eH2O absorption system please refer to Henning (2007) or ambient pressure and led to the feed water tank, from where
Otanicar et al. (2012). Fig. 6 shows one of the Robur chillers on it is fed back to the collector circuit with a piston membrane
site. pump. The secondary circuit of the chillers (violet) with
pressurised water is linked to the steam network via the
2.4. Hydraulic layout e driving circuit condenser. The two chillers are summarized as “Chiller
Block” in the figure.
The solar collector and parts of the hydraulic circuits on the In direct steam generation mode, the pressure mainte-
driving side already existed from a previous project. For this nance unit is connected to the chiller circuit, while the pres-
solar cooling plant set-up, thermal insulation, a condenser for sure in the collector circuit is maintained by the steam drum.
steam and an additional secondary circuit was added. A series The cooling of the condensate by the ventilation increases
of additional sensors and valves had to be added accordingly. the preheating power and thereby to a reduction of the steam
The two chillers are always flowed through pressurized hot production. This non-optimal configuration is the result of
water at 16 bar, also in steam mode. According to the chiller enhancing a pre-existing system with using as many of the
manufacturer’s technical requirement, a direct flow with existing components as possible. Neither the feed water tank
steam through the chillers is not possible. nor the feed pump are suited for high temperatures, so that
the system design was done as an economic compromise. The
system losses for steam operation are therefore higher than
for pressurized water mode.
For hydraulics on the hot water side, stainless steel piping
was used; most of the connections were welded or flanged.
The insulation was aluminium-coated rockwool with sheet-
metal jacket. The heat transfer fluid in the driving circuit is
water from the tap. It is treated by a softener before filling it
into the circuit. No additional de-ionisation devices was
installed, but all the piping on the driving side are made of
stainless steel of type AISI 316 Ti (1.4571) or AISI 321 (1.4541).
Since the condensate is fed back into the collector circuit,
there is no considerable water exchange during operation, so
that blow-down is not required.
2.5. Hydraulic layout e chilled water circuit
As for the driving circuit, the hydraulic layout of the chilled

Fig. 6 e One of the two NH3eH2O absorption chillers. Half of water circuit is set-up for both, serial or parallel flow through
the chiller volume is the integrated dry heat rejection via the absorption chillers. The chillers turn off as soon as the
heat exchanger grills and fan. minimum set point (10 C for most tests) is reached and is
Fig. 7 e Driving circuit e steam operation mode.
turns on again at a 2 K warmer output (8 C). This differential for pump flanges. All the press fittings showed to be leak-
can be changed. The chillers reach their maximum power proof; whereas some of the screw fittings had to be tight-
shortly after start when corresponding driving temperatures ened again.
are provided.
The nominal feed flow on the cold water side of each chiller
is 2.6 m3 h1. If the feed flow decreases below 2 m3 h1, the 2.6. Description of operation regimes
chiller is turned off automatically with an internal security
switch. The chilled water circuit comprises of four ice storages For the plant operation, test stages were determined which
(Fig. 9) and an electric load simulation unit with three resis- are detailed as follows:
tance immersion devices. Fig. 8 shows one of eight different
interconnection possibilities, the flow through the electric 1. system characteristics:
heating elements that simulate a real load. general:
The heat transfer fluid is water glycol brine with 35% by general functional tests and start-up behaviour
volume. Signal-controlled valves and pumps allow various driving circuit:
operational interconnections and strategies to be tested and operation with pressurized hot water
evaluated. According to the high quality requirements for operation with steam
the monitoring data and acquisition, sensors were posi- cold water circuit:
tioned in all relevant locations of the piping and compo- direct cooling of electric load (without storages)
nents, as can be seen in Fig. 8. Hydraulics on the chilled storage charging
water side was installed as follows: Press fittings were used storage discharging
for the copper tube connections and hemped screw fittings simultaneous storage charging & electric load cooling
Fig. 8 e Operation mode in chilled water circuit e direct cooling of load.
automatically by a superordinated control. Thus, the solar

collector and absorption chillers were controlled separately.
To keep the driving temperature constant at the chillers,
two control strategies were applied. For pressurised water
operation in the collector, the temperature at the collector
outlet is the control variable, while the number of active
mirror rows ranging from 0 to 11 is the regulating variable.
During direct steam operation, one control variable is the
steam/water pressure at the condenser with the correspond-
ing opening of the motor valve at steam exit as regulating
variable, and the other control variable is the fill level in the
steam drum with the corresponding variable speed drive
signal of the feed pump (frequency converter) as regulating
variable.
On the cold water circuit, one of the main objective was to
keep the chillers running in order to avoid system shut-down
Fig. 9 e The four ice storages with temperature sensors. for a constant operation throughout the day and test
maximum chiller capacity under different conditions. There-
fore, the electric load power was increased until a balanced
value was reached. This was done manually via the process
software. The chillers do not comprise of a oscillating function
cascade operation e sequential operation of one or two and therefore switch off when the set temperature of 10 C is
chillers for part load behaviour reached.
2. advanced operation (load profiles e industrial applications e Table 3 summarises the characteristic control variables
automated operation) and values of the cold water circuit.
system management, constant load inlet temperature Also, latter developments included PID control imple-
day/night cold storage mentation for the cold circuit at the end of the last cooling
load profile e refrigerated storage house, strategy for season. Therefore, only little usable data was produced with
covering dynamic peak demands this automated operation. First test showed a stable behav-
automated operation iour. The PID controlled values are figured in Table 4.
With control implementation using PID control, the flow
The first stage of system characteristics was tested well rates of the absorption chillers and the electrical heating load
and the system’s and components’ boundaries and potentials is maintained constant, irrespective of changes in viscosity of
were identified. Further conclusions were drawn for future glycol due to the temperature variations. The objective of the
system operation and optimisation tasks were determined in automatisation of the two main circuits is a combined auto-
2011 that were mostly applied in 2012. The performed opera- mation of the entire system.
tion experiences in the first monitoring year resulted in
developing a more defined and feasible focus for future
operation and testing sequences. Of the second stage of 3. Monitoring and data analysis
advanced system operation, the last two points e load profile
and automated operation e were not concluding tested. In The system was operated and results were produced in 2011
chapter 3.3 the optimisation actions regarding control and and 2012. In 2011, monitoring results for 40 days were pro-
hardware are explained. duced. Here, the basic system characteristics were assessed
that led to optimisations for 2012. Due to the weather condi-
2.7. Control strategy and variables tions, most of these days took place between mid of August
and end of September. In 2012, due to more complex tests and
System operation needed constant supervision as neither the weather conditions, about 30 days of usable monitoring data
collector nor the chilled water side was controlled were produced.
Table 3 e Control variables and values in cold water circuit.

Control variable Regulating value
Outlet temperature of TDC Position of mix valve

Inlet temperature of heating elements (load simulation unit) Position of mix valve
Outlet temperature of heating elements Electrical power of heating elements
Volume flow through heating elements (load simulation unit) Pump power
Constant heating power of heating elements Thermal power of load/heating circuit Q_ Load ¼ m$c
_ p $ðT_Load_out T_Load_inÞ
Table 4 e Control variables and values in cold water Table 6 e Experiments with optimized system in 2012.
circuit with PID control implemented.
Operation 2012 e advanced experiments No. experiments
Control variable Regulating value with optimised system
Flow rate through chillers Analog input to the pump 1. Pressurised water operation 17
Flow rate through heating elements Analog input to the pump 1.1 Direct cooling of electric load 14
Outlet temperature of heating Analog input for heating (without storages)
elements elements 1.2 Storage charging 3
Inlet temperature of heating Analog input for the
2. Steam generation 14
elements mix valve at load
2.1 Direct cooling of electric load 7
(without storages)
2.2 Storage charging 7
3.1. Monitoring overview 2011/2012
In the following two tables, Tables 5 and 6, an overview on the

experiments of full operation days are given. rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi !,
In 2012, focus for the driving circuit was set to: steam 1 1
w¼ C2$ðR C0Þ þ $C12 $C1 C2 (5)
4 2
operation with constant pressure at different pressure levels,
ranging from 3 barg to 16 barg, and reaching high collector The coefficients C0, C1, C2 are determined by the calibration
efficiency for pressurised water operation. procedure and used for the correct calculation of the tem-
For the cold circuit and chiller behaviour the focus was set perature with the resistance R. The flow meters were cali-
to: constant and high (180 Ce200 C) driving inlet tempera- brated with certificate by the manufacturer. With a specific
tures, maximum chillers capacity, cascading for start-up and data storage program for monitoring data bases, monitoring
shut-down behaviour. data was filtered with different ranges or due to certain testing
conditions.
3.2. Data acquisition and analysis Thermal power and energy values were measured with the
inlet and outlet temperature sensors that were calibrated as a
For high quality data acquisition and thorough monitoring, couple and corresponding volume flow by the following
measured data was written in 5-s-steps for 64 sensors and equation for heat flux:
actors. Data files started for each system or software start, in
general one data file per day. For the data analysis, averaged Q_ ¼ r$V$c
_ p $Dw (6)
values (10 s, 60 s, 1 min, 60 min) were produced to obtain It is calculated with the fluid density r, volume flow V,
representative and comparable results for further calculations specific heat capacity cp and temperature difference Dw.
and plots. The energy efficiency ratio (EER) displays the relation of
The data acquisition units are Agilent data logger and produced cooling energy Qcold to the consumed driving energy
Wago modules. Measurement accuracies of data loggers and Qheat. It is calculated by
sensors were 0.007 K for temperature measurements and 0.3%

of the measured value for flow measurements including Qcold kWhth
EERth ¼ (7)
sensor, wire, switching and transducer errors. Temperature Qheat kWhth
sensors are PT100 with accurate four-wire technique. Their For the electrical EER, two values are displayed in the dia-
signals are read and converted by an internally programmed grams of chapter 4: The EERel considering the total electricity
Linux software that generates the raw data files. consumption of the system and a corrected value (EERel,cor)
In the beginning of each cooling season, the temperature without taking into account the electricity consumption of the
sensors were calibrated with a certified reference sensor used cold distribution pump.
for collector certifications, providing reliable results. The

calibration results with differing values are included in the Qcold kWhth
EERel ¼ (8)
monitoring software that converts the temperature signals. Eelec kWhel
The calibration formula for the corrected temperature is as

follows Qcold kWhth
EERel;cor ¼ (9)
Eelec;cor kWhel
This approach is according to IEA Task 38 monitoring pro-

Table 5 e Experiments on basic system characteristics in
cedures for solar cooling plants (Napolitano et al., 2011),
2011.
considering that a conventional reference system does not
Operation 2011 e basic system No.
include this consumption, thus giving a more accurate com-
characteristics experiments
parison. Additionally, the complex circuitry of the experi-
1. General functional tests and start-up behaviour 9 mental set-up of the chilled water side needs to be considered
2. Direct cooling of electric load (without storages) 4 for the calculations, resulting in higher pressure losses and
3. Storage charging 10
pumping energy. As electric consumers are included: chiller 1,
4. Storage discharging 7
chiller 2, circulating pump in collector circuit, circulating
5. Storage and electric load cooling 7
pump in chiller circuit (for EERel). Motor activity for mirrors is
neglectible. Also, electricity for monitoring devices is excluded solar irradiation was strong during noontime hours, mirrors
as obsolete for real applications. were defocussed.
The four ice storage tanks proved to work. The measure-
3.3. Optimisation of the system performance and control ment coincided with the calculated values for melting heat
strategy considering the 334 kJ kg1 heat of fusion of water. The 1200 l
provided a storage capacity of around 110 kWh. Heat con-
Experience of the operation 2011 where combined and ductivity was decreasing clearly after short charging time. The
implemented for system optimisation in the following moni- charging temperature decreases with increasing charge level.
toring year 2012. For the chilled water circuit, the electric load The four ice storage tanks were equipped with a wooden
simulation unit was optimised to provide an overload insulation box lined with insulating material so that thermal
regarding the maximum chilling capacity of the chillers. For gains over night stayed low. This was even the case during
the driving circuit, modifications for an improved steam warm summer nights when a temperature difference between
operation were implemented. Also, further sensors were ambient temperature (w20 C) and ice storages (w0 C) was
installed for an improved collector assessment. Several high.
modifications for the control and monitoring were imple- Each of the three electric heaters is rated at 10 kW, 30 kW in
mented in the data acquisition software. total. For 2011, the maximum power did not increase more
In detail, following issues have been addressed and opti- than 18 kW in total. This was due to wire connection of the
mised in 2011 and 2012: heaters with the power controller. This resulted in part load
behaviour when both chillers were operating simultaneously.
start-up and shut-down strategy of chillers, cascading For 2012, additional electrical work was carried out in order
capacity enhancement of electric heaters, modifications of to provide full load supply. According to that, optimal oper-
electrical wiring ating results were obtained that showed a realistic load and
mirror drives and control upgraded to Industrial Solar’s operation behaviour. Both chillers were operated in full load
present state of technology easily without charging the storages, thus simulating peak
driving circuit pump replaced by a stronger pump, suitable demand behaviour.
for higher operation temperatures In Fig. 10, the dependency of the thermal EER on the
insulation of piping and balance of plant in the collector ambient temperature is shown for the whole monitoring
circuit, leading to a reduced start-up and a better heat period 2012. As a condition, the cooling power was higher than
conservation during night 6 kW for all displayed values; thus, ensuring to show smooth
automatic volume flow and magnetic valve control. operation characteristics.
As the trendline shows, the thermal EER decreases from
0.64 to 0.56 in average between 22 C and 38 C ambient
4. Operating experience and performance temperature correspondingly. If this trend is interpolated
analysis linearly for higher ambient temperatures, at an ambient
temperature of 54 C, the average thermal EER would be 0.48.
4.1. Overall system behaviour However, the chillers performance strongly depends on
further conditions, like solar irradiation, cooling load and
The installation was started up in the mornings e depending operation mode respectively.
on the radiation quality and planned test sequence e and Fig. 11 shows the distribution of the overall electricity
shut-down in the evening. A few times, the data acquisition consumption of the solar cooling plant. Depending on the
was active during night time for gathering information on operating conditions, the system electricity consumption
cooling down curves for the collector and storage behaviour. ranges between 1.5 kW and 3 kW.
For maximum collector performance, the mirrors were
cleaned manually before each test day.
Even in 65% part load, the collector provides the full driving
power for both chillers. Pressure control of the collector in
direct steam generation mode was working, but needed
further optimised during the two monitoring periods.
Troubles keeping the pressure constant in 2011, led to a
phase of manual control of a sliding gate valve in order to
closely observe the behaviour of the valve and steam pressure.
Additionally, a more suitable configuration of valve and
controller for 2013 was examined. The overflow valve was
identified as main source of difficulties. From operating
manually, a control strategy for 2013 had been identified. The
overflow valve will be replaced by another sliding gate valve.
To avoid collector overheating, the chillers need to be
operating to consume produced heat and some of the mirrors
need defocussing during noon. Usually, the main objective Fig. 10 e Sensivity analysis of thermal EER vs. ambient
was to have both chillers running with full load and only if the temperature.
Fig. 11 e Distribution of electricity consumption for typical

plant operation e pressurised hot water driven operation.
4.2. Pressurised water operation

Fig. 13 e Summary of representative operation days e
pressurised water operation e EER thermal and electric.
Figs. 12 and 13 show an overview on representative operation
days for pressurised water operation for measurements
in 2012. Here, optimised system operation was carried
out. Representative means operation under following the representative days are rather similar for all monitored
requirements: days and range between 0.53 and 0.58. A clearer difference can
be seen for the electric EER values. Those range between 5.8
long operation day e good weather conditions for constant and 9.1. Fig. 14 shows characteristic data for the collector
collector operation operation as summary of the six selected pressurised water
smooth operation without or only with short interruptions. operation days. The cumulated DNI, measured with a pyrhe-
liometer, and the cumulated thermal gain by the collector are
The bar height in Fig. 12 tends to correlate with the oper- displayed as sum in kWh over the whole operation time of the
ation duration, collector power and chillers power, e.g. on day. DNI before or after collector operation in the morning and
operation day 28.06.13 the whole plant was running for about in the evening is not included in the sum. The efficiency and
3 h whereas on operation day 23.07.13 the runtime accounted the IAM are displayed as percentage on the right axis. Here,
to more than 5 h, additionally the weather conditions were the IAM is already the product of DHI based transversal and
really good and driving temperatures were around 200 C. longitudinal IAM and the cosine of the zenith angle. These
For the EER values, the ratio of produced and consumed efficiency numbers are averaged values over the operation
energy was calculated according to Eqs. (8) and (9). The results time. Those values can be easily misinterpreted. For example
for the thermal EER, overall electric EER and corrected electric clouds in the late afternoon can lead to a higher average effi-
EER (without electricity consumption of cold distribution ciency value on a day with comparably poor measurement
pump) are displayed in Fig. 13. conditions, because during hours with lower efficiency, e.g. in
Those values were calculated with the energy values for the late afternoon, the collector was not operated. This can be
each day, starting with the heating-up phase of the collector seen, when comparing the values of 27th of July (clouds in the
that starts approximately 1.5 h before the first chiller is afternoon) with those of 23rd of July or 1st of August, which
running. Therefore, those values differ to the values that are were days with more than 8 h of collector operation.
presented for the following example days where only good The average IAM values shown in Fig. 14 are typically low
operation phases are evaluated and do not consider the because of the short length of the demonstration collector.
warm-up phase of the collector. The thermal EER values for
Fig. 12 e Summary of representative operation days e Fig. 14 e Summary of representative operation days e
pressurised water operation e thermal and electric pressurized water operation e driving circuit and collector
energies. characteristic values.
This is the main reason for low power in the early morning the thermal losses had been calculated and subtracted by
and late afternoon. Daily peak values for efficiency are using the results from measurements by DLR for thermal
reached around 1 p.m. and are usually around 20% higher losses of the used receiver pipes in the collector (Lüpfert et al.,
than the average values as can be seen in Fig. 15. 2008).
Referring to the detailed system schematic in Fig. 8, the
4.2.1. Example day for pressurised water operation circuit for the heat load power and cooling power is the same
In Figs. 15 and 16, the typical system behaviour for pressurised in the presented operation mode, even though measured with
water operation is shown and evaluated for a representative different temperature sensors and flow meters. As heat load
day in July 2012. Fig. 15 shows the measured and expected power and cooling power are calculated with DT (see Eq. (6)),
characteristic values for the collector operation. Fig. 16 shows the values are very similar, even if the temperatures change.
the characteristic parameters for the cooling system analysis, The clear peaks of the chilled water temperatures were
thermal power values and temperatures. generated through power alternating tests with the electric
For the calculation of the measured thermal collector load simulation unit. The response of the collector circuit to
power, a delay correction had been accounted for. The DNI the changes in the chilling circuit was examined, but no sig-
had been measured by a Pyrheliometer. The expected collec- nificant feedback in the collector circuit was observed. If the
tor power had been calculated by multiplying the active col- set load power is higher than the maximum cooling capacity
lector area with the measured DNI and the simulated collector of the chillers, the measured power was also higher and
efficiency. The basis for that simulation had been a raytracing temperature in the cold circuit rises: The heat load power was
model of the collector in the raytracing environment OptiCAD, adjusted manually. The temperature peaks in the cold water
where the biaxial incidence angle modifier (IAM) had been circuit were always caused by setting the electric load to
raytraced in 5 -steps in the longitudinal as well as in the 100%, starting at 11:28 h and again at 11:55 h, 12:25 h and
transversal plane. That raytracing model accounts for the 13:40 h. The chillers power was not sufficient to cover that
collector geometry, shading effects, sunshape, reflectivity cold demand and the temperatures increased. Even though
values for primary and secondary mirrors, angular dependent the temperatures change, the cooling power remains
transmissivity of the glass tube, angular dependent absorp- reasonably constant, but showing a slightly increased power
tivity of the absorber pipe, an assumption for the mirror ac- if the load was set to 100%. The system does not have a high
curacy, end losses, and cosine losses. To obtain an expected heat capacity due to short pipings and therefore, response is
thermal efficiency from the optical result of the simulation, immediate if parameters were changed. Even a small
Fig. 15 e Overview of the collector parameters for pressurised hot water operation e 23.07.2012 e 1 min values.
Fig. 16 e Overview of the monitored parameters for pressurised hot water operation e 23.07.2012 e 5 min values.
mismatch of 1 kW between cooling capacity and heat load the EER of the absorption chillers. Table 7 summarises the
leads to a temperature change rate of more than 1 K min1. measured EER values for this example day.
Since there are thermal gains from the ambience via the
thermal bridges, the cooling power in equilibrium is slightly 4.3. Steam operation
smaller than the heat load power. At 14:34 h the electric load
was set to 93%, when the temperature slowly decreases and Figs. 18 and 19 show an overview on representative operation
starts rising from 15:40 h onwards due to a lower irradiation days for steam operation for measurements in 2012.
and thus driving power for the chillers. Also for steam operation, the bar height in Fig. 18 tends to
After 16:25, one chiller was switched off as the collector correlate with the operation duration, collector power and
power was decreasing with decreasing irradiation. The vol- chillers power. The results for the thermal EER, overall electric
ume flow in the driving circuit was not adjusted as no auto- EER and corrected electric EER (without electricity consump-
mated operation had been implemented. tion of cold distribution pump) are displayed in Fig. 19.
Fig. 17 shows the values for the thermal and electrical ef- Those values were calculated with the energy values for
ficiency ratio for the same day. The average thermal EER was each day, starting with the heating-up phase of the collector
0.64. The average electric EER for the entire system is 8.9, for that starts approximately 2 h before the first chiller is running.
the average electrical EER without cold distribution the value Especially for steam operation, the heating-up phase is longer
accounts 12. for producing a constant steam flow with the experimental
Those results refer to the full operation time between plant. Here also, the presented values in Fig. 19 differ to the
11:15 he16:15 h and represent typical performances under values that are presented for the following example day for
good operating conditions. This graph was generated with the full operation phase. The thermal EER values for the
5 min average values, thus representing more liable values for representative days are rather similar for all monitored days
Fig. 17 e Thermal and electric EER values for pressurised hot water operation e 23.07.2012.
Table 7 e Average values for the thermal and electrical

energy efficiency ratio regarding good operating
conditions (11:15 he16:15 h).
EER thermal EER electrical EER electrical without
cold distribution
0.64 8.9 12
and range between 0.42 and 0.52. A clearer difference can be

seen for the electric EER values. Those range between 4.9 and
8.5. Compared to the pressurised water operation, those
values are lower as the steam typically has a lower heat ca-
pacity at the same pressure and temperature compare to
liquid water. Fig. 19 e Summary of representative operation days e
steam operation e EER thermal and electric.
4.3.1. Example day for steam operation
Monitoring data for steam operation at 6 bar is shown and
evaluated. For this operation day, both chillers were used and condenser/heat exchanger the separates the steam circuit
the chilled circuit connection enables flow through the elec- from the driving circuit of the chillers as can be seen in the
tric heat load only as shown in Fig. 8. Fig. 20 shows averaged 5- circuit schematic Fig. 7. Some losses are assigned to this
min values of the characteristic parameters for the system condenser and the ventilation unit afterwards. This inter-
analysis, thermal EER, thermal power values, volume flows connection is also due to the experimental set-up.
and temperatures. Also, the chilling power and thermal Furthermore, the chillers consume lower driving power at
output of the electric load (heat load power) are similar as the lower temperatures. The zigezag profile of the driving
hydraulic connection is direct. power between 12:00 and 14:00 h is due to the manual
However, as can be seen in the upper chart of Fig. 20, the adjustment of the collector power and pressure release
heat load power is about 1.5 kW less than the cooling power of valve in order to maintain a constant steam pressure of
the chiller. This is due to a slightly lower delta T of inlet and about 6 bars. Table 8 summarises the EER values for this
outlet temperatures accounting 0.1e0.2 K. This can be seen in steam operation day.
Fig. 21 that shows a close up view of the same temperatures in
the chilled water circuit. As the temperature difference be- 4.4. Results
tween ambient temperature and the cold water circuit is more
than 30 K, heat input into the cold water circuit is detected. 4.4.1. Chiller performance and driving circuit
With a different scaling of the chilled water temperatures e interdependency
load and chiller temperatures respectively e this temperature When the chilled water temperature reaches the fixed set
difference can be seen in profound data evaluation. point of 10 Cs, the chillers turn off. with a preset differential
Fig. 22 shows a combined view of the relevant collector of 2 K, they turn on again. Consequently, if the load is too low
parameters, such as irradiation, mirror position and steam the chillers show a clocking behaviour. Cascading smaller
power. chillers instead of using single large ones in part load showed
This shows that the steam power in the collector circuit promising results. Tests with two chillers starting below
e around 40 kW e is significantly higher than the driving 140 C driving temperature needed more time to reach effi-
energy consumed by the chillers e around 27 kW. This loss cient cold production values. At those low starting points, it
is due to several reasons. First of all, the steam passes a proved to be more efficient to use one chiller in the beginning.
Different starting points during increasing driving tempera-
tures in the morning were tested. Some example measure-
ments of the first operating hour are shown in Table 9.
Test runs with driving temperatures below 140 C showed
poor chiller performance, as can be seen in Table 9. For driving
temperatures of 120 C, the cooling power reached only 5 kW,
the thermal EER was only 0.3 for the whole operation day.
For constant operation, the collector and chillers are sta-
bilizing each other in their operating temperatures over a
certain range. If the collector power rises, the overall system
temperature is rising and the cooling power of the chillers
along with the consumed driving power is rising until a new
equilibrium temperature is reached. On the other hand, when
the collector power is decreasing, the chillers will consume
Fig. 18 e Summary of representative operation days e more power than the collector provides, so that the system
steam operation e thermal and electric energies. temperature is decreasing. Thereby, the chiller’s cooling
Fig. 20 e Overview of the chilled water circuit parameters for steam operation at 6 bar e 17.08.2012 e 5 min values.
power is decreasing as well until a new equilibrium temper- conditions. With a continuous operation, the heating-up
ature is reached. This works only within a certain range, phase of the collector reduces significantly if it has been
because when the collector power is significantly higher than operating the previous day. Also, considering the more
the maximum power demand of both chillers, the tempera- promising operation in different latitudes, the heating-up
ture increases until a determined threshold. phase would be reduced additionally. For longer collector
The presented thermal EER values for the representative strings, the end losses would be reduced and hence, the start-
days range between 0.53 and 0.58. Those values were calcu- up time would be decreased. Start-up losses are equal to the
lated with the daily energy sums for each day, starting with accumulated heat losses during night. The main part of heat
the heating-up phase of the collector that starts approxi- losses in this installation are caused by thermal bridges, like
mately 1.5 h before the first chiller is running. Therefore, those pumps and heat exchangers.
values differ to the values that are presented for the example
days where only good and constant operation phases were 4.4.2. Ice storage integration e charge and discharge
evaluated and do not consider the heating-up phase of the The ice storage integration proved to work reasonably well.
collector. Here, the EER values range between 0.4 and 0.8. Charging time with full operation accounted to 9.5 h with
Those results show the potential for good operation steam operation and parallel heat load connection. As heat
Fig. 21 e Close up view of temperatures in chilled water circuit.
Fig. 22 e Overview of the collector parameters for steam operation at 6 bar.
conductivity was decreasing clearly after short charging time, chillers until about 4 p.m. The start-up temperature and time
the electric heat load was operated in parallel to keep both depend on the heat inside the collector circuit that was stored
chillers running and thus stabilizing the steam operation of of the previous operation day and the heat losses during night
the collector. Here, the chillers are switched off automatically or non-operation days.
if the set temperature is reached, because the charging tem- During the last two monitoring periods, pressure control of
perature decreases with increasing charge level. With the the collector in direct steam generation was optimised and
wooden insulation box, thermal gains over night stayed low, data was produced for several monitoring days, testing
even during warm summer nights when a temperature dif- different pressure levels. Since problems with an overflow
ference between ambient temperature (w20 C) and ice stor- valve occurred, an automated controller for pressure control
ages (w0 C) was high. was still not working by the end of the measuring period in
The storages were steadily discharged with a thermal 2012. The operator kept the pressure constant manually,
power of 7 kW and volume flow of 2700 l h1, providing cold while recording the occuring system behaviour. For the mea-
water temperatures below 10 C for 13 h. surement period in 2013, the valve will be replaced and the
control strategy will be implemented in the existing control
4.4.3. Fresnel collector and steam mode software.
The Fresnel collector is designed for temperatures up to In direct steam generation mode, results show that the
400 C. For this specific installation and the chillers, temper- steam power in the collector was higher than the driving en-
atures between 160 C and 210 C are required for feasible ergy consumed by the chillers. This difference is partly due to
operation that are easily provided by the collector. the specific hybrid and laboratory set-up of this plant which
The highest power measured in 2012 reached more than allows different operating modes, instead of being optimized
70 kW in the early afternoon, when the IAM is largest. Since for steam operation. The biggest part of this loss is caused by a
the collector orientation is 67 from northesouth towards ventilation unit, that cools down the condensate from the
west-east, and since the longitudinal IAM is rather steep heat exchanger to roughly 60 C in order to protect the plastic
because of the short length of the demonstration collector, feed water tank and the feed water pump, which are both not
high efficiency values with an IAM above 0.8 are typically suitable for higher temperatures and pressures. At maximum
reached between 11:30 a.m. and 2:30 p.m. For this small col- steam production, the preheating of 100 l h1 from 60 C to
lector length, before 10 a.m. in the morning and after 4:30 p.m. 200 C needed 20 kW of the collector power.
in the afternoon, the IAM is below 0.5. Therefore, on an
average cloudless test day in summer, the heating-up time of
the collector is about 1.5 h and heat can be provided to both
Table 9 e Example measurements of the first operating

hour of the installation.
Table 8 e Average values for the thermal and electrical Driving Driving power Cooling power EERth Sample
energy efficiency ratio. temperature (one chiller) (one chiller) day
EER thermal EER electrical EER electrical without
120 C 12e15 kW 3.5e5 kW w0.3 09.08.2012
cold distribution
150 C 12e16 kW 7.5e9 kW w0.55 17.08.2011
0.5 4.6 5.7 200 C 20 kW 12 kW w0.64 01.08.2012
Since approximately 20 kW of the collector power was Because the usage of diffuse irradiation is not possible with
needed for preheating alone, only one of the chillers could be a Fresnel collector such systems are dedicated to be installed in
operated stably with steam. However, it could be shown that southern countries with high amount of direct irradiation.
pressure control of the collector in direct steam generation Because of the design of the collector, minimum sizes of 175 kW
mode is working. are recommended by the manufacturer for economic use. For
The feed water pump is currently being replaced by one smaller plants like the one used in this project, end losses in-
that is designed for higher temperatures and the tank may be crease significantly as well as the specific collector field cost.
replaced by a pressure vessel, that does not require cooling of The chillers produced chilled brine down to 12 C. Best
the condensate at all. Thereby, the required preheating power chiller performances were detected when the chillers were
for the feed water can be reduced and the steam production driven by pressurized hot water and the operation mode on
will be increased. the chilled side was for direct load cooling. The maximum
To achieve good measurement conditions, the mirrors are power provided by the electric heaters (which corresponds to
cleaned manually every morning. This is merely due to the R & the power of the cooling load) was around 28 kW in total for
D activities. Thus, dirt does not influence the measurement the monitoring period in 2012, resulting in good operating
quality. Because this is not the economic optimum, robots for conditions and performance assessment.
automated mirror cleaning are under development, a first The intergration of the ice storages worked well. The
prototype had been tested on site in 2012. An improved pro- storage capacity was measured with approximately 110 kWh.
totype is currently under development. It was observed that for the starting phase of storage
charging, the charging power was reasonably high. After
some time e approximately 1 h -, charging power reduces
5. Conclusions and outlook significantly. It is assumed that this behaviour is due to ice
crystal formation around the storages’ capillary tubes, which
The system showed sound operation behaviour. The main reduced in heat conductivity and thereby to a lower heat
components worked reliably. The plant was operated with a storage rate for a given temperature difference. In other
load simulation unit, so no building or real load was con- words, while charging the storages the heat transfer between
nected. Thus, the results achieved and the produced data are brine and ice storage decreased; as a consequence the tem-
of high quality as the electrical load simulation provided perature in the cold water circuit fell which ultimately lead to
defined thermal loads and energy demands. The combination a lower thermal EER of the chiller. A different operation
of a Fresnel collector with two Robur NH3eH2O chillers and strategy and/or storage design should be examined in future
flexible system design offered the possibility to analyse installations.
various operation modes. Obtained thermal power and EER Cascaded operation of the two chillers proved to be feasible
values correspond to the nominal data. The results are sum- for warm-up and cool-down phases. This also applies for
marised as follows: changing weather conditions. The cascading offers the opti-
mised operation strategy as inefficient part load behaviour
- Reliable operation in summer 2011 and 2012 for given can be avoided. An overall increase of the system perfor-
weather conditions and available direct radiation mance is feasible.
- Optimised operation in 2012 With optimized system control on both the collector and
- 50e60% efficiency of the solar system achieved (summer chilled water circuit, good efficiency and performance values
conditions) were achieved. On both the collector and chilled water cir-
- Thermal EER of the chillers: 0.4e0.8, average 0.6 cuits, automated functions were implemented that showed
- Power in the chilled water circuit between 10 and 26 kW enhanced operational results. However, there is still further
- Driving power between 12 and 45 kW, collector provided up potential for optimising the control strategy, as a fully auto-
to 70 kW matic operation has not been completed yet.
- Electric EER of the system 5.5e12 (without distribution This experience shows the importance of longer moni-
pumps), average 8 (for good operation conditions) toring periods for demonstration systems and the require-
- Electric EER of the system 4.6e8.9 (with distribution pumps), ment of a certain amount of sensors to check and evaluate the
average 5.5 (for good operation conditions) functionality of all components also in future projects.
- Collector output temperature in the range of 100e205 C Focussing the plant design on optimised behaviour and
- Chiller output temperature down to 12 C dimensioning, without taking into account existing structures
- Charging of ice storage with thermal energy of about like building hydraulics, shows that an energy efficient design
110 kWh. and performance is easily achieved. Still, this is a common
problem for real applications as many planners and installa-
Thermal and electric EER show nominal values and tion personnel are still inexperienced with novel technologies.
improved with optimized operation. In 2012, the Fresnel col- To ensure energy savings and efficient system operation,
lector was operating with an automated temperature control. updated knowledge and implementation experience is
As expected, this showed a significant improvement of the required. Additionally, it is recommended to keep the design
operation stability. The principal driving energy was pres- as simple as possible.
surized hot water. Further tests with steam operation under For economical optimisation, it is necessary to increase the
different pressure levels were carried out and worked well use of the solar thermal energy to its maximum by combining
according to described conditions. further domestic services with the solar cooling installation,
like heating and domestic hot water production. Taking into Eicker, U., Pietruschka, D., 2009. Design and performance of solar
account those aspects, the system concept provided prom- powered absorption cooling systems in office buildings.
ising results for feasible application in hot, sunny and dry Energy Build 41, 81e91.
Häberle, A., Luginsland, F., Zahler, C., Berger, M., Rommel, M.,
climates.
Henning, H.-M., Guerra, M., de Paoli, F., Motta, M., Aprile, M.,
Currently, the results and gained findings of this work 2007. A Linear Concentrating Fresnel Collector Driving a
support another project that is carried out with an industrial NH3eH2O Absorption Chiller, OTTI 2007-2nd International
partner for cold storage houses, focussing on southern Euro- Conference on Solar Air-conditioning, Tarragona, Spain.
pean and North African climates (Berger et al., 2012). Here, Henning, H.-M., 2007. Solar assisted air conditioning of buildings e
further assessment and simulation of yearly performances for an overview. Appl. Therm. Eng. 27, 1734e1749.
different load profiles are carried out. Henning, H.-M., 2011. Solar Cooling Position Paper, IEA SHC Task
38 Solar Air-conditioning and Refrigeration.
Even though a real market penetration of small systems
Henning, H.-M., Döll, J., 2012. Solar systems for heating and
has not occurred yet, the use of renewable energies is essen- cooling of buildings. Energy Proc. 30, 633e653.
tial for the future energy supply. Solar cooling is a viable so- Lüpfert, E., Riffelmann, K.J., Price, H., Burkholder, F., Moss, T.,
lution to provide cooling power in a sustainable way. Further 2008. Experimental analysis of overall thermal properties of
developments also look at bigger centralised system solutions parabolic trough receivers. Trans. ASME J. Solar Energy Eng.
like district solar heating and cooling. 130, 021007/021001e021007/021005.
MEDISCO, 2006e2009. MEDISCO e MEDiterranean Food and Agro
Industry Applications of Solar COoling Technologies. EC e
Specific Targeted Project e Contract No. 032559.
Acknowledgements Mokhtar, M., Ali, M.T., Bräuniger, S., Afshari, A., Sgouridis, S.,
Armstrong, P., Chiesa, M., 2010. Systematic comprehensive
techno-economic assessment of solar cooling technologies
The project “SOLERA” was co-funded by the EC within the FP6
using location-specific climate data. Appl. Energy 87,
supporting the topic “Cost-effective Supply of Renewable 3766e3778.
Energies”. Napolitano, A., Sparber, W., Thür, A., Finocchiaro, P., Nocke, B.,
2011. Unified monitoring procedure and performance
assessment for solar assisted heating and cooling systems. In:
references IEA SHC Task 38 Solar Air-conditioning and Refrigeration.
Otanicar, T., Taylor, R.A., Phelan, P.E., 2012. Prospects for solar
cooling e an economic and environmental assessment. Solar
Ayadi, O., Aprile, M., Motta, M., 2012. Solar cooling systems Energy 86, 1287e1299.
utilizing concentrating solar collectors e an overview. Energy PolySMART, 2006e2010. PolySMART e POLYgeneration with
Proc. 30, 875e883. Advanced Small and Medium Scale Thermally Driven Air-
Berger, M., Weckesser, M., Weber, C., Döll, J., Morgenstern, A., conditioning and Refrigeration Technology. EC e Integrated
Häberle, A., 2012. Solar driven cold rooms for industrial Project e FP6-2004-TREN-3, Contract/Proposal No. 019988.
cooling applications. Energy Proc. 30, 904e911. Saulich, S., Jakob, U., Cook, M.J., 2012. Performance analysis of a
Bermejo, P., Pino, F.J., Rosa, F., 2010. Solar absorption cooling novel control system for solar collectors coupled with sorption
plant in Seville. Solar Energy 84, 1503e1512. chillers. Int. J. Refrigeration 35, 719e728.
Cabrera, F.J., Fernández-Garcı́a, A., Silva, R.M.P., Pérez-Garcı́a, M., Schicktanz, M.D., Wapler, J., Henning, H.-M., 2011. Primary energy
2013. Use of parabolic trough solar collectors for solar and economic analysis of combined heating, cooling and
refrigeration and air-conditioning applications. Renew. power systems. Energy 36, 575e585.
Sustain. Energy Rev. 20, 103e118. SOLERA, 2007e2011. SOLERA e Integrated Small Scale Solar
Chemisana, D., López-Villada, J., Coronas, A., Rosell, J.I., Lodi, C., Heating and Cooling Systems for a Sustainable Air-
2013. Building integration of concentrating systems for solar conditioning of Buildings. EC e Integrated Project e Contract
cooling applications. Appl. Therm. Eng. 50, 1472e1479. no. TREN/07/FP6EN/S07.68923/038627.

Cooling Systeme

Uploaded by

Copyright:

Available Formats

You might also like

Cooling Systeme

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cooling Systeme

Uploaded by

Copyright:

Available Formats

i n t e r n a t i o n a l j o u r n a l o f r e f r i g e r a t i o n x x x ( 2 0 1 3 ) 1 e2 0

Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/ijrefrig

Solar cooling with watereammonia absorption

Christine Weber a,*, Michael Berger b, Florian Mehling a,

article info abstract

Nomenclature u0; u1 linear and quadratic heat loss coefficients

Froid solaire avec des refroidisseurs à absorption

1. Introduction energy balance on annual level, energy saving and energy

Fig. 1 e System schematic with main components and characteristics.

- Cold production with temperatures below 0 C.

Fig. 3 e IAM of a string of 4/8/16 IS Fresnel collector

Because of shading effects, the IAM is >1.0 for some sun

hmax ¼ 0:663ðfor sun at 5 transversal zenith angle;

Due to the use of a vacuum absorber, thermal performance

with DT being the temperature difference between absorber

2.3. The water ammonia absorption chillers

2.5. Hydraulic layout e chilled water circuit

As for the driving circuit, the hydraulic layout of the chilled

Fig. 7 e Driving circuit e steam operation mode.

Fig. 8 e Operation mode in chilled water circuit e direct cooling of load.

automatically by a superordinated control. Thus, the solar

Table 3 e Control variables and values in cold water circuit.

Outlet temperature of TDC Position of mix valve

In the following two tables, Tables 5 and 6, an overview on the

This approach is according to IEA Task 38 monitoring pro-

Fig. 11 e Distribution of electricity consumption for typical

4.2. Pressurised water operation

Table 7 e Average values for the thermal and electrical

and range between 0.42 and 0.52. A clearer difference can be

Fig. 21 e Close up view of temperatures in chilled water circuit.

Fig. 22 e Overview of the collector parameters for steam operation at 6 bar.

Table 9 e Example measurements of the first operating

You might also like