SciELO - Scientific Electronic Library Online

Home Pagealphabetic serial listing  

Services on Demand




Related links


Brazilian Journal of Chemical Engineering

Print version ISSN 0104-6632On-line version ISSN 1678-4383

Braz. J. Chem. Eng. vol. 14 no. 3 São Paulo Sept. 1997 



M.E. Vieira and H.B.C. Moreira

Chemical Engineering Department - Applied Solar Energy Laboratory, Center of Technology
Federal University of Ceara - 60.450-760 Fortaleza-CE, Brazil -
Telephone: (085) 288-9599, Fax: (085) 288-9601


(Received: March 5, 1997; Accepted: August 5, 1997)


Abstract : This work presents the principles of operation of a solar refrigerator with the following basic components: a reactor, a set of evacuated tube solar collectors, a condenser, a heat exchanger, and an evaporator. During the heating phase, solar radiation is collected and transferred to the reactor for desorption by a vapor thermal siphon loop. During the cooling phase, heat from the reactor is released to the ambient by a second water vapor loop. Ambient data collected daily during a period of 18 years were divided into hourly values and used to simulate the temperatures of the reactor, which uses salt impregnated with graphite and ammonia, during the adsorption / desorption processes. The results show that the refrigerator operates well in Fortaleza and that better results are expected for the countryside of the state of Ceara. It is concluded that only a high efficiency collector set can be used in the system.
Keywords: Solar Refrigeration, adsorption.




Solar refrigeration units can significantly improve the living conditions in regions with no or an insufficient conventional energy supply. The use of an adsorption reactor and evacuated tube collectors in a daily thermal cycle is a new way to make solar energy utilization attractive. The system uses a form of energy which is nonaggressive to the environment in contrast with the traditional refrigeration units which employ the vapor-compression cycle with a refrigerant, particularly of the chlorofluoromethane family, as the working fluid. These conventional cooling units transform electrical power, which is two or three times more expensive, into thermal power.

This cooling unit transforms solar radiation into a cooling effect in an evaporator. There are two phases in the daily cycle. During the daylight hours, evacuated tube collectors generate heat that is used to desorb ammonia, which is then liquefied and stored. During the night-time hours, cold water and ice are produced when ammonia evaporates and is adsorbed in the reactor as the energy released is transferred to the ambient. The equipment was designed to require no routine maintenance as the system has no moving parts, except for the control valve which is driven by heat. When the cold storage is loaded to 50 percent, the cooling temperature can be held constant for about three to four days, even if the insolation is negligible during this period. Its applications are the cooling of medicines and food or cold drinking water production.

The purpose of this work is to study the thermal performance of the solar refrigerator under the ambient conditions in Fortaleza by numerically simulating the heat and mass transfer processes of its components to estimate the generation, condensation, and evaporation temperatures and the cooling effect produced by the unit on typical days. The values of environmental variables such as intensity of solar radiation, ambient temperature, and wind speed were obtained from time series recorded daily for 17 years. Hourly values were estimated from daily values using the Clear Model presented by Hottel (1976) and the Erbs et. al. (1982) correlation for the hourly clearness index. The thermal properties and correlations used in the model were presented by the manufacturer, as explained in the next sections. As seen from the simulation results, the solar refrigerator operates well under the conditions found in Fortaleza and can represent an economically attractive alternative to refrigerating in remote areas.



Solar cooling cycles have been studied because of two different but still related topics, refrigeration for food preservation and comfort cooling. The units use continuous or intermittent coolers and flat-plate collectors and usually employ absorption cycles.

Most commonly, continuous absorption cycles, whose basic principles are well described in the literature [ASHRAE (1986)] are adapted to operate from solar flat-plate collectors. Because of the temperature limitations of the flat-plate collectors, commercial machines are limited to lithium bromide-water systems as operation with ammonia-water coolers is difficult due to the high temperatures needed.

Intermittent systems are associated with coolers used for food preservation. Similar to what occurs in the adsorption cycles, distillation of the refrigerant from the absorbent occurs during the regeneration stage of operation, and the refrigerant is condensed and stored. During the cooling phase, the refrigerant is evaporated and reabsorbed. Sargent and Beckman (1968) presented a theoretical study of an ammonia-sodium thiocyanate cycle (NH3-NaSCN) with ammonia as the refrigerant, using the physical chemical properties of solutions of sodium thiocyanate in ammonia presented by Blytas and Daniels (1962). The cycle appeared to have good thermodynamic properties for ice manufacturing.

Refrigeration systems using evacuated tube collectors and an adsorption reactor represent a new technology in solar cooling systems. The system operates in an intermittent cycle and experimental results for its performance in Germany were presented by the Dornier company (1993), which developed six prototypes in order to test them under different conditions. The test results indicated that ammonia production started after the reactor was preheated to about 1 kWh/m2. When the integral insolation reached 5.5 kWh/m2, the generation phase was finished and the amount of ammonia produced corresponded to a cooling energy of 1.8 kWh, or approximately 20 kg of ice. The total efficiency, defined as the cooling energy produced in relation to solar energy, up to that point, was nearly 15 %. The theoretical generation temperature under these conditions varied from 75 to 90 oC. At an ambient temperature of about 20 oC, the condensation temperature varied from 25 to 30 oC. When the reverse process started, the ammonia temperature cooled down to less than 0oC in the evaporator, producing ice and maintaining the evaporator box at a temperature between 1 and 3 oC for the whole day. The solar refrigerator was designed for ice production and cold storage at night-time ambient temperatures of 30 oC. For Fortaleza, this is a rather conservative value, particularly if radiation losses to the sky are considered.



This section describes the solar refrigeration unit and presents its fundamental working principles and the analytical procedures used to simulate its performance.

The basic components of the solar refrigerator are a reactor filled with a special salt which adsorbs and desorbs ammonia, a set of thirteen evacuated tube solar collectors, a condenser, a heat exchanger, an evaporator, and an accumulation tank. It operates in an intermittent daily cycle with a heating phase and a cooling phase. During the heating phase, solar radiation is collected and transferred to the water vapor which circulates in a thermal siphon loop. The vapor moves upwards to the reactor and transfers its collected heat into the reactor. Ammonia vapor is then desorbed, condensed, and stored in the accumulation tank. A diagram is shown in Figure 1.

In the cooling phase, the reactor releases heat to the water vapor that circulates in a second thermal siphon loop. The water vapor transfers this heat to the ambient in a single-phase heat exchanger located above the reactor. There is a control valve which changes the direction of the water vapor flow between phases. Ammonia vapor is then adsorbed in the reactor after being evaporated in the cooling box producing ice. A diagram is shown in Figure 2.

To study the performance of the solar refrigerator under ambient conditions in Fortaleza, the values of the meteorological variables such as ambient temperature, intensity of solar radiation, wind velocity, etc. are needed. Daily data gathered by the Weather Station located at the Center of Agricultural Science at the Federal University of Ceara during the period from July 1972 to June 1989 were statistically averaged for daily and monthly values. Later data were collected but are not yet available in print. The average values of insolation for the month of October were distributed hourly using Hottel’s (1976) Clear Day Model and the Erbs et. al. (1982) correlation for the hourly clearness index. A characteristic of the level of insolation in Fortaleza is its uniformity throughout the year, such that the operation of the solar refrigerator during other months is not expected to vary much. A computer program was developed for these calculations.


Figure 1: Diagram of the components and operation during the heating phase of the solar refrigerator.


Figure 2: Diagram of the components and operation during the cooling phase of the solar refrigerator.


Using the estimated hourly values for the ambient temperature, the condensation temperatures were determined, assuming a 7.5 oC temperature difference throughout the condenser. The value chosen was the average of those found in a similar solar refrigerator operated in Germany, as cited in the literature review. Using the estimated values of the saturation temperatures, T, the saturation pressures, P, were estimated using the correlations presented by the manufacturer, equation (1).


The connecting pipe between the condenser and the reactor is 0.7 m in total length, and the pressures in the reactor were assumed to have the same values as those in the condenser. Using the correlation supplied by the manufacturer, the reactor or adsorption/desorption temperatures were determined using equation (2).


In order to have an estimate of the hours in the day when generation began and stopped, the same amounts of accumulated energy per square meter found in Germany were used. This is an acceptable assumption for a rough estimation because the daily values for the ambient conditions used varied from 0 to 950 W/m2 of insolation on a horizontal surface and an ambient temperature of from 10 to 22 oC. In Fortaleza similar values of insolation, but higher values of ambient temperature, are found and the estimation is conservative because there would be less ambient loss during the heating phase.

The average temperatures of the collector throughout the day were calculated considering an average value of 25 oC as the difference between the desorption temperature and the collector temperature. This assumption was based on the temperature difference values found for operation in Germany, which varied from a few degrees just as generation started to approximately forty degrees at the peak hour. The average values remained true for most periods. The values of the temperatures at the collector outlet during the heating phase determine the type and average efficiency needed in solar radiation collection.

The cooling phase starts after the generation phase, as the sun sets. During this adsorption phase, the temperature in the reactor remains approximately constant at 55 oC above ambient temperature.



Figure 3 shows the simulated results for the ambient, condenser, generator, and collector temperatures which would to be found were the solar refrigerator to operate in Fortaleza during the month of October.

Figure 4 shows the simulated hourly values, estimated from the daily values, for insolation in Fortaleza during the month of October. Using the average experimentally measured value of 15 % for the thermal efficiency of the refrigerator, and considering the collector field area of 2.1 m2, the cooling energy expected will be 1.78 kWh, which is equivalent to 19.2 kg of ice. Also, if the same experimental values for the accumulated energy starting and ending the generation process that were found in Germany are used, the starting and ending hours will be 9:15 and 17:00, respectively.


Figure 3: Plot of the simulated values for the ambient, condenser, reactor, and collector temperatures were the solar refrigerator to operate in Fortaleza in October.


Figure 4: Plot of the simulated hourly values of insolation estimated from daily values for Fortaleza during the month of October.



From the simulation results, it is seen that the solar refrigerator will operate well in the ambient conditions found in Fortaleza. In fact, the heating phase can accumulate more energy, that is, produce ammonia, than estimated above because of the availability of solar radiation and the somewhat high ambient temperature during the heating phase. These factors contribute to more solar collection and less loss to the ambient. In regard to the cooling phase, the temperature difference between the reactor and ambient air at night, an average value of 55 oC, seems to be sufficient to assure that all energy gathered in the system can be effectively lost at night.

Higher efficiency values than those estimated for Fortaleza are expected in the operation of the solar refrigerator in the countryside of the state of Ceara, where higher values of insolation and lower ambient night temperatures are found. These will be found useful in the economical analysis of the unit, in addition to the fact that these regions use either imported electrical energy or burn wood.

In the simulations performed, the hourly values of insolation and ambient temperature needed were estimated from daily values, as reported in the data base. Due to a lack of simulation models, ambient temperatures throughout the day were estimated based on insolation distribution. Presently, a data acquisition system for global solar radiation on a horizontal surface, direct insolation, ambient temperature, and wind speed is installed in the Applied Solar Energy Laboratory at the Federal University of Ceara campus. The system employs precision pyranometers, pyrheliometer, and temperature and wind velocity sensors. These data are scanned every 2 seconds and gathered every 5 minutes. As the daily distributions of these variables become available, more accurate simulation results will be obtained.

From the simulated results presented in Figure 3, the collector temperatures throughout the heating phase need be around 130 oC. This result implies that only very high-efficiency collectors, such as those of the evacuated tube type, can be used in the system.



We would like to thank the Department of Alternative Energy of the Energy Company [COELCE] of the State of Ceara for lending us the refrigeration unit to develop this research work.

We also acknowledge the aid received from the KFA, Technische Universität München and Dr. Eng. H. Kleinemeier who worked with the refrigerator at Dornier.



P Pressure, bar

T Temperature, oC



Ashare Handbook, Refrigeration Systems and Applications (1986).

Blytas, G.C. and Daniels, F., Concentrated Solutions of NaSCN in Liquid Ammonia: Solubility, Density, Vapor Pressure, Viscosity, Thermal Conductance, Heat of Solution, and Heat Capacity, Journal of the American Chemical Society, 84, 1075-1083 (1962).

Dornier-Deutsch Aerospace, Solar Cooling - Solid Absorption, the New Technology for Solar Cooling Units (1993).

Erbs, D.G.; Klein, S.A. and Duffie, J.A., Estimation of the Diffuse Radiation Fraction for Hourly, Daily and Monthly-Average Global Radiation, Solar Energy, 28, 293 (1982).

Geankoplis, C.J., Transport Processes and Unit Operations (1993).

Hottel, H.C., A Simple Method for Estimating the Transmittance of Direct Solar Radiation Through Clear Atmospheres, Solar Energy, 18, 129 (1976).

Kulaki, S.; Vattano, P. and Brothers, P., Performance Evaluation of a Solar Heating and Cooling System Consisting of an Evacuated-Tube Heat Pipe Collector and Air Cooled Lithium Bromide Chiller, SAN-30569-32, USA (1984).

Sargent, S.L. and Beckman W.A., Theoretical Performance of an Ammonia-Sodium Thiocyanate Intermittent Absorption Refrigeration Cycle, Solar Energy, 12, 137 (1968).

Treybal, R.E., Mass Transfer Operations, McGraw-Hill (1980).

Ward, D.S., Solar Absorption Cooling Feasibility, Solar Energy, 22, 3, 259-268 (1979).

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License