#### Abstract

The objective of this paper is to investigate theoretically a solar driven 60 kW absorption cooling system. The system is constituted of a combined ejector single-effect absorption cycle coupled with a linear Fresnel solar concentrator and using water/lithium bromide as working fluid. The combined ejector single-effect absorption cycle exhibits high performances, almost equal to that of double-effect absorption device. However, higher driving heat temperatures are required than in the case of conventional single-effect machines. A mathematical model is set up to analyze the optical performance of the linear Fresnel concentrator. Simulations are carried out to study the overall system performance and the performances of the combined absorption machine for generator driving temperatures and pressures in the ranges 180°C – 210°C and 198 kPa – 270 kPa, respectively. Further, the effect of operating parameters such as the cooling medium and chilled water temperatures is investigated. A maximum cycle performance of 1.03 is found for a generator pressure of 272 kPa and chilled and cooling water temperatures of 7°C and 25°C, respectively. A case study is investigated for a typical summer Tunisian day, from 8:00 to 18:00. The effect of ambient temperature and solar radiation on cycle and system performances is simulated. The optical performances of the concentrator are also analyzed. Simulation results show that between 11:00 and 14:00 the collector efficiency is 0.61 and that the reaches values always higher than 0.9 and the is larger than 0.55. Globally the performances of the investigated cycle are similar to those of double-effect conventional absorption system.

#### 1. Introduction

In summer, the demand of electrical energy for air-conditioning and refrigeration becomes important, particularly in hot climate zones like the Mediterranean region. In many countries in this climate zone, the electric energy consumption reaches, and sometimes exceeds, in hot summer days the limit of production capacity. Consequently, the overall energy consumption and costs and the environmental impact become important. To reduce these impacts, solar cooling technologies may constitute a promising and attractive solution.

Solar energy applications and practices are investigated for different domains [1] such as pyrolysis [2], refrigeration and cooling [3], power generation [4], drying and various other industrial uses [5, 6]. In view of the increased interest in high temperature solar applications, many studies focus on the amelioration of collector configurations and concentrator techniques in order to improve their optical efficiency and the heat transfer characteristics of the absorber by using for instance nanofluids as a heat transfer fluids [7].

In the domain of refrigeration and air-conditioning, absorption techniques for production of cold are considered particularly suitable for solar applications. Single-effect absorption systems can be activated by flat plate collectors and/or evacuated tube collectors [8, 9] since they require relatively low temperatures of the driving heat sources. They have however low performances. Therefore, investigators focus attention on double-effect absorption cycles that need higher temperature heat sources provided by parabolic through collectors or linear Fresnel concentrators. Several investigations consider the technologies of double-effect absorption systems coupled with parabolic through collectors. An 11 kW solar absorption cooling system was studied by Florides* et al*. [10] in Nicosia and optimized using TRNSYS. The required parabolic collector area was 15m^{2} and the capacity of hot water storage, 600 L. Shirazi* et al. *[11] investigated the feasibility of producing cooling via the connection of LiBr-H_{2}O single, double, and triple-effect absorption cycles to solar collectors under various weather conditions, namely, 1023 kW single-effect absorption chiller coupled with evacuated tube collectors (ETCs), parabolic trough collectors (PTC) coupled with 1163 kW double-effect and triple-effect units. The authors concluded that the energetic and economic performance of multieffect absorption machines coupled with concentrating collectors in low direct normal irradiation (DNI) fraction (less than 50%) climate zones is poor due to large solar field sizes required and high payback time compared to a single-effect chiller system driven by heat from evacuated tube collector array. On the other hand, in climatic regions with very high DNI faction (more than 65%), the combination of triple-effect chiller and parabolic trough collectors results in the most energy-efficient and cost-effective plant configuration, achieving the smallest solar field and the lowest payback period. Chemisana* et al.* [12] compared cooling systems connected and not connected to solar collectors. The solar cooling devices were basic single-effect absorption system that used heat from evacuated tube collectors, double-effect absorption system coupled with a Fresnel reflective collector incorporated on a building façade. These authors found that the solar absorber area of the Fresnel collector system was lower than that of standard solar thermal machine. Beltagy* et al.* [13] investigated theoretically and experimentally a solar cooling prototype with Fresnel collectors. The setup was designed basing on developed theoretical model, and experimental data was compared to numerical results. It was found that the measured and estimated produced thermal power from collectors and the efficiency per day were comparable. In addition, they found that the efficiency exceeds 40% per day for the 250 kW thermal prototype.

A 174 kW solar/gas double-effect LiBr/water absorption plant was tested at Seville in Spain [14]. The system was driven by a pressurized hot water flow using a 352 m^{2} heated by a linear Fresnel collector. A direct-fired natural gas burner was used as auxiliary heat source. For 75% of generator’s total heat input covered by solar energy, the average collector efficiency was found to vary between 0.35 and 0.4, and the mean coefficient of performance of the chiller, between 1.1 and 1.25.

Major drawbacks of the solar cooling systems are low efficiency in case of use of solar collectors with single-effect absorption systems and huge areas of thermal collectors required in case of double-effect absorption cycles. New machine configurations with high efficiency are needed. In literature, it is reported that some modified designs of single-effect chillers might exhibit coefficients of performance similar to that machines applying double-effect cycles [15, 16]. In these configurations, ejectors are incorporated in between generator and condenser.

The aim of this paper is to investigate a solar combined single-effect cycle using LiBr-H_{2}O as working fluid and coupled to solar collector. Because high temperature levels are needed to active the installation, a linear Fresnel collector is considered for the production of high temperature steam required to regenerate the refrigerant in the desorber. Driving temperature and chilled water temperature as well as cooling water temperature affect the performance of the chiller, , and the overall system performance, . Following simulations analyze thoroughly the effect of these three parameters. Also a case study of a typical Tunisian summer day is performed.

#### 2. Description of the System

The proposed solar 60 kW cooling system is coupled with a 164 m^{2} linear Fresnel collector, as schematically illustrated in Figure 1. The adopted collector (type IS-LF11 [17]) is used to produce steam at temperatures between 180°C and 230°C (pressures of 10 bar and 28 bar, respectively) for driving the generator of the combined cooling cycle. Cold is produced at temperatures between 12°C and 4°C.

The combined ejector single-effect cycle looks like an ordinary absorption single-effect cycle where an ejector is inserted between the generator and the condenser. The basic operational principle of the system is as follows: In the generator, the aqueous salt solution (H_{2}O/LiBr) is brought to boiling by heat provided by the solar source, in order to vaporize the refrigerant (water) and separate it from the solution. These high-pressure refrigerant vapors are introduced in the ejector nozzle as primary flow where they are accelerated and strongly depressurized. A fraction of the vapor exiting the evaporator is entrained by the primary flow into the ejector. The two gases are then well mixed in the mixing chamber and compressed in the diffuser before feeding the condenser where they liquefy by rejecting condensation heat to the cooling water (cw). The condensate expands through an adiabatic expansion valve and proceeds to the evaporator where it evaporates by absorbing heat from the chilled water*. *While in the basic cycle, all the refrigerant vapor leaving the evaporator proceeds to the absorber, only a part of it flows into the absorber in the combined cycle where it dissolves in the salt rich solution. The rest of the vapor is sucked in the ejector as secondary flow. The salt solution that leaves the absorber is pumped to the generator , after preheating in the solution heat exchanger.

The generator is activated by the heat liberated by condensation of hot saturated steam (15) which is then pumped back as liquid (16) to the steam drum. A primary thermal transfer fluid loop (17-18) vehicles heat from solar collector to steam drum.

##### 2.1. Solar Linear Fresnel Collector

IS-LF11 solar linear Fresnel collectors considered here are combined with boiler and can be used for industrial processes. Process heat in the range of 100 kW to 10 MW at standard pressure of 40 bar is generated, but up to 120 bar and temperatures up to 400°C are possible [17]. A “module” is one block of collectors associated with 4.06 m long vacuum absorber tubes. Figure 2 presents two modules of the adopted industrial collector. It is possible to use different kinds of heat transfer fluids (HTF): pressurized water, thermal oil, superheated steam, etc. Table 1 gives the manufacturer data of the commercial industrial Fresnel collector used in the simulations.

##### 2.2. LiBr-H_{2}O Combined Ejector Single-Effect Absorption Chiller

In the investigated combined ejector single-effect absorption cycle the ejector is located between condenser and generator. It uses the vapor produced in the generator to entrain and compress the refrigerant vapor coming from the evaporator before flowing in the condenser. Consequently, the pressure in generator and condenser—equal in a conventional single-effect cycle—are now distinct: intermediate pressure in condenser, equal to the high pressure in the original cycle, and a higher pressure in the generator. Hence, higher driving source temperatures are now required, between 180°C and 210°C. Table 2 gives the assumed operating conditions in the simulations. In particular, the nominal cooling power of the installation is kW, available in form of chilled water at temperatures between 4°C and 12°C for driving temperatures and pressures in the range 180°C – 210°C and 198 kPa – 270 kPa, respectively.

##### 2.3. Ejector

A typical ejector configuration is shown in Figure 3. A steam ejector is composed of four zones: primary nozzle, suction chamber, mixing chamber (convergent duct followed by constant section area), and diffuser. Table 3 gives the ejector geometrical data [16].

#### 3. Theoretical Model

##### 3.1. Linear Fresnel Collector Model

The optical efficiency of the collector depends on the incidence angle of the direct irradiation. As reference of is its value under ideal conditions: clean reflectors and absorber tubes, sun at zenith. For the adopted collector and (sun at 5° transversal zenith angle, and 0° longitudinal zenith angle). For arbitrary incidence angle is multiplied by a correction factor, the so-called incidence angle modifier (). has two components: a field independent transversal component, , and a longitudinal component, , that depends on the number of in string assembled modules (one in Figure 2).

The thermal losses of the reflector, , are dominated by radiative losses and approximately estimated by stands for the active aperture area of the collector and for the temperature difference between that of the heat transfer flow HTF and the ambient, as expressed in The thermal loss coefficients of the IS-LF11 are and [17]. Equation (1) reduces then to The collected thermal power writes as where is the zenith angle, , the angle between absorber pipe and the north-south direction, and , the direct normal irradiance (DNI).

The thermal efficiency of the collector is the ratio of the usable thermal power and the solar irradiance collected , detailed in Figure 4 depicts this correlation, * versus*, for the adopted collector and various values of [17]. It is worthy noticing that knowing thermal efficiency, irradiation, and actual collector size, the expected gross heat produced by the collector can be predicted.

##### 3.2. Combined Ejector Single-Effect Absorption Cycle Model

The mathematical model of the cycle is based on mass and energy balances written for each machine element. For a steady state process and neglecting kinetic and potential energy variations, balance equations for global mass (equation (6)), salt mass (equation (7)), and energy (equation (8)) write, respectively,

and are heat and transfer rates between machine element and its surroundings. The of the absorption system is expressed as

##### 3.3. Ejector Model

The adopted 1D model of the ejector is detailed in reference [15]. It is based on quadratic equations of each zone and enables the prediction of the ejector performance, i.e., the entrainment ratio (ratio of mass flow rate of the primary flow and secondary flow rate ) for given ejector geometry and inlet and outlet pressures and temperatures.

#### 4. Validations

A computer programs are developed using the software Engineering Equation Solver (*EES*) [20] in order to validate thermodynamically the models of combined ejector single-effect absorption unit and steam ejector.

##### 4.1. Validation of Combined Ejector Single-Effect Model

The simulations are run under following conditions, according to the data of [16]:(i)Evaporator temperature: 5°C(ii)Condenser temperature: 28°C(iii)Concentrations of salt solution leaving absorber: 62.3%(iv)Heat exchanger effectiveness: 0.8(v)Solution flow rate leaving generator: 0.393 kg/min(vi)Generator pressure: 198 kPa and temperature: 192°C

Table 4 compares the obtained simulations results with those reported by Aphornratana* et al*. [16]. As can be noted, deviations are generally small, less than 1%, with a maximum of 6% for the mass flow rate of refrigerant leaving the evaporator. This can be attributed to the simplifying assumption of perfect gas behavior for the steam in the ejector model. Deviations of driving power and cooling load are respectively 0.3% and 6%. Thus, the proposed model for the simulation of the chiller is fairly validated.

##### 4.2. Validation of Ejector Model

In the experimental work of Sriveerakul* et al*. [19], various combinations of ejector geometry were tested for different operating temperatures and pressures. Primary and secondary flows were vapors. The 1D model used for the machine simulations is tested against the data of this experimental work for validation. Calculated results are compared with data reported in [19] in Table 5.

Deviations for backpressure, entrainment ratio, and ejector area ratio are determined asMinimum absolute percent deviation of backpressure is 3% and maximum, 23%. For entrainment ratio the corresponding values are 0.3% and 18%, and for ejector area ratio, 0% and 13%. Based on the simulation results and percent deviations reported in Table 5, it can be concluded that the proposed 1D ejector model describes fairly well the experimental data.

#### 5. Solar Chiller Simulation Results and Discussion

Simulations of a 60 kW solar chiller are performed in order to analyze thermodynamically and predict the performances of a combined ejector single-effect absorption refrigeration cycle coupled with linear Fresnel collector modules. The computer program was developed using the software Engineering Equation Solver (*EES*) [20]. Thermophysical properties of the refrigerant system LiBr-H_{2}O are determined by applying internal procedures of the software. Tunisian climate and meteorological data are taken into consideration in the program.

A solar absorption chiller is characterized by an overall coefficient of performance, , a key factor to describe the efficiency of the entire system, expressed in (11), and is defined as the product of the efficiency of the solar collector and the thermal coefficient of performance of the absorption refrigerant cycle [21]:

##### 5.1. Effect of Generator Temperature and Pressure on System Performance (for Constant Chilled Water Temperature)

Simulations are performed by varying gene rotor temperature and pressure as well as the cooling medium temperature, for fixed chilled outlet-temperature, °C. The results are represented graphically in Figures 5(a), 5(b) and 5(c), for, respectively, three values of the generator pressure 198 kPa, 232 kPa, and 272 kPa. The coefficient of performance of the chiller together with the performance of the whole installation , including the solar facility, is depicted* versus* generator temperature for three values of the generator pressure and cooling medium temperature.

**(a)**

**(b)**

**(c)**

Given these findings, we note that(i)The increases with increasing generator temperature, . For fixed , it decreases with decreasing cooling water temperature.(ii)The reaches values ranging from 0.4 to 1.03, reaching the performances of double- effect absorption cycles, thus larger than what can be obtained in case conventional single-effect without ejector [22].(iii)The is almost half due to the limited efficiency of the solar collector installation. Even in the case of favorable conditions, in gray high-lightened zone in Figure 4 and representing a common situation in Tunisian summer, this efficiency is ranging between 0.5 and 0.6.(iv)It is observed that the maximum (and consequently the ) attains higher values when the generator pressure is larger. The highest performances are reached for the maximum generator temperature considered in this study, = 210°C. Table 6 summarizes these results. In fact, for minimal cooling medium temperature (25°C) and highest generator temperature (210°C), the = 0.99, 1.02, 1.03 and the = 0.57, 0.59, 0.6 for generator pressures at 198 kPa, 232 kPa, 272 kPa, respectively. It is noticed further that(v)For a fixed generator pressure, the maximum decreases when the temperature of the cooling medium rises. So when the pressure in the generator is at 198 kPa, the maximum = 0.99, 0.85, 0.75 for cooling medium temperature, respectively, set at 25°C, 30°C, 35°C, with the corresponding values for = 0.57, 0.52, 0.46.(vi)The performances of cycle and system are optimal for the highest generator pressure (270 kPa), lowest cooling medium temperature (25°C), and largest chilled water temperature (12°C).

##### 5.2. Effect of Generator Temperature and Pressure on System Performance (for Constant Cooling Medium Temperature)

Similar simulations are run by maintaining constant the cooling medium temperature °C and varying the chilled water temperature Figures 6(a), 6(b) and 6(c) present the evolution of and of the solar combined single-effect absorption refrigeration system for three values of the generator pressure 198 kPa, 232 kPa, and 272 kPa, respectively. It can be noted that(i)Again, both and increase when the generator temperature is increased.(ii) and increase with the driving temperature, similarly to the behavior already observed in Figure 5. Table 7 gives the maximum of cycle and system performances, when the driving temperature reaches its highest value, 210°C. For = 12°C reaches 0.87, 0.91, and 0.95 for generator pressure 198 kPa, 232 kPa, and 270 kPa, respectively.(iii)The performances of cycle and system ( and the ) increase with increasing chilled water temperature. For example, as shown in Table 7, for generator pressure equal to 270 kPa, the maximum = 0.84, 0.91, 0.95 for chilled water temperature, respectively, set at 7°C, 10°C, 12°C, with corresponding values for = 0.51, 0.55, 0.58.(iv) is 60% lower than . It ranges between 0.1 and 0.6, comparable with the results other authors [22, 23].(v)When the cooling medium temperature is set at 32°C, the performances of the cycle and the system reach their maximum at the highest pressure in the generator (270 kPa) and for a chilled water temperature = 12°C.

**(a)**

**(b)**

**(c)**

##### 5.3. Case Study: Hourly Local Data

Belghouthi* et al*. [18] performed a thermal test of weather and solar conditions during July 2008 at a research centre in Borj Cedria, near Tunis. The research centre is located at 36°43’ latitude and 10°25’ longitude. Figure 7 shows the data for a particular day (16th of July 2008), of this Mediterranean climate zone characterized by a high level of solar resources. It receives an average of 17 kJ/m^{2}/day with a total insolation period of 3500 h/year and 350 sunny days per year. Ambient temperature varies between 39°C and 27°C, and solar radiation in the range 250 W/m^{2}–900 W/m^{2}. Maximum solar radiation of 958 W/m^{2} is observed at 12:00. Maximum ambient temperature of 37°C is found at 14:00. Figure 8 reproduces the solar chart for the test day.

The efficiency of the Fresnel collector as expressed in (5) depends on the quadratic temperature difference (equation (2)) between that of heat transfer flow and ambient. In the high-lightened gray interval in Figure 4, the temperature affects more the collector efficiency than the irradiance . Considering the model equations (1)-(5) and Figure 4, the lower is the higher is the collector efficiency. Due to the compensating effects of solar radiation and ambient temperature during the time interval 11:00-15:00, the efficiency of the collector is at its maximum at this period as shown in Figures 9 and 10.

Figure 11 shows the evolution and during the test summer day. The reaches always values higher than 0.9 and the is higher than 0.55. The cycle performances are independent on ambient temperature, set at 25°C; this is the reason of the constant The maximum is equal to 0.99, 1.02, 1.03 for generator pressure set, respectively, at 198 kPa, 232 kPa, and 270 kPa. The evolution of is similar to that of the collector efficiency; it reaches its maximum in the middle of the day from 12:00 to 14 pm, namely, 0.61, 0.63, and 0.63 for generator pressure set, respectively, at 198 kPa, 232 kPa, and 270 kPa.

#### 6. Conclusion

This paper investigated solar combined ejector single-effect absorption refrigeration cycle. The considered chiller is driven by condensation of steam at temperatures between 180°C and 210°C, using pressurized water heated by a linear Fresnel collector.

The model of the collector is detailed and simulations of the overall system and the cooling machine are performed. The produced water chilled water is varied between 7°C and 12°C. In regard to the simulation results, it can be concluded that(i)The increases with the driving the heat temperature increases, decreases when cooling medium temperature gets higher, and increases with the chilled water temperature.(ii)The of this particular single-effect cycle can reach 1.03, a value comparable to that of a conventional double-effect cycle.(iii)For fixed cooling medium temperature, the maximum reaches higher values when the generator pressure is larger,* e.g*., for a cooling medium set at 25°C and solution generator temperature at 210°C, the = 0.99, 1.02, 1.03 and the = 0.57, 0.59, 0.6, respectively, for generator pressure is at 198 kPa, 232 kPa, and 272 kPa.(iv) is almost half of due to the limited efficiency of the solar collector compartment, between 0.5 and 0.6.(v)For a fixed generator pressure, the maximum decreases when the temperature of the cooling medium rises. For a generator pressure of 198 kPa and chilled water temperature at 7°C, the maximum is 0.99, 0.85, and 0.75 for cooling medium temperature of 25°C, 30°C, and 35°C, respectively. The corresponding values of are 0.57, 0.52, and 0.46.(vi)For a cooling medium temperature of 32°C and by fixed chilled water temperature, the and the increase with the driving heat temperature, reaching a maximum for °C: 0.87, 0.91, and 0.95 for, respectively, 198 kPa, 232 kPa, and 270 kPa generator pressure and at 12°C chilled water temperature.(vii)The performances of cycle and system are optimal for the highest generator pressure (270 kPa in this study) and chilled water temperature (12°C).

A case study of a typical summer day is simulated; results showed that(i)The average efficiency of linear Fresnel collector considered in the simulations varies between 0.6 and 0.62.(ii)In the test period, reaches values always higher than 0.9 and larger than 0.55. The cycle performances are independent from ambient temperature with a maximum equal to 0.99, 1.02, and 1.03 for generator pressure, respectively, at 198 kPa, 232 kPa, and 270 kPa.(iii)The system performance evolves parallel to that of the collector efficiency, reaching its maximum in the middle of the day from 12:00 to 14 pm.

#### Data Availability

Previously reported solar collector data used to support this study are available at [W. Christine, B. Michael, M. Florian, H. Alexander, N. Tomas, “Solar cooling with water-ammonia absorption chillers and concentrating solar collector-operational experience”, International Journal of refrigeration, vol. 39, pp. 57-76, 2014.]. These prior studies and datasets are cited at relevant places within the text as [17]. Also the local metrological data were used to support this study and are available at [M. Balghouthi, M. H. Chahbani, and A. Guizani, “Investigation of a solar cooling installation in Tunisia,” Applied Energy, vol. 98, pp. 138–148, 2012.]. These prior studies and datasets are cited at relevant places within the text as [18].

#### Conflicts of Interest

The authors declare that they have no conflicts of interest.