Introduction

The current report has been prepared under the project Cold storage driven by solar and wind
energy and is financed by the Danish Energy Authority (EFP07-II no 33033-0297).
The objective of the project is to develop and demonstrate a grid-independent cold storage system
for perishable food, medicine or other goods, with a special focus on the need for such systems in
developing countries with a sparse and unreliable supply of electricity. The project is directly based
on the result from the international SolarChill project where a unique battery less solar driven
vaccine refrigerator was developed by Vestfrost in cooperation with Danish Technological Institute.
The project partners are:
- Danish Technological Institute (Project manager)
- Danfoss
- Grundfos
- Fresvik (Norway)
- Karise Klejnsmedie
Teknologisk Institut
Energi & Klima
December 2010

Final report December 2010

Table of content

1.
2.
3.
4.
5.

Background.............................................................................................................................. 4
Requirements for cooling and storage of perishable goods ....................................................... 5
Literature survey on cold stores................................................................................................ 7
Experiences with hybrid photovoltaic/wind power supply ...................................................... 11
Simulation of hybrid power supply ........................................................................................ 14
5.1. Summary of simulations ................................................................................................. 19
6. Concept study of refrigeration systems ................................................................................... 20
6.1. Temperature control and energy storage .......................................................................... 20
6.2. Compressor options ........................................................................................................ 21
6.3. Configuration structure ................................................................................................... 21
7. Design of prototype................................................................................................................ 29
8. Experiments with a PV powered refrigerator/freezer cluster ................................................... 33
Energy consumption.................................................................................................................. 42
9. Evaluation and perspectives ................................................................................................... 44
9.1 Spin-off projects ............................................................................................................. 45
9.2 Perspectives .................................................................................................................... 46
Annex 1. SolarChill B - Prototype and test results ......................................................................... 47
Annex 2. Abstracts of selected project references: ......................................................................... 53
Annex 3. Existing products of interest for the project: ................................................................... 57
Annex 4. Presentations, SolarChill B............................................................................................. 62
Annex 4. Milk Cooling ................................................................................................................. 70

Final report December 2010

1. Background
The current project is directly inspired by the so-called SolarChill project, which is an ongoing
international project aiming at the development and demonstration of battery less solar PV
(photovoltaic) refrigerators. The SolarChill project has already been successful in the demonstration
and dissemination of a large number of PV driven vaccine refrigerators of chest type named
SolarChill A. The core technology is a DC driven soft-start compressor, developed by Danfoss, that
allows direct connection to a relatively small PV array. By avoiding the battery, the system
becomes inherently robust, and cheaper than competitive products. The energy is stored as latent
heat in icepacks. SolarChill A is now being produced by the Danish company Vestfrost, and has
recently been approved by WHO.
This principle has been further refined and extended to a so-called SolarChill B refrigerator
intended for the general market of households and small restaurants and shops. This product is still
not in mass production, but some prototypes have been demonstrated in practical application.
The consortium behind SolarChill has received more awards for the project, and has received
numerous requests from interested customers and partners. Danish Technological Institute has been
heading the technical part of this project, including product development and testing.
In 2009 the World Bank joined the SolarChill partnership and infused additional funding (through a
GEF project) into the SolarChill project. This project intends to carry out extended field tests in
Colombia and Kenya and will transfer the technology to potential producers in the two countries.
Inspired by this success, the current project aims at scaling up the same principles to a typical
community or commercial type of cold store facility. A large scale system would benefit from a
better volume/surface ratio and a relatively lower cost per unit of storage volume. The focus is on
the areas of the world without access to regular grid power, but with plenty of natural solar and/or
wind resources for grid independent power supply.

Final report December 2010

2. Requirements for cooling and storage of perishable goods

When designing a standalone refrigeration system, it is obviously very important to define the
performance in terms of temperature level, maximum temperature fluctuations, freezing capacity,
storage volume etc. Various sources have been used to find recommended storage temperatures for
some typical goods which may represent users needs in developing countries. The following is
from the web site [1].
Temperatures and ventilation
Different applications have different requirements for temperature control and ventilation. Figure 16
shows the temperatures needed for the storage of butter, meat, fresh fish and milk. Very often
storage of vegetables is complicated by the need for careful ventilation to remove unwanted gases,
and to avoid humidity conditions, which would spoil the produce. Relative humidity requirements
vary depending on the moisture content of the produce. A simple method of increasing humidity is
to sprinkle water on the floor. In vaccine and blood storage very careful temperature control is
required.

Figure16 Maximum storage time vs temperature for some products.

Source: REFRIGERATION FOR DEVELOPING COUNTRIES. Practical Action, UK

Temperature
All fruits and vegetables have a 'critical temperature' below which undesirable and irreversible
reactions or 'chill damage' takes place. Carrots for example blacken and become soft, and the cell
structure of potatoes is destroyed. The storage temperature always has to be above this critical
temperature. Even though the thermostat is set at a temperature above the critical temperature, one

Final report December 2010

has to be careful that the thermostatic oscillation in temperature does not result in storage
temperature falling below the critical temperature. Even 0.5C below the critical temperature can
result in chill.
Maximum storage
Temperature Relative
time recommended
C
humidity % (ASHRAE
handbook 1982)

Storage time in
cold stores for
vegetables in
tropical countries

Apple

0-4

90-95

2-6m

Beetroot

95-99

Cabbage

95-99

5-6m

2m

Carrots

98-99

5-9m

2m

Cauliflower 0

95

2-4w

1w

Cucumber 10-13

90-95

Eggplant

8-10

90-95

Lettuce

95-99

Leeks

95

1-3m

1m

Oranges

0-4

85-90

3-4m

Pears

90-95

2-5m

Pumpkin

10-13

70-75

Spinach

95

Tomatoes

13-21

85-90

1-2w

1w

Table: Ideal storage conditions for various goods

It can be seen from the table that there are basically three groups of fruit and vegetables: those
stored at 0 - 4C; those stored at 4 - 8C; and those that require a storage temperature above 8C. It
is often more convenient to concentrate on one of these groups.
Source:
http://www.appropedia.org/Cold_storage_of_fruits_and_vegetables_(Practical_Action_Brief)

Final report December 2010

3. Literature survey on cold stores

In this chapter, some of the highlights found in the literature and on the internet regarding solar &
wind powered refrigeration systems are presented.
The idea of using renewable energy for refrigeration purposes is not a new one; in particular solar
energy has been suggested as a power source, because it is globally available and generally in phase
with the refrigeration demand. Previous refrigeration projects have mostly been based on solar
thermal energy from concentrating or flat plate solar collectors. The heat would power an
adsorption or absorption refrigerator, working dependent on the collector outlet temperature, cold
store and ambient temperatures. The main advantage of this design principle is the relatively low
cost of the solar collectors as compared with photovoltaics, but it suffers from a poor COP of the
cooling circuit and does therefore require a large solar collector surface.

Fig.1. Thermally powered solar refrigerator in Burkina Faso (adsorbent: Zeolite)

With progress in solar photovoltaic module manufacturing, the cost/performance ratio has been
dramatically improved, so that solar electric power is now an option for refrigeration purposes too.
The main advantage is that standard cabinets and compressors (for DC) can be used. Solar powered
refrigeration is mostly known from medical refrigerators for use in remote areas, where no grid
power is available. These refrigerators normally use a battery for energy storage and for provision
of start current for the compressor. There are many years of field experience with solar PV
refrigeration, showing that they work well as long as the batteries are kept in a good condition.
However, the batteries are often not replaced with the original type at end of service life, or they are
stolen or misused. In the international SolarChill project, the battery has been exchanged by a
Final report December 2010

thermal storage (ice) in order to avoid the only vulnerable component in the system (the battery).
After an intensive field test, SolarChill is now being produced and marketed by Vestfrost, DK.

Fig.2 SolarChill refrigerator with direct drive compressor

To the knowledge of the author, wind energy for specific refrigeration purposes has not been
realised so far, though wind power is often used as an important power source in small isolated
electricity grids, where any appliances may be connected, thus also refrigerators. The only reference
found on wind powered cold stores is a European (EU) project on thermal storage of excess wind
power connected to the grid. In this proposal, the cold stores are controlled to switch on when wind
power is exceeding the general power demand, and off when there is lack of electricity. In this way
they serve as free energy storage for wind electricity1.
It seems from literature, that most projects have been carried out in the field of small scale solar
refrigerators for vaccine and household use, and in the market there are many representatives in this
category, for example SunDanzer and Vestfrost.
Larger systems have only been investigated sporadically, with German ILK Dresden as an
important exception. This R&D institute has developed a series of solar PV driven refrigeration
containers in 10ft or 20ft standard size for different purposes:
- General cooling of goods and foodstuffs (23 m3 / 3.4 kWp)
- Ice making container (250 kg/day / 5,1 kWp)
- Milk cooling center with 70 kWh ice storage (1000 l / 3.4 kWp)
- Water cooling (2000 l / 680 Wp)
- Medical storage (200-1000 l/ 1.7 kWp)
It is not clear how many units have been produced so far of each unit.
Both ILK, SunDanzer and SolarChill are using ice as a means of energy storage.
The US company SunDanzer has developed a refrigerated container for military purposes, see
picture below. These units are available with single or dual compartment/dual temperature. They
1

Night Wind: storage of wind energy in cold stores An EU FP6 research project

Final report December 2010

operate as coolers, refrigerators, or freezers. They are also available with or without the solar power
system and some models utilize thermal energy storage to reduce or eliminate the need for batteries.
SunDanzer solar refrigerator and freezer models feature 4.67" (12 cm) of polyurethane insulation.
The brushless DC motor compressor operates on 24 or 48VDC. The cost is 25,000-62,500 USD for
a 20ft unit depending on specification (without PV system).

Figure3. SunDanzer container for solar PV cooling. Source:

http://www.sundanzer.com/documents/SolarContainers.pdf
Another important reference system has been reported from India, where a PV driven storage for 10
tons of frozen fish has been constructed and investigated in an R&D project. The size of the storage
is 21 m3, and it is powered by a PV array of 4 kW peak. A battery is used to power the compressor
at night and on cloudy days.
The operating experience is positive, in the sense that the design was appropriate for the load.
However, from the inauguration of the plant in 1991 until reporting in 1996, a severe deterioration
of the PV modules and the batteries could be observed. The salt mist from the sea was very
corrosive to all electrical connections.

Final report December 2010

Fig.4. Diagram of the Indian cold store

Final report December 2010

10

4. Experiences with hybrid photovoltaic/wind power supply

Electricity from wind turbines and from photovoltaic modules, is by nature fluctuating in time; it is
therefore a straightforward idea to combine the two in order to reduce the peaks and lower the risk
of a total loss of power, because solar and wind tend to supplement each other. Such a system is
called a hybrid power plant, and is most often equipped with batteries as energy storage. The
batteries are relatively smaller in such a system when compared to a pure wind or a pure PV system
of same installed power value, how much depends on the daily load profile and the fluctuations of
the inputs. There are two main configurations used today for hybrid power systems:
1) DC system with a battery bus bar connecting the PV system via a charge controller , and the
wind turbine via a separate regulator
2) AC system with bi-directional battery charger/inverter and connection of PV array by a grid
inverter, and the wind turbine by AC regulator or dedicated inverter.

Fig.5 Hybrid 3 phase AC system with SMA inverters (Spain)

The advantage of the slightly more complicated AC system is that standard AC appliances can be
connected, as for example 230 V cooling compressors. The German inverter manufacturer SMA is
the main provider of this system type.
For small-scale supply a pure DC system may be the cheapest option, while the special DC
appliances will usually be more expensive than their AC counterparts. If an inverter is connected to
the battery, AC appliances may however still be used at the cost of lower total efficiency.

Final report December 2010

11

Fig.6 Hybrid system with DC bus

Fig.7 Hybrid system with AC bus

Final report December 2010

12

PV max. generator output

1.4 kW (different values on request)

Wind turbine (9 m/s wind speed)

0.75 kW (different values on request)

Battery

420 Ah / 24 V (different values on request)

Rated output

3.0 kW (different values on request)

Peak output 5 s

4.5 kW (different values on request)

Output voltage

acc. to local requirements

Type of output voltage

pure sine-wave voltage, electrically isolated

between input and output

Rated frequency

acc. to local requirements

cos-phi

any value allowed

Distortion factor for rated output

< 3%

Max. efficiency

> 95%

Fig.8. KACO reCUBE is an example of an off-grid turn-key power supply system with AC output.
In some cases the appliance can be directly driven by renewable energy sources without the use of a
battery bank. This is the case for the power supply the SQFlex pumps from Grundfos. In their
system, a special regulator converts PV and wind power directly to 3 phase AC power fed to the
pump motor. The same idea may be used to run a variable speed compressor for refrigeration
purposes if the market volume is sufficient.
A Danish study2 from 1999 has investigated the symbiosis between wind power and PV for a
hybrid energy system with battery storage. The analysis showed that for Danish conditions the
optimum distribution between wind power and PV was 1:1 (1 m2 PV modules to 1 m2 swept area of
the turbine).

PV & Wind Hybrid System, Lars Yde, Folkecenter for Renewable Energy, 1999

Final report December 2010

13

5. Simulation of hybrid power supply

Simulation of hybrid energy systems is a prerequisite for design of site specific energy solutions. A
free simulation software named HOMER for hybrid energy supply is available from NREL, USA at
https://analysis.nrel.gov/homer/ . The following analysis is based on simulation using this software
package.
It is assumed that the cooling system will in general be fed from a hybride system consisting of PV /
Wind Turbines and a battery bank. Homer is used to estimate the optimum size for each of the main
components for a given site and load pattern.
To simulate the operation under various climatic conditions with either windy or sunny conditions
the project includes calculations of the energy system based on Cape Verde Islands climatic
conditions, which is both very windy and sunny. The distribution of wind and solar energy for other
sites can then be estimated by multiplying a reduction factor on the available climate data. The
calculation is based on a 100W average consumption equal to a daily consumption of 2.4 kWh /day.
The various climatic data are obtained from the software program MeteoNorm version 4.0. An
additional simulation where the load is distributed at different times of day was also carried out.
Prices used for the financial simulations have been taken from catalogues and websites, and are
inherently subject to some uncertainty. Prices are exclusive of installation. All technical data are
derived from the Homer program. Economic data for the Homer simulation / calculation is
introduced to the best estimate based on catalogue values assigned prices for transportation and
installation. Price and the replacement interval of components is assessed from the knowledge of
stand-alone power plants in general. Especially the batteries require replacement several times
during the lifetime of the plant and the future price of these is vital for operating economy.

Figure 9. Schematic of the simulated hybrid system

In this project, the only simulation made is for a system with AC loads, meaning that an inverter is
included in the system.
Cape Verde
First simulation is for a system made in Cape Verde Island. Data in Table 1 are obtained from the
MeteoNorm database and program. This data is used to simulate the system in Homer. The
simulation result shows that a combination of PV and battery bank is the best economic solution,
despite a very high wind speed. This is due to the relatively high price of the small wind turbine and
that there is a good solar irradiance at the site. Figure 10 shows the distribution of the output of
systems with, respectively wind / battery and PV / battery.

Final report December 2010

14

January
February
March
April
May
June
July
August
September
October
November
December

Wind speed [m/s]

6.700
6.700
6.800
6.800
7.100
6.800
5.300
5.100
5.400
5.700
5.800
6.400

Clearness index
0,651
0,705
0,676
0,669
0,637
0,575
0,524
0,518
0,554
0,655
0,651
0,619

Daily irradiance [kWh/m2]

5.190
6.250
6.670
7.060
6.870
6.200
5.630
5.480
5.570
5.970
5.300
4.742

Figure 10 shows the most economic system solution as a function of average wind speed and the
daily insolation.

Final report December 2010

15

Figure 11. PV system. Annual production from a purely PV based system.

It is estimated that approximately 1% of energy needs is not covered by this system, it means there
are periods when the refrigeration system cannot maintain the temperature within the desired limits.
To also cover the last 1% would the system size would be excessive.
In Figure 11 we can see that the purchase cost of a pure PV system is cheaper than wind turbines, as
shown in Figure 1.4. Prices may fluctuate widely from company to company, so the financial results
must be interpreted with caution. Note that the figures are in Danish kroner, although the program
provides $.

Final report December 2010

16

Figure 12. PV system costs

Figure 13 Wind turbine

In figure 13 we can see the purchase cost for a pure wind-turbine system. As expected, the
windmill represents the vast majority, but the batteries weigh too heavily.
Final report December 2010

17

Figure 14
In this figure we can see power shared between a combination of PV and wind hybrid system. Here
the PV system only contributes 40% of the power production versus 60% from wind turbines. The
system is not quite optimal because there is a large untapped energy quantity of 45%. This is partly
due to the chosen big turbine, and that wind energy is fluctuating more than solar energy over the
years. This can be addressed by using multiple batteries, but it will increase the purchase price.

Final report December 2010

18

Figure 15. Cost breakdown for the hybrid PV/wind system

Using the standard assumptions for project economy in Homer, a comparison of the three systems
in Cape Verde climate gives the following electricity prices:
Pure PV + battery:
Pure wind + battery:
Hybrid System

5.1.

7.7 kr / kWh
6.7 kr / kWh
8.1 kr / kWh

Summary of simulations

Taking uncertainty about component costs into consideration, it appears that many areas of the 3rd
World will benefit most by choosing a strictly solar based system, when the alternative is very small
and expensive mills. The pure wind solution may be the cheapest option for Cape Verde, but here
are also quite exceptional wind conditions. Opting out of the wind also offers lower maintenance
costs. Only by larger wind power systems this option will be superior, provided that there is a
reasonable wind speed. This also corresponds to Grundfos' experiences with their solar / wind
powered pumps. The solar systems are by far the most widely used system type, and the hybrid
system is used only when there is good wind present.
The exercise of designing an energy supply system has proven Homer to be a useful tool, but there
are many assumptions and price data to be assessed in each case. The result depends (as always) on
the input from the user.
On the basis of this simulation exercise, it is decided that the cooling system will initially be
equipped with a pure PV system, the decision being stimulated by the recently falling prices of PV
modules (2009).
Final report December 2010

19

6. Concept study of refrigeration systems

In this chapter it is attempted to give a systematic overview of the practicable ways of construction
of a refrigeration plant driven by renewable energy.
The examples found in literature on cold stores driven by solar/wind energy are mainly constructed
from standard containers. The main advantages of this solution are that isolated containers are
widely available at an affordable cost, and that they provide good protection of the installations
during transport to their destination. Another reason for their popularity may be that storage systems
based on renewable energy are rarely very large due to high cost of the (solar) energy supply, and
the size of a standard container is therefore a good compromise.
Site built storages have the advantage that they can be made in any size or shape, either from
traditional construction materials or pre-fabricated wall elements. For large storages, it is probably
more cost effective to build a single large room than to set up several containers. It is also important
to keep in mind that the surface/volume ratio is much lower for a single large building, and the
specific energy demand is therefore lower.

6.1.

Temperature control and energy storage

With intermittent energy supply from sun and perhaps also wind, it is a prerequisite for stable
temperature that some kind of energy storage is implied, either:
A. Thermal storage system
1. Ice-slurry or cold liquid in separate tank(s)
2. Integrated ice storage
3. Solid matter storage, e.g. rock, concrete or pebble bed
or
B. Battery storage system
1. DC battery bank
2. AC battery bank (Battery with bi-directional converter)
The storage compartment temperature is generally influenced by:
- Ambient temperature
- Irradiance on the building
- Wind induced convection
- Power availability
- Goods in/out
- Thermal capacity
- Secondary energy storage (batteries)
- Thermostatic control
Some of these parameters can be controlled, others not. For optimum control, the energy storage
must cover demand in the number of days, where no power can be foreseen.

Final report December 2010

20

In SolarChill A (vaccine cooler), the temperature was only controlled by thermal design of the ice
storage together with a minimum thermostat to prevent sub zero temperatures. In the ILK design, a
tank with ice slurry was used as main energy storage, so that the cold could be released in a
controlled fashion.
Energy storage in batteries alone is a possibility, but is likely the most costly and vulnerable
solution. No examples of cold stores with long term battery supply have been found.

6.2.

Compressor options

Given that the cooling system should be made from widely available refrigeration components
based on the compression principle, there are different possibilities for compressor selection:
- small scale hermetically sealed DC or AC compressors
- larger AC compressors
- combination, for example DC compressor on solar power, AC compressor on wind power
The BD35 compressor from the SolarChill project has a special control for soft start, and gradual
increase of the speed as a function of available power. The experience has been good with direct
connection to PV modules, though some optimization is still possible.
If a battery is used, any DC compressor with good efficiency may be used, because start current is
then no longer a big issue.
AC compressors are cheaper and found in many sizes, so with an inverter based power supply such
compressors could be used as in a grid connected cold storage.
In general, compressors with variable speed are preferable due to increase of system efficiency
while the number of daily start/stop sequences is reduced. Compressors driven by permanent
magnet motors are the most efficient, but also the most expensive.

6.3.

Configuration structure

Apart from the compressor and refrigeration system, it is important to take a closer look at the type
of enclosure for the refrigerated volume itself. In the original project application, it has been
assumed that the best option was a container or small building, with a central refrigeration system.
However, the flexibility and low cost of standard freezer and refrigerator cabinets may open up the
possibility of using clusters of such devices as an alternative solution.
Structural options:
- Isolated container with one or more refrigeration systems (split units)
- Container as sunshield for installed cluster of refrigerators
One could argue that the latter solution is close to products which are already on the market, such as
individual PV refrigerators. However, a cluster of refrigerators at one site could have many
advantages such as better protection against weather, theft and vandalism, as well as the larger
capacity and flexibility in terms of differentiated storage temperatures and utilisation of the units.
The thermal loss associated with openings will also be less for chest type cabinets than for a
container with a door, because the cool and heavy air mainly remain inside the cabinet.
With an intelligent power management system, the goods with the highest value could be stored in a
refrigeration unit with high power priority, while bottled drinks and other uncritical items could be
stored in other low priority units. In the following section some of the advantages and disadvantages
of the possible solutions will be presented.
Final report December 2010

21

Thermal storage systems (examples)

Type A1. Ice-bank

Condensor &
compressor

Ice-storage
with
evaporator
Stored goods
Heat exc

Fig 16
In this case the cold storage uses a well-insulated central ice bank for thermal energy storage.
Heat exchange from the storage to the stored goods may be based on air or liquid loops with
forced or natural convection.
Pros
Energy efficient
Accurate temperature control

Final report December 2010

Cons
Specialized solution
May be expensive for small system sizes

22

Type A2. Multiple ice-lined refrigerators

Container

Ice-lined
refrigerators

Ice-lined
refrigerators

Fig 17
A container with natural ventilation of excess heat and PV roof is used for protection against direct
sunlight. The refrigerators could for example be of the type with ice-lining of the walls, which
would provide good thermal capacity. These are normally used in regions with unstable power
supply. Alternatively one could partly fill up a standard chest freezer with ice packs.
Pros
Simple solution
Standard components
Individual compartments (ownership)
Possibility for combined
refrigeration/freezing
Less infiltration of air in chest type cabinets
when compared to vertical door

Final report December 2010

Cons
Less efficient than large cold store per net volume
Larger temperature fluctuations due to large
refrigerator surface area
Less effective volume

23

Type A3. Mass storage made of stone, concrete or similar

Condensor &
compressor

Fan cooler

Stored goods

Fig 18
This type uses a heavy construction as sensible heat storage (thermal mass). The building could be
made of concrete or thick (mud)brick walls with good external insulation. Additional mass could be
added if necessary, for example in the form of water tanks.
Pros
Simple solution
Standard components
Cheap, local materials can be used

Final report December 2010

Cons
Less temperature control
More mass required than for ice storage

24

Battery storage systems

Type B1: DC battery bank
This is the traditional type of storage, typically composed of a number of 12 V lead-acid batteries.
The total battery voltage will usually be 12, 24, 26 or 48 V dc.
If the compressor is of DC type, it will run directly on the battery, otherwise an inverter is inserted
in the circuit. The inverter must be able to deliver the necessary peak current for start of the
compressor. If there are more compressors, it is possible to reduce the peak load if they do not start
up at the same time.
Pros
Simple solution
Low energy losses

Cons
Special DC loads or inverter required

Type B2: AC battery bank

The AC battery is a combination of the battery bank described above, and a bi-directional inverter
and charging unit. (see figure 7).With this type of storage, it is possible to connect AC compressors
and other electrical AC consumers directly to the terminals. If a PV system is used to charge the
storage, the DC current from the PV plant passes a standard grid-connect inverter and a rectifier, so
the voltage of the PV array and the battery are decoupled. AC generators from wind turbines or
diesel gen-sets can be connected directly to the AC bus of the system.

Pros
Uses AC standard components
Easy to expand via the AC bus
Multiple generator options

Cons
Specialized system, only few suppliers
Only suitable in larger systems

Type B3: Capacitor

This option is only for very short term storage, and is suitable for provision of the necessary start
current for compressors. It must be combined with a long term / thermal storage system. The main
advantage of the capacitor is that it does not require any maintenance.

Final report December 2010

25

System components and sizing

There are several system components that must be matched to each other and to the total expected
cooling requirement of the application. The most important are:

Gross and net volume of the cold store

Capacity of battery storage
Capacity of thermal storage
Compressor size
Size of the PV/Hybrid power system

The daily transfer of goods is the single most important design parameter, for example the
difference between the same volume of milk and lettuce results in at least a factor of 10 in
compressor average power.
Most compressors require a considerable start current, and a standalone system will therefore need a
battery or a capacitor to be able to start on PV power alone. However, there are small DC
compressors available with a very low start current, like Danfoss BD35. The main advantage is that
a vulnerable storage battery can be avoided. On the other hand, a battery is needed to keep the
voltage stable if the power is also used for lighting, control and display functions.
For cooling systems of considerable size, there are no large capacity compressors with low starting
current available on the market. It is therefore necessary to use a storage battery, or split up the
system into a cluster of smaller compressors each featuring a low starting current.
The size of the storage battery is determined by the need for continuous operation of the
compressor(s) and by the effective capacity of the thermal storage system. If a small battery is used,
the compressor will have to be bigger, so that it can work on full power during daytime. At night,
the battery will be exhausted, and the ice storage will absorb heat.
If a larger battery is used, the compressor will only have to deliver a smaller cooling capacity in
continuous operation. Some amount of ice storage may still be needed to rapidly cool the goods
when they are loaded.
Since a given energy storage capacity is much cheaper for ice than for battery storage, it is
recommended to minimize the battery to some hours of full load operation. The cooling circuit and
compressors must therefore be oversized relative to the average cooling needs. However, there is a
risk of mismatch between the PV array and the battery, meaning that the battery may run full on
sunny days, unless the compressors can absorb all the power while building up the ice volume.
Two example cases based on this strategy have been calculated as part of the project:

Final report December 2010

26

Case study for a purely PV driven milk cooler:

Capacity
PV array
Battery
Compressor
Ice storage
Regulators
Accessories

150 l milk per day (29
4C) . Total cooling need ca 5 kWh/day

1.5 kWp . Typical electricity production 5 kWh/day
1 kWh effective capacity (fx. 12 V 100 Ah deep cycle)
2 x 250 W power consumption. Runtime 7 full load hours/day
3 days capacity 15 kWh corresponds to about 150 kg ice
MPPT charge controller for optimum PV yield. Variable speed DC/AC motor drive.
Agitator with high efficient propeller + motor (motor outside insulation) or manual
stirring routine.

The system operation will depend of the state of charge (SOC) of the battery in order to maximize
the utilization of energy:

Low SOC: no action, only control+display functions

Medium SOC: One compressor + agitator if fresh milk has been loaded
High SOC: Two compressores + agitator if fresh milk has been loaded

Low temperature disconnect: When the ice has been almost fully built up, the compressors are
disconnected in order to avoid freeze damages.
Instead of two compressors, one large compressor may be used, if efficient operation can be
obtained at part load by variable speed control.
If there is insufficient sunshine, other power sources can be connected to the battery, e.g. generator
or wind turbine.
In grid connection mode, the design could be modified so that the battery and one compressor are
removed from the system.
Critical parameters for the cooling rate are first of all the initial temperature of the milk, the load
pattern, and the heat transfer ratio from the evaporator, through the ice-lining, and to the milk
container.
An hour by hour simulation in Excel has been made for different setting of these parameters.

Final report December 2010

27

Batch cooling
35

Temperature

30
25
20
15

Ice storage

10

Milk

5
0
0

4
Hours

Fig 19
Using a realistic total heat transfer rate of about 400 W/m2K for the milk side (with agitation), and a
heat transfer rate of 200 W/m2K for the ice container side, it takes 2 hours to cool down a batch of
150 litres from 29 to 4C. A heat transfer area of 1 m2 was used in this example.

Case study for general purpose cold store

Capacity
PV array

150 kg (water-equivalent) goods per day (29
4C) . Total cooling need ca 5

kWh/day
20 kWp . Typical electricity production 25 kWh/day

Ice storage

160 kg, 3 days capacity

Regulators

MPPT charge controller for optimum PV yield. Variable speed DC/AC motor drive.

Options:
Pre-cooling/night cooling of goods
Traditional refrigerators in cool building

Final report December 2010

28

7. Design of prototype
For many reasons, it has been decided to work with the standard chest-type refrigerator option.
Such units are relatively cheap, can be manufactured by many different producers, and are wellknown and easy to operate for the end user. The only severe drawbacks are the limited storage
volume and the relatively poor energy performance. These problems must therefore be dealt with.
Volume: The largest domestic freezer cabinets are about 400 liters, and can accommodate relatively
large quantities of goods. More units can be added and run in parallel operation. If there is no
requirement for the full volume, some of the units can be switched off, and thus leave more power
for the rest.
Energy consumption: As opposed to a walk-in cold store, the surface to volume ratio is poor for the
individual refrigerators, and the heat flux is thus higher. On the other hand, the cold air will tend to
stay inside a chest cabinet when the lid is opened. For the container, the cold air will try to escape
each time the door is opened. It is not evident which effect is the most important.
It is important to eject waste heat from the common enclosure when the compressors are in
operation, in order to decrease the ambient temperature around the cabinets as much as possible.
During night time (without operation) it would be better if the container is kept tight, so the heat
flux from the surroundings can be minimized.
Possible solutions:
Use refrigerators with external condenser connected to an exhaust duct. The cabinets are
stowed closely together for minimization of the effective surface area.
Build an external and ventilated enclosure for the cabinets with exhaust to the outside. A fan
is set to operate in parallel with the compressors. When all compressors are off, the air
around the cabinets will stay still, thus providing some additional insulation.

Final report December 2010

29

Cabinet

Cabinet

Fig20. Container with integrated chimney and cabinet enclosure. Pre-cooled air is taken from a
buried duct. Outer enclosure reduces temperature around the cabinets.
The temperature and heat transport for a container with four chest freezers inside has been
calculated. The heat flux to the freeze volume can be reduced during stand-still periods (night) if the
outer enclosure could also help keep the cabinets cool. It is assumed that the freezers are -18 C
inside, and that the outside air temperature is 35C. The insulation (U-value) of the container is then
varied from 0.2-10 W/m2K, and the corresponding heat flux calculated:

Dimension
Length
Width
Height
Number

Fridge

Container

1,5
0,7
0,8
4

4 m
2 m
2 m
1

Total volume

3,36

16 m3

Total surface

22,48

40 m2

Surface/volume

6,69

2,5

Fig 21
Final report December 2010

30

It can be seen that the insulation value of the container has to be rather good in order to have a
significant effect on the overall heat transfer from outside (35 degC) to the frost zone (-18 degC).
The thickness should be at least 10 cm for conventional insulation materials; a thin layer of
insulation has almost no effect at all.
With a well-insulated container it is possible to maintain a temperate climate that may be suitable
for storage of vegetables or fruit. In this example, the power required for freezing new goods are not
included, it is only reflecting the heat transfer losses.
Some possible cabinets for a prototype system have been identified and listed in Annex 3. The
listed energy consumption of comparable models are:

Whirlpool

419 liter

0,97 kWh/d (energy star)

Elcold 51 XLE

418 liter

0,81 kWh/d

Liebherr GTP 4726

441 liter

0,67 kWh/d (Elsparefonden)

Vestfrost SOE-215 DC refrigerator

213 liter

n.a.

Final report December 2010

31

Load from door openings

A good argument for choosing a chest type cabinest is the low warm air exchange when the lid is
opened, when compared to an upright type refrigeratur / cooled container. For the extreme case that
all air is exchanged if a door is opened in a refrigeration container, the following result was found:

Daily energy demand kWh

35
30
25
20

50 kg ice latent heat

15

Transmission losses

10

Air exchange 20 openings

5
0
1

Fig. 22. Cooling demand for a 16 m3 container(2) and 8 m3 distributed chest freezers(1) for an
example with 50 kg ice load/day and 20 door openings. It is assumed that the effective capacity of
the container is 8 m3, because some of the volume is reserved for walk space and installations.
Delta T is -18/35 C in the example, and wall U value 0,3 W/m2K.

The bar graph shows that the transmission losses are dominant in this case, but other cases could be
constructed with more or less input load. It is however clear that the argument on less air infiltration
load for the chest cabinets should not be exaggerated. From an energy point of view, the two
solutions are equally good.

Final report December 2010

32

8. Experiments with a PV powered refrigerator/freezer cluster

Test purpose
The purpose of the test is to evaluate the cooling and freezing capacity and energy consumption of a
solar powered refrigeration system during a simulated practical use pattern. The thermostat settings
and battery voltage thresholds could also be verified and adjusted as necessary during the test
period. As the power supply from the PV array depends on the solar irradiance, it is not possible to
run the same test twice. It is a dynamic test where the solar irradiance as well as the ambient
temperature varies as a function of time. However, the test is closer to real operation conditions than
a laboratory test with simulated power supply.

230 V AC cabinet #1
80 V PV
array 800
Wp

230 V AC cabinet #2

plc
DC/DC

DC/AC

24 V DC
cabinet #3

24 V battery
200 Ah lead-acid

(optional)

Fig 23 Lay-out of the tested system

The refrigeration system is set up at the solar energy test area of DTI, where a PV array with a
nominal power of 800 W had already been established. The batteries and charge controller are
purchased from a Danish PV system retailer. The inverter is a trapezoid 50 Hz 230 Vac inverter
from a previous project and is a robust type with high surge current. The two AC cabinets are
standard low energy household freezers, whereas the DC cabinet is a special ice-lined refrigerator
(fresh food/middle temperature) with high thermal capacity in its walls. The selection of large chest
type freezers gives low specific energy consumption due to a high volume/surface ratio and low air
infiltration. The commercial low energy cabinets are relatively inexpensive, and can operate with
an extremely low consumption if the thermostat is set to cooling mode. This idea has also been
proposed by other researchers3.

A fridge that takes only 0.1 kWh a day? Tom Chalko,Renew (Australia) issue 90 (2005)

Final report December 2010

33

Fig 24. Cabinets in the test lab and thermal storage in the freezer.
System components:
PV array: 800 Wp polycrystalline modules, 2x4 in series, system voltage approx. 80 V
Charge controller: Outback FLEXmax 60 A with MPP tracking
Inverter: Victron Atlas 24/2000 (2 kVA)with automatic sleep mode and low voltage cut-off
Battery: Vision 6FM200D-X12V 2x12 V sealed lead acid AGM batteries in series, nominal
capacity 200Ah(10h).
Cabinet#1: Frigor low energy freezer (prototype)
Cabinet#2: Elcold 31XLE low energy household freezer (run as refrigerator)
Cabinet#3: Vestfrost icelined refrigerator (DC)
Controller: Mitsubishi Alpha 2 application controller with pt100 inputs and relay outputs

Final report December 2010

34

Fig. 25. PV array with 2 x 4 modules in series and MPP charge regulator
System control
All three cabinets are controlled with a simple PLC controller (Mitsubishi Alpha2), and the
temperature range can be shifted from freezing to cooling mode if desired for the AC cabinets. If
the battery voltage becomes critically low the PLC will switch off the appliances one by one until
the battery is recovered. It has been verified by Elcold that this type of switching does not harm the
appliances, as long as it does not result in very short runtimes, where the compressor oil does not
have time to distribute on contact surfaces.

Fig 26. System controller (24V supply) and charge regulator. Screen dump of Mitsubishi block
oriented control software

Storage capacity
The energy storage is divided in battery storage and a thermal storage consisting of 15 kg icepacks.
The usable storage capacity is calculated as follows:
The 24V battery with 200Ah nominal capacity has a useful capacity of 24h*200Ah*0,5/1000 = 2,4
kWh where 50% depth of discharge is assumed to give a good balance between lifetime and
investment cost. The battery is necessary for voltage stabilization at the inverter input, but in
principle it could be much smaller and substituted by thermal storage volume. In practice it was not
possible, because it should be able to absorb the full PV array current.
Final report December 2010

35

Latent heat in cabinet#1 (15 kg of water) is 15kg*335 kJ/kg /3600s/h = 1,4 kWh. Sensible heat
between -18 and 0C gives another 15kg*2,1kJ/kgK*18K/3600s/h = 0,16 kWh. In refrigerator #3
there is a similar capacity, whereas there is almost no fixed thermal capacity in cabinet#2.
The total storage capacity corresponds to a few days of operation.

Fig. 27. Battery pack with two AGM sealed lead-acid batteries

Final report December 2010

36

Test schedule
Cabinet

Exchange of icepack-baskets: 10-21 september 2010

Mode
Freezing
Cooling
Base capacity
15 kg
(icepacks)
Input
10 kg (warm)
10 kg from #1
Output
10 kg to #2 (frozen)
10 kg to room
Exchange of icepack-baskets: 18-25 october 2010
Mode
Freezing
Cooling
Base capacity
15 kg
(icepacks)
Input
10 kg cold+10 kg
10 kg from #1
warm
Output
10 kg to #2 (frozen)
10 kg to #1

Cooling
Iceliner: 17.6 kg
Cooling
Iceliner:17.6 kg
-

Ice packs

Goods

cool

freeze
good

Ice packs

Fig. 28. Flow diagram of icepacks and consumable goods.

Most goods are assumed to be cooled for sale within short time; non sold perishable goods are
transferred to the freezer. Liquids will only be stored in the refrigerator. In the test run the real
goods have been replaced by icepacks in standard baskets with 5 kg in each basket. The idea is to
use some of the frozen icepacks to help cooling down new and warm items in a shorter time than
would otherwise be possible.

Final report December 2010

37

Results September 2010

Fig. 29. The graph shows that the warm baskets are cooled down within 24 hours to deep-freeze
temperature. When there is no load of new baskets, the temperature is even and constant around 19C. The double peaks in the basket temperature are almost certainly caused by sub-cooling of
some or all of the icepacks.

Fig. 30. Detail of the refrigerator running in standby mode, but with a high hysteresis of 5 K. The
controller was set to a hysteresis of only 2 K, so it is not evident why it is higher. However, the
basket temperature is quite constant due to the thermal mass. Runtime is about 10% of the time.

Final report December 2010

38

Fig. 31. Solar irradiance for the measurement period was sufficient to run the system. A sensor
voltage of 0.127 V corresponds to 1000 W/m2.

Performance Ratio
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
0

0,5

1,5

2,5

3,5

4,5

Daily irradiation in plane kWh/m2

Fig. 32. The power supply system was verified by calculation of the performance ratio, i.e. the ratio
between the nominal power output of the PV array and the power being charged on the battery. It
can be seen that the performance ratio decreases with daily irradiation, which should also be
expected because the battery runs full on sunny days. In average a value of 0,56 was found for PR,
quite typical for a standalone PV system.

Final report December 2010

39

Results 18-25 october 2010

Fig. 33. The first three days of the test is a stabilization phase, where the icepacks are cooled and
frozen, respectively. From the 18th the icepacks are loaded every afternoon, and frozen until next
day. The double peaks in the basket temperature are almost certainly caused by sub-cooling of some
or all of the icepacks. It looks like the basket temperature does not reach full freezing to -18C, but
all icepacks were fully frozen when unloaded. The quite low temperature of the new icepacks (when
loading) is likely because the cold air in the cabinet cools down the sensor in the basket
immediately after loading. Though the basket temperature cycles a lot, the fixed thermal storage and
inside air temperature remains almost constant, unless power is insufficient as it was the case during
the last cloudy days.

Fig. 34. The refrigerator thermostat was set to operate between 3 and 5C, but occasionally the air
temperature (shelve) becomes sub-zero. This happens when frozen icepacks from the freezer are
loaded together with warm icepacks, but only until the temperature is leveled between the two

Final report December 2010

40

baskets. The reason for temperature fluctuations may also be caused by the lack of a fixed thermal
mass in this cabinet. The warm basket is cooled down quite fast due to this temperature
equalization, and the warm packs never freezes. During the last two days, where power was cut off
some of the time, the passive heat transfer from the warm to the cold icepacks is evident, because
the cabinet temperature is warmer than the icepacks.

Fig. 35. The graph above shows in detail how the thermostat starts regulation already from 0C
after the frozen packs have been loaded together with warm packs. The on / off time distribution is
about 20/80%. A better thermostat sensor position could have improved the regulation pattern.

Fig. 36. Solar irradiance was variable and quite representative for a Danish autumn week. Due to
low solar energy levels on the 23rd and 24th, it can be seen that the battery voltage becomes low the
last days of the test. This means that the inverter has cut off some of the time, and there has
therefore been a rise in cabinet temperatures.

Final report December 2010

41

Battery voltage
28
27,5
27
26,5
26
25,5
25
0

10

12

2,5
2
1,5

kWh from battery

kWh to battery

0,5
0
0

10

12

Fig. 37. Battery cycling. Nominal battery capacity = 200 Ah *24 V = 4,8 kWh. The graph shows
that the battery is undergoing a daily 20-25% cycle of state of charge (SOC) which is acceptable for
a deep cycle lead-acid battery. On the sunniest days the charge is about 2 kWh, of which 80-85%
will be available the next day at the inverter output terminals.

Energy consumption

Daily power consumption (#1+2)

2

kWh AC

1,5
1

15C
25C

0,5

35C
0
0

4
6
8
kg/day freeze-in

10

12

Fig. 38. The electricity supply to the two AC powered cabinets was measured with an ordinary
electricity meter, and correlated with the daily exchange of icepacks in the freezer. The
measurements are corrected to three different ambient temperatures. If an average COP of 1.4 is
used, the measurements are corresponding with the theoretical energy demand for cooling and
freezing of the icepacks. The relatively low consumption is reached by a good thermal insulation
Final report December 2010

42

and the fact that the cold air exchange is very limited when the lid is opened due to the higher
gravity of cold air.
For a typical application in developing countries, the ambient temperature would be 35C and there
would be approximately 1.6 kWh of electrical energy available after a sunny day. Under these
boundary conditions it can be calculated that the freezing capacity is about 8 kg/day for the
experimental system.
If the frozen icepacks are transferred to the refrigerator and thawed before taking them out of the
system, the energy consumption is almost halved or the capacity doubled.

Final report December 2010

43

9. Evaluation and perspectives

It is very difficult to estimate the added value of the refrigeration system, but if 8 kg of goods can
be sold at an additional price of DKK 40 (for example 10% less waste of meat @ 50 DKK/kg) a
simple calculation shows that the investment pays off in less than 5 years:
System costs (hardware) for the prototype:
PV: 25000, 3 cabinets 15000, batteries 5600, regulator 4500, inverter 2000, PLC 1700.
Grand total approx 60.000 DKK ex VAT
Sale additionally DKK 40 x 300 days *10 years (battery lifetime) = 120.000 DKK
The regulator with MPP function could be saved in a simplified design, where 24V PV modules are
connected to the battery via a simpler regulator. It will also reduce the cost if only two large
refrigerators are used instead of three smaller ones. The large inverter of the prototype system could
be replaced by a much smaller one with integrated charge controller.
Proposed optimized design:

230 V AC cabinet #1
24 V PV
array 800
Wp

230 V AC cabinet #2

plc
DC/AC
+ REG

24 V battery
200 Ah lead-acid

Fig. 39 modified system

4 x 200 Wp standard PV modules in parallel

Charge regulator with integrated inverter 500 VA
Battery
PLC controller
2 x 500 liter cabinets

Estimated hardware costs < DKK 50,000.

Final report December 2010

44

9.1

Spin-off projects

As part of this quite extensive project, there have been a number of contacts with associated
activities as well as direct requests from companies operating in 3 rd world countries. The two most
important cases have been a milk-cooling project in Uganda and a World Bank GEF project
regarding improved storage methods for vaccines.
Milk has a relatively high sales value and a very limited storage time if it is not cooled effectively
after milking. In many developing countries there is a high loss percentage, or the milk is simply
not being produced and distributed to meet the real demand that could be fulfilled with proper
cooling and transport. It can also be the case that a central diary collects the milk at a very low
price, because they know the farmers do not have possibility to store and sell the milk at the local
market as they lack cooling facilities. If the farmers could score this profit, they could invest in
development in their local area.
The current design of the PV driven refrigeration system could be modified to milk cooling, and
this is actually being investigated by the Danish company Karise Klejnsmedie who are specialist
in stainless steel processing. They are currently designing a prototype with a cooling capacity of 75
liters/day for test at DTI. If the test results are promising, it is the intention to build a version for
field test. It is a challenge that the power consumption for cooling of 75 l milk is quite high, and the
desired cooling rate of 30C in two hours is much higher than a small compressor can cope with. It
is therefore suggested to integrate a big volume of ice around the milk containers as a buffer. The
PV array will also be bigger than the current prototype, but as the cost of PV has decreased to a
favourable level, it is not an obstacle in itself.

Fig. 40. International price development on PV modules (www.Solarbuzz.com)

The relief organization PATH is a specialist in cold chain vaccine storage and transport in
developing countries. One of the major problems with vaccines is that they cannot tolerate freezing,
but on the other hand they need cooling. Current use of icepacks in transport boxes requires precise
adjustment of icepack temperature before outreach, but still there is a risk of freezing. A new type
of icepack with a freezing temperature of 5-6C may solve this problem, but until now they are
frozen in ordinary freezers to -18C in order to be sure they are totally frozen (avoiding sub
cooling). If they could be frozen at, say, 1-2C they could be used immediately for vaccine transport
without risk of freezing the vaccine. It seems that the PLC regulated chest freezer from the current
project could be well suited for this purpose, and its size is perfect for the type of rack that is used
by PATH to hold the icepacks. The rack with icepacks will be tried in the lab by DTI in the near
future.
Final report December 2010

45

9.2

Perspectives

Solar powered refrigeration and freezing has drawn attention for decades, and numerous systems
have been developed with either thermal or electric energy input. However, the commercial success
stories have been scarce. With recent 30-40% lower cost on PV modules, it seems likely that
electric system will have advantage over thermally driven systems in general terms.
The current PV based project has deviated somewhat from its original intentions, but nevertheless
some important lessons have been learnt and useful results achieved.
First of all, the result shows that a very efficient and flexible refrigeration system can be constructed
at reasonable costs from commercially available standard components. The only draw-back of the
design is the battery that has a limited lifetime, so the system is not truly maintenance-free. On the
other hand, lead-acid batteries are available everywhere in the world at reasonable costs. In a future
system it might be possible to reduce the size of the battery further, and increase the thermal storage
capacity instead.
The system has been operating without any problems, apart from temporary cut-off during very
cloudy periods. In such periods, the controller can give priority to keep the most important cabinet
in operation on the expense of less important units.
Future application of a flexible refrigerator/freezer cluster driven by PV power is first of all in nongrid connected areas with reasonably constant daily solar irradiance, i.e. the Sunbelt, where there is
a real demand for storage of perishable goods. Typically it could be fish, meat, milk, certain
vegetables, cold drinks and ice cream. The profitability will be highest if the goods have a high
commercial value per volume and heat capacity unit.
Apart from commercial use, the system will also be very suitable for larger health centres with a
need for vaccine storage and freezing of icepacks. A version for this purpose is currently being
developed by DTI.
Finally, larger clusters of PV powered cabinets could be constructed on a community basis, where
each owner could have a small compartment on a rental basis. The system should be financed by an
investor who could earn a profit and take care of maintenance and protection. These systems could
be combined with street lighting and other services such as cell phone charging. For even larger
cold stores, it is possibly necessary to develop other configurations based on isolated containers or
buildings. In this case thermally driven solar refrigeration circuits could become an economic
option.

Final report December 2010

46

Annex 1. SolarChill B - Prototype and test results

SolarChill-B-chest type
A prototype of SolarChill B chest type has been build for the Danish Technological Institute.
Vestfrost has build the unit on basis of their existing cabinet (MK204) and the Danfoss Solar
Compressor (BD35K). The cooler is controlled by a mechanical thermostat with sensor inside the
cabinet. When the temperature inside the cooler gets to cold, the thermostat will stop the
compressor. 17.6 kg icepacks is build in the side walls of the cabinet and this ice bank keeps the
food and drinks cold during OFF-time.
The cooler is working in a similar way as SolarChill A

Photo: Test of SolarChill B in climate chamber at DTI. The cooler is powered by power supply (15
V) and power is available 10 hours a day. Temperatures are measured and recorded (8
temperatures inside the cooler and ambient temperature and humidity)

Final report December 2010

47

Photo: The prototype has 5 baskets (3 baskets in the upper part and 2 baskets in the bottom). The
volume is 137 litres and the cooler can contain a great amount of food and/or a greater number of
soft drink cans (see photo). The cooler is intended to be used for household use or small
commercial use.
At this photo it can be can seen, that the cooler is packed with 2,5 kg test packages (also used in the
European test standards for household refrigerators) in each basket and 8 soft drink cans (33 ml) in
each basket. This simulates food and drinks. This is a test method developed at DTI this springtime.
Tests have been made in a climate chamber at DTI, where also other tests are ongoing. The
temperature in the climate chamber is defined by other tests, and the temperature has been 22 C in a
period and is now 30 C and will stay at 30 C for a while.
Results at 22 C ambient temperature
The very first tests where used to find correct placement of the thermostat sensor in the upper part
of the cooler.
When this was found, stable running and hold-over time test were made.

Final report December 2010

48

Curve: Test at 22 C ambient temperature. During stable running the compressor and refrigeration
system is running about 3 hours and 40 minutes per day. The temperatures in testpackages are
between 0 and + 3 C and the air temperatures are between 1 and + 3.5 C. The hold over time is
about 2.5 days (food temperature has rised to + 7 C, air temperature to + 8 C) and after 3 days the
temperature has rised to + 10 C.

Final report December 2010

49

Results at 30 C ambient temperature

Curve: Test at 30 C ambient temperature. During stable running the running time is about 6 hours,
20 minutes per day. The temperature in test packages are between 1.5 and + 2.5 C and the
temperature in air are between -2 and + 3.0 C. The hold-over time is about 2.7 days (food is
warmed up to + 7 C) and after a little more than 3 days the testpackages are warmed up to + 10 C.
It can be seen from the curves that the ice has melted after about 2.5 days.
The DTI has the opinion, that these results are very good. Both at +22 and + 30 C ambient we have
surplus of cooling capacity, temperatures from -1.5 to + 3 C. By modifying the thermostat setting
this might be a little warmer, if this is wanted. The hold-over time is about 2.5 3 days, which also
is good.
The cooler is intended for household and small commercial use, where the claims for temperature
setting is not that big. It does not matter if the temperature is -1.5 or + 3 C. The coldest temperature
is at the bottom of the cooler and the warmer is in the upper baskets, and if wanted this can be
beneficial eg. by storing food in the bottom and drinks in the upper baskets.

Simulated use of cooler: Half reload test.

The DTI again started the power supply, and the cooler vent into stable running at 30 C ambient.
We then strted a simulated use of the cooler, by changing 20 cold cans with 20 warm cans (with
temperature of + 30 C) at sunset which is supposed to be the most critical time. The cooler
contains 40 cans, so we call the test a half reload test.
I each basket are 8 cans placed in the 4 corners. The upper cans are replaced every day, and the test
was done in four days.
Two of the cans were equipped with termocouples for measuring the temperature, one in the upper
row of baskets and one in the lower row.

Final report December 2010

50

Curve: Tempertures inside the SolarChill-B cooler during the half reload test. It can be seen that
the warm soft drink cans are quite rapidly cooled down, and after about 8 hours the warmest can is
below 7 0C.
It can be seen from the curve above that the warm cans are rapidly cooled down. It takes 8 hours for
the warmest cans to be cooled down to +7 0C.

Final report December 2010

51

Curve: Half reload test with temperatures and the effect used by the power supply. It can be seen
that the compressor is running almost all the time when power is available (10 hours a day) when
the half reload test is ungoing. This reflects that we are close to the maximum cooling capacity by
this test. The average temperature in the cooler is quite low (about + 2 0C).
The conclusion on the half reload test is, that the cooler works fine and can each day cool at least
20 soft drink cans (equal to 6,7 kg water) down from 30 0C to below + 5 C.

Final report December 2010

52

Annex 2. Abstracts of selected project references:

1) A refrigeration facility for milk cooling powered by photovoltaic solar energy

M. De Blas 1 *, J. Appelbaum 2, J. L. Torres 1, A. Garca 1, E. Prieto 1, R. Illanes 3
1Department of Projects and Rural Engineering, Edificio Los Olivos, Campus Arrosada, Public University
of Navarra, 31006 Pamplona, Navarra, Spain
2Department of Electrical Engineering Systems, Faculty of Engineering, Tel Aviv 69978, Israel
3Technical School of Forestry, Polytechnic University of Madrid, Spain
email: M. De Blas ([email protected])
*Correspondence to M. De Blas, Department of Projects and Rural Engineering, Edificio Los Olivos,
Campus Arrosada, Public University of Navarra, 31006 Pamplona, Navarra, Spain.
Keywords
refrigeration stand-alone PV systems performance
Abstract
The use of renewable energy sources is usually a reliable alternative in rural areas and developing countries,
where the grid line does not exist or is at a great distance. In this work, the characteristics and working
conditions of a refrigeration facility designed for cooling down an expected daily production of 150 l of milk
are analyzed. The facility is a stand-alone, direct-coupled system where 20 photovoltaic modules, 120 Wp
each, power two permanent magnet, direct current motors of 24 V, 650 W. Each motor drives a separate
cooling system compressor, which provides the flexibility to operate the equipment with one or two motors
and with various interconnections of the PV modules, depending on the available irradiance level and the
thermodynamic state of the system. The photovoltaic energy obtained during daylight hours is stored in the
form of sensible and latent heat of frozen water in a tank surrounding a milk container.
Thermodynamic analysis of the system shows that the autonomy of the system is 25 consecutive cloudy
days if the available stored ice energy is 80% of the nominal capacity of the water/ice tank. Results of the
refrigeration efficiency are similar to those obtained by other commercial refrigeration facilities powered by
a photovoltaic array, including batteries. Copyright 2003 John Wiley & Sons, Ltd.

2) FP5 Project Record 528 . Development of a mobile solar-driven refrigerated container for rural
areas
General Project Information
FP5 Programme Acronym: EESD
Project Reference: ENK6-CT-2000-35010 Contract Type: Exploratory awards
Start Date: 2000-07-06 End Date: 2001-07-05
Duration: 12 months Project Status: Completed
Update Date: 2005-05-03
Project Description
In countries with a high level of sunshine there is a considerable demand for cold supply. Statistics have
shown that in sough Europe nearly 10% of fresh food become worse in quality due to the lack of sufficient
refrigeration. This concerns in particular structurally weak rural areas in Southern Europe (Greece, Spain,

Final report December 2010

53

Portugal, etc.) which represent about 2/3 of the total territory in these regions. The main problems in these
thinly populated areas are often the logistics and the lack of infrastructure. Against this background there is a
major demand for new and innovative solutions for decentralised and combined cooling processes. Solar
refrigeration plants are ideal for keeping foodstuffs fresh, since the solar radiation which "causes" the need
for refrigeration is available in sufficient quantities. The COOLTAINER proposed here can be used for
different applications such, for examples the storage of fresh fish, meat and vegetables and also of medicine.
Prior to the project, the possibility of application on a large scale was tested by means of some basic systemrelated laboratory trials and the results were positive. The objective is to develop further and optimise the
COOLTAINER in the framework of a research project so that upon complete
Coordinator
Organisation Type: Other
Organisation: CRYCLE CRYOGENIC DEVELOPMENT N.V.
PO Box 4056
2003 EB HAARLEM
NETHERLANDS
Contact Person:
Name: VAN BAKKUM, Theo
Participants
Organisation: COOLINGTEC INDUSTRIES (1993) LTD.
Ramat Hashron
47264 RAMAT HA-SHARON
ISRAEL
Contact Person: ARTSIELY, Yehuda

3) Vapour Compression Cooling System Powered By Solar PV Array for Potato Storage
Mohamed A. Eltawil1 and D.V.K. Samuel2
1Agril. Engineering Division, Faculty of Agric., Box 33516, Kafr El-Sheikh Univ.,
Egypt, email: [email protected]
2 Post Harvest Technology Division, IARI, New Delhi-110012, India.
ABSTRACT
Refrigerated storage, which is believed to be best method for storing the fruits and vegetables in
fresh form, is not available in rural or remote locations where grid electricity is almost not
available. So, without having a conventional energy source at these areas, the present study was
taken up to design and fabricate a solar PV powered vapour compression refrigeration system to
attain favourable conditions for potato storage, and to evaluate its shelf life under different
operating conditions. The system is designed and fabricated in the division of Agril. Engineering
at IARI, New Delhi. It consisted of PV panel, lead-acid battery, inverter and the vapour
compression refrigeration system consists of a drier-cum-filter besides the main components:
compressor, condenser, expansion device, evaporator, exhaust and evaporator fans. The 2.50 m3
cold storage structure was constructed and insulated with proper materials. An evaporatively
cooled storage structure (1.0 m3) was used for curing process. The cured potato cultivar (Kufri
Chandermukhi) were stored for 5 months. The stored tubers were divided into two lots, one used
as control (free sprouting) and the second was manually desprouted. Measuring of moisture loss,
dry matter, sprouting, rotting, sugars content, starch and chipping quality reflected the shelf life

Final report December 2010

54

of potato. The average daily solar photovoltaic (SPV) energy output and energy consumption by
the load were 5.65 and 4.115 kWh, respectively, under full load. The obtained results indicated
that, the average daily actual COP for loaded and air circulated cold storage structure was 3.25.
The average temperature and relative humidity maintained inside the loaded and air circulated
storage structure were 283.13 K and 86 %, respectively. Solar panel can serve as an alternative
source of energy for powering cooling system.
Keywords: Solar photovoltaic, cold storage, potato, curing.

4) EXPERIMENTAL EVALUATION OF A SOLAR PV REFRIGERATOR

WITH THERMOELECTRIC, STIRLING AND VAPOR COMPRESSION HEAT PUMPS
Michael K. Ewert
NASA Johnson Space Center
Mail Code EC2
2101 NASA Rd. One
Houston, TX 77058
Et.al
ABSTRACT
Solar photovoltaic (PV) refrigeration systems which use batteries have existed for several decades but have
only been used in limited applications. Recent technology developments and experimentation in the field are
refining the solar refrigerator." Coupled with the decreasing cost of PV, this is expected to lead to more
wide-spread acceptance and use. Keys to the success of the solar refrigerator are a thermally efficient
cabinet, thermal storage, and a high-efficiency heat
pump. These elements, which are also important for aerospace refrigerators, were combined in a successful
demonstration of solar refrigerator technology at the Johnson Space Center. Three heat pump, or cooler,
technologies (thermoelectric, Stirling, and vapour compression) were experimentally evaluated in the same
vacuum insulated cabinet. Individual heat pump performance was quantified, and the entire solar refrigerator
system was evaluated by defining a system
solar COP (coefficient of performance).

5) Wind-Powered Refrigeration - US Dept. of Energy

The National Renewable Energy Lab, a branch of the US Department Of Energy has funded YES to develop
an innovative wind-powered refrigeration system. Much like clean drinking water was a major issue two
decades ago for villages, reliable refrigeration (or ice) remains a need of most of the third world. The
availability of refrigeration is in great demand both in terms of public health and for standard of living
reasons. However high electrical energy costs and lack of distribution infrastructure keeps a refrigerator of
the reach of most of the world's population.
If ice could be produced for free (or nearly free) for example, many of the world's fishing villages could then
successfully transport the fish they harvest to market and convert it to hard currency. An obvious solution is
to develop a renewable energy-based system to power refrigeration on a large, "village-wide" scale, much
like shared water wells are setup today around the world. Everywhere but in the desert, wind handedly beats
solar power in energy-recovered-per-invested-dollar. However, because wind availability generally tends to
be erratic in most parts of the world, it is difficult to efficiently couple the needs of a static refrigeration cycle
to a wind turbine. YES engineers are developing an adaptive electronic load-matching charge management
system that will permit a large-scale commercial ice machine to be directly connected to a >10kW towermounted wind turbine.

Final report December 2010

55

6) Photovoltaic-powered cold store and its performance

J.Nagaraju & al.
Dept. of instrumentation, Indian Institute of Science, Bangalore
Published in: International journal of energy research 2001:25
A photovoltaic-powerd cold store plant, the first of its kind, has been developed to store 10 tons of frozen
fish at -15 oC. It consists of a photovoltaic array (4 kWpeak), a battery bank (96 Vdc, 180 Ah), a vapour
compression refrigeration system (1 ton), electronic controls for automatic operation of plant and an
insulated cold chamber. Experiments were conducted on the system to evaluate its performance with no heat
load (frozen fish at -15 oC) and with different heat loads. It is observed that the system can be operated with a
maximum heat load of 2350 W to maintain the walk-in-cooler temperature below the freezing point of fish (2 oC). The performance studies conducted on these subsystems viz., photovoltaic array and battery bank
showed that their output has deteriorated in 5 years.

Final report December 2010

56

Annex 3. Existing products of interest for the project:

ILK Dresden
Solar-powered shelters for cold-storage of medicines at hand
Commissioning and operating them is really uncomplicated. And there is practically no need for
maintenance. As they are around the- clock automatically operating systems, which are designed for
use under high ambient temperatures and for rough operation conditions in full-year operation, they are very
well suited for use as absolutely self sufficient cold-storage depots for central remote hospitals, refugee
camps, vaccines supply networks and remote immunization stations.
Main features
exclusively solar supplied cold room shelters
for cold-storage of medicines and/or vaccines and/or blood conserves
available as 10 ft-, 20 ft- and 40 ft ISO-versions
three ranges of interior room temperature levels to be arranged in different combinations on demand
room 1: 0 C to 5C; room 2: 10 C to 15 C and room 3: about 25 to 28 C
At present it is an indisputable fact that special adapted photovoltaic based off-grid applications will be
competitive in areas with unsafe or missing power supply. Especially in remote areas with high ambient
temperatures, where demands for self-sufficient, islanding cooling systems coincide with high solar radiation
rates, and where lacks of cheap mains supply or unreliability and high costs of
decentralised, conventional energy supply systems are to state, special adapted PV supplied cooling systems
for cold-storage of any type of medicines, vaccines or blood conserves may be suited well.
Bearing this in mind ILK Dresden has designed different types of so called Solar Cooling Containers for
cold-storage of food-stuffs, which have started their market carrier two years ago. Following a two-year
demonstration of the first units we are able to present now special adapted shelters as cold rooms or depots
for medicines, vaccines and blood conserve.

Picture: ILK Container system for cold storage of perishable goods and foodstuffs
basic unit: 20 ft container
PV generator: 3.4 kWp
cooling power: 5.1 kW (-5C/45C)
cold storage room temperature: 0C to 10C (fan controlled)
3
cold storage room capacity: 23 m
special features: cold storage for 3 days, redundant design of cooling system and energy supply
According to recent communication with ILK, status of the ILK project is that the commercial production
has not yet been realized, only a number of demonstration units.
Contact information:
Jrg Waschull, PhD
Department of Applied New Technologies

Final report December 2010

57

Institut fr Luft- und Kltetechnik gemeinntzige Gesellschaft mbH

Bertolt-Brecht-Allee 20, 01309 Dresden
Phone: ++49 351 40 81 771
Fax: ++49 351 40 81 755
E-Mail [email protected]

Promethean Power Systems, Massachusets, USA

http://www.coolectrica.com/
The product is under development, and consist basically of a PV powered Peltier element supplied by a
generator.
Citation from web site:
Products Off-grid solar refrigerator: our product is an energy-efficient hybrid solar refrigerator system for
food distribution networks that operate in off-grid or partially electrified areas. The system uses clean, quiet,
solid-state technologies such as thermoelectric cooling modules and solar photovoltaic (PV) panels. A small
generator is used during extended periods of cloudy conditions to backup the solar panels. By comparison,
conventional refrigeration systems are not designed for efficient, off-grid operation. They use diesel-thirsty
power generators and freon-based refrigeration compressors that require constant maintenance.
Lowest operating & maintenance cost: weve designed this product to achieve the lowest operating &
maintenance cost. Our system takes advantage of increasing efficiencies and decreasing costs in solar and
thermoelectric refrigeration technologies to reduce consumption of fossil fuels. In addition, the system
includes a remote monitoring unit, which enables remote diagnosis that further reduces maintenance cost.
As a result, our system can provide cooling power at an operating cost that is 66% lower than the operating
cost of conventional units.
Complete, stand-alone rural refrigeration system: our vision is to develop a complete, stand-alone rural
refrigeration system that stimulates businesses, reduces dependency on fossil fuels and increases the quality
of life in emerging markets by enabling its users to reliably store food, vaccines and other perishable items.

Final report December 2010

58

Titre du document / Document title

A refrigeration facility for milk cooling powered by photovoltaic solar energy
Auteur(s) / Author(s)
DE BLAS M. (1) ; APPELBAUM J. (2) ; TORRES J. L. (1) ; GARCIA A. (1) ; PRIETO E. (1) ;
ILLANES R. (3) ;
Affiliation(s) du ou des auteurs / Author(s) Affiliation(s)
(1) Department of Projects and Rural Engineering, Edificio Los Olivos, Campus Arrosada, Public
University of Navarra, 31006 Pamplona, Navarra, ESPAGNE
(2) Department of Electrical Engineering Systems, Faculty of Engineering, Tel Aviv 69978,
ISRAEL
(3) Technical School of Forestry, Polytechnic University of Madrid, ESPAGNE
Rsum / Abstract
The use of renewable energy sources is usually a reliable alternative in rural areas and developing
countries, where the grid line does not exist or is at a great distance. In this work, the characteristics
and working conditions of a refrigeration facility designed for cooling down an expected daily
production of 1501 of milk are analyzed. The facility is a stand-alone, direct-coupled system where
20 photovoltaic modules, 120 Wp each, power two permanent magnet, direct current motors of
24V, 650 W. Each motor drives a separate cooling system compressor, which provides the
flexibility to operate the equipment with one or two motors and with various interconnections of the
PV modules, depending on the available irradiance level and the thermodynamic state of the
system. The photovoltaic energy obtained during daylight hours is stored in the form of sensible
and latent heat of frozen water in a tank surrounding a milk container. Thermodynamic analysis of
the system shows that the autonomy of the system is 2.5 consecutive cloudy days if the available
stored ice energy is 80% of the nominal capacity of the water/ice tank. Results of the refrigeration
efficiency are similar to those obtained by other commercial refrigeration facilities powered by a
photovoltaic array, including batteries.
Revue / Journal Title
Progress in photovoltaics ISSN 1062-7995
Source / Source
2003, vol. 11, no7, pp. 467-479 [13 page(s) (article)] (19 ref.)
Langue / Language
Anglais
Editeur / Publisher
Wiley, London, ROYAUME-UNI (1999) (Revue)

Final report December 2010

59

Examples of large commercial low energy chest freezers

AFS 402-628 liter chest freezer
Description & Features:
Foam filled lid for efficient temperature preservation
Temperature control panel, simple design operation easily
Embossed aluminum inner with good conduction of heat, easy to clean
Strong all-direction wheels as feet, easy to move
With lock and drain hole
External condenser is available(optional)
Certifications: CE
Model(L)
AFS-402
AFS-512
AFS-628
Fridge Capacity(L)
402
512
628
Temperature range
d-18
d-18
d-18
Climate
N/ST
N/ST
N/ST
G.W/N.W(KG)
76/68
82/72
92/68
Unit dimension(mm)
1370x675x700 1660x675x700 1940x645x700
Packing dimension(mm) 1410x750x860 1700x750x860 1990x750x860
Refrigerant
R600a/R134a R600a/R134a R600a/R134a
Whirlpool Chest Freezers

Whirlpool Chest Freezer: Energy Star Qualified

Whirlpools makes a very energy efficient chest freezer, model number EH151FXR (similar models
are EH151FXQ or EH150FXQ). This 14.8-cubic-foot freezer is rated at 354 kWhrs per year. The
average cost to run this freezer for year is $29, according to Energy Star. Though it requires manual
defrosting, it sport a number of other features including an interior light and a temperature alarm. Its
key-eject lock means that the freezer can only be opened when the key is pushed in and turned a
safety feature helpful in homes with small children. Four baskets (two upper, two lower) make it
easier to organize the contents especially on the lower level.
Its available from Amazon for $405.
Final report December 2010

60

Vestfrost

The Vestfrost SOE -215 cabinet is a highly insulated cabinet (100 mm PU-foam) equipped with a
Danfoss BD-compressor for 12V or 24 V. The net volume is 213 litres and the dimensions are
1125*600*850 mm (L*D*H).
Two options are available: refrigerator (0 - + 10 C) and freezer (-10 - -18 C).
The existing coolers are likely equipped with HFC-refrigeration system (R134a), but would
possible be modified to natural refrigerant (R600a).
Elcold XLE
Rated A+ Energy efficient, the XLE range of super low energy chest freezers will help to lower
running costs. Features include temperature warning light, defrost drain and fitted lock. Made by
Elcold.
Features and Benefits

Technical Specifications

100mm Insulation

Temperature Range: -18 / -22C

A+ Energy Rating

Refrigerant: R134A

Low Running Costs

Type of Defrost: Manual

Defrost Drain

Exterior Finish: White

Mains Indicator Light

Interior Finish: Aluminium

Temperature Warning Light

Power Required: 13 Amp Supply

Hammered Aluminium Liner

Energy Consumption at -18C: XLE30: 0.46

kwh / 24 hr XLE41: 0.57 kwh / 24 hr
XLE51: 0.630 kwh / 24 hr.

Lock
Adjustable Temperature
Internal Light

Final report December 2010

61

Annex 4. Presentations, SolarChill B

SolarChill TECHNOLOGY
Solar powered direct drive refrigerators with hydrocarbon refrigerants
MR. PER HENRIK PEDERSEN(a), MR. JNOS MAT(b)
(a)
Danish Technological Institute
Gregersensvej, DK-2630 Taastrup, Denmark
Phone: +45 7220 2513
E-mail: [email protected]
(b)

SolarChill Project Coordinator

5106 Walden Street
Vancouver, B.C., Canada V5W 2V7
Phone: +1-604-327-0943
E-mail: [email protected]
ABSTRACT
Solar powered refrigeration technology (SolarChill) has been developed in an international project
involving Greenpeace International, UNEP, the World Bank, WHO, UNICEF, GTZ, PATH,
industrial partners and Danish Technological Institute. The refrigerators can operate directly on
solar PV panels, without batteries or additional electronics and therefore they are suited for
locations where maintenance and reliable operation are in short supply. The main objective of the
SolarChill project is to help deliver vaccines and to cool and store food and beverages in rural areas
without grid power. SolarChill does not use any fluorocarbons in its cooling system or in the
insulation.
SolarChill-A is a vaccine cooler with ice storage that keeps the vaccine cold for about 5 days
without power. SolarChill-A units were used in field tests with good results and the cooler is now
produced by one supplier. SolarChill-B is for domestic and small business applications, and this
type of solar refrigerator is currently being tested in two different prototype versions. The units are
suitable for cooling and storing food and beverages in areas where grid power is non-existent or
unstable. The market potential for this type is enormous as almost 2 billion people live in areas
without grid power.
SolarChill is made from mass produced cabinets and other standard components, which results in a
favourable cost compared to other solar refrigerators.
SolarChill-A (as well as B) has a DC compressor with R600a and a special built-in controller for
solar power, including adaptive speed control. A compressor manufacturer has developed the
compressor for the SolarChill partnership.
1. INTRODUCTION
Refrigeration of vaccines and food is problematic in parts of the world where there is no electricity
or where the electricity supply is unreliable. In those regions, vaccines are maintained by either
absorption or battery-based solar refrigeration. Absorption refrigerators are also used for food
preservation. There are concerns related to absorption refrigeration as well as to the existing
generation of solar-vaccine coolers because of the constant need for fuel and new batteries.
Final report December 2010

62

The need for environmentally friendly and affordable solar vaccine coolers and refrigerators was
realized in 1998-2000 through separate discussions between the United Nations Environment
Programme (UNEP), World Health Organization (WHO) and Greenpeace International (GPI).
At about the same time, Danish Technological Institute (DTI) independently began the development
of a new solar refrigerator that bypassed the use of batteries (funded by the Danish Energy Agency).
DTI worked together with the Danish refrigerator manufacturer Vestfrost. The direct current
hydrocarbon compressor was developed by the Danfoss Company.
The first meeting of the SolarChill Project Partners was hosted by GTZ Proklima in Eschborn, Germany, on 5 May 2001. With an initial decision to proceed with the project, Greenpeace
International provided the funds for the development of the first SolarChill prototypes. These were
exhibited at the World Summit on Sustainable Development in the fall of 2002 in Johannesburg,
South Africa.
A second generation of SolarChill Vaccine Cooler prototype was field tested at the beginning of
2004 in Senegal, Indonesia and Cuba. 10 prototypes of the chest vaccine cooler were tested under a
variety of climatic conditions, 3 units in each of the countries mentioned, and 1 unit at the DTI
laboratory in Denmark. The field tests were coordinated by DTI, and overseen in Senegal and
Indonesia by PATH, and in Cuba by GTZ. The governments and Ministries of Health of the host
countries were active participants in the field tests.
Following the experience from the field tests and the implementation of new standards from WHO
(PQS, Performance Quality Safety) specifying required vaccine storage temperatures, a third
generation of Solar Vaccine Cooler prototype was tested and the unit was marketed. Prototypes of
SolarChill B for domestic and small commercial businesses were developed and tested with help
from GTZ and the Danish Energy Agency. In 2009, the World Bank joined the SolarChill
partnership. SolarChill technology is free of charge and accessible to everyone. With substantial
funding from the Global Environment Facility, SolarChill Project will conduct, with the SolarChillA, extensive demonstration and technology transfer projects in Colombia and Kenya. In addition,
the project plans to complete the development, field testing and commercialization of the
SolarChill-B refrigerators.
More details about the partnership can be found in Ref. 1.
2. ICE STORAGE VERSUS BATTERIES
It has been a wish from the relief organizations to avoid lead batteries as the main source of energy
to keep the vaccine coolers cold during night-time and during periods with minor solar power.
Previous experience has shown that additional costs are related to the batteries because frequent decharging results in a fast degradation of the batteries. That is one reason why the solar powered
coolers have been more expensive compared to kerosene or LPG-powered absorption refrigerators.
In the SolarChill project, ice batteries have been developed as an alternative source for energy
storage, and the SolarChill refrigerators use ice batteries in different versions. In the following, a
comparison is carried out between the energy storage in a typical lead battery and in ice:
One major supplier of batteries informs the following about its 12 V, 50 Ah batteries:
Dimensions: 0.175*0.190*0.221 = 0.00734 m3

Final report December 2010

63

Weight: 13.6 kg
Energy storage: 2,160,000 J and the specific energy storage is 0.159 MJ/kg or 294 MJ/m3.
If we assume a COP value of 1.49 of the refrigeration system (Danfoss BD35K, - 10 C, 2000 RPM,
CECOMAF-data, Ref. 2) then the specific cooling energy in the battery is 0.23691 MJ/kg or 438
MJ/m3.
The similar figures for ice are 0.333 MJ/kg or 333 MJ/m3.
The conclusion is that the cooling capacity for ice storage is of the same order of magnitude as for a
lead battery on volumetric and mass basis. The cooling capacity is app. 40% higher (for the ice
storage) on weight basis and app. 30% less on volume basis. In this comparison, a 100% discharge
of the battery is assumed, which should not be carried out in reality as that would harm the battery.
Another point that disfavors lead batteries is the risk of lead pollution at the end of the lifetime of
the battery.
Future commercialization of batteries with high energy density for electric vehicles might change
the role of battery storage in solar refrigeration. But so far, the project partners believe that ice
storage in SolarChill is the preferred solution. A disadvantage when using ice storage is that the ice
has to be stored inside the insulation in the cabinet and therefore part of the volume inside the
refrigerator is used.
3. CONCEPT FOR SOLARCHILL
The philosophy behind SolarChill is that the coolers must be as cheap as possible and affordable for
people living in areas without grid electricity.
The SolarChill coolers are based on existing well-insulated cabinets, which are mass produced. The
SolarChill-A vaccine cooler has a small chest cabinet with 100 mm polyurethane insulation (blown
with cyclopentane gas), the SolarChill-B-chest-type is based on an ice-liner refrigerator produced
for other purposes and also with 100 mm insulation and the SolarChill-B-upright prototype is based
on a small well-insulated upright household freezer cabinet with 80 mm insulation.
It is important for the SolarChill partnership to use natural refrigerants and a compressor
manufacturer entered the project as industrial partner and developed a DC compressor for isobutane
refrigerant (R600a). The displacement is 3 cm3. The compressor manufacturer also developed a new
integrated electronic control for the compressor, which ensures that the photovoltaic panels can be
connected directly to the compressor without external control. The electronic control also ensures a
soft start which is important when no battery is used.
The electronic control is equipped with an adaptive speed control (Adaptive Energy optimizer
AEO). By using that control, the compressor will stepwise speed up from low speed to maximum
speed in 12.5 RPM/min. If the photovoltaic panels cannot provide sufficient power, the compressor
will stop and after a short while it will try to start again. The compressor will try to start every
minute and once the power from the panels is sufficient the compressor will start at lower speed.
The first start in the morning is at app. 2500 RPM. After a compressor stop the compressor will
start up at the latest speed minus 400 RPM. The speed range is from 2000 to 3500 RPM.
The controller accepts a voltage between 10 and 45 V. The voltage from photovoltaic panels can
vary and that is a good feature for solar powered refrigerators and freezers. When using 12 V

Final report December 2010

64

modules, the compressor starting current is less than 3 A. The compressor runs continuously at
about 3 A at low speed (see figure 1). Using normal electronic control the start current would be
much higher, requiring much bigger PV-panels or require the use of a capacitor to help start the
compressor.

Figure 1: Starting current using the solar electronic control, 12 V.

The expansion device is a capillary tube with heat exchange to the suction line. In the chest type
cabinets integrated skin condensers are used as in most chest freezers.
The evaporator in the SolarChill-A vaccine cooler is a wire-on-tube-type placed in the ice storage as
shown in figure 2. The evaporator in SolarChill-B chest-type is an integrated skin-type as in most
chest freezers. The evaporator in SolarChill-B-upright-type is a box-type roll bond-aluminum
evaporator as known from old refrigerators with a small freezer compartment. The refrigerant
charge in the SolarChill-A vaccine cooler is 48 grams of R600a and the charge in the SolarChill-Bchest-type is 60 grams. The charge in SolarChill-B-upright is 48 gram.

1 Cabinet with 100 mm insulation

3 Skin condenser
5 Internal wall, insulated
7 Evaporator, wire on tube

2 Vaccine compartment
4 Lid
6 Ice storage (26 ice packs, 600 ml)
8 Compressor

Figure 2a: Figure with basic principles for the SolarChill-A vaccine cooler. Figure 2b: Photo of the solar
DC compressor. The integrated solar electronic control is placed at the left-hand side of the compressor.

Final report December 2010

65

4. SOLARCHILL-A VACCINE COOLER

The basic principle of the SolarChill-A vaccine cooler is shown in figure 2a. The evaporator is
placed in the ice storage in the right-hand side of the figure, and natural and forced convection
ensures the temperature of the vaccine stored in baskets as shown in the left-hand side of the figure.
Pedersen and Mat (2007) explain how the field test of this vaccine cooler took place in Indonesia,
Senegal and Cuba using 180 W photovoltaric panels (3*60 W peak) and tests show the hold-over
time of about 5 days without any energy available. The field test lasted about one year and the ice
banks were never totally melted except when the connections to the PV-panels were disconnected
intentionally.
WHO has now developed specifications for battery free vaccine coolers, and tests are ongoing to
validate that the SolarChill-A can fulfill the specifications. About 200 SolarChill-A vaccine coolers
have been manufactured by Vestfrost A/S and installed in many countries, and an even greater
implementation rate is expected with approval by the new WHO specifications and the involvement
of the World Bank. The present price is in the range of app. 1000 Euro for the cabinet and a little
less for the photovoltaic panels depending on the size of the panels for the specific location. The
price for both cabinet and panels is expected to decrease in the future.
5. SOLARCHILL-B UPRIGHT TYPE
SolarChill-B is a refrigerator for domestic and small commercial use. The purpose of SolarChill-B
is to help people cool and store food and drinks in small scale. Almost 2 billion people live in areas
without grid electricity and the potential need for the product is enormous. The problem is that only
part of the potential users at present can afford to buy a cooler.
The first customers to buy SolarChill-B are expected to be the more wealthy people in areas without
grid electricity and small businesses that can profit from selling cooled products. When the coolers
come into mass production the cost will decrease and more people will be able to buy them.
Danish Technological Institute built a prototype upright SolarChill-B cooler in 2004 and after
laboratory tests the cooler was placed at DTI and has been powered by 3*60 W PV-panels. Except
for in mid-winter (from November to February) at this high latitude (56 degrees north) the coolers
have been working well since 2005. The net volume of the cooler is about 100 litres.

Figure 3a: SolarChill-B upright prototype built in a well-insulated freezer cabinet with 80 mm PUinsulation. Figure 3b: The box-type roll-bond evaporator in the cooler.

Final report December 2010

66

40C

1200

30C

1000

20C

800

10C

600

0C

400

-10C

200

-20C

Voltage / [V] - Power / [W]

Temperature

"Solar Chill B.1.b": Tests @ 32C

Power Supply (230 V ac): 8 hours ON - 16 hours OFF - etc.

t_8 - ice bank

t_12 - ambient
t_13 - pack.top
t_14 - pack.mid
t_15 - pack.bot.
Volt - [V]
Watt

#. kWh - R.T.
-------------1. 0.31 - 70%
2. 0.33 - 74%
3. 0.32 - 72%
4. 0.33 - 75%

0
17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Date in September 2004

Test # 040916 + 040921

Figure 4: Lab test of SolarChill upright at 32C ambient temperature. Power is available 8 hours a day, but
the compressor runs less (about 6 hours a day). The temperature inside the compartment is between 2.5 and
7.5C which is good for food storage. Hold-over time is 3.1 days (75 h) defined until the temperature rises to
+10C. The compartment contained 10 kg of test packages.

The ice storage is placed inside the box evaporator and natural convection ensures the temperature
inside the cooler compartment. The shelves have to be open grill type to ensure convection. The
compressor is of the same type as for the SolarChill-A vaccine cooler. So far, no manufacturer has
been found for this type of SolarChill, but discussions are taking place with a potential
manufacturer.
6. SOLARCHILL-B CHEST-TYPE
The development and laboratory test of a chest-type SolarChill-B took place at Danish
Technological Institute in 2009. The cooler is based on an existing cabinet used for other purposes
(ice liners).

Figure 5: SolarChill-B chest-type prototype based on a 160 liter ice liner cabinet. The cooler is equipped
with 5 baskets (3 upper baskets and 2 lower baskets). The photo shows that the cooler is tested with 12.5 kg
test packages and 13.3 kg soft drinks (40 cans) simulating food and drinks.

Final report December 2010

67

The cabinet has been installed with the solar DC compressor of same type as in SolarChill-A. An
ice bank of about 17.5 kg is placed in the wall between the evaporator and the interior cabinet. The
cooler is controlled by a mechanical thermostat and tests were conducted in the laboratory at DTI.

Figure 6: Test at ambient temperature of 30C. The compressor runs about 6h 40 minutes per day and the
temperature of the test packages is between -1.5 and 2.5C. The air temperature is between -2 and +3 0C
which is relatively cold for a refrigerator. The hold-over time is 2.7 days (up to +7C) and 3 days (up to +
10C).

Figure 7: Half reload test. 20 warm soft drink cans were inserted inside SolarChill-B chest-type when
taking out 20 cold cans. That was done at sunset four days in a row. Power is available 10 hours a day,
temperature is measured in one can in an upper basket and one can in a lower basket and the test shows the
cooling capacity is sufficient to cool 20 cans a day. However, the cooling capacity is limited for this type of
cooler.

Final report December 2010

68

7. DISCUSSION ON FUTURE TYPES OF SOLARCHILLS

The next step for the SolarChill partnership is the commercialization of the existing coolers
mentioned above. However, in future it will be possible to develop other types.
If greater cooling capacity is needed, a compressor with higher cooling capacity is needed and so is
a larger photovoltaic area. It will be possible in the future to develop coolers for cooling a greater
number of drinks and food (e.g. 50 soft drinks a day). The technology and the components are more
or less available today.
Danish Technological Institute has been involved in discussions concerning the development of a
milk cooler for farms with only a small number of cows. Milk has to be cooled down to about +4C
in less than 2 hours and that can take place quickly by building up ice storage and thus having
sufficient cooling capacity for fast cooling of the milk. The existing prototype of SolarChill-B
chest-type might be used by inserting the milk in small containers, but a more efficient cooler
would probably need another design.
Finally, it should be mentioned that freezer type SolarChills would be useful for freezing and
conservation of fish, meat and vegetables. A freezer type would need eutectic ice storage with a
melting point around -20 0C, which could be a saturated solution of salt and water.
ACKNOWLEDGEMENTS
The authors would like to thank Greenpeace International, GTZ and the Danish Energy Agency for
financial support during the process.
The authors also would like to thank the Danfoss Company and the Vestfrost Company for
providing components, cabinets and advice during the project.
Finally, the authors would like to thank the project partners from UNEP, the World Bank, GTZ,
PATH, WHO, UNICEF and Greenpeace International for good discussions and support during the
project.
REFERENCES
Ref 1: http://www.solarchill.org/
Ref. 2: Compressor data:
http://www.danfoss.com/BusinessAreas/RefrigerationAndAirConditioning/DanfossCompressors/
Ref. 3: Pedersen P. H. , Mat J.; 2007, SolarChill Vaccine Cooler and Refrigerator: A
Breakthrough Technology, Industria Informatione, special international issue on refrigeration and
air conditioning, UNEP, ATF and Centro Studi Galileo.

Final report December 2010

69

Annex 4. Milk Cooling

Teknologisk Institut
4. februar 2009/Per Henrik Pedersen

Solcelledrevet mlkekler bygget ind i isoleret kumme

I samarbejde med Karise Kleinsmedie har Teknologisk Institut fet den id, at benytte et kabinet fra
en kummefryser til at lave en billig mlkekler til mindre kvgbestninger i visse u-lande.
Karise Kleinsmedie (KK) har i samarbejde med Danida og lokale samarbejdspartnere en
virksomhed i Uganda, som leverer rustfrie stl-produkter til mejeriindustrien.
Derfor er Teknologisk Institut i samarbejde med ovennvnte virksomhed i gang med at lbe et nyt
projekt i gang, og vi hber at Vestfrost kan have interesse i at deltage.
Groft sagt gr princippet ud p, at nedsnke et rustfrit stlkar i en kummefryser. Karret er mindre
end inderkabinettet, sledes at der er plads til en vandreservoir (is/vand-blanding) i hulrummet
imellem stlkar og inderkabinettet. Karret skal monteres, sledes at det sidder fast og ikke flyder
ovenp. Der skal endvidere vre aflb i bunden med et to tommer rustfrist rr (51 mm ydre
dimension).
Samarbejdet skal g p, at KK laver de rustfrie emner (+ omrrer) og Vestfrost (i samarbejde med
Teknologisk Institut) laver prototype af kabinet, kleanlg og styring. Vi tager udgangspunkt i
eksisterende kabinet.
Den frste prototype testes p Teknologisk Institut i Taastrup.
Karise Kleinsmedie satser p i frste omgang at lave en prototype, som kan kle 150 liter mlk fra
29 C til 4 C i lbet af max. 2 timer. Det er meningen at den skal kunne kre med energi fra
solceller, og gerne opbygge et islager i lbet af dagen.
Sekundrt kan man forestille sig et 300 liter anlg til netdrift (eventuelt ustabil netdrift).
P de nste sider er tre grove skitser fra KK, og disse giver ide om, hvordan det mske kan
principopbygges.

Final report December 2010

70

Kabinettet kunne mske vre Vestfrost Function SW284C med bruttovolumen p 296 liter eller
mske det professionelle kabinet HF396 med et bruttovolumen p 373 liter ?
Final report December 2010

71

Nedklingsprocessen
Her kommer en beregning af nedklingsprocessen m.m.:
Klekapacitet i nedklingsfasen p max 2 timer:
Der skal nedkles 150 liter mlk fra 29 C til 4 C. Vi regner p vand:
Qcm = (M*DeltaT*Cp)/Delta TID = (150 kg*25K*4186(J/kgK))/(2*3600s) = 2180 W
Det svarer til smeltning af et ismngde p:
Mis = (QCM*DeltaTID)/smeltevarme = (2180W*2*3600s)/335000(J/kg) = 46,85 kg is
Af to grunde vil det vre godt, hvis fordamperrrene ligger et stykke nede i kummen:
Det gr det muligt, at stte sttteskruer p karret (jf. tegning)
Det vil vre godt, hvis isopbygning sker lidt nede i vandreservoiret. Det vil medvirke til god
cirkulation (konvektion) samtidig med, at der er plads til is, som mtte lsrive sig og flyde
overp.
Der vil i vrigt vre luft foroven i reservoiret, og der skal kunne ske passage af luft ind og
ud af reservoiret evt. igennem et trykudligningshul p bagsiden.
Nedklingsprocessen for system A: Kling med direkte solkraft
Ved krsel med kleanlg i 7 timer kan dette genereres med en kleeffekt p:
Qcoolcapacity,SC = 2/7 * 2180 W = 623 W.
Dette kan klares med flgende kompressortyper (propan R290):
Antagelse: to = - 10 C, tc = + 45 C (mske for lavt?)
SC10CNX (lige akkurat !)
SC12CNX
SLV15CNX (variable speed, - kan mske neddrosles til 12 eller 10 cm3)
Det vil nok vre mest naturligt og passende at vlge SC12CNX i frste omgang.
Nedklingsprocessen for system B: Kling med batteridrift
Ved krsel i 20 timer kan man klare det med 2/20 * 2180 W = 280 W kleeffekt.
Det kan klares med en noget mindre og i klembler mere almindelig kompressor.
Det vil mske vre oplagt at benytte NLE15KTK, som Vestfrost i forvejen benytter i flaskeklere
(M200 R600a) og impulsklere (POS72 R600a).
Der findes ogs en Danfoss jvnstrmskompressor, som kan klare det. Det er en BD350GH, som
dog er til R134a.
Hvis det skal vre med naturlige klemidler kan en twin-lsning med BD100CN tnkes.
Konklusion: I frste omgang benyttes en NLE15KTK i den frste prototype. Kapillarrr og intern
varmeveksler fra flaskekler kan mske benyttes.
Varmeovergangsforhold p vand- og mlkesiden
Final report December 2010

72

Det antages, at der er omrring i mlketanken i nedklingsperioden. Herved fr et meget stort

varmeovergangstal i strrelsesordenen 1000 W/m2K.
P vandsiden er der opbygget is langs fordamperrr i svbet, og de er mske ogs is i toppen af
vandbeholderen, mens der er vand lngere ude mod mlketanken.
Vandtemperaturen vil vre omkring 0 C i den vre del af vandtanken, mens det vil vre ca. + 4 C
helt i bunden. Der vil i nedklingsprocessen vre naturlig cirkulation p vandsiden, og
varmeovergangstallet vurderes til at vre ca. 100 W/m2K (eller mere) i det meste af
nedklingsprocessen, men nr mlketemperaturen reduceres til omkring 4 5 C, vil den blive
mindre idet den naturlige cirkulation af vandet gr i st ved + 4 C (hvor vandet er tungest). Der vil
dog stadig vare varmeledning over et stort overfladeareal.
Det vurders at det varmeoverfrende areal er mindst 1,5 m2.
Med 1,5 m2 vil den varmeoverfrende effekt vre ca. 3750 W i starten af nedklingsperioden, ca.
1875 W, nr halvdelen af energien er overfrt og noget mindre i slutningen af perioden. Hvor meget
mindre ved jeg ikke lige nu, men i det helt ekstreme tilflde:
1,5 m2, stillestende vand, deltaT=4 K, 2 cm tyk lag vand:
Q = Lambda/delta*deltaT*A = 0,57 W/mK/0,02m*4K*1,5m2 = 171 W
Dette ekstreme tilflde ns aldrig, idet der altid vil vre omkring 0 C i reservoiret foroven og
dermed konvektion p vandsiden i den vre del af mlketanken.
Hvis det i praksis viser sig at blive et problem at f afklet de sidste par grader p grund at, at vand
har maksimal densitet ved + 4 C, s kan man komme sprit (eller andet) i vandet.
Lad os se, hvad en nedklingstest med 150 liter vand vil vise.
Styring
Hvordan skal kleanlgget styres ?
Her er et bud:
Batteridrift:
Hvis vi vlger batteridrift skal kleanlgget kre indtil ismngden er tilpas stor / istykkelsen er
tilpas stor. Herefter skal kleanlgget slukkes for ikke at nedfryse hele vandreservoiret og mlken.
Hvordan kan vi gre det med en almindelig termostat:
Hvis der benyttes en almindelig termostatlomme (som i kummefrysere) kan denne mske benyttes
sammen med en termostat, som afbryder ved f-eks. 15 C ?
Nr der er en tilpas islag i reservoiret vil fordampningstemperaturen krybe nedad og
termostatfleren vil blive koldere.
Direkte soldrift:
Som udgangspunkt skal kleanlgget kre, nr der er kraft (det m Ivans strmforsyning (solceller
+ lille batteri + styring hertil klare).
Nr ismngden er tilstrkkelig stor, skal kleanlgget stoppe. Det kan mske gres som vist
ovenfor under batteridrift.
Indpasning af mlketank:
Kan KK placere et 2 stlrr igennem bunden ?
Kan de ligeledes bore sttte-stnger igennem bunden?
Final report December 2010

73

Er kummen vandtt (udover KKs gennemboringer)?

Hvor hjt sidder de verste fordamperrr. Kan KK montere sttte-skruer foroven (som vist
p tegningen)?
KK skal have tegning for den kumme, som vi vlger

Afgivelse af varme:
Normalt afgives varmen igennem svbskondensatoren i en fryseboks. Vi har her med strre
varmeeffekter at gre; - isr hvis der satses mod af benytte SC12CN-kompressoren og brug af
direkte solkraft.
Batteridrift:
Hvis vi benytter NL15KTK.2 vil den afgivne effekt vre ca. 600 W.
Det kan mske klares med svbskondensator + overhedningsfjerner m. blser ?
Det svarer lidt til en stor iscremefryser.
Direkte solkraft:
Her har vi med strre effekter at gre. Her skal vi af med 1,3 kW, og det er mske ndvendig med
ekstern blokkondensator m. blser.

Final report December 2010

74

Markedspotentiale, notat udarbejdet af Karise Kleinsmedie, Palle Maag:

Mlkekler 150 l vurdering af afstningspotentiale i Uganda.
Karise Kleinsmedie Aps. (KK) har taget initiativ til udvikling af et mlkekleanlg drevet af
solenergi med henblik p afstning i Uganda og nabolande.
I samarbejde med Danmarks Tekniske Institut (DTI) vurderes det, at et anlg i strrelsesordenen
150 l vil vre konomisk og teknologisk forsvarligt at producere.
KK vurderer afstningsmulighederne for et 150 l anlg sledes :
1. I oktober 2008 deltog Karise Kleinsmedie Uganda Ltd. (KKU) i den rlige messe, Kampala
International Trade Fair .
KKU havde p messen en strre stand med udstilling af et udvalg af forskellige produkter
fremstillet i rustfrit str, herunder flere forskellige typer mlketanke.
Interessen for ikke mindst mlketankene var overvldende, men et stadigt tilbagevendende
sprgsml drejede sig om kling kunne KKU ikke levere tanke med kleanlg ?
2. Der findes i Kampalas omegn og i det sydvestlige Uganda store landbrugsomrder, hvor
mange isr mindre farmere holder malkekvg, der producerer mlk, som finder afstning i
lokalomrderne og leveres til strre mejerier i Kampala og Mbarara. Inden mlken kan
afhentes, er det ndvendigt at f den klet ned, og da der i stor udstrkning er tale om
smfarmere, er der hovedsageligt behov for mindre kleanlg, som oven i kbet vil kunne
betjene et fllesskab af flere farmere.
3. Dairy Development Aythority (DDA), der har den overordnede fling med og indsigt i
produktion og forarbejdning af mlkeprodukter i Uganda, er ikke i tvivl om behovet for
mindre kleanlg til smfarmere i Uganda, isr anlg, der kan drives af solenergi, hvilket
vil gre anlggene anvendelige ogs i omrder uden offentlig strmforsyning (der i vrigt
ofte er ustabil) og overfldiggre brugen af generatorer, der ger omkostningerne ved kling
af mlken bde hvad angr investering og drift.
4. Mht. afstning af mindre kleanlg i regionen, skal man holde sig for je, at det ikke blot
glder Uganda, men ogs omliggende lande, dvs. Rwanda, Kenya, Congo og Sudan, der i
vrigt alle havde besgende gster p Kampala Trade Fair.
Endelig m det formodes, at et mlkekleanlg drevet at solenergi og fremstillet til en
konkurrencedygtig pris, vil kunne afstte i mange andre udviklingslande.
Den afsluttende konklusion er, at vi ser et absolut lovende afstningspotetiale for en mindre
mlkekler drevet af solenergi og fremstillet til en pris, der kan matche traditionelle anlg
drevet at el.
Det er meget svrt at stte et prcist antal p, hvor mange vi vil kunne afstte , men vi
forventer inden for det frste r at kunne afstte mindst 100 stk. Herefter tror vi at de gode
erfaringer fra anlggene vil sprede sig blandt farmerne og dermed f eftersprgslen til at
accelerere.
Karise Kleinsmedie Aps. 16.02.2009

Final report December 2010

75